summary table research

Log in using your username and password

• Search More Search for this keyword Advanced search
• Latest content
• Current issue
• Write for Us
• BMJ Journals

You are here

• Volume 24, Issue 2
• Five tips for developing useful literature summary tables for writing review articles
• Article Text
• Article info
• Citation Tools
• Rapid Responses
• Article metrics

• http://orcid.org/0000-0003-0157-5319 Ahtisham Younas 1 , 2 ,
• http://orcid.org/0000-0002-7839-8130 Parveen Ali 3 , 4
• 1 Memorial University of Newfoundland , St John's , Newfoundland , Canada
• 2 Swat College of Nursing , Pakistan
• 3 School of Nursing and Midwifery , University of Sheffield , Sheffield , South Yorkshire , UK
• 4 Sheffield University Interpersonal Violence Research Group , Sheffield University , Sheffield , UK
• Correspondence to Ahtisham Younas, Memorial University of Newfoundland, St John's, NL A1C 5C4, Canada; ay6133{at}mun.ca

https://doi.org/10.1136/ebnurs-2021-103417

Statistics from Altmetric.com

Request permissions.

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Introduction

Literature reviews offer a critical synthesis of empirical and theoretical literature to assess the strength of evidence, develop guidelines for practice and policymaking, and identify areas for future research. 1 It is often essential and usually the first task in any research endeavour, particularly in masters or doctoral level education. For effective data extraction and rigorous synthesis in reviews, the use of literature summary tables is of utmost importance. A literature summary table provides a synopsis of an included article. It succinctly presents its purpose, methods, findings and other relevant information pertinent to the review. The aim of developing these literature summary tables is to provide the reader with the information at one glance. Since there are multiple types of reviews (eg, systematic, integrative, scoping, critical and mixed methods) with distinct purposes and techniques, 2 there could be various approaches for developing literature summary tables making it a complex task specialty for the novice researchers or reviewers. Here, we offer five tips for authors of the review articles, relevant to all types of reviews, for creating useful and relevant literature summary tables. We also provide examples from our published reviews to illustrate how useful literature summary tables can be developed and what sort of information should be provided.

Tip 1: provide detailed information about frameworks and methods

• Download figure
• Open in new tab
• Download powerpoint

Tabular literature summaries from a scoping review. Source: Rasheed et al . 3

The provision of information about conceptual and theoretical frameworks and methods is useful for several reasons. First, in quantitative (reviews synthesising the results of quantitative studies) and mixed reviews (reviews synthesising the results of both qualitative and quantitative studies to address a mixed review question), it allows the readers to assess the congruence of the core findings and methods with the adapted framework and tested assumptions. In qualitative reviews (reviews synthesising results of qualitative studies), this information is beneficial for readers to recognise the underlying philosophical and paradigmatic stance of the authors of the included articles. For example, imagine the authors of an article, included in a review, used phenomenological inquiry for their research. In that case, the review authors and the readers of the review need to know what kind of (transcendental or hermeneutic) philosophical stance guided the inquiry. Review authors should, therefore, include the philosophical stance in their literature summary for the particular article. Second, information about frameworks and methods enables review authors and readers to judge the quality of the research, which allows for discerning the strengths and limitations of the article. For example, if authors of an included article intended to develop a new scale and test its psychometric properties. To achieve this aim, they used a convenience sample of 150 participants and performed exploratory (EFA) and confirmatory factor analysis (CFA) on the same sample. Such an approach would indicate a flawed methodology because EFA and CFA should not be conducted on the same sample. The review authors must include this information in their summary table. Omitting this information from a summary could lead to the inclusion of a flawed article in the review, thereby jeopardising the review’s rigour.

Tip 2: include strengths and limitations for each article

Critical appraisal of individual articles included in a review is crucial for increasing the rigour of the review. Despite using various templates for critical appraisal, authors often do not provide detailed information about each reviewed article’s strengths and limitations. Merely noting the quality score based on standardised critical appraisal templates is not adequate because the readers should be able to identify the reasons for assigning a weak or moderate rating. Many recent critical appraisal checklists (eg, Mixed Methods Appraisal Tool) discourage review authors from assigning a quality score and recommend noting the main strengths and limitations of included studies. It is also vital that methodological and conceptual limitations and strengths of the articles included in the review are provided because not all review articles include empirical research papers. Rather some review synthesises the theoretical aspects of articles. Providing information about conceptual limitations is also important for readers to judge the quality of foundations of the research. For example, if you included a mixed-methods study in the review, reporting the methodological and conceptual limitations about ‘integration’ is critical for evaluating the study’s strength. Suppose the authors only collected qualitative and quantitative data and did not state the intent and timing of integration. In that case, the strength of the study is weak. Integration only occurred at the levels of data collection. However, integration may not have occurred at the analysis, interpretation and reporting levels.

Tip 3: write conceptual contribution of each reviewed article

While reading and evaluating review papers, we have observed that many review authors only provide core results of the article included in a review and do not explain the conceptual contribution offered by the included article. We refer to conceptual contribution as a description of how the article’s key results contribute towards the development of potential codes, themes or subthemes, or emerging patterns that are reported as the review findings. For example, the authors of a review article noted that one of the research articles included in their review demonstrated the usefulness of case studies and reflective logs as strategies for fostering compassion in nursing students. The conceptual contribution of this research article could be that experiential learning is one way to teach compassion to nursing students, as supported by case studies and reflective logs. This conceptual contribution of the article should be mentioned in the literature summary table. Delineating each reviewed article’s conceptual contribution is particularly beneficial in qualitative reviews, mixed-methods reviews, and critical reviews that often focus on developing models and describing or explaining various phenomena. Figure 2 offers an example of a literature summary table. 4

Tabular literature summaries from a critical review. Source: Younas and Maddigan. 4

Tip 4: compose potential themes from each article during summary writing

While developing literature summary tables, many authors use themes or subthemes reported in the given articles as the key results of their own review. Such an approach prevents the review authors from understanding the article’s conceptual contribution, developing rigorous synthesis and drawing reasonable interpretations of results from an individual article. Ultimately, it affects the generation of novel review findings. For example, one of the articles about women’s healthcare-seeking behaviours in developing countries reported a theme ‘social-cultural determinants of health as precursors of delays’. Instead of using this theme as one of the review findings, the reviewers should read and interpret beyond the given description in an article, compare and contrast themes, findings from one article with findings and themes from another article to find similarities and differences and to understand and explain bigger picture for their readers. Therefore, while developing literature summary tables, think twice before using the predeveloped themes. Including your themes in the summary tables (see figure 1 ) demonstrates to the readers that a robust method of data extraction and synthesis has been followed.

Tip 5: create your personalised template for literature summaries

Often templates are available for data extraction and development of literature summary tables. The available templates may be in the form of a table, chart or a structured framework that extracts some essential information about every article. The commonly used information may include authors, purpose, methods, key results and quality scores. While extracting all relevant information is important, such templates should be tailored to meet the needs of the individuals’ review. For example, for a review about the effectiveness of healthcare interventions, a literature summary table must include information about the intervention, its type, content timing, duration, setting, effectiveness, negative consequences, and receivers and implementers’ experiences of its usage. Similarly, literature summary tables for articles included in a meta-synthesis must include information about the participants’ characteristics, research context and conceptual contribution of each reviewed article so as to help the reader make an informed decision about the usefulness or lack of usefulness of the individual article in the review and the whole review.

In conclusion, narrative or systematic reviews are almost always conducted as a part of any educational project (thesis or dissertation) or academic or clinical research. Literature reviews are the foundation of research on a given topic. Robust and high-quality reviews play an instrumental role in guiding research, practice and policymaking. However, the quality of reviews is also contingent on rigorous data extraction and synthesis, which require developing literature summaries. We have outlined five tips that could enhance the quality of the data extraction and synthesis process by developing useful literature summaries.

• Aromataris E ,
• Rasheed SP ,

Twitter @Ahtisham04, @parveenazamali

Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

Competing interests None declared.

Patient consent for publication Not required.

Provenance and peer review Not commissioned; externally peer reviewed.

Read the full text or download the PDF:

Drafting a summary table

Contributor: Logan Miller

A summary table allows you to compare common research methods, findings, limitations, etc. You can order the entries in any way that you find useful; consider ordering your research alphabetically, by timeliness, or even by grouping similar study aims, models, or results.

Once compiled, you can use this table to compare studies side by side. Such comparison can help you see trends in findings, identify gaps in the research, and rank each study by relative strength. In short, it helps you organize information on a broad topic, which is a crucial first step in synthesizing that information within a research paper.

Summary areas might include

Authors / date : If a paper has numerous authors, consider the level of detail you require to identify a given study.

Aim of study / paper : What were the researchers hoping to learn? This section may include research questions or hypotheses.

Type of study / information : These might be systematic reviews, randomized controlled trials, etc. If you’re less familiar with what these designs entail, writing a short description can be useful.

Main findings / conclusions : The level of detail you employ will come down to necessity and experience, but in listing specific findings, you may see trends or discrepancies across studies.

Strengths / limitations : Strengths may include good research design or data-based conclusions. Remember, a study may mention its limitations explicitly, but many limitations require careful inquiry to uncover.

*Azzopardi, D., Patel, K., Jaunky, T., Santopietro, S., Camacho, O. M., McAughey, J., Gaça, M. (2016). Electronic cigarette aerosol induces significantly less cytotoxicity than tobacco smoke. Toxicology Mechanisms and Methods 26(6), 477-497, doi: 10.1080/15376516.2016.1217112

Nursing and Allied Health: Building a Summary Table or Synthesis Matrix

• Nursing Library Services
• Library Support for Nursing Accreditation
• Academic Research Libraries (ACRL) Information Literacy Standards for Science
• ACRL Guidelines for Distance Learning
• AACN Resources: Information Literacy
• The Essentials: Core Competencies for Professional Nursing Education
• Accreditation Commision for Education
• ACEN Learning Resources | Definitions & Standards
• Tutoring & Paper Review: College of Nursing & Allied Health Students
• Technology and Collaboration
• Affordable Learning Louisiana
• NIH Virtual Reality Funding
• Assistive Technology Resources
• Web Accessibility Initiative WC3
• Healthcare Links for Persons With Disabilities
• Accessibility
• Accessibility Tools: Videos
• Braile Institute: Blind and Low Vision Web Accessibility
• Braile Institute: Assistive Technology for the Visually Impaired This link opens in a new window
• Mental Health Resources for Students & Faculty
• Student Activities & Organizations
• Anatomage FAQ's
• APA Reference Worksheet With Examples This link opens in a new window
• APA Style and Grammar Guidelines, 7.0
• What's New in APA Guide, 7.0 This link opens in a new window
• APA Instructional Aids for Professors/Instructors
• APA Handouts & Guides This link opens in a new window
• Sample Papers for Students and Instructors
• Academic Writer Tutorial: Basics This link opens in a new window
• APA Styling Your Paper
• When to cite or not cite a database
• Video: A Step-By-Step Guide for APA Style Student Papers
• Video: Citing Works in Text Using APA 7.0 This link opens in a new window
• Video: Creating References Using APA Guide, 7.0 This link opens in a new window
• Journal Article Reporting Standards
• Digital Information Literacy
• Tips Sheet: Copyright Essentials for Higher Education This link opens in a new window
• Citing National Patient Safety Goals (Joint Commission)
• Citing Agency for Healthcare Research and Quality (AHRQ) in APA format
• Best Nursing Apps
• Writing an Abstract for Your PILT or Special Project
• PILT Poster Presentation Archive
• Academic Success
• Healthcare Career Salaries
• Industry Statistics: Healthcare
• Nursing Organizations
• Radiology Organizations
• Controlled Medical Vocabularies
• COVID-19 Current Vaccines (Medscape)
• COVID-19 Current and Candidate Vaccine Landscape and Tracker (WHO)
• COVID-19: Clinician Care Guidance (CDC.gov) This link opens in a new window
• COVID -19 Contract Tracing & Patient Privacy (Cornell) This link opens in a new window
• COVID-19: Coronavirus Dashboard (Johns Hopkins Epidemiology)
• COVID-19: Coronavirus Guidelines (Up-To-Date) This link opens in a new window
• COVID-19: Critical Care Evidence (Cochrane) This link opens in a new window
• COVID-19: Diagnosis & Treatment (JAMA) This link opens in a new window
• COVID-19: Free Video Access (Jove) This link opens in a new window
• COVID-19: Healthcare Hub (Elsevier) This link opens in a new window
• COVID-19: Keeping Up With A Moving Target (Johns Hopkins Nursing Videos)
• COVID-19: LitCovid Daily Update (NLM)
• COVID 19 - Long Term Health Effects Stemming From COVID-19
• COVID-19: Louisiana Department of Health
• COVID-19: Novel Coronavirus Information Center (Elsevier) This link opens in a new window
• COVID-19 Nursing Resources (Medscape)
• COVID-19: Red Book - Recent Pediatric Updates (AAP)
• COVID-19: Washing Your Hands Thoroughly (NHS)
• COVID-19: Well-Being Initiative (ANF)
• Properly Putting On Your Facemask & Getting a Good Seal (Dr. Scheiner) This link opens in a new window
• Creating Personal Accounts
• Creating a CINAHL MyFolder Account
• Creating a PubMed | MyNCBI Personal Account
• Creating a ProQuest Nursing & Allied Health Premium | MyResearch Folder
• Creating an OVID MyWorkspace Personal Account
• Mobile APPS | CINAHL for Apple and Android Mobile Devices
• My Circulation Login
• Interlibrary Loan Personal Account
• Data Visualization Products
• International Classification of Diseases (ICD)
• Diagnostic and Statistical Manual of Mental Disorders | DSM-5-TR
• [C1] Infections | Infectious Diseases
• [C04] Neoplasms
• [C05] Musculoskeletal Diseases
• [C06] Digestive System Diseases
• [C07] Stomatognathic Diseases
• [C08] Respiratory Tract Diseases
• [C09] Otorhinolaryngologic Diseases
• [C10] Nervous System Diseases
• [C11] Eye Diseases
• [C12] Urogenital Diseases
• [C14] Cardiovascular Diseases
• [C15] Hemic and Lymphatic Diseases
• [C16] Congenital, Hereditary, and Neonatal Diseases and Abnormalities
• [C17] Skin and Connective Tissue Diseases
• [C18] Nutritional and Metabolic Diseases
• [C19] Endocrine System Diseases
• [C20] Immune System Diseases
• [C21] Disorders of Environmental Origin
• [C22] Animal Diseases [Zoonotic diseases]
• [C23] Pathological Conditions, Signs and Symptoms
• [C24] Occupational Diseases
• [C25] Chemically-Induced Disorders
• [C26] Wounds and Injuries
• WHO Drug Information [Relative to Diseases]
• NDDK Patient Education Tool Kit
• Clinical Tools & Patient Education
• NDDK Resources on Medline Plus
• NDDK Open Research
• Dietary Guidelines for Americans
• Physical Activity Guidelines for Americans
• Move Your Way Community Resources
• National Youth Sports Strategy
• President’s Council on Sports, Fitness & Nutrition
• White House Conference on Hunger, Nutrition, and Health
• Equitable Long-Term Recovery and Resilience
• National Health Observances
• Finding Clinical Trials
• NIH News in Health
• Dosage Calculations & Pharmacology
• PICO - EBM Video This link opens in a new window
• PICO Slides
• Fillable CONSAH Pico Form
• Evidence Based Practice for Nursing
• Evidence-Based Nursing: -Step 2
• Evidence Appraisal - Step 3
• Evidence Application - Step 4
• Outcome Evaluation - Step 5
• Evidence Translation - Step 6
• Google Advanced Search for EBM
• Nursing Research Methods
• Faculty Book Request
• Proctor Request Form
• Peer Reviewed Literature: Assessment Goals
• Full Text Finder
• EBSCO e-Books Nursing 2021
• EBSCO eBooks Anesthesia
• EBSCO eBooks Radiology Science & Allied Health
• EBSCO eBooks: Writing About Nursing and Allied Health
• Alzheimers and Dementia
• Statistics on Aging
• CDC Bibliography: Alzheimers & Aging This link opens in a new window
• Health Conditions
• Health Behaviors
• Populations
• Settings and Systems
• Social Determinants of Health
• ILL Interlibrary Loan
• Gestational Diabetes and Fast Foods (MeSH)
• Mobile Resources
• Nursing Theory
• Psychiatric Nursing Journals
• Display, sort and; navigate
• Similar articles
• Cite, save and share
• Citations in PubMed
• All About PubMed Filters
• PubMed Quick Tours and Tutorials
• Evidence Based Practice Tutorial (PubMed) This link opens in a new window
• Developing a Clinical Question This link opens in a new window
• Using PubMed to Find Relevant Articles This link opens in a new window
• Next Steps This link opens in a new window
• Scenario (practice) This link opens in a new window
• Radiology Books and e-Books
• History of Radiology & Radiography
• Radiology: Anatomage This link opens in a new window
• Radiology Anatomy Atlas Viewer
• Advanced Radiographic Research
• Diagnostic Imaging Selected Articles
• Faculty and Administrative Resources
• Radiology Tech and MRI Salaries (Bureau of Labor Statistics)
• Radiology Technician Demand by State
• Review Tools for Graduate Students
• Training & Videos
• Register for an Online Meeting
• Joining an Online Meeting
• Training Videos & Search Examples
• Training Survey
• Sources for Health Statistics
• Ebola and Infectious Diseases
• Nursing Sites
• Writing Research Papers

Building a Summary Table or Synthesis Matrix

Quick Links | Nursing & Allied Health

• Library Home
• Campus Computer Labs
• Circulation Policies
• Catalog Search (books/ebooks)
• Contact Your Librarian For Research Assistance             (link sends email)
• Database Directory >> Nursing & Allied Health
• Faculty Materials Request Form
• Holidays & University Closures
• Interlibrary Loan Request Form
• Journals Full Text Finder
• LibGuides (Online Help)
• Review My Paper**
• Shreveport Proctor Request Form
• Shreveport Main Phone:  (318) 677-3007
• Shreveport Librarian Phone:  (318) 677-3013
• Student Help Desk :  Phone: (318) 357-6696  [email protected]
• The 20-20-20 Rule
• Training Videos

** Submit your research paper for grammar review

   GOOGLE MAP  |  LINKED

• Simplifying Synthesis | Download the Article PDF Copy
• Writing a Literature Review and Using a Synthesis Matrix

What a Summary Table or Synthesis Matrix looks like

Use the "Literature Review Matrix Template" as a guideline to help you sort through your thoughts, note important points and think through the similarities and differences: 

You are organizing the review by ideas and not by sources .  The literature review is not just a summary of the already published works.  Your synthesis should show how various articles are linked. 

A summary table is also called a synthesis matrix.  The table helps you organize and compare information for your systematic review, scholarly report, dissertation or thesis

Synthesis Matrix.

A summary table is also called a synthesis matrix . A summary table helps you record the main points of each source and document how sources relate to each other. After summarizing and evaluating your sources, arrange them in a matrix to help you see how they relate to each other, and apply to each of your themes or variables.

Faculty who typically guide students find it challenging to help students learn how to synthesize material (Blondy, Blakesless, Scheffer, Rubenfeld, Cronin, & Luster-Turner, 2016; Kearney, 2015) .  Writers  can easily summarize material but seem to struggle to adequately synthesize knowledge about their topic and express that in their writing. So, whether you are writing a student papers, dissertations, or scholarly report it is necessary to learn a few tips and tricks to organize your ideas.

Building a summary table and developing solid synthesis skills is important for nurses, nurse practitioners, and allied health researchers.  Quality evidence-based practice initiatives and nursing care and medicine are based on understanding and evaluating the resources and research available, identifying gaps, and building a strong foundation for future work.

Good synthesis is about putting the data gathered, references read, and literature analyzed together in a new way that shows connections and relationships. ( Shellenbarger, 2016 ). The Merriam-Webster dictionary defines synthesis as something that is made by combining different things or the composition or combination of parts or elements so as to form a whole (Synthesis, n.d.).  

In other words, building a summary table or synthesis matrix  involves taking information from a variety of sources, evaluating that information and forming new ideas or insights in an original way.  This can be a new and potentially challenging experience for students and researchers who are used to just repeating what is already in the literature.

Visit Our Libraries

   Interlibrary Loan    |    Shreveport Education Center Library   |    Eugene P. Watson Memorial Library    |     NSU Leesville Library 

Cammie G. Henry Research Center    |   Prince Music Media Library

• << Previous: Best Nursing Apps
• Next: Writing an Abstract for Your PILT or Special Project >>
• Last Updated: Jul 29, 2024 2:38 PM
• URL: https://libguides.nsula.edu/NursingandAlliedHealth

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• J Med Libr Assoc
• v.109(1); 2021 Jan 1

Use of a search summary table to improve systematic review search methods, results, and efficiency

Alison c. bethel.

1 [email protected] , Information Specialist, Evidence Synthesis Team, University of Exeter Medical School, Exeter, United Kingdom

Morwenna Rogers

2 [email protected] , Evidence Synthesis Team, National Institute for Health Research Applied Research Collaboration South West Peninsula, University of Exeter Medical School, Exeter, United Kingdom

Rebecca Abbott

3 [email protected] , Evidence Synthesis Team, National Institute for Health Research Applied Research Collaboration South West Peninsula,, University of Exeter Medical School, Exeter, United Kingdom

Associated Data

Background:.

Systematic reviews are comprehensive, robust, inclusive, transparent, and reproducible when bringing together the evidence to answer a research question. Various guidelines provide recommendations on the expertise required to conduct a systematic review, where and how to search for literature, and what should be reported in the published review. However, the finer details of the search results are not typically reported to allow the search methods or search efficiency to be evaluated.

Case Presentation:

This case study presents a search summary table, containing the details of which databases were searched, which supplementary search methods were used, and where the included articles were found. It was developed and published alongside a recent systematic review. This simple format can be used in future systematic reviews to improve search results reporting.

Conclusions:

Publishing a search summary table in all systematic reviews would add to the growing evidence base about information retrieval, which would help in determining which databases to search for which type of review (in terms of either topic or scope), what supplementary search methods are most effective, what type of literature is being included, and where it is found. It would also provide evidence for future searching and search methods research.

Systematic reviews are designed to be comprehensive, robust, inclusive, transparent, and reproducible when bringing together the evidence to answer a research question. Depending on the field and topic, they may be large and time consuming with many included studies, or they can contain no relevant studies at all, finding that the area urgently requires primary research [ 1 ]. The timescales to publication can vary widely [ 2 ], and some systematic reviews are regularly updated, particularly if there is new relevant evidence in the field being researched [ 3 ]. However, research consistently shows that search strategies are not recorded well enough to allow them to be reproduced [ 4 – 6 ].

Systematic review guidelines recommend that the systematic review team include expertise in systematic review methods, including information retrieval [ 7 – 10 ]. Information retrieval is a core competency for librarians and information specialists who are involved in systematic reviews [ 11 ], and having a librarian or information specialist as part of the team is associated with significantly higher-quality search strategies [ 12 ]. The role of a librarian or information specialist in a systematic review can vary, ranging from the more limited role of checking searches written by others in the team to taking on many, if not all, aspects of search development and information retrieval [ 7 , 9 , 13 ].

Once a protocol is in place, one of the first tasks undertaken by the librarian or information specialist is to create search strategies for the predefined bibliographic databases listed in the protocol. Ideally, this task is then followed up by supplementary searching, such as forward and backward citation searching or table of contents searching in journals that are relevant to the topic [ 14 ]. Searching for grey literature is also recognized as an important part of a comprehensive search strategy for systematic reviews, and previous literature describes which resources are best suited to finding it and the contribution it can make to a systematic review [ 15 – 18 ]. Finally, depending on how long the systematic review takes to publication, update searches may also be part of the process [ 3 ].

Guidelines and guidance for conducting systematic reviews are available from various organizations, including Cochrane, the Campbell Collaboration, the Centre for Reviews and Dissemination, and the Joanna Briggs Institute. These guidelines all include some detail about how the searching should be undertaken, but there is no clear consensus about how many or which databases should be searched. Similarly, tools for assessing the quality of systematic reviews vary somewhat in their recommendations of what reviewers should look for in the search methods. Table 1 highlights a few of the different guidelines and checklists for undertaking, reporting, and appraising the searching component of systematic reviews, organized by the tool or author or organization.

Recommendations for systematic review searching from guidelines and checklists

As well as understanding where and how to search for information, it is important to understand how well the search strategies perform. Cooper et al. suggest there are six summative metrics of search effectiveness: sensitivity, specificity, precision, accuracy, number needed to read (NNR), and yield [ 23 ]. However, while there are suggested standards for reporting search methods and strategies [ 24 – 26 ], there currently are no requirements to report on these effectiveness metrics. Some research has been published on search effectiveness, but this research seems to be restricted to systematic reviews of certain conditions [ 27 – 32 ] or from specific organizations [ 33 – 35 ]. By reporting search effectiveness as well as search methods in more detail, evidence about information retrieval would accumulate, which could then inform guidelines about how many and which databases to search and which supplementary search methods to use for particular topics or types of evidence synthesis.

The aim of this study was to develop a search summary table (SST) that could report on search methods as well as search effectiveness. The authors demonstrate what an SST could look like and how it can be used. In the suggested SST, the only metric not covered by Cooper et al. [ 23 ] is “specificity,” because this requires a known number of references (e.g., when developing a search filter).

CASE PRESENTATION

The SST was tested by the Evidence Synthesis Team at the University of Exeter in a systematic review, “‘They've Walked the Walk': A Systematic Review of Quantitative and Qualitative Evidence for Parent-to-Parent Support for Parents of Babies in the Neonatal Unit” by Hunt et al. [ 36 ]. The database search strategies used for the review are provided in supplemental Appendix A , and a blank SST template is provided in supplemental Appendix B .

Completion of a search summary table (SST)

The SST was completed in two stages. In stage one, all the references that were downloaded or exported from every electronic database, including all duplicates, were recorded and saved in an EndNote library. Every record included a code for the database name where the record was found. As per traditional systematic review methods, the number of records screened at both the title-and-abstract stage and full-text stage were recorded as well as the final number of included references and which supplementary search methods were undertaken. Stage two involved rerunning the searches in those databases where most included references had been found in order to discover whether references that were not found during the original search were in the database and, if they were, whether they were retrieved by the search.

The SST presents the search information used to inform the PRISMA flow diagram, the search methods, and additional information gathered by the librarian or information specialist in their search log. Completion of stage one took approximately forty minutes and completion of stage two approximately one hour. Using this format to present the information allows calculation of various search effectiveness metrics.

Table 2 shows the key features of the SST. The first five metrics (numbered 1 to 5) are summative metrics of effective searching suggested by Cooper et al. [ 23 ]. Three additional metrics (numbered 6 to 8) provide further useful search-related information for the librarian or information specialist.

Metrics used in the search summary table (SST)

Sensitivity/recall and precision calculations are given in the SST for each database searched and overall, using the total number of references that have been found from database searching, the number of included (i.e., relevant) references from database searching, and the total number of included (i.e., relevant) references from all search methods. Reporting these metrics in this manner shows the effectiveness of search strategies for each individual database as well as database searching as a whole.

NNR usually indicates the number of references needed to screen at the title and abstract stage to find one included reference. However, the value of splitting this metric into two additional metrics can be seen: (1) number needed to screen (NNS), which is the number of references that needed to be screened during title and abstract screening to identify one reference to undergo full-text screening; and (2) number needed to read at full text (NNR FT), which is the number of references that needed to be read during full-text screening to include one reference in the systematic review. Reporting these three metrics separately increases the transparency of the searching and selection process.

Table 3 shows the SST for Hunt et al.'s systematic review [ 36 ].

Completed SST for Hunt et al.'s systematic review, ‘“They've Walked the Walk': A Systematic Review of Quantitative and Qualitative Evidence for Parent-To-Parent Support for Parents of Babies in Neonatal Care” [ 36 ]

Codes: x=found from the search; y=in database, found when search strategy rerun; n=not in the database; z=in the database, not found using the search strategy; qL=qualitative; qT=quantitative.

Format codes: jnl=journal article; ths=PhD thesis.

Supplementary search codes: fcs=forward citation search; bcs=backward citation search; hs=hand search; wss=website search; org=from contacting organizations.

Databases listed: ASSIA=Applied Social Sciences Index and Abstracts; BNI=British Nursing Index; HMIC=Health Management Information Consortium; PQDT=ProQuest Dissertations and Theses; SPP=Social Policy and Practice; WoS=Web of Science.

Contextual consideration of the findings

Key findings can be surmised from the metrics reported in the SST for this example systematic review, which involved a search for both qualitative and quantitative evidence.

Grey literature.

Two doctoral (PhD) theses were included in the systematic review, and both were found by searching in PsycINFO. One was also found by searching CINAHL. This was surprising, because grey literature searching is often seen as separate from the database searching process, yet these theses were found by searching bibliographic databases as opposed to Proquest Dissertations and Theses Global.

Search strategy comprehensiveness.

Three included references were found from citation searching, two of which were in both EMBASE and MEDLINE but were not retrieved by the database search strategies. If the search strategies had included the free-text search term “council*” ( supplemental Appendix A ), then these references would have been retrieved. This was an extremely valuable learning point for the information specialist in the team and reaffirmed the purpose of supplementary searching.

Unique references.

The only database to retrieve unique references (n=2) was PsycINFO, demonstrating the high degree of duplication among bibliographic databases.

Supplementary searching.

Hand searching, website searching, and organization searching was carried out but found no additional relevant references. Although forward citation searching found two additional relevant references, both of these (and a third additional relevant reference) were also found by backward citation searching. The time spent on these methods of supplementary searching was not recorded but might be useful in the future.

Qualitative references.

The CINAHL search retrieved only two of the eight qualitative references and did not retrieve any unique qualitative references. This was surprising, because previous research showed that this database was a good source of qualitative studies [ 37 ].

Quantitative references.

All the quantitative references were found from searching MEDLINE and citation searching.

Number needed to read.

Reporting the overall NNR as well as splitting this metric into two metrics—NNS and NNR FT—allowed more accurate and transparent reporting of the screening stages. Concerning the NNS, for every thirty-nine references that underwent title and abstract screening, one underwent full-text screening. Concerning the NNR, for every ten references read in full-text screening, one was included in the systematic review.

Often MEDLINE and EMBASE are suggested as the basic minimum for searching on health care topics [ 7 , 19 ]; however, in this case, neither database provided unique records, although MEDLINE had higher sensitivity and precision than EMBASE. For this particular systematic review, searching MEDLINE and PsycINFO along with backward citation searching would have found all of the included references. If this had been done, then the maximum number of references to screen would have been 2,835 (total number of references downloaded), a reduction of over 1,600.

The optimal number of databases that need to be searched varies depending on the review question. However, it is commonly agreed that searching only one database is not sufficient, and supplementary searching in some form is also needed. In this case, backwards citation searching found additional studies.

An SST can report data related to database and search performance and effectiveness in terms of sensitivity, precision, recall, NNR, yield, and number of unique records. Furthermore, additional information gathered during supplementary searching (e.g., citation searching and hand-searching) indicates the effectiveness of search strategies for individual databases and which methods of supplementary searching were most useful. This information could allow librarians and information specialists to be more selective when choosing databases and supplementary search methods. Publishing an SST as part of a systematic review would help to develop and make more explicit, rather than tacit, the model of the literature search process as described by Cooper et al. [ 38 ].

Future systematic reviews on a similar condition or population could use a related SST that is already available, either in-house or one that has been published, to enable a more evidence-based approach to database and search methods selection. For rapid reviews [ 39 ] and scoping reviews [ 40 ], in which searching might not be as exhaustive, this information could provide evidence about where to focus the search. SSTs would be particularly valuable in updating a systematic review [ 3 ]; if, for example, two databases that are always searched consistently do not contain any of the included studies, then perhaps they need not be searched in the future.

An SST only provides evidence for one particular systematic review; teams using them for future systematic reviews might not be fully confident that the same results would be produced for their specific questions. However, if all systematic reviews completed and published an SST as standard, then there would be more evidence available for making evidence-based decisions on which databases to search, which supplementary search methods would be most valuable, and which search strategies and terms would find the most relevant references for specific questions. Broad generalizations on searching cannot be made until more SSTs are available, but they can still be a valuable learning tool for all those involved in searching for systematic reviews, as their creation requires reflection on what was done and why, which can be carried on into the next systematic review. SSTs can also provide evidence for other purposes of searching, such as update searches [ 3 ] or scoping and preliminary searches [ 40 ].

SSTs can be useful to librarians and information specialists in several ways. First, for individuals who are new to the topic area or to systematic reviews, they provide a valuable source of evidence on which to base database and search method choices and recommendations. Second, they provide evidence about which databases are essential for undertaking specific systematic reviews, which could be useful for groups or individuals in negotiating database access with their institutions. Third, SSTs could help librarians and information specialists audit their database selections and search strategies, as they would show whether a database contains a reference and whether it would be captured by their search strategy. Fourth, a librarian or information specialist's knowledge would be built up more quickly, because completing an SST would help them reflect on their search strategies, search methods, and database selection.

Another area in which SSTs could be useful is in search methods and information retrieval research. If SSTs are published as part of a systematic review, then the searching becomes more transparent, replicable, and open, which is a fundamental component of good quality systematic reviews. Librarians and information specialists could use the data provided in SSTs to perform more thorough analyses on where studies are likely to be found and which databases suit particular topics. Trends might be observed, such as country-specific biases in database selection and use, and knowledge about specific databases could be shared in an easy format.

One specific area for monitoring is grey literature. By reviewing and analyzing SSTs, librarians and information specialists would be able to determine the extent to which grey literature publications are included in systematic reviews and how they are found, which would help to focus search time and energy. Future research following from this project may include finding a simple way to retrospectively evaluate search strategies, which could help improve future search strategy or search methods development or aid in the creation of a repository where all SSTs could be shared and accessed.

Cooper et al.'s systematic review identified fifty studies of the effectiveness of literature searching, which was a representative sample of the available literature [ 23 ]. SSTs would add to this literature and help move forward the discussion about what constitutes an effective search for a systematic review.

The SST is a simple way to collate the search information generated from a systematic review. Creating and reporting an SST as part of a systematic review would add to the knowledgebase on database selection and supplementary search methods and provide evidence for future searching and search methods research.

SUPPLEMENTAL FILES

• Linguistics
• Composition Studies

Five tips for developing useful literature summary tables for writing review articles

• Evidence-Based Nursing 24(2)

• Memorial University of Newfoundland

• The University of Sheffield

Discover the world's research

• 25+ million members
• 160+ million publication pages
• 2.3+ billion citations
• Heart Lung Circ

• Alyne Santos Borges
• Paola Pugian Jardim

• Pearse McCusker

• Hamzah Shahid Rafiq

• Creativ Nurs

• Richard Booth

• Bethany Jackson
• Esther Weir
• Jonathan Mead

• Maryanna Cruz da Costa e Silva Andrade
• Juliana de Melo Vellozo Pereira Tinoco
• Isabelle Andrade Silveira

• Recruit researchers
• Join for free
• Login Email Tip: Most researchers use their institutional email address as their ResearchGate login Password Forgot password? Keep me logged in Log in or Continue with Google Welcome back! Please log in. Email · Hint Tip: Most researchers use their institutional email address as their ResearchGate login Password Forgot password? Keep me logged in Log in or Continue with Google No account? Sign up

• Privacy Policy

Home » Tables in Research Paper – Types, Creating Guide and Examples

Tables in Research Paper – Types, Creating Guide and Examples

Table of Contents

Tables in Research Paper

Definition:

In Research Papers , Tables are a way of presenting data and information in a structured format. Tables can be used to summarize large amounts of data or to highlight important findings. They are often used in scientific or technical papers to display experimental results, statistical analyses, or other quantitative information.

Importance of Tables in Research Paper

Tables are an important component of a research paper as they provide a clear and concise presentation of data, statistics, and other information that support the research findings . Here are some reasons why tables are important in a research paper:

• Visual Representation : Tables provide a visual representation of data that is easy to understand and interpret. They help readers to quickly grasp the main points of the research findings and draw their own conclusions.
• Organize Data : Tables help to organize large amounts of data in a systematic and structured manner. This makes it easier for readers to identify patterns and trends in the data.
• Clarity and Accuracy : Tables allow researchers to present data in a clear and accurate manner. They can include precise numbers, percentages, and other information that may be difficult to convey in written form.
• Comparison: Tables allow for easy comparison between different data sets or groups. This makes it easier to identify similarities and differences, and to draw meaningful conclusions from the data.
• Efficiency: Tables allow for a more efficient use of space in the research paper. They can convey a large amount of information in a compact and concise format, which saves space and makes the research paper more readable.

Types of Tables in Research Paper

Most common Types of Tables in Research Paper are as follows:

• Descriptive tables : These tables provide a summary of the data collected in the study. They are usually used to present basic descriptive statistics such as means, medians, standard deviations, and frequencies.
• Comparative tables : These tables are used to compare the results of different groups or variables. They may be used to show the differences between two or more groups or to compare the results of different variables.
• Correlation tables: These tables are used to show the relationships between variables. They may show the correlation coefficients between variables, or they may show the results of regression analyses.
• Longitudinal tables : These tables are used to show changes in variables over time. They may show the results of repeated measures analyses or longitudinal regression analyses.
• Qualitative tables: These tables are used to summarize qualitative data such as interview transcripts or open-ended survey responses. They may present themes or categories that emerged from the data.

How to Create Tables in Research Paper

Here are the steps to create tables in a research paper:

• Plan your table: Determine the purpose of the table and the type of information you want to include. Consider the layout and format that will best convey your information.
• Choose a table format : Decide on the type of table you want to create. Common table formats include basic tables, summary tables, comparison tables, and correlation tables.
• Choose a software program : Use a spreadsheet program like Microsoft Excel or Google Sheets to create your table. These programs allow you to easily enter and manipulate data, format the table, and export it for use in your research paper.
• Input data: Enter your data into the spreadsheet program. Make sure to label each row and column clearly.
• Format the table : Apply formatting options such as font, font size, font color, cell borders, and shading to make your table more visually appealing and easier to read.
• Insert the table into your paper: Copy and paste the table into your research paper. Make sure to place the table in the appropriate location and refer to it in the text of your paper.
• Label the table: Give the table a descriptive title that clearly and accurately summarizes the contents of the table. Also, include a number and a caption that explains the table in more detail.
• Check for accuracy: Review the table for accuracy and make any necessary changes before submitting your research paper.

Examples of Tables in Research Paper

Examples of Tables in the Research Paper are as follows:

Table 1: Demographic Characteristics of Study Participants

This table shows the demographic characteristics of 200 participants in a research study. The table includes information about age, gender, and education level. The mean age of the participants was 35.2 years with a standard deviation of 8.6 years, and the age range was between 21 and 57 years. The table also shows that 46% of the participants were male and 54% were female. In terms of education, 10% of the participants had less than a high school education, 30% were high school graduates, 35% had some college education, and 25% had a bachelor’s degree or higher.

Table 2: Summary of Key Findings

This table summarizes the key findings of a study comparing three different groups on a particular variable. The table shows the mean score, standard deviation, t-value, and p-value for each group. The asterisk next to the t-value for Group 1 indicates that the difference between Group 1 and the other groups was statistically significant at p < 0.01, while the differences between Group 2 and Group 3 were not statistically significant.

Purpose of Tables in Research Paper

The primary purposes of including tables in a research paper are:

• To present data: Tables are an effective way to present large amounts of data in a clear and organized manner. Researchers can use tables to present numerical data, survey results, or other types of data that are difficult to represent in text.
• To summarize data: Tables can be used to summarize large amounts of data into a concise and easy-to-read format. Researchers can use tables to summarize the key findings of their research, such as descriptive statistics or the results of regression analyses.
• To compare data : Tables can be used to compare data across different variables or groups. Researchers can use tables to compare the characteristics of different study populations or to compare the results of different studies on the same topic.
• To enhance the readability of the paper: Tables can help to break up long sections of text and make the paper more visually appealing. By presenting data in a table, researchers can help readers to quickly identify the most important information and understand the key findings of the study.

Advantages of Tables in Research Paper

Some of the advantages of using tables in research papers include:

• Clarity : Tables can present data in a way that is easy to read and understand. They can help readers to quickly and easily identify patterns, trends, and relationships in the data.
• Efficiency: Tables can save space and reduce the need for lengthy explanations or descriptions of the data in the main body of the paper. This can make the paper more concise and easier to read.
• Organization: Tables can help to organize large amounts of data in a logical and meaningful way. This can help to reduce confusion and make it easier for readers to navigate the data.
• Comparison : Tables can be useful for comparing data across different groups, variables, or time periods. This can help to highlight similarities, differences, and changes over time.
• Visualization : Tables can also be used to visually represent data, making it easier for readers to see patterns and trends. This can be particularly useful when the data is complex or difficult to understand.

About the author

Muhammad Hassan

Researcher, Academic Writer, Web developer

You may also like

Dissertation Methodology – Structure, Example...

Informed Consent in Research – Types, Templates...

Research Project – Definition, Writing Guide and...

Research Methodology – Types, Examples and...

Research Contribution – Thesis Guide

APA Table of Contents – Format and Example

• Resources Home 🏠
• Try SciSpace Copilot
• Search research papers
• Add Copilot Extension
• Try AI Detector
• Try Paraphraser
• Try Citation Generator
• April Papers
• June Papers
• July Papers

How To Write A Research Summary

It’s a common perception that writing a research summary is a quick and easy task. After all, how hard can jotting down 300 words be? But when you consider the weight those 300 words carry, writing a research summary as a part of your dissertation, essay or compelling draft for your paper instantly becomes daunting task.

A research summary requires you to synthesize a complex research paper into an informative, self-explanatory snapshot. It needs to portray what your article contains. Thus, writing it often comes at the end of the task list.

Regardless of when you’re planning to write, it is no less of a challenge, particularly if you’re doing it for the first time. This blog will take you through everything you need to know about research summary so that you have an easier time with it.

What is a Research Summary?

A research summary is the part of your research paper that describes its findings to the audience in a brief yet concise manner. A well-curated research summary represents you and your knowledge about the information written in the research paper.

While writing a quality research summary, you need to discover and identify the significant points in the research and condense it in a more straightforward form. A research summary is like a doorway that provides access to the structure of a research paper's sections.

Since the purpose of a summary is to give an overview of the topic, methodology, and conclusions employed in a paper, it requires an objective approach. No analysis or criticism.

Research summary or Abstract. What’s the Difference?

They’re both brief, concise, and give an overview of an aspect of the research paper. So, it’s easy to understand why many new researchers get the two confused. However, a research summary and abstract are two very different things with individual purpose. To start with, a research summary is written at the end while the abstract comes at the beginning of a research paper.

A research summary captures the essence of the paper at the end of your document. It focuses on your topic, methods, and findings. More like a TL;DR, if you will. An abstract, on the other hand, is a description of what your research paper is about. It tells your reader what your topic or hypothesis is, and sets a context around why you have embarked on your research.

Getting Started with a Research Summary

Before you start writing, you need to get insights into your research’s content, style, and organization. There are three fundamental areas of a research summary that you should focus on.

• While deciding the contents of your research summary, you must include a section on its importance as a whole, the techniques, and the tools that were used to formulate the conclusion. Additionally, there needs to be a short but thorough explanation of how the findings of the research paper have a significance.
• To keep the summary well-organized, try to cover the various sections of the research paper in separate paragraphs. Besides, how the idea of particular factual research came up first must be explained in a separate paragraph.
• As a general practice worldwide, research summaries are restricted to 300-400 words. However, if you have chosen a lengthy research paper, try not to exceed the word limit of 10% of the entire research paper.

How to Structure Your Research Summary

The research summary is nothing but a concise form of the entire research paper. Therefore, the structure of a summary stays the same as the paper. So, include all the section titles and write a little about them. The structural elements that a research summary must consist of are:

It represents the topic of the research. Try to phrase it so that it includes the key findings or conclusion of the task.

The abstract gives a context of the research paper. Unlike the abstract at the beginning of a paper, the abstract here, should be very short since you’ll be working with a limited word count.

Introduction

This is the most crucial section of a research summary as it helps readers get familiarized with the topic. You should include the definition of your topic, the current state of the investigation, and practical relevance in this part. Additionally, you should present the problem statement, investigative measures, and any hypothesis in this section.

Methodology

This section provides details about the methodology and the methods adopted to conduct the study. You should write a brief description of the surveys, sampling, type of experiments, statistical analysis, and the rationality behind choosing those particular methods.

Create a list of evidence obtained from the various experiments with a primary analysis, conclusions, and interpretations made upon that. In the paper research paper, you will find the results section as the most detailed and lengthy part. Therefore, you must pick up the key elements and wisely decide which elements are worth including and which are worth skipping.

This is where you present the interpretation of results in the context of their application. Discussion usually covers results, inferences, and theoretical models explaining the obtained values, key strengths, and limitations. All of these are vital elements that you must include in the summary.

Most research papers merge conclusion with discussions. However, depending upon the instructions, you may have to prepare this as a separate section in your research summary. Usually, conclusion revisits the hypothesis and provides the details about the validation or denial about the arguments made in the research paper, based upon how convincing the results were obtained.

The structure of a research summary closely resembles the anatomy of a scholarly article . Additionally, you should keep your research and references limited to authentic and  scholarly sources only.

Tips for Writing a Research Summary

The core concept behind undertaking a research summary is to present a simple and clear understanding of your research paper to the reader. The biggest hurdle while doing that is the number of words you have at your disposal. So, follow the steps below to write a research summary that sticks.

1. Read the parent paper thoroughly

You should go through the research paper thoroughly multiple times to ensure that you have a complete understanding of its contents. A 3-stage reading process helps.

a. Scan: In the first read, go through it to get an understanding of its basic concept and methodologies.

b. Read: For the second step, read the article attentively by going through each section, highlighting the key elements, and subsequently listing the topics that you will include in your research summary.

c. Skim: Flip through the article a few more times to study the interpretation of various experimental results, statistical analysis, and application in different contexts.

Sincerely go through different headings and subheadings as it will allow you to understand the underlying concept of each section. You can try reading the introduction and conclusion simultaneously to understand the motive of the task and how obtained results stay fit to the expected outcome.

2. Identify the key elements in different sections

While exploring different sections of an article, you can try finding answers to simple what, why, and how. Below are a few pointers to give you an idea:

• What is the research question and how is it addressed?
• Is there a hypothesis in the introductory part?
• What type of methods are being adopted?
• What is the sample size for data collection and how is it being analyzed?
• What are the most vital findings?
• Do the results support the hypothesis?

Discussion/Conclusion

• What is the final solution to the problem statement?
• What is the explanation for the obtained results?
• What is the drawn inference?
• What are the various limitations of the study?

3. Prepare the first draft

Now that you’ve listed the key points that the paper tries to demonstrate, you can start writing the summary following the standard structure of a research summary. Just make sure you’re not writing statements from the parent research paper verbatim.

Instead, try writing down each section in your own words. This will not only help in avoiding plagiarism but will also show your complete understanding of the subject. Alternatively, you can use a summarizing tool (AI-based summary generators) to shorten the content or summarize the content without disrupting the actual meaning of the article.

SciSpace Copilot is one such helpful feature! You can easily upload your research paper and ask Copilot to summarize it. You will get an AI-generated, condensed research summary. SciSpace Copilot also enables you to highlight text, clip math and tables, and ask any question relevant to the research paper; it will give you instant answers with deeper context of the article..

4. Include visuals

One of the best ways to summarize and consolidate a research paper is to provide visuals like graphs, charts, pie diagrams, etc.. Visuals make getting across the facts, the past trends, and the probabilistic figures around a concept much more engaging.

5. Double check for plagiarism

It can be very tempting to copy-paste a few statements or the entire paragraphs depending upon the clarity of those sections. But it’s best to stay away from the practice. Even paraphrasing should be done with utmost care and attention.

Also: QuillBot vs SciSpace: Choose the best AI-paraphrasing tool

6. Religiously follow the word count limit

You need to have strict control while writing different sections of a research summary. In many cases, it has been observed that the research summary and the parent research paper become the same length. If that happens, it can lead to discrediting of your efforts and research summary itself. Whatever the standard word limit has been imposed, you must observe that carefully.

7. Proofread your research summary multiple times

The process of writing the research summary can be exhausting and tiring. However, you shouldn’t allow this to become a reason to skip checking your academic writing several times for mistakes like misspellings, grammar, wordiness, and formatting issues. Proofread and edit until you think your research summary can stand out from the others, provided it is drafted perfectly on both technicality and comprehension parameters. You can also seek assistance from editing and proofreading services , and other free tools that help you keep these annoying grammatical errors at bay.

8. Watch while you write

Keep a keen observation of your writing style. You should use the words very precisely, and in any situation, it should not represent your personal opinions on the topic. You should write the entire research summary in utmost impersonal, precise, factually correct, and evidence-based writing.

9. Ask a friend/colleague to help

Once you are done with the final copy of your research summary, you must ask a friend or colleague to read it. You must test whether your friend or colleague could grasp everything without referring to the parent paper. This will help you in ensuring the clarity of the article.

Once you become familiar with the research paper summary concept and understand how to apply the tips discussed above in your current task, summarizing a research summary won’t be that challenging. While traversing the different stages of your academic career, you will face different scenarios where you may have to create several research summaries.

In such cases, you just need to look for answers to simple questions like “Why this study is necessary,” “what were the methods,” “who were the participants,” “what conclusions were drawn from the research,” and “how it is relevant to the wider world.” Once you find out the answers to these questions, you can easily create a good research summary following the standard structure and a precise writing style.

You might also like

Consensus GPT vs. SciSpace GPT: Choose the Best GPT for Research

Literature Review and Theoretical Framework: Understanding the Differences

Types of Essays in Academic Writing - Quick Guide (2024)

Have a language expert improve your writing

Run a free plagiarism check in 10 minutes, generate accurate citations for free.

• Knowledge Base
• Working with sources
• How to Write a Summary | Guide & Examples

How to Write a Summary | Guide & Examples

Published on November 23, 2020 by Shona McCombes . Revised on May 31, 2023.

Summarizing , or writing a summary, means giving a concise overview of a text’s main points in your own words. A summary is always much shorter than the original text.

There are five key steps that can help you to write a summary:

• Read the text
• Break it down into sections
• Identify the key points in each section
• Write the summary
• Check the summary against the article

Writing a summary does not involve critiquing or evaluating the source . You should simply provide an accurate account of the most important information and ideas (without copying any text from the original).

Table of contents

When to write a summary, step 1: read the text, step 2: break the text down into sections, step 3: identify the key points in each section, step 4: write the summary, step 5: check the summary against the article, other interesting articles, frequently asked questions about summarizing.

There are many situations in which you might have to summarize an article or other source:

• As a stand-alone assignment to show you’ve understood the material
• To keep notes that will help you remember what you’ve read
• To give an overview of other researchers’ work in a literature review

When you’re writing an academic text like an essay , research paper , or dissertation , you’ll integrate sources in a variety of ways. You might use a brief quote to support your point, or paraphrase a few sentences or paragraphs.

But it’s often appropriate to summarize a whole article or chapter if it is especially relevant to your own research, or to provide an overview of a source before you analyze or critique it.

In any case, the goal of summarizing is to give your reader a clear understanding of the original source. Follow the five steps outlined below to write a good summary.

Prevent plagiarism. Run a free check.

You should read the article more than once to make sure you’ve thoroughly understood it. It’s often effective to read in three stages:

• Scan the article quickly to get a sense of its topic and overall shape.
• Read the article carefully, highlighting important points and taking notes as you read.
• Skim the article again to confirm you’ve understood the key points, and reread any particularly important or difficult passages.

There are some tricks you can use to identify the key points as you read:

• Start by reading the abstract . This already contains the author’s own summary of their work, and it tells you what to expect from the article.
• Pay attention to headings and subheadings . These should give you a good sense of what each part is about.
• Read the introduction and the conclusion together and compare them: What did the author set out to do, and what was the outcome?

To make the text more manageable and understand its sub-points, break it down into smaller sections.

If the text is a scientific paper that follows a standard empirical structure, it is probably already organized into clearly marked sections, usually including an introduction , methods , results , and discussion .

Other types of articles may not be explicitly divided into sections. But most articles and essays will be structured around a series of sub-points or themes.

Now it’s time go through each section and pick out its most important points. What does your reader need to know to understand the overall argument or conclusion of the article?

Keep in mind that a summary does not involve paraphrasing every single paragraph of the article. Your goal is to extract the essential points, leaving out anything that can be considered background information or supplementary detail.

In a scientific article, there are some easy questions you can ask to identify the key points in each part.

If the article takes a different form, you might have to think more carefully about what points are most important for the reader to understand its argument.

In that case, pay particular attention to the thesis statement —the central claim that the author wants us to accept, which usually appears in the introduction—and the topic sentences that signal the main idea of each paragraph.

Now that you know the key points that the article aims to communicate, you need to put them in your own words.

To avoid plagiarism and show you’ve understood the article, it’s essential to properly paraphrase the author’s ideas. Do not copy and paste parts of the article, not even just a sentence or two.

The best way to do this is to put the article aside and write out your own understanding of the author’s key points.

Examples of article summaries

Let’s take a look at an example. Below, we summarize this article , which scientifically investigates the old saying “an apple a day keeps the doctor away.”

Davis et al. (2015) set out to empirically test the popular saying “an apple a day keeps the doctor away.” Apples are often used to represent a healthy lifestyle, and research has shown their nutritional properties could be beneficial for various aspects of health. The authors’ unique approach is to take the saying literally and ask: do people who eat apples use healthcare services less frequently? If there is indeed such a relationship, they suggest, promoting apple consumption could help reduce healthcare costs.

The study used publicly available cross-sectional data from the National Health and Nutrition Examination Survey. Participants were categorized as either apple eaters or non-apple eaters based on their self-reported apple consumption in an average 24-hour period. They were also categorized as either avoiding or not avoiding the use of healthcare services in the past year. The data was statistically analyzed to test whether there was an association between apple consumption and several dependent variables: physician visits, hospital stays, use of mental health services, and use of prescription medication.

Although apple eaters were slightly more likely to have avoided physician visits, this relationship was not statistically significant after adjusting for various relevant factors. No association was found between apple consumption and hospital stays or mental health service use. However, apple eaters were found to be slightly more likely to have avoided using prescription medication. Based on these results, the authors conclude that an apple a day does not keep the doctor away, but it may keep the pharmacist away. They suggest that this finding could have implications for reducing healthcare costs, considering the high annual costs of prescription medication and the inexpensiveness of apples.

However, the authors also note several limitations of the study: most importantly, that apple eaters are likely to differ from non-apple eaters in ways that may have confounded the results (for example, apple eaters may be more likely to be health-conscious). To establish any causal relationship between apple consumption and avoidance of medication, they recommend experimental research.

An article summary like the above would be appropriate for a stand-alone summary assignment. However, you’ll often want to give an even more concise summary of an article.

For example, in a literature review or meta analysis you may want to briefly summarize this study as part of a wider discussion of various sources. In this case, we can boil our summary down even further to include only the most relevant information.

Using national survey data, Davis et al. (2015) tested the assertion that “an apple a day keeps the doctor away” and did not find statistically significant evidence to support this hypothesis. While people who consumed apples were slightly less likely to use prescription medications, the study was unable to demonstrate a causal relationship between these variables.

Citing the source you’re summarizing

When including a summary as part of a larger text, it’s essential to properly cite the source you’re summarizing. The exact format depends on your citation style , but it usually includes an in-text citation and a full reference at the end of your paper.

You can easily create your citations and references in APA or MLA using our free citation generators.

APA Citation Generator MLA Citation Generator

Finally, read through the article once more to ensure that:

• You’ve accurately represented the author’s work
• You haven’t missed any essential information
• The phrasing is not too similar to any sentences in the original.

If you’re summarizing many articles as part of your own work, it may be a good idea to use a plagiarism checker to double-check that your text is completely original and properly cited. Just be sure to use one that’s safe and reliable.

If you want to know more about ChatGPT, AI tools , citation , and plagiarism , make sure to check out some of our other articles with explanations and examples.

• ChatGPT vs human editor
• ChatGPT citations
• Is ChatGPT trustworthy?
• Using ChatGPT for your studies
• What is ChatGPT?
• Chicago style
• Paraphrasing

 Plagiarism

• Types of plagiarism
• Self-plagiarism
• Avoiding plagiarism
• Academic integrity
• Consequences of plagiarism
• Common knowledge

A summary is a short overview of the main points of an article or other source, written entirely in your own words. Want to make your life super easy? Try our free text summarizer today!

A summary is always much shorter than the original text. The length of a summary can range from just a few sentences to several paragraphs; it depends on the length of the article you’re summarizing, and on the purpose of the summary.

You might have to write a summary of a source:

• As a stand-alone assignment to prove you understand the material
• For your own use, to keep notes on your reading
• To provide an overview of other researchers’ work in a literature review
• In a paper , to summarize or introduce a relevant study

To avoid plagiarism when summarizing an article or other source, follow these two rules:

• Write the summary entirely in your own words by paraphrasing the author’s ideas.
• Cite the source with an in-text citation and a full reference so your reader can easily find the original text.

An abstract concisely explains all the key points of an academic text such as a thesis , dissertation or journal article. It should summarize the whole text, not just introduce it.

An abstract is a type of summary , but summaries are also written elsewhere in academic writing . For example, you might summarize a source in a paper , in a literature review , or as a standalone assignment.

All can be done within seconds with our free text summarizer .

Cite this Scribbr article

If you want to cite this source, you can copy and paste the citation or click the “Cite this Scribbr article” button to automatically add the citation to our free Citation Generator.

McCombes, S. (2023, May 31). How to Write a Summary | Guide & Examples. Scribbr. Retrieved August 13, 2024, from https://www.scribbr.com/working-with-sources/how-to-summarize/

Is this article helpful?

Shona McCombes

Other students also liked, how to paraphrase | step-by-step guide & examples, how to quote | citing quotes in apa, mla & chicago, the basics of in-text citation | apa & mla examples, "i thought ai proofreading was useless but..".

I've been using Scribbr for years now and I know it's a service that won't disappoint. It does a good job spotting mistakes”

Want to create or adapt books like this? Learn more about how Pressbooks supports open publishing practices.

Part 1: Introduction to research

5. Writing your literature review

Chapter outline.

• Reading results (16 minute read)
• Synthesizing information (16 minute read)
• Writing a literature review (18 minute read)

Content warning: examples in this chapter contain references to domestic violence and details on types of abuse, drug use, poverty, mental health, sexual harassment and details on harassing behaviors, children’s mental health, LGBTQ+ oppression and suicide, obesity, anti-poverty stigma, and psychotic disorders.

5.1 Reading results

Learning objectives.

Learners will be able to…

• Describe how statistical significance and confidence intervals demonstrate which results are most important
• Differentiate between qualitative and quantitative results in an empirical journal article

If you recall from section 3.1 , empirical journal articles are those that report the results of quantitative or qualitative data analyzed by the author. They follow a set structure—introduction, methods, results, discussion/conclusions. This section is about reading the most challenging section: results.

Read beyond the abstract

At this point, I have read hundreds of literature reviews written by students. One of the challenges I have noted is that students will report the results as summarized in the abstract, rather than the detailed findings laid out in the results section of the article. This poses a problem when you are writing a literature review because you need to provide specific and clear facts that support your reading of the literature. The abstract may say something like: “we found that poverty is associated with mental health status.” For your literature review, you want the details, not the summary. In the results section of the article, you may find a sentence that states: “children living in households experiencing poverty are three times more likely to have a mental health diagnosis.” This more specific statistical information provides a stronger basis on which to build the arguments in your literature review.

Using the summarized results in an abstract is an understandable mistake to make. The results section often contains figures and tables that may be challenging to understand. Often, without having completed more advanced coursework on statistical or qualitative analysis, some of the terminology, symbols, or diagrams may be difficult to comprehend. This section is all about how to read and interpret the results of an empirical (quantitative or qualitative) journal article. Our discussion here will be basic, and in parts three and four of the textbook, you will learn more about how to interpret results from statistical tests and qualitative data analysis.

Remember, this section only addresses empirical articles. Non-empirical articles (e.g., theoretical articles, literature reviews) don’t have results. They cite the analysis of raw data completed by other authors, not the person writing the journal article who is merely summarizing others’ work.

Quantitative results

Quantitative articles often contain tables, and scanning them is a good way to begin reading the results. A table usually provides a quick, condensed summary of the report’s key findings. Tables are a concise way to report large amounts of data. Some tables present descriptive information about a researcher’s sample (often the first table in a results section). These tables will likely contain frequencies (N) and percentages (%). For example, if gender happened to be an important variable for the researcher’s analysis, a descriptive table would show how many and what percent of all study participants are of a particular gender. Frequencies or “how many” will probably be listed as N, while the percent symbol (%) might be used to indicate percentages.

In a table presenting a causal relationship, two sets of variables are represented. The independent variable , or cause, and the dependent variable , the effect. We’ll go into more detail on variables in Chapter 8 . Independent variable attributes are typically presented in the table’s columns, while dependent variable attributes are presented in rows. This allows the reader to scan a table’s rows to see how values on the dependent variable change as the independent variable values change. Tables displaying results of quantitative analysis will also likely include some information about which relationships are significant or not. We will discuss the details of significance and p-values later in this section.

Let’s look at a specific example: Table 5.1. It presents the causal relationship between gender and experiencing harassing behaviors at work. In this example, gender is the independent variable (the cause) and the harassing behaviors listed are the dependent variables (the effects). [1] Therefore, we place gender in the table’s columns and harassing behaviors in the table’s rows.

Reading across the table’s top row, we see that 2.9% of women in the sample reported experiencing subtle or obvious threats to their safety at work, while 4.7% of men in the sample reported the same. We can read across each of the rows of the table in this way. Reading across the bottom row, we see that 9.4% of women in the sample reported experiencing staring or invasion of their personal space at work while just 2.3% of men in the sample reported having the same experience. We’ll discuss  p values later in this section.

While you can certainly scan tables for key results, they are often difficult to understand without reading the text of the article. The article and table were meant to complement each other, and the text should provide information on how the authors interpret their findings. The table is not redundant with the text of the results section. Additionally, the first table in most results sections is a summary of the study’s sample, which provides more background information on the study than information about hypotheses and findings. It is also a good idea to look back at the methods section of the article as the data analysis plan the authors outline should walk you through the steps they took to analyze their data which will inform how they report them in the results section.

Statistical significance

The statistics reported in Table 5.1 represent what the researchers found in their sample. The purpose of statistical analysis is usually to generalize from a the small number of people in a study’s sample to a larger population of people. Thus, the researchers intend to make causal arguments about harassing behaviors at workplaces beyond those covered in the sample.

Generalizing is key to understanding statistical significance . According to Cassidy and colleagues, (2019) [2] 89% of research methods textbooks in psychology define statistical significance incorrectly. This includes an early draft of this textbook which defined statistical significance as “the likelihood that the relationships we observe could be caused by something other than chance.” If you have previously had a research methods class, this might sound familiar to you. It certainly did to me!

But statistical significance is less about “random chance” than more about the null hypothesis . Basically, at the beginning of a study a researcher develops a hypothesis about what they expect to find, usually that there is a statistical relationship between two or more variables . The null hypothesis is the opposite. It is the hypothesis that there is no relationship between the variables in a research study. Researchers then can hopefully reject the null hypothesis because they find a relationship between the variables.

For example, in Table 5.1 researchers were examining whether gender impacts harassment. Of course, researchers assumed that women were more likely to experience harassment than men. The null hypothesis, then, would be that gender has no impact on harassment. Once we conduct the study, our results will hopefully lead us to reject the null hypothesis because we find that gender impacts harassment. We would then generalize from our study’s sample to the larger population of people in the workplace.

Statistical significance is calculated using a p-value which is obtained by comparing the statistical results with a hypothetical set of results if the researchers re-ran their study a large number of times. Keeping with our example, imagine we re-ran our study with different men and women from different workplaces hundreds and hundred of times and we assume that the null hypothesis is true that gender has no impact on harassment. If results like ours come up pretty often when the null hypothesis is true, our results probably don’t mean much. “The smaller the p-value, the greater the statistical incompatibility with the null hypothesis” (Wasserstein & Lazar, 2016, p. 131). [3] Generally, researchers in the social sciences have used 0.05 as the value at which a result is significant (p is less than 0.05) or not significant (p is greater than 0.05). The p-value 0.05 refers to if 5% of those hypothetical results from re-running our study show the same or more extreme relationships when the null hypothesis is true. Researchers, however, may choose a stricter standard such as 0.01 in which only 1% of those hypothetical results are more extreme or a more lenient standard like 0.1 in which 10% of those hypothetical results are more extreme than what was found in the study.

Let’s look back at Table 5.1. Which one of the relationships between gender and harassing behaviors is statistically significant? It’s the last one in the table, “staring or invasion of personal space,” whose p-value is 0.039 (under the p<0.05 standard to establish statistical significance). Again, this indicates that if we re-ran our study over and over again and gender did not  impact staring/invasion of space (i.e., the null hypothesis was true), only 3.9% of the time would we find similar or more extreme differences between men and women than what we observed in our study. Thus, we conclude that for staring or invasion of space only , there is a statistically significant relationship.

For contrast, let’s look at “being pushed, hit, or grabbed” and run through the same analysis to see if it is statistically significant. If we re-ran our study over and over again and the null hypothesis was true, 48% of the time (p=.48) we would find similar or more extreme differences between men and women. That means these results are not statistically significant.

This discussion should also highlight a point we discussed previously: that it is important to read the full results section, rather than simply relying on the summary in the abstract. If the abstract stated that most tests revealed no statistically significant relationships between gender and harassment, you would have missed the detail on which behaviors were and were not associated with gender. Read the full results section! And don’t be afraid to ask for help from a professor in understanding what you are reading, as results sections are often not written to be easily understood.

Statistical significance and p-values have been critiqued recently for a number of reasons, including that they are misused and misinterpreted (Wasserstein & Lazar, 2016) [4] , that researchers deliberately manipulate their analyses to have significant results (Head et al., 2015) [5] , and factor into the difficulty scientists have today in reproducing many of the results of previous social science studies (Peng, 2015). [6] For this reason, we share these principles, adapted from those put forth by the American Statistical Association, [7]  for understanding and using p-values in social science:

• P-values provide evidence against a null hypothesis.
• P-values do not indicate whether the results were produced by random chance alone or if the researcher’s hypothesis is true, though both are common misconceptions.
• Statistical significance can be detected in minuscule differences that have very little effect on the real world.
• Nuance is needed to interpret scientific findings, as a conclusion does not become true or false when the p-value passes from p=0.051 to p=0.049.
• Real-world decision-making must use more than reported p-values. It’s easy to run analyses of large datasets and only report the significant findings.
• Greater confidence can be placed in studies that pre-register their hypotheses and share their data and methods openly with the public.
• “By itself, a p-value does not provide a good measure of evidence regarding a model or hypothesis. For example, a p-value near 0.05 taken by itself offers only weak evidence against the null hypothesis. Likewise, a relatively large p-value does not imply evidence in favor of the null hypothesis; many other hypotheses may be equally or more consistent with the observed data” (Wasserstein & Lazar, 2016, p. 132).

Confidence intervals

Because of the limitations of p-values, scientists can use other methods to determine whether their models of the world are true. One common approach is to use a confidence interval , or a range of values in which the true value is likely to be found. Confidence intervals are helpful because, as principal #5 above points out, p-values do not measure the size of an effect (Greenland et al., 2016). [8] Remember, something that has very little impact on the world can be statistically significant, and the values in a confidence interval would be helpful. In our example from Table 5.1, imagine our analysis produced a confidence interval that women are 1.2-3.4x more likely to experience “staring or invasion of personal space” than men. As with p-values, calculation for a confidence interval compares what was found in one study with a hypothetical set of results if we repeated the study over and over again. If we calculated 95% confidence intervals for all of the hypothetical set of hundreds and hundreds of studies, that would be our confidence interval. 

Confidence intervals are pretty intuitive. As of this writing, my wife and are expecting our second child. The doctor told us our due date was December 11th. But the doctor also told us that December 11th was only their best estimate. They were actually 95% sure our baby might be born any time in the 30-day period between November 27th and December 25th. Confidence intervals are often listed with a percentage, like 90% or 95%, and a range of values, such as between November 27th and December 25th. You can read that as: “we are 95% sure your baby will be born between November 27th and December 25th because we’ve studied hundreds of thousands of fetuses and mothers, and we’re 95% sure your baby will be within these two dates.”

Notice that we’re hedging our bets here by using words like “best estimate.” When testing hypotheses, social scientists generally phrase their findings in a tentative way, talking about what results “indicate” or “support,” rather than making bold statements about what their results “prove.” Social scientists have humility because they understand the limitations of their knowledge. In a literature review, using a single study or fact to “prove” an argument right or wrong is often a signal to the person reading your literature review (usually your professor) that you may not have appreciated the limitations of that study or its place in the broader literature on the topic. Strong arguments in a literature review include multiple facts and ideas that span across multiple studies.

You can learn more about creating tables, reading tables, and tests of statistical significance in a class focused exclusively on statistical analysis. We provide links to many free and openly licensed resources on statistics in Chapter 16 . For now, we hope this brief introduction to reading tables will improve your confidence in reading and understanding the results sections in quantitative empirical articles.

Qualitative results

Quantitative articles will contain a lot of numbers and the results of statistical tests demonstrating associations between those numbers. Qualitative articles, on the other hand, will consist mostly of quotations from participants. For most qualitative articles, the authors want to put their results in the words of their participants, as they are the experts. Articles that lack quotations make it difficult to assess whether the researcher interpreted the data in a trustworthy, unbiased manner. These types of articles may also indicate how often particular themes or ideas came up in the data, potentially reflective of how important they were to participants.

Authors often organize qualitative results by themes and subthemes. For example, see this snippet from the results section in Bonanno and Veselak (2019) [9] discussion parents’ attitudes towards child mental health information sources.

Data analysis revealed four themes related to participants’ abilities to access mental health help and information for their children, and parents’ levels of trust in these sources. These themes are: others’ firsthand experiences family and friends with professional experience, protecting privacy, and uncertainty about schools as information sources. Trust emerged as an overarching and unifying concept for all of these themes. Others’ firsthand experiences. Several participants reported seeking information from other parents who had experienced mental health struggles similar to their own children. They often referenced friends or family members who had been or would be good sources of information due to their own personal experiences. The following quote from Adrienne demonstrates the importance of firsthand experience: [I would only feel comfortable sharing concerns or asking for advice] if I knew that they had been in the same situation. (Adrienne) Similarly, Michelle said: And I talked to a friend of mine who has kids who have IEPs in the district to see, kind of, how did she go about it. (Michelle) … Friends/family with professional experience . Several respondents referred to friends or family members who had professional experience with or knowledge of child mental health and suggested that these individuals would be good sources of information. For example, Hannah said: Well, what happened with me was I have an uncle who’s a psychiatrist. Sometimes if he’s up in (a city to the north), he’s retired, I can call him sometimes and get information. (Hannah) Michelle, who was in nursing school, echoed this sentiment: At this point, [if my child’s behavioral difficulties continued], I would probably call one of my [nursing] professors. That’s what I’ve done in the past when I’ve needed help with certain things…I have a professor who I would probably consider a friend who I would probably talk to first. She has a big adolescent practice. (Michelle) (p. 402-403)

The terms in bold above refer to the key themes (i.e., qualitative results) that were present in the data. Researchers will state the process by which they interpret each theme, providing a definition and usually some quotations from research participants. Researchers will also draw connections between themes, note consensus or conflict over themes, and situate the themes within the study context.

Qualitative results are specific to the time, place, and culture in which they arise, so you will have to use your best judgment to determine whether these results are relevant to your study. For example, students in my class at Radford University in Southwest Virginia may be studying rural populations. Would a study on group homes in a large urban city transfer well to group homes in a rural area?

Maybe. But even if you were using data from a qualitative study in another rural area, are all rural areas the same? How is the client population and sociocultural context in the article similar or different to the one in your study? Qualitative studies have tremendous depth, but researchers must be intentional about drawing conclusions about one context based on a study in another context.

Key Takeaways

• The results section of empirical articles are often the most difficult to understand.
• To understand a quantitative results section, look for results that were statistically significant and examine the confidence interval, if provided.
• To understand a qualitative results section, look for definitions of themes or codes and use the quotations provided to understand the participants’ perspective.

Select a quantitative empirical article related to your topic.

• Write down the results the authors identify as statistically significant in the results section.
• How do the authors interpret their results in the discussion section?
• Do the authors provide enough information in the introduction for you to understand their results?

Select a qualitative empirical article relevant to your topic.

• Write down the key themes the authors identify and how they were defined by the participants.

5.2 Organizing information

• Describe how to use summary tables to organize information from empirical articles
• Describe how to use topical outlines to organize information from the literature reviews of articles you read
• Create a concept map that visualizes the key concepts and relationships relevant to your working question
• Use what you learn in the literature search to revise your working question

This section will introduce you to three tools scholars use to organize and synthesize (i.e., weave together) information from multiple sources. First, we will discuss how to build a summary table containing information from empirical articles that are highly relevant—from literature review, to methods and results—to your entire research proposal. These are articles you will need to know the details of back-to-front because they are so highly related to your proposed study.

Second, we’ll discuss what to do with the other articles you’ve downloaded. As we’ve discussed previously, you’re not going to read most of the sources you download from start-to-finish. Instead, you’ll look at the author’s literature review, key ideas, and skim for any relevant passages for your project. As you do so, you should create a topical outline that organizes all relevant facts you might use in your literature that you’ve collected from the abstract, literature review, and conclusion of the articles you’ve found. Of course, it is important to note the original source of the information you are citing.

Finally, we will revisit concept mapping as a technique for visualizing the concepts in your study. Altogether, these techniques should help you create intermediary products—documents you are not likely to show to anyone or turn in for a grade—but that are vital steps to a final research proposal.

Organizing empirical articles using a summary table

Your research proposal is an empirical project. You will collect raw data and analyze it to answer your question. Over the next few weeks, identify about 10 articles that are empirically similar to the study you want to conduct. If you plan on conducting surveys of practitioners, it’s a good idea for you to read in detail other articles that have used similar methods (sampling, measures, data analysis) and asked similar questions to your proposal. A summary table can help you organize these Top 10 articles: empirical articles that are highly relevant to your proposal and working question.

Using the annotations in Section 4.2 as a guide, create a spreadsheet or Word table with your annotation categories as columns and each source as new row. For example, I was searching for articles on using a specific educational technique in the literature. I wanted to know whether other researchers found positive results, how big their samples were, and whether they were conducted at a single school or across multiple schools. I looked through each empirical article on the topic and filled in a summary table. At the end, I could do an easy visual analysis and state that most studies revealed no significant results and that there were few multi-site studies. These arguments were then included in my literature review. These tables are similar to those you will find in a systematic review article.

A basic summary table is provided in Figure 5.1. A more detailed example is available from Elaine Gregersen’s blog , and you can download an Excel template from Raul Pacheco-Vega’s blog . Remember, although “the means of summarizing can vary, the key at this point is to make sure you understand what you’ve found and how it relates to your topic and research question” (Bennard et al., 2014, para. 10). [10] As you revisit and revise your working question over the next few weeks, think about those sources that are so relevant you need to understand every detail about them.

A good summary table will also ensure that when you cite these articles in your literature review, you are able to provide the necessary detail and context for readers to properly understand the results. For example, one of the common errors I see in student literature reviews is using a small, exploratory study to represent the truth about a larger population. You will also notice important differences in how variables are measured or how people are sampled, for instance, and these details are often the source of a good critical review of the literature.

• Using your folder of article PDFs from you’ve downloaded in previous exercises, identify which articles are likely to be most relevant to your proposed study. This may change as you revise your working question and study design over the next few weeks. Create a list of 10 articles that are highly relevant to the extent that you will need to remember key details from each section of the article.
• Create a spreadsheet for your summary table and save it in your project folder on your hard drive. Using one of the templates linked in this chapter, fill in the columns of your spreadsheet. Enter the information from one of the articles you’ve read so far. As you finalize your research question over the next few weeks, fill in your summary table with the 5 most relevant empirical articles on your topic.

Synthesizing facts using a topical outline

If we’re only reading 10 articles in detail, what do we do with the others? Raul Pacheco-Vega recommends using the AIC approach : read the abstract, introduction, and conclusion (and the discussion section, in empirical articles). For non-empirical articles, it’s a little less clear but the first few pages and last few pages of an article usually contain the author’s reading of the relevant literature and their principal conclusions. You may also want to skim the first and last sentence of each paragraph. Only read paragraphs in which you are likely to find information relevant to your working question. Skimming like this gives you the general point of the article, though you should read in detail the most valuable resource of all—another author’s literature review.

It’s impossible to read all of the literature about your topic. You will read about 10 articles in detail. For a few dozen more (there is no magic number), you will read the abstract, introduction, and conclusion, skim the rest of the article, but ultimately never read everything. Make the most out of the articles you do read by extracting as many facts as possible from each. You are starting your research project without a lot of knowledge of the topic you want to study, and by using the literature reviews provided in academic journal articles, you can gain a lot of knowledge about a topic in a short period of time. This way, by reading only a small number of articles, you are also reading their citations and synthesis of dozens of other articles as well.

As you read an article in detail, we suggest copying any facts you find relevant in a separate word processing document. Another idea is to copy anything you’ve annotated as background information in Section 4.2 into an outline. Copying and pasting from PDF to Word can be difficult because PDFs are image files, not documents. To make that easier, use the HTML version of the article, convert the PDF to Word in Adobe Acrobat or another PDF reader, or use the “paste special” command to paste the content into Word without formatting. If it’s an old PDF, you may have to simply type out the information you need. It can be a messy job, but having all of your facts in one place is very helpful when drafting your literature review.

You should copy and paste any fact or argument you consider important. Some good examples include definitions of concepts, statistics about the size of the social problem, and empirical evidence about the key variables in the research question, among countless others. It’s a good idea to consult with your professor and the course syllabus to understand what they are looking for when reading your literature review. Facts for your literature review are principally found in the introduction, results, and discussion section of an empirical article or at any point in a non-empirical article. Copy and paste into your notes anything you may want to use in your literature review.

Importantly, you must make sure you note the original source of each bit of information you copy. Nothing is worse than needing to track down a source for fact you read who-knows-where. If you found a statistic that the author used in the introduction, it almost certainly came from another source that the author cited in a footnote or internal citation. You will want to check the original source to make sure the author represented the information correctly. Moreover, you may want to read the original study to learn more about your topic and discover other sources relevant to your inquiry.

Assuming you have pulled all of the facts out of multiple articles, it’s time to start thinking about how these pieces of information relate to each other. Start grouping each fact into categories and subcategories as shown in Table 5.2. For example, a statistic stating that single adults who are homeless are more likely to be male may fit into a category of gender and homelessness. For each topic or subtopic you identify during your critical analysis of each paper, determine what those papers have in common. Likewise, determine which differ. If there are contradictory findings, you may be able to identify methodological or theoretical differences that could account for these contradictions. For example, one study may sample only high-income earners or those living in a rural area. Determine what general conclusions you can report about the topic or subtopic, based on all of the information you’ve found.

Create a separate document containing a topical outline that combines your facts from each source and organizes them by topic or category. As you include more facts and more sources in your topical outline, you will begin to see how each fact fits into a category and how categories are related to one another. Keep in mind that your category names may change over time, as may their definitions. This is a natural reflection of the learning you are doing.

A complete topical outline is a long list of facts arranged by category. As you step back from the outline, you should assess which topic areas for which you have enough research support to allow you to draw strong conclusions. You should also assess which areas you need to do more research in before you can write a robust literature review. The topical outline should serve as a transitional document between the notes you write on each source and the literature review you submit to your professor. It is important to note that they contain plagiarized information that is copied and pasted directly from the primary sources. In this case, it is not problematic because these are just notes and are not meant to be turned in as your own ideas. For your final literature review, you must paraphrase these sources to avoid plagiarism. More importantly, you should keep your voice and ideas front-and-center in what you write as this is your analysis of the literature. Make strong claims and support them thoroughly using facts you found in the literature. We will pick up the task of writing your literature review in section 5.3.

• In your folder full of article PDFs, look for the most relevant review articles. If you don’t have any, try to look for some. If there are none in your topic area, you can also use other non-empirical articles or empirical articles with long literature reviews (in the introduction and discussion sections).
• Create a word processing document for your topical outline and save it in your project folder on your hard drive. Using a review article, start copying facts you identified as Background Information or Results into your topical outline. Try to organize each fact by topic or theme. Make sure to copy the internal citation for the original source of each fact. For articles that do not use internal citations, create one using the information in the footnotes and references. As you finalize your research question over the next few weeks, skim the literature reviews of the articles you download for key facts and copy them into your topical outline.

Putting the pieces together: Building a concept map

Developing a concept map or mind map around your topic can be helpful in figuring out how the facts fit together. We talked about concept mapping briefly in Chapter 2 , when we were first thinking about your topic and sketching out what you already know about it. Concept mapping during the literature review stage of a research project builds on this foundation of knowledge and aims to improve the “description of the breadth and depth of literature in a domain of inquiry. It also facilitates identification of the number and nature of studies underpinning mapped relationships among concepts, thus laying the groundwork for systematic research reviews and meta-analyses” (Lesley, Floyd, & Oermann, 2002, p. 229). [11] Its purpose, like other question refinement methods, is to help you organize, prioritize, and integrate material into a workable research area—one that is interesting, answerable, feasible, objective, scholarly, original, and clear.

Think about the topics you created in your topic outline. How do they relate to one another? Within each topic, how do facts relate to one another? As you write down what you have, think about what you already know. What other related concepts do you not yet have information about? What relationships do you need to investigate further? Building a conceptual map should help you understand what you already know, what you need to learn next, and how you can organize a literature review.

This technique is illustrated in this YouTube video about concept mapping . You may want to indicate which concepts and relationships you’ve already found in your review and which ones you think might be true but haven’t found evidence of yet. Once you get a sense of how your concepts are related and which relationships are important to you, it’s time to revise your working question.

• Create a concept map using a pencil and paper.
• Identify the key ideas inside the literature, how they relate to one another, and the facts you know about them.
• Reflect on those areas you need to learn more about prior to writing your literature review.
• As you finalize your research question over the next few weeks, update your concept map and think about how you might organize it into a written literature review.
• Refer to the topics and headings you use in your topical outline and think about what literature you have that helps you understand each concept and relationship between them in your concept map.

Revising your working question

You should be revisiting your working question throughout the literature review process. As you continue to learn more about your topic, your question will become more specific and clearly worded. This is normal, and there is no way to shorten this process. Keep revising your question in order to ensure it will contribute something new to the literature on your topic, is relevant to your target population, and is feasible for you to conduct as a student project.

For example, perhaps your initial idea or interest is how to prevent obesity. After an initial search of the relevant literature, you realize the topic of obesity is too broad to adequately cover in the time you have to do your project. You decide to narrow your focus to causes of childhood obesity. After reading some articles on childhood obesity, you further narrow your search to the influence of family risk factors on overweight children. A potential research question might then be, “What maternal factors are associated with toddler obesity in the United States?” You would then need to return to the literature to find more specific studies related to the variables in this question (e.g. maternal factors, toddler, obesity, toddler obesity).

Similarly, after an initial literature search for a broad topic such as school performance or grades, examples of a narrow research question might be:

• “To what extent does parental involvement in children’s education relate to school performance over the course of the early grades?”
• “Do parental involvement levels differ by family social, demographic, and contextual characteristics?”
• “What forms of parent involvement are most highly correlated with children’s outcomes? What factors might influence the extent of parental involvement?” (Early Childhood Longitudinal Program, 2011). [12]

In either case, your literature search, working question, and understanding of the topic are constantly changing as your knowledge of the topic deepens. A literature review is an iterative process, one that stops, starts, and loops back on itself multiple times before completion. As research is a practice behavior of social workers, you should apply the same type of critical reflection to your inquiry as you would to your clinical or macro practice.

There are many ways to approach synthesizing literature. We’ve reviewed the following: summary tables, topical outlines, and concept maps. Other examples you may encounter include annotated bibliographies and synthesis matrices. As you are learning how to conduct research, find a method that works for you. Reviewing the literature is a core component of evidence-based practice in social work. See the resources below if you need some additional help:

Literature Reviews: Using a Matrix to Organize Research  / Saint Mary’s University of Minnesota

Literature Review: Synthesizing Multiple Sources  / Indiana University

Writing a Literature Review and Using a Synthesis Matrix  / Florida International University

Sample Literature Reviews Grid  / Complied by Lindsay Roberts

Literature review preparation: Creating a summary table . (Includes transcript) / Laura Killam

• You won’t read every article all the way through. For most articles, reading the abstract, introduction, and conclusion are enough to determine its relevance. It’s expected that you skim or search for relevant sections of each article without reading the whole thing.
• For articles where everything seems relevant, use a summary table to keep track of details. These are particularly helpful with empirical articles.
• For articles with literature review content relevant to your topic, copy any relevant information into a topical outline, along with the original source of that information.
• Use a concept map to help you visualize the key concepts in your topic area and the relationships between them.
• Revise your working question regularly. As you do, you will likely need to revise your search queries and include new articles.
• Look back at the working question for your topic and consider any necessary revisions. It is important that questions become clearer and more specific over time. It is also common that your working question shift over time, sometimes drastically, as you explore new lines of inquiry in the literature. Return to your working question regularly and make sure it reflects the focus of your inquiry. You will continue to revise your working question until we formalize it into a research question at the end of Part 2 of this textbook.

5.3 Writing your literature review

• Describe the components of a literature review
• Begin to write your literature review
• Identify the purpose of a problem statement
• Apply the components of a formal argument to your topic
• Use elements of formal writing style, including signposting and transitions
• Recognize commons errors in literature reviews

Congratulations! By now, you should have discovered, retrieved, evaluated, synthesized, and organized the information you need for your literature review. It’s now time to turn that stack of articles, papers, and notes into a literature review–it’s time to start writing!

Writing about research is different than other types of writing. Research writing is not like a journal entry or opinion paper. The goal here is not to apply your research question to your life or growth as a practitioner. Research writing is about the provision and interpretation of facts. The tone should be objective and unbiased, and personal experiences and opinions are excluded. Particularly for students who are used to writing case notes, research writing can be a challenge. That’s why its important to normalize getting help! If your professor has not built in peer review, consider setting up a peer review group among your peers. You should also reach out to your academic advisor to see if there are writing services on your campus available to graduate students. No one should feel bad for needing help with something they haven’t done before, haven’t done in a while, or were never taught how to do. 

If you’ve followed the steps in this chapter, you likely have an outline, summary table, and concept map from which you can begin the writing process. But what do you need to include in your literature review? We’ve mentioned it before, but to summarize, a literature review should:

• Introduce the topic and define its key terms.
• Establish the importance of the topic.
• Provide an overview of the important literature related to the concepts found in the research question.
• Identify gaps or controversies in the literature.
• Point out consistent findings across studies.
• Synthesize that which is known about a topic, rather than just provide a summary of the articles you read.
• Discuss possible implications and directions for future research.

Do you have enough facts and sources to accomplish these tasks? It’s a good time to consult your outlines and notes on each article you plan to include in your literature review. You may also want to consult with your professor on what is expected of you. If there is something you are missing, you may want to jump back to section 2.3 where we discussed how to search for literature. While you can always fill in material, there is the danger that you will start writing without really knowing what you are talking about or what you want to say. For example, if you don’t have a solid definition of your key concepts or a sense of how the literature has developed over time, it will be difficult to make coherent scholarly claims about your topic.

There is no magical point at which one is ready to write. As you consider whether you are ready, it may be useful to ask yourself these questions:

• How will my literature review be organized?
• What section headings will I be using?
• How do the various studies relate to each other?
• What contributions do they make to the field?
• Where are the gaps or limitations in existing research?
• And finally, but most importantly, how does my own research fit into what has already been done?

The problem statement

Scholarly works often begin with a problem statement, which serves two functions. First, it establishes why your topic is a social problem worth studying. Second, it pulls your reader into the literature review. Who would want to read about something unimportant?

A problem statement generally answers the following questions, though these are far from exhaustive:

• Why is this an important problem to study?
• How many people are affected by this problem?
• How does this problem impact other social issues relevant to social work?
• Why is your target population an important one to study?

A strong problem statement, like the rest of your literature review, should be filled with empirical results, theory, and arguments based on the extant literature. A research proposal differs significantly from other more reflective essays you’ve likely completed during your social work studies. If your topic were domestic violence in rural Appalachia, I’m sure you could come up with answers to the above questions without looking at a single source. However, the purpose of the literature review is not to test your intuition, personal experience, or empathy. Instead, research methods are about gaining specific and articulable knowledge to inform action. With a problem statement, you can take a “boring” topic like the color of rooms used in an inpatient psychiatric facility, transportation patterns in major cities, or the materials used to manufacture baby bottles, and help others see the topic as you see it—an important part of the social world that impacts social work practice.

The structure of a literature review

In general, the problem statement belongs at the beginning of the literature review. We usually advise students to spend no more than a paragraph or two for a problem statement. For the rest of your literature review, there is no set formula by which it needs to be organized. However, a literature review generally follows the format of any other essay—Introduction, Body, and Conclusion.

The introduction to the literature review contains a statement or statements about the overall topic. At a minimum, the introduction should define or identify the general topic, issue, or area of concern. You might consider presenting historical background, mentioning the results of a seminal study, and providing definitions of important terms. The introduction may also point to overall trends in what has been previously published on the topic or on conflicts in theory, methodology, evidence, conclusions, or gaps in research and scholarship. We also suggest putting in a few sentences that walk the reader through the rest of the literature review. Highlight your main arguments from the body of the literature review and preview your conclusion. An introduction should let the reader know what to expect from the rest of your review.

The body of your literature review is where you demonstrate your synthesis and analysis of the literature. Again, do not just summarize the literature. We would also caution against organizing your literature review by source—that is, one paragraph for source A, one paragraph for source B, etc. That structure will likely provide an adequate summary of the literature you’ve found, but it would give you almost no synthesis of the literature. That approach doesn’t tell your reader how to put those facts together, it doesn’t highlight points of agreement or contention, or how each study builds on the work of others. In short, it does not demonstrate critical thinking.

Organize your review by argument

Instead, use your outlines and notes as a guide what you have to say about the important topics you need to cover. Literature reviews are written from the perspective of an expert in that field. After an exhaustive literature review, you should feel as though you are able to make strong claims about what is true—so make them! There is no need to hide behind “I believe” or “I think.” Put your voice out in front, loud and proud! But make sure you have facts and sources that back up your claims.

I’ve used the term “ argument ” here in a specific way. An argument in writing means more than simply disagreeing with what someone else said, as this classic Monty Python sketch demonstrates. Toulman, Rieke, and Janik (1984) identify six elements of an argument:

• Claim: the thesis statement—what you are trying to prove
• Grounds: theoretical or empirical evidence that supports your claim
• Warrant: your reasoning (rule or principle) connecting the claim and its grounds
• Backing: further facts used to support or legitimize the warrant
• Qualifier: acknowledging that the argument may not be true for all cases
• Rebuttal: considering both sides (as cited in Burnette, 2012) [13]

Let’s walk through an example. If I were writing a literature review on a negative income tax, a policy in which people in poverty receive an unconditional cash stipend from the government each month equal to the federal poverty level, I would want to lay out the following:

• Claim: the negative income tax is superior to other forms of anti-poverty assistance.
• Grounds: data comparing negative income tax recipients to people receiving anti-poverty assistance in existing programs, theory supporting a negative income tax, data from evaluations of existing anti-poverty programs, etc.
• Warrant: cash-based programs like the negative income tax are superior to existing anti-poverty programs because they allow the recipient greater self-determination over how to spend their money.
• Backing: data demonstrating the beneficial effects of self-determination on people in poverty.
• Qualifier: the negative income tax does not provide taxpayers and voters with enough control to make sure people in poverty are not wasting financial assistance on frivolous items.
• Rebuttal: policy should be about empowering the oppressed, not protecting the taxpayer, and there are ways of addressing taxpayer spending concerns through policy design.

Like any effective argument, your literature review must have some kind of structure. For example, it might begin by describing a phenomenon in a general way along with several studies that provide some detail, then describing two or more competing theories of the phenomenon, and finally presenting a hypothesis to test one or more of the theories. Or, it might describe one phenomenon, then describe another that seems inconsistent with the first, then propose a theory that resolves the inconsistency, and finally present a hypothesis to test that theory. In applied research, it might describe a phenomenon or theory, then describe how that phenomenon or theory applies to some important real-world situation, and finally, may suggest a way to test whether it does, in fact, apply to that situation.

Use signposts

Another important issue is  signposting . It may not be a term you are familiar with, but you are likely familiar with the concept. Signposting refers to the words used to identify the organization and structure of your literature review to your reader. The most basic form of signposting is using a topic sentence at the beginning of each paragraph. A topic sentence introduces the argument you plan to make in that paragraph. For example, you might start a paragraph stating, “There is strong disagreement in the literature as to whether psychedelic drugs cause people to develop psychotic disorders, or whether psychotic disorders cause people to use psychedelic drugs.” Within that paragraph, your reader would likely assume you will present evidence for both arguments. The concluding sentence of your paragraph should address the topic sentence, discussing how the facts and arguments from the paragraph you’ve written support a specific conclusion. To continue with our example, I might say, “There is likely a reciprocal effect in which both the use of psychedelic drugs worsens pre-psychotic symptoms and worsening psychosis increases the desire to use psychedelic drugs.”

Signposting also involves using headings and subheadings. Your literature review will use APA formatting, which means you need to follow their rules for bolding, capitalization, italicization, and indentation of headings. Headings help your reader understand the structure of your literature review. They can also help if the reader gets lost and needs to re-orient themselves within the document. We often tell our students to assume we know nothing (they don’t mind) and need to be shown exactly where they are addressing each part of the literature review. It’s like walking a small child around, telling them “First we’ll do this, then we’ll do that, and when we’re done, we’ll know this!”

Another way to use signposting is to open each paragraph with a sentence that links the topic of the paragraph with the one before it. Alternatively, one could end each paragraph with a sentence that links it with the next paragraph. For example, imagine we wanted to link a paragraph about barriers to accessing healthcare with one about the relationship between the patient and physician. We could use a transition sentence like this: “Even if patients overcome these barriers to accessing care, the physician-patient relationship can create new barriers to positive health outcomes.” A transition sentence like this builds a connection between two distinct topics. Transition sentences are also useful within paragraphs. They tell the reader how to consider one piece of information in light of previous information. Even simple transitional words like ‘however’ and ‘similarly’ can help demonstrate critical thinking and link each building block of your argument together.

Many beginning researchers have difficulty incorporating transitions into their writing. Let’s look at an example. Instead of beginning a sentence or paragraph by launching into a description of a study, such as “Williams (2004) found that…,” it is better to start by indicating something about why you are describing this particular study. Here are some simple examples:

• Another example of this phenomenon comes from the work of Williams (2004)…
• Williams (2004) offers one explanation of this phenomenon…
• An alternative perspective has been provided by Williams (2004)…

Now that we know to use signposts, the natural question is “What goes on the signposts?” First, it is important to start with an outline of the main points that you want to make, organized in the order you want to make them. The basic structure of your argument should then be apparent from the outline itself. Unfortunately, there is no formula we can give you that will work for everyone, but we can provide some general pointers on structuring your literature review.

The literature review tends to move from general to more specific ideas. You can build a review by identifying areas of consensus and areas of disagreement. You may choose to present historical studies—preferably seminal studies that are of significant importance—and close with the most recent research. Another approach is to start with the most distantly related facts and literature and then report on those most closely related to your research question. You could also compare and contrast valid approaches, features, characteristics, theories – that is, one approach, then a second approach, followed by a third approach.

Here are some additional tips for writing the body of your literature review:

• Start broad and then narrow down to more specific information.
• When appropriate, cite two or more sources for a single point, but avoid long strings of references for a single idea.
• Use quotes sparingly. Quotations for definitions are okay, but reserve quotes for when something is said so well you couldn’t possible phrase it differently. Never use quotes for statistics.
• Paraphrase when you need to relay the specific details within an article
• Include only the aspects of the study that are relevant to your literature review. Don’t insert extra facts about a study just to take up space.
• Avoid reflective, personal writing. It is traditional to avoid using first-person language (I, we, us, etc.).
• Avoid informal language like contractions, idioms, and rhetorical questions.
• Note any sections of your review that lack citations from the literature. Your arguments need to be based in empirical or theoretical facts. Do not approach this like a reflective journal entry.
• Point out consistent findings and emphasize stronger studies over weaker ones.
• Point out important strengths and weaknesses of research studies, as well as contradictions and inconsistent findings.
• Implications and suggestions for further research (where there are gaps in the current literature) should be specific.

The conclusion should summarize your literature review, discuss implications, and create a space for further research needed in this area. Your conclusion, like the rest of your literature review, should make a point. What are the important implications of your literature review? How do they inform the question you are trying to answer?

You should consult with your professor and the course syllabus about the final structure your literature review should take. Here is an example of one possible structure:

• Establish the importance of the topic
• Number and type of people affected
• Seriousness of the impact
• Physical, psychological, economic, social, or spiritual consequences of the problem
• Definitions of key terms
• Supporting evidence
• Common findings across studies, gaps in the literature
• Research question(s) and hypothesis(es)

Editing your literature review

Literature reviews are more than a summary of the publications you find on a topic. As you have seen in this brief introduction, literature reviews represent a very specific type of research, analysis, and writing. We will explore these topics further in upcoming chapters. As you begin your literature review, here are some common errors to avoid:

• Accepting a researcher’s finding as valid without evaluating methodology and data
• Ignoring contrary findings and alternative interpretations
• Using findings that are not clearly related to your own study or using findings that are too general
• Dedicating insufficient time to literature searching
• Reporting statistical results from a single study, rather than synthesizing the results of multiple studies to provide a comprehensive view of the literature on a topic
• Relying too heavily on secondary sources
• Overusing quotations
• Not justifying arguments using specific facts or theories from the literature

For your literature review, remember that your goal is to construct an argument for the importance of your research question. As you start editing your literature review, make sure it is balanced. Accurately report common findings, areas where studies contradict each other, new theories or perspectives, and how studies cause us to reaffirm or challenge our understanding of your topic.

It is acceptable to argue that the balance of the research supports the existence of a phenomenon or is consistent with a theory (and that is usually the best that researchers in social work can hope for), but it is not acceptable to ignore contradictory evidence. A large part of what makes a research question interesting is uncertainty about its answer (University of Minnesota, 2016). [14]

In addition to subjectivity and bias, writer’s block can obstruct the completion of your literature review. Often times, writer’s block can stem from confusing the creating and editing parts of the writing process. Many writers often start by simply trying to type out what they want to say, regardless of how good it is. Author Anne Lamott (1995) [15] terms these “shitty first drafts,” and we all write them. They are a natural and important part of the writing process.

Even if you have a detailed outline from which to work, the words are not going to fall into place perfectly the first time you start writing. You should consider turning off the editing and critiquing part of your brain for a while and allow your thoughts to flow. Don’t worry about putting a correctly formatted internal citation (as long as  you know which source you used there) when you first write. Just get the information out. Only after you’ve reached a natural stopping point might you go back and edit your draft for grammar, APA style, organization, flow, and more. Divorcing the writing and editing process can go a long way to addressing writer’s block—as can picking a topic about which you have something to say!

As you are editing, keep in mind these questions adapted from Green (2012): [16]

• Content: Have I clearly stated the main idea or purpose of the paper? Is the thesis or focus clearly presented and appropriate for the reader?
• Organization: How well is it structured? Is the organization spelled out and easy to follow for the reader ?
• Flow: Is there a logical flow from section to section, paragraph to paragraph, sentence to sentence? Are there transitions between and within paragraphs that link ideas together?
• Development: Have I validated the main idea with supporting material? Are supporting data sufficient? Does the conclusion match the introduction?
• Form: Are there any APA style issues, redundancy, problematic wording and terminology (always know the definition of any word you use!), flawed sentence constructions and selection, spelling, and punctuation?

Social workers use the APA style guide to format and structure their literature reviews. Most students know APA style only as it relates to internal and external citations. If you are confused about them, consult this amazing APA style guide from the University of Texas-Arlington library. Your university’s library likely has resources they created to help you with APA style, and you can meet with a librarian or your professor to talk about formatting questions you have. Make sure you budget in a few hours at the end of each project to build a correctly formatted references page and check your internal citations.

Of course, APA style is about much more than knowing there is a period after “et al.” or citing the location a book was published. APA style is also about what the profession considers to be good writing. If you haven’t picked up an APA publication manual because you use citation generators, know that I did the same thing when I was in school. Purchasing the APA manual can help you with a common problem we hear about from students. Every professor (and every website about APA style) seems to have their own peculiar idea of “correct” APA style that you can, if needed, demonstrate is not accurate.

Here are some additional resources, if you would like more guidance on writing your literature review.

Doing a literature review  / University of Leicester

Get lit: The literature review  / Texas A&M Writing Centre

Guidebook for social work literature reviews / by Rebecca Mauldin and Matthew DeCarlo

• A literature review is not a book report. Do not organize it by article, with one paragraph for each source in your references. Instead, organize it based on the key ideas and arguments.
• The problem statement draws the reader into your topic by highlighting the importance of the topic to social work and to society overall.
• Signposting is an important component of academic writing that helps your reader follow the structure of your argument and of your literature review.
• Transitions demonstrate critical thinking and help guide your reader through your arguments.
• Editing and writing are separate processes.
• Consult with an APA style guide or a librarian to help you format your paper.

Look at your professor’s prompt for the literature review component of your research proposal (or if you don’t have one, use the example question provided in this section).

• Write 2-3 facts you would use to address each question or component in the prompt.
• Reflect on which questions you have a lot of information about and which you need to gather more information about in order to answer adequately.

Outline the structure of your literature review using your concept map from Section 5.2 as a guide.

• Identify the key arguments you will make and how they are related to each other.
• Reflect on topic sentences and concluding sentences you would use for each argument.

Media Attributions

• Numbers © Pop and Zebra is licensed under a CC0 (Creative Commons Zero) license
• summary table © Laura Frederiksen is licensed under a Public Domain license
• problem-2731501_1920 © Geralt is licensed under a CC0 (Creative Commons Zero) license
• sign-2080927_1920 © MariaMichelle is licensed under a CC0 (Creative Commons Zero) license
• It wouldn’t make any sense to say that people’s workplace experiences cause  their gender, so in this example, the question of which is the independent variable and which are the dependent variables has a pretty obvious answer. ↵
• Cassidy, S. A., Dimova, R., Giguère, B., Spence, J. R., & Stanley, D. J. (2019). Failing grade: 89% of introduction-to-psychology textbooks that define or explain statistical significance do so incorrectly. Advances in Methods and Practices in Psychological Science ,  2 (3), 233-239. ↵
• Wasserstein, R. L., & Lazar, N. A. (2016). The ASA statement on p-values: context, process, and purpose. The American Statistician, 70 , p. 129-133. ↵
• Head, M. L., Holman, L., Lanfear, R., Kahn, A. T., & Jennions, M. D. (2015). The extent and consequences of p-hacking in science. PLoS biology, 13 (3). ↵
• Peng, R. (2015), The reproducibility crisis in science: A statistical counterattack. Significance , 12 , 30–32. ↵
• Greenland, S., Senn, S. J., Rothman, K. J., Carlin, J. B., Poole, C., Goodman, S. N., & Altman, D. G. (2016). Statistical tests, P values, confidence intervals, and power: a guide to misinterpretations.  European journal of epidemiology ,  31 (4), 337-350. ↵
• Bonanno, R., & Veselak, K. (2019). A matter of trust: Parents attitudes towards child mental health information sources.  Advances in Social Work ,  19 (2), 397-415. ↵
• Bernnard, D., Bobish, G., Hecker, J., Holden, I., Hosier, A., Jacobson, T., Loney, T., & Bullis, D. (2014). Presenting: Sharing what you’ve learned. In Bobish, G., & Jacobson, T. (eds.)  The information literacy users guide: An open online textbook .  https://milnepublishing.geneseo.edu/the-information-literacy-users-guide-an-open-online-textbook/chapter/present-sharing-what-youve-learned/ ↵
• Leslie, M., Floyd, J., & Oermann, M. (2002). Use of MindMapper software for research domain mapping. Computers, informatics, nursing,  20(6), 229-235. ↵
• Early Childhood Longitudinal Program. (2011).  Example research questions .  https://nces.ed.gov/ecls/researchquestions2011.asp ↵
• Burnett, D. (2012). Inscribing knowledge: Writing research in social work. In W. Green & B. L. Simon (Eds.),  The Columbia guide to social work writing  (pp. 65-82). New York, NY: Columbia University Press. ↵
• University of Minnesota Libraries Publishing. (2016). This is a derivative of  Research Methods in Psychology  by a publisher who has requested that they and the original author not receive attribution, which was originally released and is used under CC BY-NC-SA. This work, unless otherwise expressly stated, is licensed under a  Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License ↵
• Lamott, A. (1995). Bird by bird: Some instructions on writing and life . New York, NY: Penguin. ↵
• Green, W. Writing strategies for academic papers. In W. Green & B. L. Simon (Eds.),  The Columbia guide to social work writing  (pp. 25-47). New York, NY: Columbia University Press. ↵

report the results of a quantitative or qualitative data analysis conducted by the author

a quick, condensed summary of the report’s key findings arranged by row and column

causes a change in the dependent variable

a variable that depends on changes in the independent variable

(as in generalization) to make claims about a large population based on a smaller sample of people or items

"Assuming that the null hypothesis is true and the study is repeated an infinite number times by drawing random samples from the same populations(s), less than 5% of these results will be more extreme than the current result" (Cassidy et al., 2019, p. 233).

the assumption that no relationship exists between the variables in question

“a logical grouping of attributes that can be observed and measured and is expected to vary from person to person in a population” (Gillespie & Wagner, 2018, p. 9)

summarizes the incompatibility between a particular set of data and a proposed model for the data, usually the null hypothesis. The lower the p-value, the more inconsistent the data are with the null hypothesis, indicating that the relationship is statistically significant.

a range of values in which the true value is likely to be, to provide a more accurate description of their data

a statement about what you think is true backed up by evidence and critical thinking

the words used to identify the organization and structure of your literature review to your reader

what a researcher hopes to accomplish with their study

Graduate research methods in social work Copyright © 2021 by Matthew DeCarlo, Cory Cummings, Kate Agnelli is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License , except where otherwise noted.

Share This Book

Literature Review Basics

• What is a Literature Review?
• Synthesizing Research
• Using Research & Synthesis Tables
• Additional Resources

About the Research and Synthesis Tables

Research Tables and Synthesis Tables are useful tools for organizing and analyzing your research as you assemble your literature review. They represent two different parts of the review process: assembling relevant information and synthesizing it. Use a Research table to compile the main info you need about the items you find in your research -- it's a great thing to have on hand as you take notes on what you read! Then, once you've assembled your research, use the Synthesis table to start charting the similarities/differences and major themes among your collected items.

We've included an Excel file with templates for you to use below; the examples pictured on this page are snapshots from that file.

• Research and Synthesis Table Templates This Excel workbook includes simple templates for creating research tables and synthesis tables. Feel free to download and use!

Using the Research Table

This is an example of a  research table,  in which you provide a basic description of the most important features of the studies, articles, and other items you discover in your research. The table identifies each item according to its author/date of publication, its purpose or thesis, what type of work it is (systematic review, clinical trial, etc.), the level of evidence it represents (which tells you a lot about its impact on the field of study), and its major findings. Your job, when you assemble this information, is to develop a snapshot of what the research shows about the topic of your research question and assess its value (both for the purpose of your work and for general knowledge in the field).

Think of your work on the research table as the foundational step for your analysis of the literature, in which you assemble the information you'll be analyzing and lay the groundwork for thinking about what it means and how it can be used.

Using the Synthesis Table

This is an example of a  synthesis table  or  synthesis matrix , in which you organize and analyze your research by listing each source and indicating whether a given finding or result occurred in a particular study or article ( each row lists an individual source, and each finding has its own column, in which X = yes, blank = no). You can also add or alter the columns to look for shared study populations, sort by level of evidence or source type, etc. The key here is to use the table to provide a simple representation of what the research has found (or not found, as the case may be). Think of a synthesis table as a tool for making comparisons, identifying trends, and locating gaps in the literature.

How do I know which findings to use, or how many to include?  Your research question tells you which findings are of interest in your research, so work from your research question to decide what needs to go in each Finding header, and how many findings are necessary. The number is up to you; again, you can alter this table by adding or deleting columns to match what you're actually looking for in your analysis. You should also, of course, be guided by what's actually present in the material your research turns up!

• << Previous: Synthesizing Research
• Next: Additional Resources >>
• Last Updated: Sep 26, 2023 12:06 PM
• URL: https://usi.libguides.com/literature-review-basics

The Sheridan Libraries

• Write a Literature Review
• Sheridan Libraries
• Evaluate This link opens in a new window

Not every source you found should be included in your annotated bibliography or lit review. Only include the most relevant and most important sources.

Get Organized

• Lit Review Prep Use this template to help you evaluate your sources, create article summaries for an annotated bibliography, and a synthesis matrix for your lit review outline.

Summarize your Sources

Summarize each source: Determine the most important and relevant information from each source, such as the findings, methodology, theories, etc.  Consider using an article summary, or study summary to help you organize and summarize your sources.

Paraphrasing

• Use your own words, and do not copy and paste the abstract
• The library's tutorials about plagiarism are excellent, and will help you with paraphasing correctly

Annotated Bibliographies

     Annotated bibliographies can help you clearly see and understand the research before diving into organizing and writing your literature review.        Although typically part of the "summarize" step of the literature review, annotations should not merely be summaries of each article - instead, they should be critical evaluations of the source, and help determine a source's usefulness for your lit review.  

Definition:

A list of citations on a particular topic followed by an evaluation of the source’s argument and other relevant material including its intended audience, sources of evidence, and methodology

• Explore your topic.
• Appraise issues or factors associated with your professional practice and research topic.
• Help you get started with the literature review.
• Think critically about your topic, and the literature.

Steps to Creating an Annotated Bibliography:

• Find Your Sources
• Read Your Sources
• Identify the Most Relevant Sources
• Cite your Sources
• Write Annotations

Annotated Bibliography Resources

• Purdue Owl Guide
• Cornell Annotated Bibliography Guide
• << Previous: Evaluate
• Next: Synthesize >>
• Last Updated: Jul 30, 2024 1:42 PM
• URL: https://guides.library.jhu.edu/lit-review

About Systematic Reviews

Summary of Findings Table in a Systematic Review

Automate every stage of your literature review to produce evidence-based research faster and more accurately.

What is a summary of findings table.

The Cochrane Review defines the “summary of findings table” as a structured tabular format in which the primary findings of a review, particularly information related to the quality of evidence, the magnitude of the effects of the studied interventions, and the aggregate of available data on the main outcomes, are presented. It includes multiple pieces of data derived from both quantitative and qualitative data analysis in systematic reviews . These include information about the main outcomes, the type and number of studies included, the estimates (both relative and absolute) of the effect or association, and important comments about the review, all written in a plain-language summary so that it’s easily interpreted. It also includes a grade of the quality of evidence; i.e., a rating of its certainty.

Most systematic reviews are expected to have one summary of findings table. But some studies may have multiple, if the review addresses more than one comparison, or deals with substantially different populations that require separate tables. The studies in a table can also be grouped in terms of applied intervention type, type of outcome measure, the type of participants, the study design etc..

How Do You Make A Summary Of Findings Table For A Systematic Review?

Learn more about distillersr.

(Article continues below)

What Does A Summary Of Findings Table Include?

A summary of findings table typically includes the following information:

• A description of the population and setting addressed by the available evidence
• A description of comparisons addressed in the table, including all interventions
• A list of the most important outcomes, whether desirable or undesirable (limited to seven)
• A measure of the burdens of each outcome
• The magnitude of effect measured for each outcome (both absolute and relative)
• The participants and studies analyzed for each outcome
• An assessment of the certainty of the evidence for each outcome (typically using GRADE)
• Explanations

It’s best to include evidence profiles, i.e. additional tables that support the data in the summary of findings, to which the review may be linked. It also may be neat to have a study descriptor table different from a results table. The study descriptor table shows information about the characteristics of included studies, like study design, study region, participant information, etc. The results table mostly contains outcomes, outcome measures, study results, etc. These can help provide readers with more context about the review, and its conclusions.

Final Takeaway

3 reasons to connect.

Be the boss of your literature review

Download this free article summary table template.

When dealing with the literature, summarise the articles you read as you go along. This will ensure that you don't read and forget. Using the Article Summary Table template, you can neatly add a summary of each study to a table. This table is handy because you can easily refer to a specific article without searching through piles of pdfs.

Get the Article Summary Table template in Microsoft Word AND Microsoft Excel - for FREE

Jump to navigation

Cochrane Training

Chapter 9: summarizing study characteristics and preparing for synthesis.

Joanne E McKenzie, Sue E Brennan, Rebecca E Ryan, Hilary J Thomson, Renea V Johnston

Key Points:

• Synthesis is a process of bringing together data from a set of included studies with the aim of drawing conclusions about a body of evidence. This will include synthesis of study characteristics and, potentially, statistical synthesis of study findings.
• A general framework for synthesis can be used to guide the process of planning the comparisons, preparing for synthesis, undertaking the synthesis, and interpreting and describing the results.
• Tabulation of study characteristics aids the examination and comparison of PICO elements across studies, facilitates synthesis of these characteristics and grouping of studies for statistical synthesis.
• Tabulation of extracted data from studies allows assessment of the number of studies contributing to a particular meta-analysis, and helps determine what other statistical synthesis methods might be used if meta-analysis is not possible.

Cite this chapter as: McKenzie JE, Brennan SE, Ryan RE, Thomson HJ, Johnston RV. Chapter 9: Summarizing study characteristics and preparing for synthesis. In: Higgins JPT, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA (editors). Cochrane Handbook for Systematic Reviews of Interventions version 6.4 (updated August 2023). Cochrane, 2023. Available from www.training.cochrane.org/handbook .

9.1 Introduction

Synthesis is a process of bringing together data from a set of included studies with the aim of drawing conclusions about a body of evidence. Most Cochrane Reviews on the effects of interventions will include some type of statistical synthesis. Most commonly this is the statistical combination of results from two or more separate studies (henceforth referred to as meta-analysis) of effect estimates.

An examination of the included studies always precedes statistical synthesis in Cochrane Reviews. For example, examination of the interventions studied is often needed to itemize their content so as to determine which studies can be grouped in a single synthesis. More broadly, synthesis of the PICO (Population, Intervention, Comparator and Outcome) elements of the included studies underpins interpretation of review findings and is an important output of the review in its own right. This synthesis should encompass the characteristics of the interventions and comparators in included studies, the populations and settings in which the interventions were evaluated, the outcomes assessed, and the strengths and weaknesses of the body of evidence.

Chapter 2 defined three types of PICO criteria that may be helpful in understanding decisions that need to be made at different stages in the review:

• The review PICO (planned at the protocol stage) is the PICO on which eligibility of studies is based (what will be included and what excluded from the review).
• The PICO for each synthesis (also planned at the protocol stage) defines the question that the specific synthesis aims to answer, determining how the synthesis will be structured, specifying planned comparisons (including intervention and comparator groups, any grouping of outcome and population subgroups).
• The PICO of the included studies (determined at the review stage) is what was actually investigated in the included studies.

In this chapter, we focus on the PICO for each synthesis and the PICO of the included studies , as the basis for determining which studies can be grouped for statistical synthesis and for synthesizing study characteristics. We describe the preliminary steps undertaken before performing the statistical synthesis. Methods for the statistical synthesis are described in Chapter 10 , Chapter 11 and Chapter 12 .

9.2 A general framework for synthesis

Box 9.2.a A general framework for synthesis that can be applied irrespective of the methods used to synthesize results

Box 9.2.a provides a general framework for synthesis that can be applied irrespective of the methods used to synthesize results. Planning for the synthesis should start at protocol-writing stage, and Chapter 2 and Chapter 3 describe the steps involved in planning the review questions and comparisons between intervention groups. These steps included specifying which characteristics of the interventions, populations, outcomes and study design would be grouped together for synthesis (the PICO for each synthesis: stage 1 in Box 9.2.a ).

This chapter primarily concerns stage 2 of the general framework in Box 9.2.a . After deciding which studies will be included in the review and extracting data, review authors can start implementing their plan, working through steps 2.1 to 2.5 of the framework. This process begins with a detailed examination of the characteristics of each study (step 2.1), and then comparison of characteristics across studies in order to determine which studies are similar enough to be grouped for synthesis (step 2.2). Examination of the type of data available for synthesis follows (step 2.3). These three steps inform decisions about whether any modification to the planned comparisons or outcomes is necessary, or new comparisons are needed (step 2.4). The last step of the framework covered in this chapter involves synthesis of the characteristics of studies contributing to each comparison (step 2.5). The chapter concludes with practical tips for checking data before synthesis (Section 9.4 ).

Steps 2.1, 2.2 and 2.5 involve analysis and synthesis of mainly qualitative information about study characteristics. The process used to undertake these steps is rarely described in reviews, yet can require many subjective decisions about the nature and similarity of the PICO elements of the included studies. The examples described in this section illustrate approaches for making this process more transparent.

9.3 Preliminary steps of a synthesis

9.3.1 summarize the characteristics of each study (step 2.1).

A starting point for synthesis is to summarize the PICO characteristics of each study (i.e. the PICO of the included studies, see Chapter 3 ) and categorize these PICO elements in the groups (or domains) pre-specified in the protocol (i.e. the PICO for each synthesis). The resulting descriptions are reported in the ‘Characteristics of included studies’ table, and are used in step 2.2 to determine which studies can be grouped for synthesis.

In some reviews, the labels and terminology used in each study are retained when describing the PICO elements of the included studies. This may be sufficient in areas with consistent and widely understood terminology that matches the PICO for each synthesis. However, in most areas, terminology is variable, making it difficult to compare the PICO of each included study to the PICO for each synthesis, or to compare PICO elements across studies. Standardizing the description of PICO elements across studies facilitates these comparisons. This standardization includes applying the labels and terminology used to articulate the PICO for each synthesis ( Chapter 3 ), and structuring the description of PICO elements. The description of interventions can be structured using the Template for Intervention Description and Replication (TIDIeR) checklist, for example (see Chapter 3 and Table 9.3.a ).

Table 9.3.a illustrates the use of pre-specified groups to categorize and label interventions in a review of psychosocial interventions for smoking cessation in pregnancy (Chamberlain et al 2017). The main intervention strategy in each study was categorized into one of six groups: counselling, health education, feedback, incentive-based interventions, social support, and exercise. This categorization determined which studies were eligible for each comparison (e.g. counselling versus usual care; single or multi-component strategy). The extract from the ‘Characteristics of included studies’ table shows the diverse descriptions of interventions in three of the 54 studies for which the main intervention was categorized as ‘counselling’. Other intervention characteristics, such as duration and frequency, were coded in pre-specified categories to standardize description of the intervention intensity and facilitate meta-regression (not shown here).

Table 9.3.a Example of categorizing interventions into pre-defined groups

* The definition also specified eligible modes of delivery, intervention duration and personnel.

While this example focuses on categorizing and describing interventions according to groups pre-specified in the PICO for each synthesis, the same approach applies to other PICO elements.

9.3.2 Determine which studies are similar enough to be grouped within each comparison (step 2.2)

Once the PICO of included studies have been coded using labels and descriptions specified in the PICO for each synthesis, it will be possible to compare PICO elements across studies and determine which studies are similar enough to be grouped within each comparison.

Tabulating study characteristics can help to explore and compare PICO elements across studies, and is particularly important for reviews that are broad in scope, have diversity across one or more PICO elements, or include large numbers of studies. Data about study characteristics can be ordered in many different ways (e.g. by comparison or by specific PICO elements), and tables may include information about one or more PICO elements. Deciding on the best approach will depend on the purpose of the table and the stage of the review. A close examination of study characteristics will require detailed tables; for example, to identify differences in characteristics that were pre-specified as potentially important modifiers of the intervention effects. As the review progresses, this detail may be replaced by standardized description of PICO characteristics (e.g. the coding of counselling interventions presented in Table 9.3.a ).

Table 9.3.b illustrates one approach to tabulating study characteristics to enable comparison and analysis across studies. This table presents a high-level summary of the characteristics that are most important for determining which comparisons can be made. The table was adapted from tables presented in a review of self-management education programmes for osteoarthritis (Kroon et al 2014). The authors presented a structured summary of intervention and comparator groups for each study, and then categorized intervention components thought to be important for enabling patients to manage their own condition. Table 9.3.b shows selected intervention components, the comparator, and outcomes measured in a subset of studies (some details are fictitious). Outcomes have been grouped by the outcome domains ‘Pain’ and ‘Function’ (column ‘Outcome measure’ Table 9.3.b ). These pre-specified outcome domains are the chosen level for the synthesis as specified in the PICO for each synthesis. Authors will need to assess whether the measurement methods or tools used within each study provide an appropriate assessment of the domains ( Chapter 3, Section 3.2.4 ). A next step is to group each measure into the pre-specified time points. In this example, outcomes are grouped into short-term (<6 weeks) and long-term follow-up (≥6 weeks to 12 months) (column ‘Time points (time frame)’ Table 9.3.b ).

Variations on the format shown in Table 9.3.b can be presented within a review to summarize the characteristics of studies contributing to each synthesis, which is important for interpreting findings (step 2.5).

Table 9.3.b Table of study characteristics illustrating similarity of PICO elements across studies

BEH = health-directed behaviour; CON = constructive attitudes and approaches; EMO = emotional well-being; ENG = positive and active engagement in life; MON = self-monitoring and insight; NAV = health service navigation; SKL = skill and technique acquisition. ANCOVA = Analysis of covariance; CI = confidence interval; IQR = interquartile range; MD = mean difference; SD = standard deviation; SE = standard error, NS = non-significant. Pain and function measures: Dutch AIMS-SF = Dutch short form of the Arthritis Impact Measurement Scales; HAQ = Health Assessment Questionnaire; VAS = visual analogue scale; WOMAC = Western Ontario and McMaster Universities Osteoarthritis Index. 1 Ordered by type of comparator; 2 Short-term (denoted ‘immediate’ in the review Kroon et al (2014)) follow-up is defined as <6 weeks, long-term follow-up (denoted ‘intermediate’ in the review) is ≥6 weeks to 12 months; 3 For simplicity, in this example the available data are assumed to be the same for all outcomes within an outcome domain within a study. In practice, this is unlikely and the available data would likely vary by outcome; 4 Indicates that an effect estimate and its standard error may be computed through imputation of missing statistics, methods to convert between statistics (e.g. medians to means) or contact with study authors. *Indicates the selected outcome when there was multiplicity in the outcome domain and time frame.

9.3.3 Determine what data are available for synthesis (step 2.3)

Once the studies that are similar enough to be grouped together within each comparison have been determined, a next step is to examine what data are available for synthesis. Tabulating the measurement tools and time frames as shown in Table 9.3.b allows assessment of the potential for multiplicity (i.e. when multiple outcomes within a study and outcome domain are available for inclusion ( Chapter 3, Section 3.2.4.3 )). In this example, multiplicity arises in two ways. First, from multiple measurement instruments used to measure the same outcome domain within the same time frame (e.g. ‘Short-term Pain’ is measured using the ‘Pain VAS’ and ‘Pain on walking VAS’ scales in study 3). Second, from multiple time points measured within the same time frame (e.g. ‘Short-term Pain’ is measured using ‘Pain VAS’ at both 2 weeks and 1 month in study 6). Pre-specified methods to deal with the multiplicity can then be implemented (see Table 9.3.c for examples of approaches for dealing with multiplicity). In this review, the authors pre-specified a set of decision rules for selecting specific outcomes within the outcome domains. For example, for the outcome domain ‘Pain’, the selected outcome was the highest on the following list: global pain, pain on walking, WOMAC pain subscore, composite pain scores other than WOMAC, pain on activities other than walking, rest pain or pain during the night. The authors further specified that if there were multiple time points at which the outcome was measured within a time frame, they would select the longest time point. The selected outcomes from applying these rules to studies 3 and 6 are indicated by an asterisk in Table 9.3.b .

Table 9.3.b also illustrates an approach to tabulating the extracted data. The available statistics are tabulated in the column labelled ‘Data’, from which an assessment can be made as to whether the study contributes the required data for a meta-analysis (column ‘Effect & SE’) ( Chapter 10 ). For example, of the seven studies comparing health-directed behaviour (BEH) with usual care, six measured ‘Short-term Pain’, four of which contribute required data for meta-analysis. Reordering the table by comparison, outcome and time frame, will more readily show the number of studies that will contribute to a particular meta-analysis, and help determine what other synthesis methods might be used if the data available for meta-analysis are limited.

Table 9.3.c Examples of approaches for selecting one outcome (effect estimate) for inclusion in a synthesis.* Adapted from López-López et al (2018)

9.3.4 Determine if modification to the planned comparisons or outcomes is necessary, or new comparisons are needed (step 2.4)

The previous steps may reveal the need to modify the planned comparisons. Important variations in the intervention may be identified leading to different or modified intervention groups. Few studies or sparse data, or both, may lead to different groupings of interventions, populations or outcomes. Planning contingencies for anticipated scenarios is likely to lead to less post-hoc decision making ( Chapter 2 and Chapter 3 ); however, it is difficult to plan for all scenarios. In the latter circumstance, the rationale for any post-hoc changes should be reported. This approach was adopted in a review examining the effects of portion, package or tableware size for changing selection and consumption of food, alcohol and tobacco (Hollands et al 2015). After preliminary examination of the outcome data, the review authors changed their planned intervention groups. They judged that intervention groups based on ‘size’ and those based on ‘shape’ of the products were not conceptually comparable, and therefore should form separate comparisons. The authors provided a rationale for the change and noted that it was a post-hoc decision.

9.3.5 Synthesize the characteristics of the studies contributing to each comparison (step 2.5)

A final step, and one that is essential for interpreting combined effects, is to synthesize the characteristics of studies contributing to each comparison. This description should integrate information about key PICO characteristics across studies, and identify any potentially important differences in characteristics that were pre-specified as possible effect modifiers. The synthesis of study characteristics is also needed for GRADE assessments, informing judgements about whether the evidence applies directly to the review question (indirectness) and analyses conducted to examine possible explanations for heterogeneity (inconsistency) (see Chapter 14 ).

Tabulating study characteristics is generally preferable to lengthy description in the text, since the structure imposed by a table can make it easier and faster for readers to scan and identify patterns in the information presented. Table 9.3.b illustrates one such approach. Tabulating characteristics of studies that contribute to each comparison can also help to improve the transparency of decisions made around grouping of studies, while also ensuring that studies that do not contribute to the combined effect are accounted for.

9.4 Checking data before synthesis

Before embarking on a synthesis, it is important to be confident that the findings from the individual studies have been collated correctly. Therefore, review authors must compare the magnitude and direction of effects reported by studies with how they are to be presented in the review. This is a reasonably straightforward way for authors to check a number of potential problems, including typographical errors in studies’ reports, accuracy of data collection and manipulation, and data entry into RevMan. For example, the direction of a standardized mean difference may accidentally be wrong in the review. A basic check is to ensure the same qualitative findings (e.g. direction of effect and statistical significance) between the data as presented in the review and the data as available from the original study.

Results in forest plots should agree with data in the original report (point estimate and confidence interval) if the same effect measure and statistical model is used. There are legitimate reasons for differences, however, including: using a different measure of intervention effect; making different choices between change-from-baseline measures, post-intervention measures alone or post-intervention measures adjusted for baseline values; grouping similar intervention groups; or making adjustments for unit-of-analysis errors in the reports of the primary studies.

9.5 Types of synthesis

The focus of this chapter has been describing the steps involved in implementing the planned comparisons between intervention groups (stage 2 of the general framework for synthesis ( Box 9.2.a )). The next step (stage 3) is often performing a statistical synthesis. Meta-analysis of effect estimates, and its extensions have many advantages. There are circumstances under which a meta-analysis is not possible, however, and other statistical synthesis methods might be considered, so as to make best use of the available data. Available summary and synthesis methods, along with the questions they address and examples of associated plots, are described in Table 9.5.a . Chapter 10 and Chapter 11 discuss meta-analysis (of effect estimate) methods, while Chapter 12 focuses on the other statistical synthesis methods, along with approaches to tabulating, visually displaying and providing a structured presentation of the findings. An important part of planning the analysis strategy is building in contingencies to use alternative methods when the desired method cannot be used.

Table 9.5.a Overview of available methods for summary and synthesis

9.6 Chapter information

Authors: Joanne E McKenzie, Sue E Brennan, Rebecca E Ryan, Hilary J Thomson, Renea V Johnston

Acknowledgements: Sections of this chapter build on Chapter 9 of version 5.1 of the Handbook , with editors Jonathan Deeks, Julian Higgins and Douglas Altman. We are grateful to Julian Higgins, James Thomas and Tianjing Li for commenting helpfully on earlier drafts.

Funding: JM is supported by an NHMRC Career Development Fellowship (1143429). SB and RR’s positions are supported by the NHMRC Cochrane Collaboration Funding Program. HT is funded by the UK Medical Research Council (MC_UU_12017-13 and MC_UU_12017-15) and Scottish Government Chief Scientist Office (SPHSU13 and SPHSU15). RJ’s position is supported by the NHMRC Cochrane Collaboration Funding Program and Cabrini Institute.

9.7 References

Chamberlain C, O’Mara-Eves A, Porter J, Coleman T, Perlen SM, Thomas J, McKenzie JE. Psychosocial interventions for supporting women to stop smoking in pregnancy. Cochrane Database of Systematic Reviews 2017; 2 : CD001055.

Hollands GJ, Shemilt I, Marteau TM, Jebb SA, Lewis HB, Wei Y, Higgins JPT, Ogilvie D. Portion, package or tableware size for changing selection and consumption of food, alcohol and tobacco. Cochrane Database of Systematic Reviews 2015; 9 : CD011045.

Kroon FPB, van der Burg LRA, Buchbinder R, Osborne RH, Johnston RV, Pitt V. Self-management education programmes for osteoarthritis. Cochrane Database of Systematic Reviews 2014; 1 : CD008963.

López-López JA, Page MJ, Lipsey MW, Higgins JPT. Dealing with effect size multiplicity in systematic reviews and meta-analyses. Research Synthesis Methods 2018; 9 : 336–351.

For permission to re-use material from the Handbook (either academic or commercial), please see here for full details.

Overview of Summary Tables

• November 22, 2023 03:08

Summary tables typically describe the responses to each of the questions in a survey. They are the first step in understanding what the data means and in checking and cleaning data. 

Frequency tables

The simplest of all the summary tables shows the number of responses (cases) for each possible answer to a question in a survey. Such a table of counts  is shown below. These are known as frequency tables .

These tables can occasionally be useful - we return to this in our more detailed exploration of data cleaning. However, they are rarely the best place to start.

Summary tables showing percentages and means

The most useful summary tables are those that show percentages of people that choose each option, or if the data is numeric rather than categorical, the average. Examples of each are below.

Summaries of sets of variables

The left-most and right-most tables above look to have similar structures. However, if you carefully check the numbers, you will see that the different categories of race add up to 105%. Why is this? It is because this survey permitted people to choose multiple races.

In the raw data, sex is represented as a single variable (the first column below). By contrast, with the race data, we have five binary variables representing the data, one for each of the categories (as this is required to deal with people being permitted to choose multiple categories). The percentages that are shown in the table are the number of times the 1s appear in each of the variables. 

The sex and age summary tables above are summaries of single variables. The race table shows the data from a variable set of related variables. The race example is the most simple type of variable set that commonly appears in surveys .

Summaries of multiple categorical variables

In the table of raw data above, the two right-most columns represent how likely people said they would be to buy a product where the description contained no price, and, how likely they were to buy it after being shown the price.

This data is summarized in the table below. Here, each row summarizes the data of one of the columns. The percentages in the table are labeled as Row %, to make it easier for the reader to ascertain how to read the table.

Summary tables of variable sets that create a table with multiple rows and columns are often called  grids,  where the term can also be used for the question used to collect the data as well.

Grids of binary variables

The table below shows brand imagery for different cola brands. For example, we can see that 6% of people regard Coke as Feminine, 2% regard Coke as Health-conscious, etc. At first glance, the table below looks similar to the one above. However, the underlying data is completely different. Whereas the table above summarized six categorical variables, the table below summarizes 63 binary variables. 

The topic is discussed in more detail in Interpreting Grids of Binary Variables .

Grids of numeric variables

Summary tables of numeric variables can also be constructed as grids. The table below shows consumption per week of different brands in two locations,  'out and about'  and 'at home' .

The basic structure and interpretation of these tables are the same as with binary variables, except that the comparison is of averages rather than percentages and the nets are sums rather than averages. 

With grids of binary variables the default statistic, the percentage is usually the most interesting statistic. With numeric grids, it can be the case that other statistics are more useful. When there are outliers, more robust statistics like the  median can be better. And, sometimes it is more relevant to convert the data into percentages. The table below shows percentages rather than averages, and this makes it more readily interpretable. For example, we can see that:

• 68% of the consumption is 'at home'.
• Coca-Cola  and  Coke Zero  make up more than half the market (i..e., 31% + 29%).

Summary tables and data structures

In the summary tables above, multiple different statistics have been shown (e.g., counts, row %, and the average). In older data analysis programs, such as SPSS Statistics, the user chooses which statistic to display when. For example, the counts are produced using Analyze > Descriptive Statistics > Frequencies , the means via  Analyze > Statistics > Descriptives, and the row percentages like the ones above using Analyze > Custom Tables .

More modern analysis programs, such as Q , Displayr , R, and Tableau, instead automatically select the best statistic to show based on the structure of the data. For example, in Displayr, if the user creates a table of a nominal variable set, percentages are shown, and if the table contains numeric data it automatically shows means. In such programs, the user changes the setup of the data to generate the desired summary table.

Generating lots of summary tables

Typically, the analysis starts by generating lots of summary tables automatically and then reading through them. 

Related articles

• Frequency Tables
• Optimizing Product Ranges using Correspondence Analysis
• Understanding Nets and Sums on Tables
• Best Market Research Tools & Software in 2024
• Banner Tables

Please sign in to leave a comment.

MAXQDA 24 Manual

• MAXQDA Manual
• Other versions of MAXQDA
• How to Start the Program
• Project File Backups
• The Main Menu and the Four Main Windows
• The "Document System"
• The "Document Browser"
• The "Code System"
• The "Retrieved Segments" Window
• Managing Table Overviews
• General Preferences
• Supported Data Types
• Import and Group your Data
• Text Documents
• Create New Text or Table Documents
• External Files
• PDF Documents
• Image Documents
• Table Documents
• Audio and Video
• Transcripts
• Focus Group Transcripts
• Survey Data from Excel
• Survey Data from SurveyMonkey
• Survey Data from SPSS
• YouTube Data
• Subtitle Data (SRT)
• Bibliographic Data (Endnote, Zotero etc.)
• Structured Documents (Preprocessor)
• View Documents
• Edit Text Documents and Tables
• Line Numbers in Text Documents
• Assign a Color to a Document
• The Document Properties Window
• About Codes and the Code System
• Create a New Code
• Edit Code Properties
• Edit the Code System
• Delete Codes
• Create Code Sets
• The Overview of Codes
• Transfer the "Code System" to Other Project
• Import Codes and Code Memos from Excel
• How to code
• Display of Coded Segments in the "Document Browser"
• Display of Coded Segments in the "Code System"
• Further Ways of Coding
• Open Coding Mode
• Color Coding (Highlighting)
• Emoticode: Coding with Emojis and Symbols
• Comments for Coded Segments
• Weight Scores for Coded Segments
• Intercoder Agreement
• The Smart Coding Tool
• Remove Code Assignments
• Modify Coded Segments
• Move or Copy all Coded Segments of a Code
• Compare Codes
• Code Frequencies
• Code Coverage
• Code Patterns
• The Code Explorer
• The Overview of Documents for a Code
• Frequency Tables and Charts for Top-Level Codes and Subcodes
• The Idea behind Creative Coding
• Start Creative Coding and Select Codes
• Organize Codes
• Quit Creative Coding and Transfer Changes to Code System
• Purchase and Redeeem Transcription Time
• MAXQDA Transcription within MAXQDA
• MAXQDA Transcription via Webpage
• Transcribe Manually
• Link Transcript to Recording with Timestamps
• Play Transcript and Recording Synchronously
• The Overview of Timestamps
• About Memos
• Opening and Editing Memos
• Opening Multiple Memos at Once
• Memos in Documents
• Memos in the Document System Window
• The Memo Manager
• The Overview of Memos
• Search within Memos
• Linking Memo Content with Content in Documents
• Link Memos to Code(s) or Coded Segments
• Convert Memos into Documents
• The Logbook
• Links in MAXQDA
• Internal Links
• External Links
• The Overview of Links
• Local Text Search
• Global Text Search
• Search Results
• Autocode Search Results
• Export Search Results
• Extended Text Search
• The Word Explorer
• Activation: The Principle of Segment Retrieval
• The Overview of Coded Segments
• Export and Print Retrieved Segments
• Code Retrieved Segments
• Simple Coding Query
• Complex Coding Query
• Complex Retrieval Functions
• Include Subcodes in Retrieval
• Use Weight Filter
• Code Matrix Browser (Codes by Cases)
• Code Relations Browser (Code Co-occurrences)
• Code Map (Similarity of Codes)
• Document Map (Similarity of Documents)
• Document Comparison Chart (Sequence of Codings)
• Profile Comparison Chart (Code Frequencies and Variable Values)
• Word Trends (Frequencies of Words)
• Document Portrait (Sequence of Codings)
• Codeline (Score of Codings)
• Word Cloud (Frequencies of Words)
• Code Cloud (Frequencies of Codes)
• Code Trends (Frequencies of Codes)
• What Does MAXMaps Do?
• The MAXMaps Interface
• Create a Map and Add Objects
• Customize Objects in MAXMaps
• The Object Library
• Linking Objects
• Print, Export, and Organize Maps
• Synchronize Map With Project Data
• Code Co-Occurrence Model
• Code Theory Model
• Code-Subcodes-Segments Model
• Single-Case Model (Coded Segments)
• > Single-Code Model (Coded Segments)
• Single-Case Model (Focus Group Speaker)
• Single-Case Model (Code Hierarchy)
• Two-Cases Model
• Hierarchical Code-Subcodes Model
• Single-Case Model (Summaries)
• Single-Case Model (Paraphrases)
• Single-Code Model (Summaries)
• Code Distribution Model
• Options in MAXMaps
• About Paraphrasing in MAXQDA
• Paraphrase Text Passages and Image Segments
• Paraphrase Video and Audio Segments
• The “Paraphrased Segments” Code in Your “Code System”
• Categorize Paraphrases
• Paraphrases Matrix
• The Overview of Paraphrases
• The Overview of Paraphrased Segments
• Summary Grid (Create and Edit Summaries)

Summary Tables (Compile and Present Summaries)

• Summary Explorer (Compare Summaries for Cases or Groups)
• Code and Document Summaries
• Overview of Summaries
• Qualitative Comparison (Coded Segments)
• Quantitative Comparison (Code Frequencies)
• Document and Code Variables in MAXQDA
• Create and Edit Variables
• The Data Editor
• Importing Data from Excel and SPSS
• Exporting Data to Excel and SPSS
• Displaying Document Variables for Documents and Coded Segments
• Frequency Tables and Charts for Variables
• About Mixed Methods in MAXQDA
• Activate Documents by Variables
• Segment Matrix
• Typology Table
• Similarity Analysis for Documents
• Side-by-side Display of Results
• Qualitative Themes by Quantitative Groups
• Statistics by Qualitative Groups
• Transform a Code into a Document Variable (Quantitizing)
• Transform a Code into a Categorical Document Variable
• What Does MAXQDA Stats offer?
• Using MAXQDA Stats
• The Variable List
• The Output Viewer
• Analyze Variables and Codes from a MAXQDA Project - Start MAXQDA Stats
• Work with External Data in SPSS or Excel Format - Open SPSS data
• Transform Data
• Frequency Tables
• Analyze Multiple Responses
• Analyze Matrix Questions
• Descriptive Statistics
• t-Test and U-Test
• Hierarchical Cluster Analysis
• One-Way Analysis of Variance
• Correlation
• Limits and Technical Notes
• What Does MAXDictio Offer?
• Basic Terms
• Word Frequencies: Analyze Word Frequencies
• Word Frequencies: Table of Results
• Word Frequencies for Words in the Go Word List
• Word Frequencies for Dictionary Words
• Lemmatization
• Stop Word Lists
• Go Word Lists
• Keyword-in-Context
• Word Combinations
• Word Matrix Browser
• Managing Dictionaries
• Schematic Course
• Determine Frequency of Dictionary Categories
• Results Table
• Analyze Structured Texts
• The Category Matrix Browser
• Category Trends
• Autocode Documents with Dictionary Categories
• Topic Modeling
• Analyzing Japanese texts
• Limits to MAXDictio
• The Idea behind "Questions - Themes - Theories"
• Create a Worksheet
• Add Elements to a QTT Worksheet
• Export Worksheet
• Video Analysis with MAXQDA
• Import Videos
• The Multimedia Browser
• Code Videos
• Transcribe Video Clips
• Paraphrase Videos
• Add Memos and Links to Videos
• Export Video Clips
• Analyze Coded Videos
• How Does MAXQDA Support Focus Group Data Analysis?
• Text Search in Speaker’s Contributions
• Coding Query for Coded Segments
• View Coded Segments per Speaker
• The Overview of Focus Group Speakers
• Crosstab for Focus Groups
• Categorize Survey Data
• Search and Autocode Survey Responses
• Autocode Survey Responses with a Dictionary
• Analyze Survey Data with MAXQDA Stats
• Analyze Sentiments in Survey Responses
• Analyze Tweets
• Autocode Tweets with Hashtags and Authors
• Analyze Sentiments of Tweets
• Smart Publisher for Coded Segments
• Codebook with Category Definitions
• Summaries with Coded Segments
• Project Information
• Document Profiles
• Export and Print Documents
• Export Adjacency Matrix for Network Analysis
• Exporting Project Components as Excel file
• Anonymize Projects
• Save Project in MAXQDA 2022 Format
• Export and Import REFI-QDA Projects
• Archiving for Data Reuse
• How Does MAXQDA Support Teamwork?
• Copy Codings, Memos, Variables, etc. Into Another Project
• Hand over a MAXQDA Project to Other Team Members
• Merge Two MAXQDA Projects
• User Access Management
• What is MAXQDA TeamCloud?
• Requirements & License Management
• Log into TeamCloud from within MAXQDA
• Project Roles & Permissions
• LEAD: Initiate team project & start cycle 1
• MEMBER: Download and/or open project
• MEMBER: Upload project for team lead
• LEAD: Import teamwork data
• LEAD: Start new cycle or end project
• TeamCloud Storage
• Best Practice – Things to know for optimal teamwork
• The dashboard
• Detail page: Project > Files
• Detail page: Project > Team members
• Detail page: Project > Logbook
• Team overview
• Backup of TeamCloud Project
• Switching from MAXQDA 2022 to MAXQDA 24
• What is AI Assist?
• AI Summary: Document Summary
• AI Summary: Highlighted Text
• AI Summary: Coded Segment
• AI Summary: Code Summary
• AI Summary: Summary Grid
• AI Paraphrase
• AI Subcodes Suggestions
• AI New Codes Suggestions
• AI Explain This
• AI Chat With One Document
• AI Chat with Coded Segments
• Accessibility
• Checking for Updates
• Keyboard Shortcuts
• File Management
• Limitations and Technical Information

A Summary Table is an overview table of summaries and document variables for selected documents and codes. It serves as a compilation of the summaries for selected topics. Summary Tables are a useful tool for case-based and cross-case analyses, as well for presentations and publications.

To create Summary Tables and to view already created iones, click the Summary Tables tab in either

• the Analysis menu tab of the main MAXQDA window or
• the Start menu tab of the Summary Grid window.

This will open the Summary Tables window:

To create your first Summary Table, click the button in the middle of the window. If you already have created a Summary Table, you can click the New Summary Table button in the toolbar. A dialog window will appear:

• Using the tab “Choose documents and codes”, you must decide whether to include all documents or only activated documents. Usually, before creating a new table, you will activate the documents for which you have written summaries.
• On the same tab, you must codes to be included in the Summary Table. Usually, you will only select codes for which you have written summaries.
• Using the tab “Choose variables”, you can optionally choose which document variables should be output as supplementary information. All variables that are placed in the “Variables for first column” pane will appear in the first column with the document name. All variables that are placed in the “Variables in own column” pane will appear in additional columns next to the codes.

By clicking OK , the table with the selected documents, codes, and variables will be created:

The selected documents form the rows of the Summary Table and the selected codes form the columns. In the code columns, the summaries for the selected codes are listed for each document. The first column indicates the respective document by presenting its document group and document name.

If you have selected document variables for the first column, their variable values are displayed underneath each document name. If you have selected variables to be displayed in own columns, additional columns will be added to the table using the variable name as the column header. You can click on such a column head to easily sort the table by the variable values.

Additionally to sorting the table by clicking on a column header, you can use the following functions to adjust the view of the table:

• You can change the order of columns by clicking and dragging with the mouse.
• You can hide or display columns (right-click on the column header and choose Select columns ).
• You can order columns alphabetically by clicking on a column header.

The table is in plain text format, so the entire table is formatted with the same font.

Retrieving the Coded Segments Associated with a Summary

Each cell of the Summary Table is connected to the summarized coded segments. These can be retrieved as a list by right-clicking in the cell of the Summary Table and choosing Display Summarized Coded Segments .

Display coded segments instead of summaries

You can display the underlying summarized coded segments for a code instead of the self-written summaries, for example, because the coded segments are very short and they should be integrated in the table directly. To do this, right-click on the column header and select the Display Coded Segments option. Please note that it is not necessary that you have written summaries for a code to use this option. That is, when creating the table, you can also select codes for which no summaries exist and then switch the corresponding column in the Summary table to show the coded segment instead of empty cells.

You can switch back to view the summaries at any time by clicking the column header again and this time selecting Display Summaries .

The next time you open the MAXQDA project, the summaries (not the coded segments) will be displayed for all code columns.

Highlight documents according to their color

If you select the Highlight Rows in Document Color option in the Start menu tab, the rows will be highlighted according to the assigned document color. This function is very useful for grouping cases, for example, in the context of creating typologies: you read the content of each row, compare it with the other rows, and give the row a corresponding color to assign it to a newly formed or existing group. The colored background then makes it easy to recognize cases that belong together. In the simplest case, two colors are used, for example, red for ‘negative’ or ‘low’ and green for ‘positive’ or ‘high’.

To change the color of a document, that is, the background color of a row, right-click in the first column on the respective document icon:

If you change the color of a document in a Summary Table, the document color in the "Document System" will also be changed – and vice versa.

Saving the Summaries in a Column as a New Document Variable

Let’s say one column of your Summary Table contains short and similar summaries for a topic in an interview study. This analytic work can be used as a document variable in MAXQDA, for example, to activate interviewees for whom you have written the same summary: right-click on the column heading and choose Transform into Document Variable . MAXQDA will create a new document variable that contains the summaries as values and that will be named as the column. If the variable already exists, MAXQDA will add a number to the name. Please note that the values will be shortened to 63 characters. The column in the Summary Table remains.

Adding an Empty Column for Additional Summaries

Let’s assume when working with an existing Summary Table you would like to add a new column in which you would like to write integrative summaries (as described, for example, in Kuckartz & Rädiker , 2019, pp. 147). For this purpose, it is necessary to add another code to the "Code System", which will then serve as a "container" for the integrating summaries. Click on New Code in the Start menu tab to create a new code and add it directly to the currently displayed summary table as a new column for additional summaries.

Swap Rows and Columns of a Summary Table

You can use the icon Swap Rows and Column in the menu tab of the Summary Table to rotate the view by 90°, which means you swap the rows and the columns for display. The documents will be displayed in the columns then and the codes (and the variables with their own column) in the rows.

Open a Summary Table

To open an existing Summary Table, click on My Summary Tables on the menu tab and select a table.

Rename a Summary Table

To rename the displayed Summary Table, click on the name displayed above the table on the left:

Delete a Summary Table

To delete the displayed summary table, click Delete Summary in the menu tab. This operation cannot be undone. Deleting a table does not affect its contents, that is, the summaries and variable values are still present in the project and the table can be recreated if required.

Store Summary Table in a QTT Worksheet

Insert summary table as a table document in the “document system”.

The displayed Summary Table can be stored as a table document so that you can then analyze its contents with all the available MAXQDA functions. To do this, click on Export in the menu tab at the top right and select the option Insert Summary Table as Table Document into Document System .

Export Summary Table

The displayed summary table can be exported as a Word document or Excel table. To do this, click on Export in the menu tab at the top right and then select the corresponding entry.

Was this article helpful?

Table Setup

Tables are visual displays composed of columns and rows in which numbers, text, or a combination of numbers and text are presented. There are many common kinds of tables, including demographic characteristics tables, correlation tables, factor analysis tables, analysis of variance tables, and regression tables.

This page addresses the basics of table setup, including table components, principles of table construction (including the use of borders and how to handle long or wide tables), and placement of tables in the paper. Note that tables and figures have the same overall setup.

View the sample tables to see these guidelines in action.

Table components

APA Style tables have the following basic components:

• number: The table number (e.g., Table 1) appears above the table title and body in bold font. Number tables in the order in which they are mentioned in your paper.
• title: The table title appears one double-spaced line below the table number. Give each table a brief but descriptive title, and capitalize the table title in italic title case .
• headings: Tables may include a variety of headings depending on the nature and arrangement of the data. All tables should include column headings, including a stub heading (heading for the leftmost, or stub, column). The heading “Variable” is often used for the stub column if no other heading is suitable. Some tables also include column spanners, decked heads, and table spanners; these are described in the Publication Manual . Center column headings and capitalize them in sentence case .
• The table body may be single-spaced, one-and-a-half-spaced, or double-spaced.
• Left-align the information in the leftmost column or stub column of the table body (but center the heading).
• In general, center information in all other cells of the table. However, left-align the information if doing so would improve readability, particularly when cells contain lots of text.
• note: Three types of notes (general, specific, and probability) appear below the table as needed to describe contents of the table that cannot be understood from the table title or body alone (e.g., definitions of abbreviations, copyright attribution, explanations of asterisks used to indicate p values). Include table notes only as needed.

This diagram is an illustration of the basic table components.

Tables are covered in the seventh edition APA Style manuals in the Publication Manual Sections 7.8 to 7.21 and the Concise Guide Sections 7.8 to 7.21

Related handout

• Student Paper Setup Guide (PDF, 3MB)

Principles of table construction

The most important principle to follow when creating a table is to present information in a way that is easy for readers to understand. Provide sufficient information in the table itself so that readers do not need to read the text to understand it.

When creating a table, place entries that are to be compared next to each other. In general, place different indices (e.g., means and standard deviations) in different columns rather than in the same column. Use the same font in tables as in the rest of your paper.

Use the tables feature of your word-processing program to create tables in APA Style papers. Do not use the tab key or space bar to manually create the look of a table.

Table borders

Limit the use of borders or lines in a table to those needed for clarity. In general, use a border at the top and bottom of the table, beneath column headings (including decked heads), and above column spanners. You may also use a border to separate a row containing totals or other summary information from other rows in the table.

Do not use vertical borders to separate data, and do not use borders around every cell in a table. Use spacing between columns and rows and strict alignment to clarify relations among the elements in a table.

Long or wide tables

If a table is longer than one page, use the tables feature of your word-processing program to make the headings row repeat on the second and any subsequent pages. No other adjustments are necessary. If a table is too wide to fit on one page, use landscape orientation on the page with the wide table. It does not matter if the page header also moves when switching to landscape orientation.

Placement of tables in a paper

There are two options for the placement of tables (and figures) in a paper. The first is to embed tables in the text after each is first mentioned (or “called out”); the second is to place each table on a separate page after the reference list.

An embedded table may take up an entire page; if the table is short, however, text may appear on the same page as the table. In that case, place the table at either the top or bottom of the page rather than in the middle. Also add one blank double-spaced line between the table and any text to improve the visual presentation.

View the sample tables for more information on tables.

• Open access
• Published: 06 August 2024

Adaptation and validation of the evidence-based practice profile (EBP 2 ) questionnaire in a Norwegian primary healthcare setting

• Nils Gunnar Landsverk 1 ,
• Nina Rydland Olsen 2 ,
• Kristine Berg Titlestad 4 ,
• Are Hugo Pripp 3 &
• Therese Brovold 1  

BMC Medical Education volume  24 , Article number:  841 ( 2024 ) Cite this article

203 Accesses

Metrics details

Access to valid and reliable instruments is essential in the field of implementation science, where the measurement of factors associated with healthcare professionals’ uptake of EBP is central. The Norwegian version of the Evidence-based practice profile questionnaire (EBP 2 -N) measures EBP constructs, such as EBP knowledge, confidence, attitudes, and behavior. Despite its potential utility, the EBP 2 -N requires further validation before being used in a cross-sectional survey targeting different healthcare professionals in Norwegian primary healthcare. This study assessed the content validity, construct validity, and internal consistency of the EBP 2 -N among Norwegian primary healthcare professionals.

To evaluate the content validity of the EBP 2 -N, we conducted qualitative individual interviews with eight healthcare professionals in primary healthcare from different disciplines. Qualitative data was analyzed using the “text summary” model, followed by panel group discussions, minor linguistic changes, and a pilot test of the revised version. To evaluate construct validity (structural validity) and internal consistency, we used data from a web-based cross-sectional survey among nurses, assistant nurses, physical therapists, occupational therapists, medical doctors, and other professionals ( n  = 313). Structural validity was tested using a confirmatory factor analysis (CFA) on the original five-factor structure, and Cronbach’s alpha was calculated to assess internal consistency.

The qualitative interviews with primary healthcare professionals indicated that the content of the EBP 2 -N was perceived to reflect the constructs intended to be measured by the instrument. However, interviews revealed concerns regarding the formulation of some items, leading to minor linguistic revisions. In addition, several participants expressed that some of the most specific research terms in the terminology domain felt less relevant to them in clinical practice. CFA results exposed partial alignment with the original five-factor model, with the following model fit indices: CFI = 0.749, RMSEA = 0.074, and SRMR = 0.075. Cronbach’s alphas ranged between 0.82 and 0.95 for all domains except for the Sympathy domain (0.69), indicating good internal consistency in four out of five domains.

The EBP 2 -N is a suitable instrument for measuring Norwegian primary healthcare professionals’ EBP knowledge, attitudes, confidence, and behavior. Although EBP 2 -N seems to be an adequate instrument in its current form, we recommend that future research focuses on further assessing the factor structure, evaluating the relevance of the items, and the number of items needed.

Registration

Retrospectively registered (prior to data analysis) in OSF Preregistration. Registration DOI: https://doi.org/10.17605/OSF.IO/428RP .

Peer Review reports

Evidence-based practice (EBP) integrates the best available research evidence with clinical expertise, patient characteristics, and preferences [ 1 ]. The process of EBP is often described as following the five steps: ask, search, appraise, integrate, and evaluate [ 1 , 2 ]. Practicing the steps of EBP requires that healthcare professionals hold a set of core competencies [ 3 , 4 ]. Lack of competencies such as EBP knowledge and skills, as well as negative attitudes towards EBP and low self-efficacy, may hinder the implementation of EBP in clinical practice [ 5 , 6 , 7 , 8 , 9 , 10 ]. Measuring of EBP competencies may assist organizations in defining performance expectations and directing professional practice toward evidence-based clinical decision-making [ 11 ].

Using well-designed and appropriate measurement instruments in healthcare research is fundamental for gathering precise and pertinent data [ 12 , p. 1]. Access to valid and reliable instruments is also essential in the field of implementation science, where conducting consistent measurements of factors associated with healthcare professionals’ uptake of EBP is central [ 13 ]. Instruments measuring the uptake of EBP should be comprehensive and reflect the multidimensionality of EBP; they should be valid, reliable, and suitable for the population and setting in which it is to be used [ 14 ]. Many instruments measuring different EBP constructs are available today [ 15 , 16 , 17 , 18 , 19 , 20 , 21 , 22 ]. However, the quality of these instruments varies, and rigorous validation studies that aim to build upon and further develop existing EBP instruments are necessary [ 13 , 16 ].

The authors of this study conducted a systematic review to summarize the measurement properties of existing instruments measuring healthcare professionals’ EBP attitudes, self-efficacy, and behavior [ 16 ]. This review identified 34 instruments, five of which were translated into Norwegian [ 23 , 24 , 25 , 26 , 27 ]. Of these five instruments, only the Evidence-based practice profile questionnaire (EBP 2 ) was developed to measure various EBP constructs, such as EBP knowledge, confidence, attitudes, and behavior [ 28 ]. In addition, EBP 2 was developed to be trans-professional [ 28 ]. Although not exclusively demonstrating high-quality evidence for all measurement properties, the review authors concluded that the EBP 2 was among the instruments that could be recommended for further use and adaption for use among different healthcare disciplines [ 16 ].

EBP 2 was initially developed by McEvoy et al. in 2010 and validated for Australian academics, practitioners, and students from different professions (physiotherapy, podiatry, occupational therapy, medical radiation, nursing, human movement) [ 28 ]. The instrument was later translated into Chinese and Polish and further tested among healthcare professionals in these countries [ 29 , 30 , 31 , 32 ]. The instrument was also translated into Norwegian and cross-culturally adapted into Norwegian [ 27 ]. The authors assessed content validity, face validity, internal consistency, test-retest reliability, measurement error, discriminative validity, and structural validity among bachelor students from nursing and social education and health and social workers from a local hospital [ 27 ]. Although the authors established the content validity of the EBP 2 -Norwegian version (EBP 2 -N), they recommended further linguistic improvements. Additionally, while they found the EBP 2 -N valid and reliable for three subscales, the original five-factor model could not be confirmed using confirmatory factor analysis. Therefore, they recommended further research on the instrument measurement properties [ 27 ].

We recognized the need for further assessment of measurement properties of the EBP 2 -N before using this instrument in a planned cross-sectional survey targeting physical therapists, occupational therapists, nurses, assistant nurses, and medical doctors working with older people in Norwegian primary healthcare [ 33 ]. As our target population differed from the population studied by Titlestad et al. [ 27 ], the EBP 2 -N should be validated again, assessing content validity, construct validity and internal consistency [ 12 , p. 152]. The assessment of content validity evaluates whether the content of an instrument is relevant, comprehensive, and understandable for a specific population [ 34 ]. Construct validity, including structural validity and cross-cultural validity, can provide evidence on whether an instrument measures what it intends to do [ 12 , p. 169]. Furthermore, the degree of interrelatedness among the items (internal consistency) should be assessed when evaluating how items of a scale are combined [ 35 ]. Our objectives were to comprehensively assess content validity, structural validity, and internal consistency of the EBP 2 -N among Norwegian primary healthcare professionals. We hypothesized that the EBP 2 -N was a valid and reliable instrument suitable for use in Norwegian primary healthcare settings.

Study design

This study was conducted in two phases: Phase 1 comprised a qualitative assessment of the content validity of the EBP 2 -N, followed by minor linguistic adaptions and a pilot test of the adapted version. Phase 2 comprised an assessment of structural validity and internal consistency of the EBP 2 -N based on the result from a web-based cross-sectional survey.

The design and execution of this study adhered to the COSMIN Study Design checklist for patient-reported outcome measurement instruments, as well as the methodology for assessing the content validity of self-reported outcome measures [ 34 , 36 , 37 ]. Furthermore, this paper was guided by the COSMIN Reporting guidelines for studies on measurement properties of patient-reported outcome measures [ 38 ].

Participants and setting

Participants eligible for inclusion in both phases of this study were health personnel working with older people in primary healthcare in Norway, such as physical therapists, occupational therapists, nurses, assistant nurses, and medical doctors. Proficiency in reading and understanding Norwegian was a prerequisite for inclusion. This study is part of a project called FALLPREVENT, a research project that aims to bridge the gap between research and practice in fall prevention in Norway [ 39 ].

Instrument administration

The EBP 2 -N consists of 58 self-reported items that are divided into five different domains: (1) Relevance (items 1–14), which refers to the value, emphasis, and importance respondents place on EBP; (2) Sympathy (items 15–21) which refers to the perceived compatibility of EBP with professional work; (3) Terminology (items 22–38), which refers to the understanding of common research terms; (4) Practice (items 39–47), which refers to the use of EBP in clinical practice and; (5) Confidence (items 48–58), which relates to respondents perception of their EBP skills [ 28 ]. All the items are rated on a five-point Likert scale (1 to 5) (see questionnaire in Additional file 1 ). Each domain is summarized, with higher scores indicating a higher degree of the construct measured in the domain in question. The items in the Sympathy domain are negatively phrased and need to be reversed before being summarized. The possible range in summarized scores (min-max) per domain are as follows: Relevance (14–70), Sympathy (7-35) , Terminology (17–85), Practice (9-45) , and Confidence (11–55).

Phase 1: content validity assessment

Recruitment and participant characteristics.

Snowball sampling was used to recruit participants in Eastern Norway, and possible eligible participants were contacted via managers in healthcare settings. The number of participants needed for the qualitative content validity interviews was based on the COSMIN methodology recommendations and was set to at least seven participants [ 34 , 37 ]. We recruited and included eight participants. All participants worked with older people in primary healthcare, and included two physical therapists, two occupational therapists, two assistant nurses, one nurse, and one medical doctor. The median age (min-max) was 35 (28–55). Two participants held upper secondary education, four held a bachelor’s degree, and two held a master’s degree. Six participants reported that they had some EBP training from their education or had attended EBP courses, and two had no EBP training.

Qualitative interviews

Before the interviews, a panel of four members (NGL, TB, NRO, and KBT) developed a semi-structured interview guide. Two panel members were EBP experts with extensive experience in EBP research and measurement (NRO and KBT). KBT obtained consent from the developer of the original EBP 2 questionnaire and translated the questionnaire into Norwegian in 2013 [ 27 ].

To evaluate the content validity of the EBP 2 -N for use among different healthcare professionals working in primary healthcare in Norway, we conducted individual interviews with eight healthcare professionals from different disciplines. Topics in the interview guide were guided by the standards of the COSMIN study design checklist and COSMIN criteria for good content validity, which include questions related to the following three aspects [ 34 , 37 ]: Whether the items of the instrument were perceived relevant (relevance), whether all key concepts were included (comprehensiveness), and whether the instructions, items, and response options were understandable (comprehensibility) [ 34 ]. The interview guide is presented in Additional File 2 . Interview preparations and training included a review of the interview guide and a pilot interview with a physical therapist not included in the study.

Eight interviews were conducted by the first author (NGL) in May and June 2022. All interviews were conducted in the participant’s workplaces. The interviews followed a “think-aloud” method [ 12 , p. 58, 40 , p. 5]. Hence, in the first part of the interview, the participants were asked to complete the questionnaire on paper while simultaneously saying aloud what they were thinking while responding to the questionnaire. Participants also had to state their choice of answer aloud and make a pen mark on the items or responses that either were difficult to understand or did not feel relevant to them. In the second part of the interviews, participants were asked to elaborate on why items were marked as difficult to understand or irrelevant, focusing on relevance and comprehensibility. In addition, the participants were asked to give their overall impression of the instrument and state if they thought any essential items (comprehensiveness) were missing. Only the second part of the interviews were audio-recorded.

Analysis and panel group meetings

After conducting the individual interviews, the first author immediately transcribed the recorded audio data. The subsequent step involved gathering and summarizing participants’ comments into one document that comprised the questionnaire instructions, items, and response options. Using the “text summary” model [ 41 , p.61], we summarized the primary “themes” and “problems” identified by participants during the interviews. These were then aligned with the specific item or section of the questionnaire to which the comments were related. For example, comments on the items’ comprehensibility were identified as one “theme”, and the corresponding “problem” was that the item was perceived as too academically formulated or too complex to understand. Comments on an item’s relevance was another “theme” identified, and an example of a corresponding “problem” was that the EBP activity presented in the item was not recognized as usual practice for the participant. The document contained these specific comments and summarized the participants’ overall impression of the instrument. Additionally, it included more general comments addressing the instrument’s relevance, comprehensibility, and comprehensiveness.

Next, multiple rounds of panel group discussions took place, and the final document with a summary of participants’ comments served as the foundation for these discussions. The content validity of the items, instructions, and response options underwent thorough examinations by the panel members. Panel members discussed aspects, such as relevance, comprehensiveness, and comprehensibility, drawing upon insights from interview participants’ comments and the panel members’ extensive knowledge about EBP.

Finally, the revised questionnaire was pilot tested on 40 master’s students (physical therapists) to evaluate the time used to respond, and the students were invited to make comments in free text adjacent to each domain in the questionnaire. The pilot participants answered a web-based version of the questionnaire.

Phase 2: Assessment of structural validity and internal consistency

Recruitment and data collection for the cross-sectional survey.

Snowball sampling was used to recruit participants. The invitation letter, with information about the study and consent form, was distributed via e-mail to healthcare managers in over 37 cities and municipalities representing the eastern, western, central, and northern parts of Norway. The managers forwarded the invitation to eligible employees and encouraged them to respond to the questionnaire. The respondents that consented to participation automatically received a link to the online survey. Our approach to recruitment made it impossible to keep track of the exact number of potential participants who received invitations to participate. As such, we were unable to determine a response rate.

Statistical methods

Statistical analyses were performed using STATA [ 42 ]. We tested the structural validity and internal consistency of the 58 domain items of the EBP 2 -N, using the same factor structure as in the initial evaluation [ 28 ] and the study that translated the questionnaire into Norwegian [ 27 ]. Structural validity was assessed using confirmatory factor analysis with maximum likelihood estimation to test if the data fit the predetermined original five-factor structure. Model fit was assessed by evaluating the comparative fit index (CFI), root mean square error of approximation (RMSEA), and the standardized root mean square residual (SRMR). Guidelines suggest that a good-fitting model should have a CFI of around 0.95 or higher, RMSEA of around 0.06 or lower, and SRMR of around 0.08 or lower [ 43 ]. Cronbach’s alpha was calculated for each of the five domains to evaluate whether the items within the domains were interrelated. It has been proposed that Cronbach’s alpha between 0.70 and 0.95 can be considered good [ 44 ].

The sample size required for a factor analysis was set based on COSMIN criteria for at least an “adequate” sample size, which is at least five times the number of items and > 100 [ 45 , 46 ]. Accordingly, the sample size required in our case was > 290 respondents. Regarding missing data, respondents with over 25% missing items on domain items were excluded from further analysis. Respondents with over 20% missing on one domain were excluded from the analysis of that domain. The Little’s MCAR test was conducted to test whether data were missing completely at random. Finally, for respondents with 20% or less missing data on one domain, the missing values were substituted with the respondent’s mean of other items within the same domain.

Ethical approval and consent to participate

The Norwegian Agency for Shared Services in Education and Research (SIKT) approved the study in March 2022 (ref: 747319). We obtained written informed consent from the participants interviewed and the cross-sectional survey participants.

The findings for Phase 1 and Phase 2 will be presented separately. Phase 1 will encompass the results of the qualitative content validity assessment, adaptions, and pilot testing of the EBP 2 -N. Phase 2 will encompass the results of assessing the structural validity and internal consistency of the EBP 2 -N.

Phase 1: Results of the content validity assessment

Comprehensiveness: whether key concepts are missing.

Only a few comments were made on comprehensiveness. Notably, one participant expressed the need for additional items addressing clinical experience and user perspectives.

Relevance: whether the items are perceived relevant

Overall, the participants commented that they perceived the instrument as relevant to their context. However, several participants pointed out some items that felt less relevant. The terminology domain emerged as a specific area of concern, as most participants expressed that this subscale contained items that felt irrelevant to clinical practice. Comments such as “I do not feel it’s necessary to know all these terms to work evidence-based,” and “The more overarching terms like RCT, systematic review, clinical relevance, and meta-analysis I find relevant, but not the more specific statistical terms,” captured the participants’ perspectives on the relevance of the terminology domain.

Other comments related to the terminology domain revealed that these items could cause feelings of demotivation or inadequacy: “One can become demotivated or feel stupid because of these questions” and “Many will likely choose not to answer the rest of the form, as they would feel embarrassed not knowing”. Other comments on relevance were related to items in other subscales, for example, critical appraisal items (i.e., items 20, 42, and 55), which were considered less relevant by some participants. One participant commented: “If one follows a guideline as recommended, there is no need for critical assessment”.

Comprehensibility: Whether instructions, items, and response options are understandable

All eight participants stated that they understood what the term EBP meant. The predominant theme from the participant’s comments was related to the comprehensibility of the EBP 2 -N. Most of the comments on comprehensibility revolved around the formulation of items. Participants noted challenges related to comprehensibility in 35 out of 58 items, either due to difficulty in understanding, readability issues, the length of items, lack of clarity, or overly academic language. For instance, item five in the Relevance domain, “I intend to develop knowledge about EBP”, received comments that expressed uncertainty about whether “EBP” referred to the five steps of EBP or evidence-based clinical interventions/practices (e.g., practices following recommendations in evidence-based guidelines). Items that were perceived as overly academic included phrases such as “intend to apply”, “intend to develop”, or “convert your information needs”. For these phrases, participants suggested simpler formulations in layperson’s Norwegian. Some participants deemed the instrument “too advanced,” “on a too high level,” or “too abstract”, and others expressed that they understood most of the instrument’s content, indicating a divergence among participants.

Examples of items considered challenging to read, too complex, or overly lengthy were items six and 12 in the relevance domain, 16 and 20 in the sympathy domain, and 58 in the confidence domain. The typical comments from participants revealed a preference for shorter, less complex items with a clear and singular focus. In addition, some comments referred to the formulation of response options. For instance, two response options in the confidence domain, “Reasonably confident” and “Quite confident”, were perceived as too similar in Norwegian. In the practice subscale, a participant pointed out that the term “monthly or less” lacked precision, as it could cover any frequency from once to twelve times a year, thus being perceived as imprecise.

Panel group meetings and instrument revision

The results of the interviews were discussed during several rounds of panel group meetings. After thoroughly examining the comments, 33 items underwent revisions during the panel meetings. These revisions primarily involved minor linguistic adjustments to preserve the original meaning of the items. For example, the Norwegian version of item 8 was considered complex and overly academically formulated and underwent revision. The phrase “I intend to apply” was replaced by “I want to use”, as the panel group considered this phrase easier to understand in Norwegian. Another example involved the term “Framework,” which some participants found vague or difficult to understand (i.e., in item 3, “my profession uses EBP as a framework”). The term “framework” was replaced with “way of thinking and working”, considered more concrete and understandable in Norwegian. The phrase “way of thinking and working” was also added to item 5 to clarify that “EBP” referred to the five steps of EBP, not interventions in line with evidence-based recommendations. Additionally, it was challenging to revise items that participants considered challenging to read, too complex, or overly lengthy (i.e., 6, 12, 16, 20, and 58), as it was difficult to shorten them without losing their original meaning. However, replacing overly academic words with simpler formulations made these examples less complex and more readable.

In terms of relevance of the items, no items were removed, and the terminology domain was retained despite comments regarding its relevance. Changing this domain would have impeded the opportunity to compare results from future studies using this questionnaire with previous studies using the same questionnaire. Regarding comprehensiveness, the panel group reached a consensus that the domains included all essential items concerning the constructs that the original instrument states to measure. Further, examples of minor linguistic changes and additional details on item revisions are reported in Additional File 3 .

The median time to answer the questionnaire was nine minutes. Students made no further comments to the questionnaire.

Participants’ characteristics and mean domain scores

A total of 313 responded to the survey. The respondents’ mean age (SD) was 42.7 years (11.4).The sample included 119 nurses, 74 assistant nurses, 64 physical therapists, 38 occupational therapists, three medical doctors, and 15 other professionals, mainly social educators. In total, 63.9% ( n  = 200) of the participants held a bachelor’s degree, 11.8% ( n  = 37) held a master’s degree, and 0.3% ( n  = 1) held a Ph.D. Moreover, 10.5% ( n  = 33) of the participants had completed upper secondary education, and 13.1% ( n  = 41) had tertiary vocational education. One hundred and eighty-five participants (59.1%) reported no formal EBP training, while among the 128 participants who had undergone formal EBP training, 31.5% had completed over 20 h of EBP training. The mean scores (SD) for the different domains were as follows: Relevance 80.2 (7.3), Sympathy 21.2 (3.6), Terminology 44.5 (15.3), Practice 22.2 (5.8), and Confidence 31.2 (9.2).

Missing data

Out of 314 respondents, one was excluded due to over 25% missing domain items, and three were excluded due to more than 20% missing data in specific domains. Twenty-six respondents had under 20% missing data on one domain, and these missing values were substituted with the respondent’s mean of the other items within the same domain. In total, 313 responses were included in the final analysis. Each domain item had at most 1.3% missing items in total. The percentage of missing data per domain was low and relatively similar across the five domains ( Relevance  = 0.05%, Sympathy  = 0.2%, Terminology  = 0.4%, Practice  = 0.6%, Confidence  = 0.6%). The Little’s MCAR test showed p-values higher than 0.05 for all domains, indicating that data was missing completely at random.

Structural validity results

A five-factor model was estimated based on the original five-factor structure (Fig.  1 ). The model was estimated using the maximum likelihood method. A standardized solution was estimated, constraining the variance of latent variables to 1. Correlation among latent variables was allowed. The results of the CFA showed the following model fit indices: CFI = 0.749, RMSEA = 0.074, and SRMR = 0.075. The CFI and RMSEA results did not meet the criteria for a good-fitting model set a priori (CFI of around 0.95 or higher, RMSEA of around 0.06 or lower). However, the SRMR value met the criteria around 0.08 or lower. All standardized factor loadings were over 0.32, and only five items loaded under 0.5. The range of standardized factor loadings was the following in the different domains: Relevance  = 0.47–0.79; Terminology  = 0.51–0.80; Practice  = 0.35–0.70, Confidence  = 0.43–0.86, and Sympathy  = 0.32–0.65 (Fig.  1 ).

Confirmatory factor analysis, standardized solution of the EBP2-N. ( n  = 313). Note: Large circles = latent variables, Rectangles = measured items, small circles = residual variance

Internal consistency results

As reported in Table  1 , Cronbach’s alphas ranged between 0.82 and 0.95 for all domains except for the Sympathy domain, where Cronbach’s alpha was 0.69. Results indicate good internal consistency for four domains and close to the cut-off of good internal consistency (> 0.70) on Sympathy.

In this study, we aimed to assess the measurement properties of the EBP 2 -N questionnaire. The study population of interest was healthcare professionals working with older people in Norwegian primary healthcare, including physical therapists, occupational therapists, nurses, assistant nurses, and medical doctors. The study was conducted in two phases: content validity was assessed in Phase 1, and construct validity and internal consistency were assessed in phase 2.

The findings from Phase 1 and the qualitative interviews with primary healthcare professionals indicated that the content of the EBP 2 -N was perceived to reflect the constructs intended to be measured by the instrument [ 28 ]. However, the interviews also revealed different perceptions regarding the relevance and comprehensibility of certain items. Participants expressed concerns about the formulation of some items, and we decided to make minor linguistic adjustments, aligning with previous recommendations to refine item wording through interviews [ 27 ]. Lack of content validity can have adverse consequences [ 34 ]. Irrelevant or incomprehensible items may make respondents tired of answering, leading to potentially biased answers [ 47 , 48 , p. 139]. Analysis of missing data showed that possible irrelevant or incomprehensible items did not lead to respondent fatigue, as the overall percentage of missing items was low (at most 1.3%), and the percentage of missing data did not vary across the domains. Irrelevant items may also impact other measurement properties, such as structural validity and internal consistency [ 34 ]. We believe that the minor linguistic revisions we made to some items made the questionnaire easier to understand. This assumption was supported by the pilot test of 40 master’s students, where no further comments regarding comprehensibility were added.

The overall relevance of the instruments was perceived positively. However, several participants expressed concerns about the terminology domain as some of the most specific research terms felt irrelevant to them in clinical practice. Still, the panel group decided to keep all items in the terminology domain to allow comparison of results among future studies on the same instrument and subscales. In addition, this decision was based on the fact that knowledge about research terminology, such as “types of data,” “measures of effect,” and “statistical significance,” are essential competencies to perform step three of the EBP process (critical appraisal) [ 3 ]. Leaving out parts of the terminology domain could, therefore, possibly make our assessment of the EBP constructs less comprehensive and complete [ 14 ]. However, since the relevance of some items in the terminology domain was questioned, we cannot fully confirm the content validity of this domain, and we recommend interpreting it with caution.

The confirmatory factor analysis (CFA) in Phase 2 of this study revealed that the five-factor model only partially reflected the dimensionality of the constructs measured by the instrument. The SRMR was the only model fit indices that completely met the criteria for a good-fitting model set a priori, yielding a value of 0.075. In contrast, the CFI at 0.749 and RMSEA at 0.074 fell short of the criteria for a good-fitting model (CFI ≥ 0.95, RMSEA ≤ 0.06). However, our model fit indices were closer to the criteria for a good-fitting model compared to Titlestad et al. (2017) [ 27 ] who demonstrated a CFI of 0.69, RMSEA of 0.089, and SRMR of 0.095. This tendency toward better fit in our study may be related to the larger sample size, in agreement with established recommendations of a minimum of 100–200 participants and at least 5–10 times the number of items to ensure the precision of the model and overall model fit [ 46 , p. 380].

Although our sample size met COSMIN’s criteria for an “adequate” sample size [ 45 ], the partially adequate fit indices suggest that the original five-factor model might not be the best-fitting model. A recent study on the Chinese adaptation of the EBP 2 demonstrated that item reduction and using a four-factor structure improved model fit (RMSEA = 0.052, CFI = 0.932) [ 30 ]. The same study removed eighteen items based on content validity evaluation (four from relevance , seven from terminology , and seven from sympathy ) [ 30 ]. In another study where the EBP 2 was adapted for use among Chinese nurses, thirteen items (two from sympathy , eight from terminology , one from practice , and two from confidence ) were removed, and an eight-factor structure was identified [ 29 ]. However, compared to our study, noticeably improved model fit was not demonstrated in this study [ 29 ]. The model fit indices of their 45-item eight-factor structure were quite similar to the one found in our study (RMSEA = 0.065, SRMR = 0.077, CFI = 0.884) [ 29 ]. The results from the two above mentioned studies suggest that a model including fewer items and another factor structure potentially could have applied to our population as well. Although the five-factor model only partially reflects the constructs measured by the EBP 2 -N in our population, it contributes valuable insights into the instrument’s performance in a specific healthcare setting.

Cronbach’s alpha results in this study indicate good internal consistency for four domains, being over 0.82. However, the alpha of 0.69 in the sympathy did not reach the pre-specified cut-off of good internal consistency (0.70) [ 44 ]. A tendency of relatively lower Cronbach’s alpha values on the sympathy domain, compared to the other four domains, has also been identified in previous similar studies [ 27 , 28 , 31 , 32 ]. Titlestad et al. (2017) reported Cronbach’s alpha to be 0.66 in the sympathy domain and above 0.90 in the other domains [ 27 ]. McEvoy et al. (2010), Panczyk et al. (2017), and Belowska et al. (2020) reported Cronbach’s alphas of 0.76–0.80 for the sympathy domain, and 0.85–0.97 for the other domains [ 28 , 31 , 32 ]. In these three cases, Cronbach’s alphas of the sympathy domain were all over 0.70, but the same tendency of this domain demonstrating lower alphas than the other four domains was evident. The relatively lower alpha values in the sympathy domain may be related to the negative phrasing of items [ 49 ], the low number of items in this domain compared to the others ( n  = 7) [ 12 , p. 84, 47 , p. 86], and a possible heterogeneity in the construct measured [ 47 , p. 232]. The internal consistency results of our study indicate that the items in the sympathy domain are less interrelated than the other domains. However, having a Cronbach’s alpha value of 0.69 indicates that the items do not entirely lack interrelatedness.

Limitations

Methodological limitations that could potentially introduce bias into the results should be acknowledged. Although the eight participants involved in the qualitative content validity interviews in Phase 1 covered all healthcare disciplines and education levels aimed to be included in the survey in Phase 2, it remains uncertain whether these eight participants demonstrated all potential variations in the population of interest. It is possible that those that agreed to participate in qualitative interviews regarding an EBP instrument held more positive attitudes toward EBP than the general practitioner would do. Another possible limitation pertains to the qualitative interviews and the fact that the interviewer (NGL) had limited experience facilitating “think-aloud” interviews. To reduce the potential risk of bias related to the interviewer, the panel group with extensive experience in EBP research took part in the interview preparation, and a pilot interview was conducted before the interviews to ensure training.

Furthermore, using a non-random sampling method and the unknown response rate in Phase 2 may have led to biased estimates of measurement properties and affected the representativeness of the sample included. Additionally, the characteristics of non-responders remain unknown, making it challenging to assess whether they differ from the responders and if the final sample adequately represents the variability in the construct of interest. Due to potential selection bias and non-response bias, there may be uncertainty regarding the accuracy of the measurement property assessment and whether the study sample fully represents the entire population of interest [ 50 , p. 205].

Conclusions

The EBP 2 -N is suitable for measuring Norwegian primary healthcare professionals’ EBP knowledge, attitudes, confidence, and behavior. Researchers can use the EBP 2 -N to increase their understanding of factors affecting healthcare professional’s implementation of EBP and to guide the development of tailored strategies for implementing EBP.

This study revealed positive perceptions of the content validity of the EBP 2 -N, though with nuanced concerns about the relevance and comprehensibility of certain items and uncertainty regarding the five-factor structure of the EBP 2 -N. The minor linguistic revisions we made to some items made the questionnaire more understandable. However, when EBP 2 -N is used in primary healthcare, caution should be exercised when interpreting the results of the terminology domain, as the relevance of some items has been questioned.

Future research should focus on further assessing the factor structure of the EBP 2 -N, evaluating the relevance of the items, and exploring the possibility of reducing the number of items, especially when applied in a new setting or population. Such evaluations could further enhance our understanding of the instrument’s measurement properties and potentially lead to improvements in the measurement properties of the EBP 2 -N.

Data availability

The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.

Abbreviations

• Evidence-based practice

The Evidence-based practice profile

The Norwegian version of the Evidence-based practice profile questionnaire

Consensus-based Standards for the Selection of Health Measurement Instruments

Confirmatory factor analysis

Comparative fit index

Root mean square error of approximation

Standardized square residual

The Norwegian Agency for Shared Services in Education and Research

Dawes M, Summerskill W, Glasziou P, Cartabellotta A, Martin J, Hopayian K, et al. Sicily statement on evidence-based practice. BMC Med Educ. 2005;5(1):1.

Article   Google Scholar  

Straus SE, Glasziou P, Richardson WS, Haynes RB, Pattani R, Veroniki AA. Evidence-based medicine: how to practice and teach EBM. Edinburgh: Elsevier; 2019.

Google Scholar  

Albarqouni L, Hoffmann T, Straus S, Olsen NR, Young T, Ilic D, et al. Core competencies in evidence-based practice for Health professionals: Consensus Statement based on a systematic review and Delphi Survey. JAMA Netw Open. 2018;1(2):e180281.

Straus S, Glasziou P, Richardson W, Haynes R. Evidence-based medicine: how to practice and teach EBM. Fifth edition ed: Elsevier Health Sciences; 2019.

Paci M, Faedda G, Ugolini A, Pellicciari L. Barriers to evidence-based practice implementation in physiotherapy: a systematic review and meta-analysis. Int J Qual Health Care. 2021;33(2).

Sadeghi-Bazargani H, Tabrizi JS, Azami-Aghdash S. Barriers to evidence-based medicine: a systematic review. J Eval Clin Pract. 2014;20(6):793–802.

da Silva TM, Costa Lda C, Garcia AN, Costa LO. What do physical therapists think about evidence-based practice? A systematic review. Man Ther. 2015;20(3):388–401.

Grol R, Wensing M. What drives change? Barriers to and incentives for achieving evidence-based practice. Med J Aust. 2004;180(S6):S57–60.

Saunders H, Gallagher-Ford L, Kvist T, Vehviläinen-Julkunen K. Practicing Healthcare professionals’ evidence-based practice competencies: an overview of systematic reviews. Worldviews Evid Based Nurs. 2019;16(3):176–85.

Salbach NM, Jaglal SB, Korner-Bitensky N, Rappolt S, Davis D. Practitioner and organizational barriers to evidence-based practice of physical therapists for people with stroke. Phys Ther. 2007;87(10):1284–303.

Saunders H, Vehvilainen-Julkunen K. Key considerations for selecting instruments when evaluating healthcare professionals’ evidence-based practice competencies: a discussion paper. J Adv Nurs. 2018;74(10):2301–11.

de Vet HCW, Terwee CB, Mokkink LB, Knol DL. Measurement in Medicine: a practical guide. Cambridge: Cambridge: Cambridge University Press; 2011.

Book   Google Scholar  

Tilson JK, Kaplan SL, Harris JL, Hutchinson A, Ilic D, Niederman R, et al. Sicily statement on classification and development of evidence-based practice learning assessment tools. BMC Med Educ. 2011;11:78.

Roberge-Dao J, Maggio LA, Zaccagnini M, Rochette A, Shikako K, Boruff J et al. Challenges and future directions in the measurement of evidence-based practice: qualitative analysis of umbrella review findings. J Eval Clin Pract. 2022.

Shaneyfelt T, Baum KD, Bell D, Feldstein D, Houston TK, Kaatz S, et al. Instruments for evaluating education in evidence-based practice: a systematic review. JAMA. 2006;296(9):1116–27.

Landsverk NG, Olsen NR, Brovold T. Instruments measuring evidence-based practice behavior, attitudes, and self-efficacy among healthcare professionals: a systematic review of measurement properties. Implement Science: IS. 2023;18(1):42.

Hoegen PA, de Bot CMA, Echteld MA, Vermeulen H. Measuring self-efficacy and outcome expectancy in evidence-based practice: a systematic review on psychometric properties. Int J Nurs Stud Adv. 2021;3:100024.

Oude Rengerink K, Zwolsman SE, Ubbink DT, Mol BW, van Dijk N, Vermeulen H. Tools to assess evidence-based practice behaviour among healthcare professionals. Evid Based Med. 2013;18(4):129–38.

Leung K, Trevena L, Waters D. Systematic review of instruments for measuring nurses’ knowledge, skills and attitudes for evidence-based practice. J Adv Nurs. 2014;70(10):2181–95.

Buchanan H, Siegfried N, Jelsma J. Survey instruments for Knowledge, skills, attitudes and Behaviour related to evidence-based practice in Occupational Therapy: a systematic review. Occup Ther Int. 2016;23(2):59–90.

Fernández-Domínguez JC, Sesé-Abad A, Morales-Asencio JM, Oliva-Pascual-Vaca A, Salinas-Bueno I, de Pedro-Gómez JE. Validity and reliability of instruments aimed at measuring evidence-based practice in physical therapy: a systematic review of the literature. J Eval Clin Pract. 2014;20(6):767–78.

Belita E, Squires JE, Yost J, Ganann R, Burnett T, Dobbins M. Measures of evidence-informed decision-making competence attributes: a psychometric systematic review. BMC Nurs. 2020;19:44.

Egeland KM, Ruud T, Ogden T, Lindstrom JC, Heiervang KS. Psychometric properties of the Norwegian version of the evidence-based practice attitude scale (EBPAS): to measure implementation readiness. Health Res Policy Syst. 2016;14(1):47.

Rye M, Torres EM, Friborg O, Skre I, Aarons GA. The evidence-based practice attitude Scale-36 (EBPAS-36): a brief and pragmatic measure of attitudes to evidence-based practice validated in US and Norwegian samples. Implement Science: IS. 2017;12(1):44.

Grønvik CKU, Ødegård A, Bjørkly S. Factor Analytical Examination of the evidence-based practice beliefs scale: indications of a two-factor structure. scirp.org; 2016.

Moore JL, Friis S, Graham ID, Gundersen ET, Nordvik JE. Reported use of evidence in clinical practice: a survey of rehabilitation practices in Norway. BMC Health Serv Res. 2018;18(1):379.

Titlestad KB, Snibsoer AK, Stromme H, Nortvedt MW, Graverholt B, Espehaug B. Translation, cross-cultural adaption and measurement properties of the evidence-based practice profile. BMC Res Notes. 2017;10(1):44.

McEvoy MP, Williams MT, Olds TS. Development and psychometric testing of a trans-professional evidence-based practice profile questionnaire. Med Teach. 2010;32(9):e373–80.

Hu MY, Wu YN, McEvoy MP, Wang YF, Cong WL, Liu LP, et al. Development and validation of the Chinese version of the evidence-based practice profile questionnaire (EBP < sup > 2 Q). BMC Med Educ. 2020;20(1):280.

Jia Y, Zhuang X, Zhang Y, Meng G, Qin S, Shi WX, et al. Adaptation and validation of the evidence-based Practice Profile Questionnaire (EBP(2)Q) for clinical postgraduates in a Chinese context. BMC Med Educ. 2023;23(1):588.

Panczyk M, Belowska J, Zarzeka A, Samolinski L, Zmuda-Trzebiatowska H, Gotlib J. Validation study of the Polish version of the evidence-based Practice Profile Questionnaire. BMC Med Educ. 2017;17(1):38.

Belowska J, Panczyk M, Zarzeka A, Iwanow L, Cieślak I, Gotlib J. Promoting evidence-based practice - perceived knowledge, behaviours and attitudes of Polish nurses: a cross-sectional validation study. Int J Occup Saf Ergon. 2020;26(2):397–405.

Knowledge A. Confidence, and Behavior Related to Evidence-based Practice Among Healthcare Professionals Working in Primary Healthcare. Protocol of a Cross-sectional Survey [Internet]. OSF. 2023. Available from: https://doi.org/10.17605/OSF.IO/428RP

Terwee CB, Prinsen CAC, Chiarotto A, Westerman MJ, Patrick DL, Alonso J, et al. COSMIN methodology for evaluating the content validity of patient-reported outcome measures: a Delphi study. Qual Life Res. 2018;27(5):1159–70.

Prinsen CAC, Mokkink LB, Bouter LM, Alonso J, Patrick DL, de Vet HCW, et al. COSMIN guideline for systematic reviews of patient-reported outcome measures. Qual Life Res. 2018;27(5):1147–57.

Mokkink LB, de Vet HCW, Prinsen CAC, Patrick DL, Alonso J, Bouter LM, et al. COSMIN Risk of Bias checklist for systematic reviews of patient-reported outcome measures. Qual Life Res. 2018;27(5):1171–9.

Mokkink LB, Prinsen CA, Patrick D, Alonso J, Bouter LM, Vet HCD et al. Cosmin Study design checklist for patient-reported outecome measurement instruments [PDF]. 2019. https://www.cosmin.nl/tools/checklists-assessing-methodological-study-qualities/ . https://www.cosmin.nl/wp-content/uploads/COSMIN-study-designing-checklist_final.pdf

Gagnier JJ, Lai J, Mokkink LB, Terwee CB. COSMIN reporting guideline for studies on measurement properties of patient-reported outcome measures. Qual Life Res. 2021;30(8):2197–218.

Bjerk M, Flottorp SA, Pripp AH, Øien H, Hansen TM, Foy R, et al. Tailored implementation of national recommendations on fall prevention among older adults in municipalities in Norway (FALLPREVENT trial): a study protocol for a cluster-randomised trial. Implement Science: IS. 2024;19(1):5.

Presser S, Couper MP, Lessler JT, Martin E, Martin J, Rothgeb JM et al. Methods for Testing and Evaluating Survey Questions. Methods for Testing and Evaluating Survey Questionnaires2004. pp. 1–22.

Willis GB. Analysis of the cognitive interview in Questionnaire Design. Cary: Cary: Oxford University Press, Incorporated;; 2015.

StataCorp. Stata Statistical Software. 18 ed. College Station, TX: StataCorp; 2023.

Hu L-t, Bentler PM. Cutoff criteria for fit indexes in covariance structure analysis: conventional criteria versus new alternatives. Struct Equ Model. 1999;6(1):1–55.

Terwee CB, Bot SD, de Boer MR, van der Windt DA, Knol DL, Dekker J, et al. Quality criteria were proposed for measurement properties of health status questionnaires. J Clin Epidemiol. 2007;60(1):34–42.

Mokkink LB, Prinsen CA, Patrick DL, Alonso J, Bouter LM, de Vet HC et al. COSMIN methodology for systematic reviews of Patient-Reported Outcome Measures (PROMs) – user manual. 2018. https://www.cosmin.nl/tools/guideline-conducting-systematic-review-outcome-measures/

Brown TA. Confirmatory Factor Analysis for Applied Research. New York: New York: Guilford; 2015.

Streiner DL, Norman GR, Cairney J. Health measurement scales: a practical guide to their development and use. New York, New York: Oxford University Press; 2015.

de Leeuw ED, Hox JJ, Dillman DA. International handbook of survey methodology. New York, NY: Taylor & Francis Group/Lawrence Erlbaum Associates; 2008. x, 549-x, p.

Solís Salazar M. The dilemma of combining positive and negative items in scales. Psicothema. 2015;27(2):192–200.

Bowling A. Research methods in health: investigating health and health services. 4th ed. ed. Maidenhead: Open University, McGraw-Hill;; 2014.

Download references

Acknowledgements

The authors would like to thank all the participants of this study, and partners in the FALLPREVENT research project.

Open access funding provided by OsloMet - Oslo Metropolitan University. Internal founding was provided by OsloMet. The funding bodies had no role in the design, data collection, data analysis, interpretation of the results or decision to submit for publication.

Open access funding provided by OsloMet - Oslo Metropolitan University

Author information

Authors and affiliations.

Department of Rehabilitation Science and Health Technology, Faculty of Health Science, Oslo Metropolitan University, Oslo, Norway

Nils Gunnar Landsverk & Therese Brovold

Department of Health and Functioning, Faculty of Health and Social Sciences, Western Norway University of Applied Sciences, Bergen, Norway

Nina Rydland Olsen

Faculty of Health Sciences, OsloMet - Oslo Metropolitan University, Oslo, Norway

Are Hugo Pripp

Department of Welfare and Participation, Faculty of Health and Social Sciences, Western Norway University of Applied Sciences, Bergen, Norway

Kristine Berg Titlestad

You can also search for this author in PubMed   Google Scholar

Contributions

NGL, TB, and NRO initiated the study and contributed to the design and planning. NGL managed the data collection (qualitative interviews and the web-based survey) and conducted the data analyses. NGL, TB, NRO, and KBT formed the panel group that developed the interview guide, discussed the results of the interviews in several meetings, and made minor linguistic revisions to the items. AHP assisted in planning the cross-sectional survey, performing statistical analyses, and interpreting the results of the statistical analyses. NGL wrote the manuscript draft, and TB, NRO, and KBT reviewed and revised the text in several rounds. All authors contributed to, reviewed, and approved the final manuscript.

Corresponding author

Correspondence to Nils Gunnar Landsverk .

Ethics declarations

Ethics approval and consent to participate, consent for publication.

Not Applicable.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary Material 1:

The EBP2-N questionnaire

Supplementary Material 2:

The interview guide

Supplementary Material 3:

Details on item revisions

Supplementary Material 4:

Reporting guideline

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ . The Creative Commons Public Domain Dedication waiver ( http://creativecommons.org/publicdomain/zero/1.0/ ) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Cite this article.

Landsverk, N.G., Olsen, N.R., Titlestad, K.B. et al. Adaptation and validation of the evidence-based practice profile (EBP 2 ) questionnaire in a Norwegian primary healthcare setting. BMC Med Educ 24 , 841 (2024). https://doi.org/10.1186/s12909-024-05842-z

Download citation

Received : 09 April 2024

Accepted : 30 July 2024

Published : 06 August 2024

DOI : https://doi.org/10.1186/s12909-024-05842-z

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Healthcare professional
• Primary healthcare
• Content validity
• Construct validity
• Structural validity
• Internal consistency
• Self-efficacy

BMC Medical Education

ISSN: 1472-6920

• Submission enquiries: [email protected]
• General enquiries: [email protected]

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• Explore content
• About the journal
• Publish with us
• Sign up for alerts
• Data Descriptor
• Open access
• Published: 13 August 2024

ENTRANT: A Large Financial Dataset for Table Understanding

• Elias Zavitsanos   ORCID: orcid.org/0000-0002-2417-3307 1 ,
• Dimitris Mavroeidis 1 ,
• Eirini Spyropoulou 1 ,
• Manos Fergadiotis 1 &
• Georgios Paliouras   ORCID: orcid.org/0000-0001-9629-2367 1  

Scientific Data volume  11 , Article number:  876 ( 2024 ) Cite this article

Metrics details

• Information technology

Tabular data is a way to structure, organize, and present information conveniently and effectively. Real-world tables present data in two dimensions by arranging cells in matrices that summarize information and facilitate side-by-side comparisons. Recent research efforts aim to train large models to understand structured tables, a process that enables knowledge transfer in various downstream tasks. Model pre-training, though, requires large datasets, conveniently formatted to reflect cell and table characteristics. This paper presents ENTRANT, a financial dataset that comprises millions of tables, which are transformed to reflect cell attributes, as well as positional and hierarchical information. Hence, they facilitate, among other things, pre-training tasks for table understanding with deep learning methods. The dataset provides table and cell information along with the corresponding metadata in a machine-readable format. We have automated all data processing and curation and technically validated the dataset through unit testing of high code coverage. Finally, we demonstrate the use of the dataset in a pre-training task of a state-of-the-art model, which we use for downstream cell classification.

Similar content being viewed by others

Enhanced analysis of tabular data through Multi-representation DeepInsight

Table2Vec-automated universal representation learning of enterprise data DNA for benchmarkable and explainable enterprise data science

A tabular data generation framework guided by downstream tasks optimization

Background & summary.

Tables constitute fundamental structures to organize and represent data. They are widely used in web pages, spreadsheets, reports, and other types of documents. Tabular data are usually arranged in cells of bi-dimensional matrices, with stylistic formatting that effectively reflects important regions and data in the table. Tables are typically made to ease readability by humans; they are usually assumed to be read left-to-right and top-to-bottom, while related items are assumed to appear nearby. They present information using spacing, lines, rules, font style, weight, and indentation. In order to represent data as effectively as possible, real-world tables often imply hierarchical structures shaped by merged cells and indents to organize and summarize the data in a compact way that facilitates quick lookup and side-by-side comparisons. In the simplest case, a table can be organized in a horizontal or vertical relational form, as shown in Fig.  1(a) , while in a more complex case, it may take the form of a matrix. The latter case is depicted in Fig.  1(b) , where two hierarchies are implied simultaneously. A top-down hierarchy of three levels: a root node whose children are “Incidence” and “Mortality” at level 1, and their children of the last two are “Males” and “Females” at level 2. Likewise, there is a left-to-right hierarchy also of three levels: a root node whose children are the cells in bold text at level 1, and their corresponding children at level 2, such as “Melanoma of skin”, “Non-melanoma skin cancer” and so on.

Examples of structured tables: ( a ) relational web table, ( b ) spreadsheet table.

Tables are used to store valuable information that can be used, in turn, to answer questions, populate knowledge bases, generate natural language descriptions, and various other applications. Most of the time, though, tables do not include machine-readable semantics for the automated processing of the information they contain. This lack of semantics is a challenge that table understanding aims to address. In particular, table understanding aims at recovering the missing semantics that enables the extraction of facts from tables and facilitates interoperability, knowledge extraction, and information processing. This topic covers many challenging tasks, from table detection in documents and relational data extraction 1 to semantic table interpretation and representation 2 . In this work, we are primarily interested in the part of table understanding that concerns table representation and interpretation.

Increased attention towards table representation from the research community has led to various approaches to table understanding and its subtasks. In particular, several approaches have been proposed to solve structure-based tasks, such as table segmentation, cell role prediction, block detection, and join identification. Moreover, approaches focusing on knowledge alignment tasks deal with semantic typing and modeling, schema mapping, entity linking, knowledge extraction and augmentation. Finally, several downstream tasks concern cell type classification, table type classification, formula prediction, summarization, question answering, and validation 3 . In other words, table understanding is a significant area of research with several constituent tasks. The understanding of table structure is a foundational step toward knowledge-centric tabular information processing. In recent years, the academic community has proposed several structural models in this direction, from traditional exploitation of cell-level features to models that incorporate contextual information and use deep learning.

Early attempts in the field started with the exploitation of cell-level features. An immediate extension was to use sequence-based features 4 integrating table content and layout features, using conditional random fields to find table cell boundaries, to classify cells as data or labels, and to associate data cells with their corresponding label cells. Using traditional supervised learning methods, Koci et al . 5 rely on content, stylistic, font, and spatial features to infer the layout of tables. This task is often referred to as table-type classification.

More recent approaches in table representation learning rely on cell and table embeddings to capture the underlying semantics. Initial approaches have used continuous bag-of-words and skip-gram models to learn cell embeddings 6 . Table2Vec 7 , among the popular approaches, uses skip-gram neural networks to train word embeddings, aiming at capturing the semantics of table cells. Moreover, TabNet 8 relies on LSTMs to learn token embeddings, considering at the same time spatial information of the table, and more recently, Gunther et al . 9 proposed an embedding technique to be pre-trained directly on a large web table corpus to learn table-level embeddings.

As already mentioned, tables usually arrange similar items in nearby groups. Based on that, and since the value of a cell may be related to the neighborhood of that cell, many approaches rely on both semantic and spatial information and try to capture them jointly. In this direction, several neural network architectures have been proposed. For instance, CNNs have been used to model web and spreadsheet tables 10 , 11 , RNNs and LSTMs have been employed to capture the order of rows and columns 6 , 8 , 12 , 13 , and CNNs and RNNs have been proposed for column type classification 14 . Finally, besides recursive and convolutional network architectures, graph neural networks have also been explored for table understanding and particularly for question answering 15 , 16 , 17 .

The most recent work on table understanding, though, relies heavily on table pre-training, which is actively and widely studied, aiming to build encoders suitable for structured tabular data and natural language context. In this direction, TaBERT 18 aims to recover column names and cell values from unmasked table cells. TaPas 19 , on the other hand, relies on conventional Masked Language Modeling (MLM) to pre-train a model that provides a basic understanding of tables that can be used in downstream table question-answering tasks. STRUG 20 uses different pre-training objectives and focuses on explicitly learning the alignment between text and tables via a classification task, which improves schema link performance in downstream text-to-SQL tasks. In addition, TURL 21 learns representations from relational tables to enhance table knowledge matching and augmentation.

Finally, among the most recent approaches that are also closely related since they use the same pre-training data are TUTA 22 and FORTAP 23 . TUTA is a transformer-based method for understanding general table structure, which pre-trains on general tables that may contain hierarchical structures. The pre-trained model can be used as a basis for table question answering, formulae prediction, cell type classification, table type classification, and other downstream tasks. FORTAP, on the other hand, has been suggested as a method to fill the gap in inner-table numeric reasoning capabilities, which are essential for downstream tasks that involve calculations and reference resolution. For this reason, FORTAP needs to pre-train on data containing spreadsheet formulae, among others.

The increased interest of the research community in such large models for table understanding requires the use of large datasets that comprise millions of tables. These tables should be represented in a convenient and machine-readable form. The most widely used datasets are based on WikiTables 24 . For instance, TURL 21 relies on a subset of WikiTables containing 1.6 million tables, while TaPas 19 pre-trains on another subset of 6.2 million tables. WikiTables has also been used in combination with other sources. StruBERT 25 , for instance, has been pre-trained on a dataset comprising data from WikiTables and data from a subset of PubMed Central (PMC) 26 . TABBIE 27 and TaBERT 18 used Wikipedia tables and tables from Common Crawl ( https://commoncrawl.org/ ) to compile a pre-training dataset of 26.6 million tables.

Moving to larger pre-training datasets, Gunther et al . 9 use the DWTC Web Table corpus 28 , comprising 125 million tables. Finally, TUTA 22 and FORTAP 23 combine web tables from Wikipedia with the WDC WebTable corpus 29 and web-crawled spreadsheets to compile a final dataset of 57.9 million tables. The latter spreadsheets dataset is among the very few financial data used in the domain, together with the dataset used in TAT-QA 30 , which is particularly small and contains 20K financial tables (downloaded from https://www.annualreports.com/ ).

Although there are publicly available datasets that can be used for pre-training, they mainly come from general-purpose sources, such as Wikipedia, Common Crawl, and Web data (WDC), that additionally require much pre-processing to extract the tables and the corresponding cell features in a machine-ready form that large models can easily consume. Moreover, datasets with tables from spreadsheets that contain both textual and numerical values are hard to find. For instance, the spreadsheets used in TUTA and FORTAP are not publicly available. On the other hand, the financial tables used in TAT-QA constitute a relatively small dataset for pre-training purposes. In order to fill this gap, we compile a large dataset that we call ENTRANT, consisting of millions of financial tables with both textual and numerical information from the EDGAR database, which provides free public access to corporate information. In addition, we transform the raw data into a convenient format that can be consumed by state-of-the-art models and provides both table and cell characteristics, including several cell attributes, positional information, and hierarchical information. Although the primary aim of this paper is to introduce a large dataset to enhance model pre-training, this work also contributes the following:

We identify additional usages of the data that enable the application of machine learning methods in financial use cases.

We release the code to encourage the augmentation of the dataset with financial tables from EDGAR.

We enable the automated transformation of other corporate spreadsheets to the machine-readable format by releasing the code scripts that transform excel files to the proposed format.

In what follows, we describe our methodology for data gathering and transformation. Next, we present the data structure and provide an overview of the data files and their format. Last, we provide the analysis we followed to support and validate the dataset’s technical quality. As an additional verification, we demonstrate the usage of the compiled dataset for pre-training a state-of-the-art model that we later use to experiment in a downstream classification task.

The data construction process comprised two main phases. The first concerned the raw data gathering, and the second the data transformation to produce a machine-readable format ready to be consumed by large machine learning architectures. The following sections describe in detail the steps followed in each phase.

Raw data gathering

At a high level, the dataset constitutes a collection of financial statements and disclosures in tabular format. Although this information is freely available for public companies, gathering large amounts of tabular data from the public statements for training is not practical. For this reason, we crawl the Electronic Data Gathering, Analysis, and Retrieval system (EDGAR) ( https://www.sec.gov/edgar/about ). EDGAR is the primary system for companies submitting reports, statements, and documents under the US Securities and Exchange Commission (SEC) ( https://www.sec.gov/ ). It contains millions of company and individual filings, ensuring transparency and fairness in the securities markets and providing benefits for investors, corporations, and researchers. According to EDGAR, the system processes a few thousand filings daily and accommodates 40 thousand new files per year on average. For these reasons, it constitutes an ideal source of financial information that is updated constantly and contains various types of spreadsheet data, depending on the type of filing.

Regarding accessing the EDGAR data, as stated in the access policy, anyone can access and download the company information for free or query it through a variety of EDGAR public services ( https://www.sec.gov/os/accessing-edgar-data ). To ensure that everyone has equitable access to EDGAR we use efficient scripting to guarantee fair access by (a) limiting the maximum request rate to ten requests per second and (b) declaring the user agent in the request headers, as per EDGAR instructions.

EDGAR contains filings from the ’90s up to this day. Many systems, standards, regulations, and filing requirements have changed during this period. Historical filings from the ’90s and early ’00s are available in plain text or HTML, which makes it difficult or inefficient to extract the financial tables robustly. Such an endeavor would require enormous human effort to ensure that all tables have been extracted correctly and that each table cell contains the correct data. In 2009, SEC adopted XBRL ( https://www.xbrl.org/ ), an open international standard for digital business reporting. XBRL provides a language in which reporting terms can be authoritatively defined. In other words, XBRL delivers human-readable financial statements in a machine-readable, structured data format. Most importantly, publicly traded companies are required to create financial statements and assign an XBRL tag to every number, table, accounting policy, statement, and note. Hence, XBRL largely standardized the filing format, and nowadays, financial reports are structured, and financial statements appear inside not only HTML reports but also as separate spreadsheets, annotated and ready to be processed in an automated way.

For the compilation of our dataset, we focus on reports that are filed in XBRL format, ensuring that the spreadsheets we crawl are free from inconsistencies and that they can be parsed and processed by programming libraries in an automated way. The data gathering phase consists of two main steps; (a) checking for valid XBRL representations to filter out reports that are not appropriately tagged, and (b) crawling the reports from EDGAR, as shown in Fig.  2 .

EDGAR crawling workflow. Each company is uniquely identified by a central index key and an accession number that is used to build an endpoint for retrieving the report. Each endpoint is used to get the company spreadsheets, as long as the filing is in XBRL format.

In particular, each company in EDGAR is uniquely identified by a Central Index Key (CIK), which is used in the SEC computer systems to identify corporations and individuals, who have filed a disclosure with the SEC. EDGAR uses the CIK to construct the Accession number, a unique identifier assigned automatically to an accepted submission by EDGAR. With each company’s Accession number, we can construct the API endpoint of EDGAR to request the company filings (Listing 1 provides an example of an endpoint). If a company has filed a report that is not XBRL tagged, the report is filtered out as the corresponding spreadsheets cannot be retrieved robustly.

To ensure that most retrieved reports conform to XBRL, we crawl EDGAR for reports filed from 2013 to 2021. We stopped at 2021 due to the use of the data for our research on material misstatement detection 31 . Material misstatement can take - on average- two-to-three years to be discovered 31 , 32 . Using this cutoff, we are confident that we miss only a few misstatements and we minimize the risk of accounting inconsistencies. Under these assumptions, we generated 330 thousand API endpoints with actual files. These include annual and quarterly financial reports, current reports, registration statements, and more.

SEC specifies that each kind of report is filed in a different form. For instance, a 10-K report is an annual report with consolidated financial statements, and a 10-Q report is a quarterly report with financial statements. Table  1 presents the types of reports gathered in this phase. Gathering reports of different types is particularly important because, eventually, the compiled dataset not only includes tables with financial indices and numbers but also tables and cells with textual information, including standard XBRL concepts, such as “Net Income”, “Current Liabilities”, as well as short disclosures and notes that accompany significant events. We, therefore, produce a financial dataset primarily based on financial statements but with a rich vocabulary.

The above-described crawling phase gathered about 330 thousand reports from the generated API endpoints, grouped by report type, as Table  2 shows. As we will see in the following section, each spreadsheet contains several financial tables, resulting in the final dataset comprising millions of tables.

Data transformation

The second phase of the methodology transforms the data to a machine-readable format that provides as much exploitable information as possible. This process takes as input a raw data file from the previous phase, i.e., a workbook, and extracts the tables that are contained within the worksheets (or spreadsheets) of the workbook. Then, it proceeds to the processing of the contents of the tables in order to annotate cells with valuable attributes. Finally, it extracts the implied hierarchies of the table (if any) and stores the transformed workbook in a key-value format as a JSON file. Figure  3 provides a high-level overview of this phase.

Data transformation workflow. For each spreadsheet, the table boundaries and the merged regions are detected. Then, headers and cell values and attributes are extracted. At the third step, the top-down and left-right hierarchies are constructed and finally, the transformed data are stored in JSON format. A test suite that provides unit testing of all steps covers the functionality of the entire process.

In more detail, the procedure that we follow includes the following steps:

Extract the worksheets of the given workbook: A given workbook may contain several worksheets (i.e., spreadsheets) and thus contain several tables. For instance, if the workbook is from an annual report, it contains a different spreadsheet for each financial statement, a spreadsheet with company information, and some with notes, small disclosures, and highlighted economic characteristics. This step extracts all the worksheets of each workbook. Then, it identifies the different tables by retrieving the table dimensions and validating that there is actual content in the enclosed cells for the retrieved table range, not just empty space. Also, for each table, we identify the merged regions.

Process table: Given the tables identified at the previous step, each table is processed to extract useful information. To reduce noise and keep what is useful, we drop tiny tables that may include just a few cells containing footnotes or remarks that are not directly related to a broader context. In addition, we drop broken tables, meaning they have sparse values and many empty rows in between. The actual processing of each table starts by iterating over table cells and assessing whether they belong in a top or left header. Then, the values and possible stylistic attributes of each cell are extracted (i.e., whether there are borders, italics, foreground and background color, alignment, orientation, and more). Finally, we keep the position of the cell in the table as its spatial coordinates.

Bi-tree extraction: This step extracts the implied top-down and left-right hierarchies of the table. The hierarchical organization of a table is typically implied by merged regions and indentation. In this step we enrich the information already extracted from each cell with the cell’s position in a semantic hierarchy. A prerequisite for this process is to identify how many rows (starting from the top of the table) and how many columns (starting from the left) are headers, in order to determine the depth of each hierarchy. Figure  4 shows an example, where we want to encode the information that a data region cell belongs to node A8 in the hierarchy, and this node has a sibling (A7), and their parent node is A6.

Dataset compilation: The last step of the process integrates all the extracted information in a key-value memory structure for each table. This allows quick and efficient lookup of the table-related information. Finally, all tables of the given workbook are exported in JSON format. Thus, at the end of this process, we have a separate JSON file for each workbook, including all the extracted tables.

It is important to note that throughout the data transformation process, a test suite framework, comprising more than 40 unit tests that cover the complete functionality of the process, ensures the validity of the extracted information. Executing the process on the gathered data, we produced 6.7 million tables (6,735,407 to be precise). The entire data transformation phase is fully automated using Python scripts. In particular, we use OpenPyXL ( https://openpyxl.readthedocs.io/en/stable/ ) to process the spreadsheets and extract content and attributes.

Regarding the Bi-tree extraction, the related work assumes hierarchies of maximum depth equal to four 22 , 23 . We follow the same assumption as it covers all the table types we downloaded in the data gathering phase. To represent each hierarchy (in memory and in the JSON format), we use a key-value dictionary. Each cell in the hierarchy constitutes a node that is again represented as a dictionary, containing the row and column indices (positional coordinates) and a list of its children nodes. The children nodes follow the same key-value representation, including in turn their positional coordinates and their children nodes. This way a nested structure that represents the hierarchy is created, as shown in Listing 2. The coordinates of the root node are (-1,-1) since the root is assumed to lie “outside” the table (see Fig.  4 ). Figure  5 provides an example of a financial table and illustrates how the top hierarchy is represented after the data transformation process.

Implied hierarchies in table: both the top-down hierarchy and the left-to-right hierarchy contain three levels each. Cell A6 is the parent node of cells A7 and A8, and its siblings are A3 and A9. D2 is a child of DE1 and its sibling is E2. (Image inspired from Wang et al . 22 ).

An example of the top hierarchy representation at the data transformation phase.

Data Records

The dataset is available at Zenodo 33 , with this section being the primary source of information on the availability and content of the data being described.

As mentioned above, the compiled dataset consists of about 330 thousand JSON files. Each file corresponds to a company filing (e.g., a report) and contains several tables. The files are organized in a directory structure that includes different folders for different types of report, as shown in Fig.  6 . This structure facilitates file system handling and allows selective focus based on different types of report and company filing. Therefore, the parent directory contains as many subfolders as the different type of reports (see Table  1 ), e.g., a subfolder named “10-K” containing the data extracted from “10-K” reports, a subfolder named “10-Q” containing the data extracted from “10-Q” reports, and so forth. Each subfolder is compressed as a zip file and uploaded to Zenodo.

Directory structure of the dataset and analysis of JSON contents.

The name convention of the JSON files follows the pattern <  C I K  > _ <  Y E A R  > _ <  T Y P E  > _ <  A C C E S S I O N _ N U M B E R  > .  j s o n , where C I K is the central index key of the company, Y E A R is the year of the filing submission, T Y P E is the type of report (e.g., 10-K), and A C C E S S I O N _ N U M B E R is the unique identifier by EDGAR.

Each JSON file contains a list. Each element of that list corresponds to a table represented as a dictionary. Figure  6 shows an example of a JSON file containing three dictionaries with the first one expanded. For each table, we store metadata, such as the storage description, the language, and the spreadsheet name. We also include table-specific information that comprises the dimensions of the table (“RangeAddress”), its title, its content (“Cells”), the identified merged regions, the number of top and left headers (“TopHeaderRowsNumber” and “LeftHeaderColumnsNumber” respectively) and the Bi-tree (“TopTreeRoot” and “LeftTreeRoot”).

All the table content with its attributes is under the key “Cells”. This structure is essentially a list of lists that represent the rows and columns of the table. Thus, the elements of the outer list are the rows, and each row is a list, the elements of which are the cells of the corresponding columns. As shown in Fig.  6 , each cell is a dictionary comprising the following information:

“T”: the value of the cell as string

“V”: the value of the cell

“is_header”: whether the cell participates in a header row

“is_attribute”: whether the cell participates in a header column

“coordinates”: the spatial coordinates of the cell ([row index, column index])

"HF”: whether the cell has formula

“A1”: formula specific: absolute cell reference

“R1”: formula specific: relative cell reference

“font_name”: the name of the font

“font_size”: the size of the font

“wrap_text”: whether the text is wrapped within the cell dimension

“BC”: whether there is a non-white background

“FC”: whether there is a non-black font color

“FB”: whether the font is bold

“I”: whether the font is italic

“NS”: the number format of the cell

“DT”: the data type of the cell

“LB”: whether the cell has a left border

“TB”: whether the cell has a top border

“BB”: whether the cell has a bottom border

“RB”: whether the cell has a right border

“O”: the orientation of the cell

“HA”: whether the cell is horizontally aligned

“VA”: whether the cell is vertically aligned

The remaining table attributes are the number of top header rows and the number of the left header columns, the two hierarchies presented in the previous section (see also Fig.  5 as an example), and the “MergedRegions” that contains all the identified merged areas of the table. The latter is a list of dictionaries, one for every identified merged region, indicating the starting row and column and the ending row and column of the region.

Regarding dataset statistics, there are about 6.7 million tables in total. On average, there are 20 tables per filing, i.e., per JSON file. Each table has approximately 25 rows on average and five columns. The number of columns does not vary much. On the other hand, the number of rows may deviate much more from the mean. Finally, the number of cells per table also varies largely, with a mean value of 130 and with many tables containing hundreds or even thousands of cells. Table  3 summarizes the essential statistical information of the dataset. The heavy-tail distribution of the rows per table (and consequently for the cells per table) is due to the variety of types of report in the dataset. Reports of financial statements often include additional highlights with notes and numerical indices across years that tend to produce tables with many rows and cells. All the above contribute to the large variety in tables and vocabulary this dataset offers.

Technical Validation

This article is the first public release of the ENTRANT dataset and the code used to produce it. Thus, the research community has not yet had the opportunity to review it and provide feedback. However, all users are encouraged to fork the data and code, report issues, and make pull requests with their modifications.

Having said that, we produced the ENTRANT dataset intending to be readily consumable by large models, such as TUTA 22 and FORTAP 23 , as well as to be easily exploitable and manageable for other use cases. The code used to build ENTRANT was version-controlled and well tested. For this purpose, a test suite framework was implemented, comprising more than 40 unit tests that cover all the functionality of the code. The unit tests focus on data quality and representational inconsistencies that may occur in the data transformation process due to either human error or noise in the initial raw data. Specifically, the unit and sanity tests cover the following aspects:

Table identification: validates that all tables that could have been extracted from a given workbook should have been extracted. Additionally, we test for the correct extraction of the table titles and the table dimensions. Regarding the latter, we ensure that the number of rows and columns is consistent in the table and matches the identified dimensions. Moreover, we check for the correct header identification, and we perform these tests for several types of workbook (i.e., workbooks that refer to 10-K, 10-Q, and 8-K reports).

Merged region identification: validates that all merged regions and table cells have been identified and encoded correctly in the final JSON representation. It also ensures that the number of the identified merged regions matches the size of the data structure that is used for the representation.

Cell attribute extraction: validates that all cell attributes have been computed. We test for cells that participate in header rows and those that do not participate in table headers to ensure that the expected number of attributes has been calculated and that the corresponding attribute values have been correctly extracted from the worksheet. Again, we perform these tests for several types of workbook.

Bi-tree extraction: validates that the two implied hierarchies have been extracted correctly. This set includes tests on the number of top headers and the number of left headers. We also check for the correct coordinates of the top tree root and the left tree root. Finally, we ensure each node contains the correct positional coordinates and the corresponding list of its children nodes.

At the time of the code publication, all unit tests are successfully completed, as shown in Listing 3.

Besides the unit tests, as an additional validation step, a post-processing procedure checks that there are no empty regions, rows and columns in each extracted table. This step ensures that all corresponding values have been successfully retrieved and that there are no empty spaces from the initial workbooks that have been considered as parts of the extracted tables. As a final technical validation step, all produced JSON files have been parsed an loaded to memory as dictionaries successfully, using the JSON Python module, in order to endure that all data can be parsed and loaded correctly.

Regarding the actual usage of the dataset, we used it to pre-train TUTA 22 , a large model for table understanding that exploits all the information that is encoded in the data, including cell attributes, table attributes, and bi-trees. We used the code published ( https://github.com/microsoft/TUTA_table_understanding ) by the authors of the model to pre-train it and then employed it on a downstream task that focuses on cell type classification.

In the pre-training phase, we trained two models. The first used three large datasets, namely the WikiTable (provided by the authors of TUTA), WDC (also provided by the authors), and ENTRANT. The second model consumed only the ENTRANT data. The downstream task aimed to identify fine-grained cell types. Cell type classification has been widely studied 6 , 11 , 34 , 35 , 36 with several smaller datasets, and it is a task that requires models to capture both semantic and structural information, which is something that ENTRANT encodes successfully.

The datasets used for the downstream task are DeEx 35 and SAUS 6 , which are collected from various domains (financial, educational, public health) and thus contain tables with various structures and semantics. DeEx and SAUS categorize cells into general types: “metadata”, “notes”, “data”, “top attribute”, “left attribute”, and “derived”. We, therefore, evaluated the models on a multi-class classification setting. Table  4 provides the classification results in terms of Macro F1 score. The models we included in this task are those mentioned above, i.e., the two pre-trained TUTA models, plus the original TUTA that was pre-trained using WikiTable, WDC, and web-crawled spreadsheets that are not publicly available.

The most important conclusion of this experiment, for the purposes of the present paper, was that the compiled dataset is easily consumable and useful for state-of-the-art models. Additionally, the pre-trained model that uses only the ENTRANT data provides results that do not deviate much from the other two models. When combined with other public data (WikiTables and WDC) it leads to state-of-the-art results. Therefore, although the dataset is focused on the financial domain, it provides a rich vocabulary and useful semantic and structural information.

Usage Notes

All researchers and interested parties can access the dataset freely at Zenodo 33 . The dataset is based on publicly and freely available information from SEC’s EDGAR, enriched with metadata and table and cell attributes to facilitate research and experiments that rely on financial information. According to EDGAR, access to its public database is free, allowing everyone to retrieve public companies’ financial information and operations, by reviewing the filings the companies make with the SEC.

The way the dataset is organized facilitates quick lookups for tables of specific filing types and specific years. The JSON format facilitates quick data loading and processing, as well as interoperability between applications, without explicit requirements for specialized libraries.

The dataset can be conveniently processed and used for various tasks. The primary aim of this research was to compile an extensive dataset for pre-training large models for table understanding. One direct usage, therefore, is to pre-train a model like TUTA 22 or FORTAP 23 . We have already demonstrated this by pre-training TUTA.

In addition, the data can be used as a source to extract financial indices from the financial statements, which can in turn be used as features for machine learning methods in financial use cases. For example, financial misstatement detection and distress or bankruptcy prediction models rely heavily on such features in combination with market data or other operational and categorical variables.

Finally, the variety of the tables that this dataset contains allows the selection or sampling of tables that meet specific criteria for table classification, cell classification, or other tasks, in order to create smaller datasets for direct training and testing. For instance, the dataset could be used to create smaller and focused datasets for argument mining tasks or table question-answering. Tables that correspond to specific filings (e.g., annual reports) of companies can be combined with textual information or the auditor’s opinion that appears in the reports to support argument mining regarding the concerns that the auditors express, given the financial status of the company 37 .

Code availability

We automated both phases of the data compilation process presented in this work, using the Python programming language and corresponding libraries that are freely distributed by the Python Package Index (PyPI). The data resources described in this paper, including Python scripts, can be accessed without restrictions at GitHub ( https://github.com/iit-Demokritos/entrant ). Anyone can browse the content of the repository, and everyone is encouraged to fork or create pull requests with improvements and enhancements. The code is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as appropriate credit to the original author(s) and the source are given, a link to the Creative Commons license is provided, and any changes are clearly indicated.

By releasing the code for this work, we aim to enable further augmentation of the dataset, as anyone can use it to fetch more reports from EDGAR and transform them in the described format. An additional advantage is that the transformation process alone is a powerful tool, since it allows the transformation of excel files that follow similar table formats to the described JSON format.

Eberius, J. et al . Deexcelerator: A framework for extracting relational data from partially structured documents. In Proceedings of the 22nd ACM International Conference on Information & Knowledge Management , CIKM ’13, 2477-2480, https://doi.org/10.1145/2505515.2508210 (Association for Computing Machinery, New York, NY, USA, 2013).

Shigarov, A. Table understanding: Problem overview. WIREs Data Mining and Knowledge Discovery 13 , e1482, https://doi.org/10.1002/widm.1482 (2023).

Article   Google Scholar  

Pujara, J., Szekely, P., Sun, H. & Chen, M. From tables to knowledge: Recent advances in table understanding. In KDD ’21: Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining , 4060–4061, https://doi.org/10.1145/3447548.3470809 (2021).

Pinto, D., McCallum, A., Wei, X. & Croft, W. Table extraction using conditional random fields. In SIGIR’03 , https://doi.org/10.1145/860435.860479 (Toronto, Canada, 2003).

Koci, E., Thiele, M., Romero, O. & Lehner, W. A machine learning approach for layout inference in spreadsheets. In International Conference on Knowledge Discovery and Information Retrieval , https://doi.org/10.5220/0006052200770088 (2016).

Gol, M. G., Pujara, J. & Szekely, P. Tabular cell classification using pre-trained cell embeddings. In 2019 IEEE International Conference on Data Mining (ICDM) , 230–239, https://doi.org/10.1109/ICDM.2019.00033 (2019).

Zhang, L., Zhang, S. & Balog, K. Table2vec: Neural word and entity embeddings for table population and retrieval. In Proceedings of International ACM SIGIR Conference on Research and Development in Information Retrieval , 1029-1032, https://doi.org/10.1145/3331184.3331333 (ACM, 2019).

Nishida, K., Sadamitsu, K., Higashinaka, R. & YoshihiroMatsuo. Understanding the semantic structures of tables with a hybrid deep neural network architecture. In Thirty-First AAAI Conference on Artificial Intelligence , https://doi.org/10.1609/aaai.v31i1.10484 (2017).

Günther, M., Thiele, M., Gonsior, J. & Lehner, W. Pre-trained web table embeddings for table discovery. In Proceedings of the Fourth International Workshop on Exploiting Artificial Intelligence Techniques for Data Management , aiDM ’21, 24-31, https://doi.org/10.1145/3464509.3464892 (Association for Computing Machinery, New York, NY, USA, 2021).

Chen, J., Jiménez-Ruiz, E., Horrocks, I. & Sutton, C. Colnet: Embedding the semantics of web tables for column type prediction. In Proceedings of the AAAI Conference on Artificial Intelligence , vol. 33, 29–36, https://doi.org/10.48550/arXiv.1811.01304 (2019).

Dong, H., Liu, S., Fu, Z., Han, S. & Zhang, D. Semantic structure extraction for spreadsheet tables with a multi-task learning architecture. In Workshop on Document Intelligence at NeurIPS , https://openreview.net/forum?id=r1x3GTq5IB (2019).

Fetahu, B., Anand, A. & Koutraki, M. Tablenet: An approach for determining fine-grained relations for wikipedia tables. In The World Wide Web Conference , 2736-2742, https://doi.org/10.48550/arXiv.1902.01740 (2019).

Kardas, M. et al . AxCell: Automatic extraction of results from machine learning papers. In Webber, B., Cohn, T., He, Y. & Liu, Y. (eds.) Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP) , 8580–8594, https://doi.org/10.18653/v1/2020.emnlp-main.692 (Association for Computational Linguistics, Online, 2020).

Chen, J., Jiménez-Ruiz, E., Horrocks, I. & Sutton, C. Learning semantic annotations for tabular data. In Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence (IJCAI-19) , https://doi.org/10.48550/arXiv.1906.00781 (2019).

Koci, E., Thiele, M., Lehner, W., & Romero, O. Table recognition in spreadsheets via a graph representation. In 2018 13th IAPR International Workshop on Document Analysis Systems (DAS) , 139-144, https://doi.org/10.1109/DAS.2018.48 (IEEE, 2018).

Zayats, V., Toutanova, K. & Ostendorf, M. Representations for question answering from documents with tables and text. In Merlo, P., Tiedemann, J. & Tsarfaty, R. (eds.) Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume , 2895–2906, https://doi.org/10.18653/v1/2021.eacl-main.253 (Association for Computational Linguistics, Online, 2021).

Zhang, X. et al . A graph representation of semi-structured data for web question answering. In Proceedings of the 28th International Conference on Computational Linguistics , 51–61, https://doi.org/10.48550/arXiv.2010.06801 (2020).

Yin, P., Neubig, G., Yih, W.-t. & Riedel, S. TaBERT: Pretraining for joint understanding of textual and tabular data. In Jurafsky, D., Chai, J., Schluter, N. & Tetreault, J. (eds.) Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics , 8413–8426, https://doi.org/10.18653/v1/2020.acl-main.745 (Association for Computational Linguistics, Online, 2020).

Herzig, J., Nowak, P. K., Müller, T., Piccinno, F. & Eisenschlos, J. TaPas: Weakly supervised table parsing via pre-training. In Jurafsky, D., Chai, J., Schluter, N. & Tetreault, J. (eds.) Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics , 4320–4333, https://doi.org/10.18653/v1/2020.acl-main.398 (Association for Computational Linguistics, Online, 2020).

Deng, X. et al . Structure-grounded pretraining for text-to-SQL. In Toutanova, K. et al . (eds.) Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies , 1337–1350, https://doi.org/10.18653/v1/2021.naacl-main.105 (Association for Computational Linguistics, Online, 2021).

Deng, X., Sun, H., Lees, A., Wu, Y. & Yu, C. Turl: Table under-standing through representation learning, https://doi.org/10.48550/arXiv.2006.14806 (2020). 2006.14806.

Wang, Z. et al . Tuta: Tree-based transformers for generally structured table pre-training. In Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining , KDD ’21, 1780-1790, https://doi.org/10.1145/3447548.3467434 (Association for Computing Machinery, New York, NY, USA, 2021).

Cheng, Z. et al . Fortap: Using formulas for numerical-reasoning-aware table pretraining. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics , 1150–1166, https://doi.org/10.48550/arXiv.2109.07323 (2022).

Bhagavatula, C., Noraset, T. & Downey, D. Tabel: Entity linking in web tables. In Proceedings of the International Semantic Web Conference , https://doi.org/10.1007/978-3-319-25007-6_25 (2015).

Trabelsi, M., Chen, Z., Zhang, S., Davison, B. D. & Heflin, J. Strubert: Structure-aware bert for table search and matching. In Proceedings of the ACM Web Conference 2022 (WWW ‘22) , https://doi.org/10.1145/3485447.3511972 (ACM, Virtual Event, Lyon, France, 2022).

Habibi, M., Starlinger, J. & Leser, U. Tabsim: A siamese neural network for accurate estimation of table similarity. In 2020 IEEE International Conference on Big Data (Big Data) , 930–937, https://doi.org/10.1109/BigData50022.2020.9378077 (2020).

Iida, H., Thai, D., Manjunatha, V. & Iyyer, M. TABBIE: Pretrained representations of tabular data. In Toutanova, K. et al . (eds.) Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies , 3446–3456, https://doi.org/10.18653/v1/2021.naacl-main.270 (Association for Computational Linguistics, Online, 2021).

Eberius, J. et al . Building the dresden web table corpus: A classification approach. In 2015 IEEE/ACM 2nd International Symposium on Big Data Computing (BDC) , 41–50, https://doi.org/10.1109/BDC.2015.30 (2015).

Lehmberg, O., Ritze, D., Meusel, R. & Bizer, C. A large public corpus of web tables containing time and context metadata. In Proceedings of the 25th International Conference Companion on World Wide Web , WWW ’16 Companion, 75-76, https://doi.org/10.1145/2872518.2889386 (International World Wide Web Conferences Steering Committee, Republic and Canton of Geneva, CHE, 2016).

Zhu, F. et al . TAT-QA: A question answering benchmark on a hybrid of tabular and textual content in finance. In Zong, C., Xia, F., Li, W. & Navigli, R. (eds.) Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers) , 3277–3287, https://doi.org/10.18653/v1/2021.acl-long.254 (Association for Computational Linguistics, Online, 2021).

Zavitsanos, E. et al . Financial misstatement detection: a realistic evaluation. In 2nd ACM International Conference on AI in Finance (ICAIF`21) , 1–9, https://doi.org/10.1145/3490354.3494453 (Association for Computing Machinery, November 3-5, 2021, Virtual Event, USA., 2021).

Dyck, A., Morse, A. & Zingales, L. Who blows the whistle on corporate fraud? The Journal of Finance 65 , 2213–2253, https://doi.org/10.1111/j.1540-6261.2010.01614.x (2010).

Zavitsanos, E., Mavroeidis, D., Spyropoulou, E., Fergadiotis, M. & Paliouras, G. Entrant: A large financial dataset for table understanding. Zenodo https://doi.org/10.5281/zenodo.10667088 (2024).

Gonsior, J. et al . Active learning for spreadsheet cell classification. In EDBT/ICDT Workshops (2020).

Koci, E., Thiele, M., Rehak, J., Romero, O. & Lehner, W. Deco: A dataset of annotated spreadsheets for layout and table recognition. In International Conference on Document Analysis and Recognition , https://doi.org/10.1109/ICDAR.2019.00207 (IEEE, 2019).

Sun, K., Rayudu, H. & Pujara, J. A hybrid probabilistic approach for table understanding. In Proceedings of the AAAI Conference on Artificial Intelligence , vol. 35, 4366–4374, https://doi.org/10.1609/aaai.v35i5.16562 (2021).

Bougiatiotis, K., Zavitsanos, E. & Paliouras, G. Identifying going concern issues in auditor opinions: link to bankruptcy events. In 2023 IEEE International Conference on Big Data (BigData) , 2805–2813, https://doi.org/10.1109/BigData59044.2023.10386452 (2023).

Download references

Acknowledgements

We appreciate the support of “SKEL ∣ The AI Lab” of the Institute of Informatics and Telecommunications, NCSR “Demokritos” for providing computational and storage resources. We would also like to acknowledge the financial support of Qualco SA. The opinions of the authors expressed herein do not necessarily state or reflect those of Qualco SA.

Author information

Authors and affiliations.

Institute of Informatics and Telecommunications, National Centre for Scientific Research “Demokritos”, Aghia Paraskevi, 15341, Greece

Elias Zavitsanos, Dimitris Mavroeidis, Eirini Spyropoulou, Manos Fergadiotis & Georgios Paliouras

You can also search for this author in PubMed   Google Scholar

Contributions

Work conception and design: E.Z., E.S., M.F., D.M., G.P. Data acquisition: D.M. Data transformations and analysis: E.Z. Revisions of code, workflow and unit testing: E.Z., D.M., M.F. Data usability: M.F., D.M. Wrote the paper: E.Z. All authors reviewed the manuscript. All authors read and approved the final manuscript.

Corresponding author

Correspondence to Elias Zavitsanos .

Ethics declarations

Competing interests.

The authors declare no competing interests.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/ .

Reprints and permissions

About this article

Cite this article.

Zavitsanos, E., Mavroeidis, D., Spyropoulou, E. et al. ENTRANT: A Large Financial Dataset for Table Understanding. Sci Data 11 , 876 (2024). https://doi.org/10.1038/s41597-024-03605-5

Download citation

Received : 23 February 2024

Accepted : 04 July 2024

Published : 13 August 2024

DOI : https://doi.org/10.1038/s41597-024-03605-5

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

Sign up for the Nature Briefing: AI and Robotics newsletter — what matters in AI and robotics research, free to your inbox weekly.

• Open access
• Published: 07 August 2024

The cGAS-STING pathway in COPD: targeting its role and therapeutic potential

• Kexin Liao 1 ,
• Fengshuo Wang 3 ,
• Chenhao Xia 2 ,
• Sen Zhong 2 ,
• Wenqi Bi 1 &
• Jingjing Ruan 2  

Respiratory Research volume  25 , Article number:  302 ( 2024 ) Cite this article

182 Accesses

Metrics details

Chronic obstructive pulmonary disease(COPD) is a gradually worsening and fatal heterogeneous lung disease characterized by airflow limitation and increasingly decline in lung function. Currently, it is one of the leading causes of death worldwide. The consistent feature of COPD is airway inflammation. Several inflammatory factors are known to be involved in COPD pathogenesis; however, anti-inflammatory therapy is not the first-line treatment for COPD. Although bronchodilators, corticosteroids and roflumilast could improve airflow and control symptoms, they could not reverse the disease. The cyclic GMP-AMP synthase-stimulator of interferon genes (cGAS-STING) signaling pathway plays an important novel role in the immune system and has been confirmed to be a key mediator of inflammation during infection, cellular stress, and tissue damage. Recent studies have emphasized that abnormal activation of cGAS-STING contributes to COPD, providing a direction for new treatments that we urgently need to develop. Here, we focused on the cGAS-STING pathway, providing insight into its molecular mechanism and summarizing the current knowledge on the role of the cGAS-STING pathway in COPD. Moreover, we explored antagonists of cGAS and STING to identify potential therapeutic strategies for COPD that target the cGAS-STING pathway.

Introduction

COPD is a progressive and debilitating respiratory disease that affects millions of people worldwide and poses a considerable medical and financial burden [ 1 , 2 ]. Traditionally, COPD is considered as an inflammatory response elicited by cigarette smoking(CS) in older males [ 3 ]. In addition, other factors, such as air pollution, occupational particles and aging, have also been found to trigger lung inflammation, with COPD subsequently accompanied by inflammation [ 2 ].

Pattern recognition receptors (PRRs) are a significant component of the innate immune system and constitute the first line of defense in organisms. As a member of the PPR family, the cGAS protein acts as an innate nucleic acid sensor recognizing exogenous DNA generated by viral or bacterial infection or in the cytoplasm and converts ATP and GTP into 2’3’-cyclic GMP-AMP (cGAMP), which can be used to monitor pathogen infection or cellular stress [ 4 ]. cGAMP binds to the adapter protein stimulator of interferon genes (STING) localized at the endoplasmic reticulum (ER) membrane [ 5 ] and initiates a downstream immune response. Increasing research suggests that the cGAS-STING pathway plays an important role in the development of many diseases through its involvement in autoimmunity, cellular senescence and anti-inflammation [ 6 ].

The Global Initiative of Obstructive Lung Disease (GOLD) has suggested guidelines for COPD management. However, symptomatic treatment involving bronchodilators continues to be the mainstay in COPD management, despite the understanding of inflammation as a key driver of COPD progression. There is currently no cure for COPD. Recent efforts have tended to focus on the molecular mechanisms underlying COPD to explore therapeutic targets for COPD. Studies have shown that cGAS-STING contributes to COPD, especially under exposure to cigarette smoking [ 7 ] or air pollutants including silica [ 8 ] and PM 2.5 [ 9 ]. Moreover, targeting the cGAS-STING pathway can circumvent cellular senescence [ 10 , 11 ], which has also been shown to contribute to the accelerated aging process in COPD patients [ 12 ]. Delving into the structure and function of the cGAS-STING pathway may enable the development of selective small-molecule inhibitors to manage the inflammation associated with COPD. In this review, we discussed the role of the cGAS-STING pathway in the pathogenesis of COPD, as well as antagonists of this pathway, with a focus on its therapeutic potential for COPD. Our aim is to contribute to the optimization of fundamental therapies for COPD, ultimately improving patient prognosis.

Inflammation-associated mechanisms in the pathogenesis of COPD

Airway inflammation is a consistent feature of COPD and plays an important role in the disease pathogenesis, progression and mortality [ 2 , 13 ]. Inflammation has many manifestations. In this paragraph, we describe neutrophil-associated and eosinophil-associated inflammation in COPD as well as some relevant inflammatory signaling pathways.

Neutrophil-associated airway inflammation in COPD

Neutrophil inflammation is the key inflammatory phenotype in the pathogenesis of COPD, with increased neutrophils in sputum and blood being a characteristic feature of all COPD patients. Studies have reported that neutrophil count is a marker of COPD severity and patients with higher sputum neutrophil percentages have greater dyspnea scores [ 2 , 14 , 15 ]. When stimulated by inflammation, neutrophils leave the circulation to congregate in lungs. The aggregation of neutrophils produces a large amount of reactive oxygen species (ROS), which can destroy lung tissues [ 16 ]. Moreover, neutrophils produce the inflammatory factor IL-6, which induces the production of elastase and oxygen free radicals, thereby increasing pulmonary vascular permeability and exacerbating lung tissue destruction [ 17 ]. Neutrophils accumulate in the airways of COPD patients [ 18 ] and can secrete serine proteases including matrix metalloproteinase (MMP) and neutrophil elastase (NE) [ 19 ]. MMP is significantly increased in patients with COPD and destroys the structural components of extracellular matrix (ECM), contributing to alveolar destruction [ 20 ]. In animal models, dominant-negative MafB was shown to downregulated MMP, thereby suppressing porcine pancreatic elastase-induced emphysema [ 21 ]. NE is a neutrophil-derived serine proteinase and has proven to be involved in lung damage. A study revealed that NE deficiency in mice protects them from emphysema after exposure to cigarette smoke (CS) [ 22 ]. The underlying mechanism may be that NE can also degrade the structural components of ECM and cooperate with MMPs to amplify the degradation [ 23 ]. In addition, NE is effective in stimulating mucus secretion from submucosal glands and thrush cells, leading to airway obstruction [ 24 ]. All these findings indicate the contribution of neutrophil inflammation to the development of COPD.

Eosinophilic-associated airway inflammation in COPD

Although neutrophil-associated COPD is the most common inflammatory phenotype, it has been recognized that eosinophils may also be involved in the inflammatory response in COPD. Approximately 10-40% of COPD patients demonstrate increased eosinophilic inflammation in the sputum or blood [ 25 ]. Eosinophilic airway inflammation occurs in COPD exacerbations. Clinical research has shown that patients with high eosinophil count but a low percentage of macrophages exhibit the greatest decline in lung function during an exacerbation and a greater exacerbation frequency. This group of patients has persistent eosinophilic inflammation due to defective macrophage efferocytosis, which contributes to the severity of the disease [ 26 ]. Like in asthma, recruitment of eosinophils to the airway in COPD is mediated via CCR3 chemokines, which play a critical role together with other eosinophil chemoattractants, such as prostaglandin (PG)D2 [ 27 , 28 ]. Inflammatory cues prompt the recruitment of eosinophils into the lungs, where the secretion of a variety of chemokines (e.g., CCL5, CCL11, CCL13), cytokines (e.g., IL-2, IL-3, IL-4, IL-5, IL-10, IL-12, IL-13, IL-16, IL-25) and cytotoxic granular products (major basic protein, eosinophil cationic protein, eosinophil peroxidase, eosinophil-derived neurotoxin) contributes to inflammation [ 29 ]. An increase in eosinophilic inflammation in peripheral blood and sputum samples from COPD patients is associated with an increased risk of severe deterioration in the future [ 30 ]. However, the etiology of eosinophilic inflammation in COPD is not completely understood.

Inflammation-associated pathways in COPD

The pathogenesis of COPD involves the activation of diverse inflammatory pathways. The NF-κB pathway is activated by the ubiquitination of IκB [ 31 ]. As a result, NF-κB is released from the NF-κB/IκB complex and are able to bind to target genes, thereby initiating the expression of target genes, such as TNF-α and IL-1, and causing an inflammatory response [ 32 ]. In addition, the inhalation of ozone and cigarette smoke results in the migration of neutrophils into the lungs to generate ROS, which is another factor in the activation of NF-κB [ 33 ]. A study in mouse models of COPD demonstrated that NF-kB pathway is essential for inflammation in smoking-induced bronchiolitis [ 34 ]. Moreover, hypomethylation of NF-κB-mediated pathway genes has also been conformed to contribute to COPD exacerbation [ 35 ].

The mitogen-activated protein kinase (MAPK) pathway participates in stress adaptation and inflammatory responses and its activation can stimulate cytokines, neurotransmitters, serine proteases, and oxidative stress [ 36 ]. Haemophilus influenzae is a common pathogen of COPD, and it was found to upregulate MUC gene transcription through the activation of MAPK signaling pathway [ 37 ]. In addition, IL-8 and TNF-α are key factors in COPD development and are are also regulated by p38MAPK [ 38 ].

Many other pathways including EGFR signaling pathway [ 39 ], MARCKS protein signaling pathway [ 40 ], SNARE protein signaling pathway [ 41 ], and Ciliophagy signaling pathway [ 42 ] are associated with airway mucus hypersecretion, which is recognized as one of the main pathophysiological changes in COPD patients. The cGAS-STING signaling pathway is also highly involved, and we elaborate on this pathway in this review.

Overview of the cGAS-STING pathway

Sun et al. identified cGAS through isolation and purification in 2013 [ 43 ], and revealed a novel immune signaling pathway, namely, the cGAS-STING signaling pathway. This pathway occurs within cells and is highly important for immune systems that can sense double-stranded DNA (dsDNA) to defend against extracellular or intracellular pathogens [ 4 ]. The cGAS-STING pathway has emerged as a critical mechanism for the induction of powerful innate immune defense programs [ 44 ].

In many organisms, the detection of foreign DNA is a key factor in immunity. In mammalian cells, this process is largely facilitated by the cGAS, which has become an important mechanism for combining DNA perception with the induction of powerful innate immune defense strategies [ 45 , 46 ]. PRRs are essential components of the innate immune system that can recognize biomolecules such as pathogen-associated molecular patterns (PAMPs) and DAMPs. PAMPs contain double- or single-stranded DNA and RNA generated by viral or bacterial infection or in the cytoplasm [ 47 ]. The strongest response after DNA stimulation is initiated by cGAS, a member of the PRR family, which is activated after binding to dsDNA [ 43 ] in a minimal 2:2 complex to induce conformational changes that allow cGAS to catalyze ATP and GTP into 2’,3’-cGAMP [ 48 , 49 , 50 ]. The sugar phosphate backbone of DNA-binds to a nucleotide transferase domain (catalytic part) in the C-terminus of cGAS, which includes positively charged DNA binding sites, a primary site, and two additional sites [ 45 ]. cGAMP binds to STING, inducing its phosphorylation of STING and causing its conformational changes, activating downstream signal transduction. During this process, STING undergoes high-order oligomerization to form tetramers [ 51 , 52 ] and is transferred from the endoplasmic reticulum to the intermediate compartment of the endoplasmic reticulum Golgi apparatus. For the past several years, structural studies have shown that the tetramerization of STING in the Golgi complex is a signaling platform for recruiting and activating dimeric TANK-binding kinase 1 (TBK1) dimers through phosphorylation [ 53 ]. Conversely, TBK1 transphosphorylates the C-terminal domain of STING to recruit interferon regulatory factor 3 (IRF3) for activation [ 54 ], where IRF3 translocates to the nucleus. This gene plays a transcriptional role in the expression of immune stimulating genes (ISGs) and type 1 interferons (IFNs) [ 43 , 50 ]. Moreover, STING also activated IκB kinase (IKK)-mediated induction of NF-κB-driven inflammatory genes. After activation, STING is transported to the inner lysosome for degradation, while NF-κB translocates to the nucleus, where it triggers the the expression of proinflammation cytokines(e.g., TNF and IL-6) [ 55 , 56 ]. Activated STING passes through signal transduction pathways, ultimately leading to the production of a large amount of interferon and other immune related cytokines, thereby triggering an immune response. In addition, the binding of cGAS to DNA is independent of the DNA sequence [ 57 ]. Therefore, theoretically, self DNA from mitochondria or nuclei can also act as a cGAS ligand, activating the cGAS-STING pathway and triggering inflammatory responses [ 58 ]. Recent studies have also shown that endogenous cGAS is tightly bound to the nucleus and prevents its self response to its own DNA [ 59 , 60 ]. Moreover, other studies have shown that cGAS inhibits homologous recombination mediated DNA repair and promotes genomic instability, micronucleus generation, and cell death under genomic stress conditions in a manner independent of STING by other studies [ 61 ]. With increasing researches on cGAS and STING, the cGAS-STING pathway has been revealed to be involved in autoimmunity, cellular senescence and inflammation inhibition, indicating that it plays an important role in the occurrence of inflammation and many diseases [ 6 , 62 , 63 ].(Fig. 1 ).

Overview of the cGAS–STING pathway

A schematic detailing the cGAS-STING signaling pathway. Upon binding dsDNA, cGAS dimers assemble on dsDNA to generate 2’-3’cGAMP. cGAMP binds to STING, leading to the translocation of STING from the ER to the Golgi and ER-Golgi intermediate compartment (ERGIC). This activation of STING recruits TBK1 and IKK, promoting their autophosphorylation and triggering the phosphorylation of IRF3 and IκB. Phosphorylated IRF3 translocates to the nucleus, where it results in the gene expression of type I interferons. Phosphorylated IκB recruits NF-κB, induces expression of genes encoding proinflammatory cytokines, and is subsequently degraded.

The cGAS-STING pathway in COPD

The cgas-sting pathway in cigarette smoke-induced copd.

The significance of the cGAS-STING pathway in the inflammatory response is well known. As mentioned earlier, smoking is a key factor that induces COPD. In 2016, Pouwels revealed that exposure to a smoke environment can induce the release of dsDNA and mtDNA in mouse bronchoalveolar tissues in vivo, leading to cell death in human bronchial epithelial cells, and increased release of dsDNA and mtDNA was detected in the extracellular environment [ 64 ]. Moreover, sensing of these two types of DNA opens up the cGAS-STING immune pathway. In a 2019 study, Sears reported that exposure to CS triggers cGAS and STING expression at the mRNA and protein levels. Importantly, DNA damage repair defects are related to the pathogenesis of COPD, indicating that DNA release and sensing play a crucial role [ 65 ]. In mice and COPD patients, macrophage uptake of nanoparticulate carbon black induces DNA repair enzymes, leading to dsDNA breakage and an inflammatory response through activation of the cGAS-STING signaling pathway [ 66 ]. Moreover, in 2019, Nascimento, using gene-deficient mouse strains, reported that the absence of cGAS or STING can lead to reduced lung inflammation. In this study, the BALF of COPD model mice showed overexpression of cGAS and STING, while the DNA content increased, and neutrophil recruitment increased. Compared with wild-type mice, mice with STING and cGAS gene knockout showed a significant decrease in dsDNA content in the BALF and in the production of the downstream chemokine CXCL10 and relatively mild lung inflammation, while TLR-9 gene knockout mice showed no significant changes. This finding suggested that the self-DNA released after exposure to cigarette smoke is recognized by cGAS rather than by TLR-9, which activates the STING pathway and leads to increased secretion of type I IFN, promoting pulmonary inflammation [ 7 ]. In addition, the release of some self-dsDNA is dependent on STING, indicating that lung injury induces de novo cell death and self-dsDNA-dependent lung inflammation through amplification loops [ 67 ].

The cGAS-STING pathway in particulate matter-induced COPD

The long-term inhalation of particulate matter, such as silica and PM 2.5, is another major cause of COPD and can also lead to chronic lung inflammation. Exposure to silica particles can induce the release of proinflammatory and profibrotic factors (e.g., IL-6, TNF-α, and TGF-β), which contribute to the acceleration of lung inflammation, and the activation of the cGAS-STING signaling pathway is involved in this process. Benmerzoug reported that mitochondria can be a source of self-dsDNA triggering DNA sensor activation after exposure to silica particles, triggering the type I IFN pathway and inducing cell death in the lungs [ 8 ]. After silica exposure, both the STING and NLRP3 pathways were activated, leading to cell death and the release of proinflammatory cytokines. This process leads to necrosis and apoptosis in a STING-dependent manner. Another study conducted by Wang in 2022 [ 9 ] revealed that PM2.5-induced aging is regulated by an inflammatory response that is activated by the cGAS/STING/NF-κB pathway, which is closely related to DNA damage. Their study also showed that pretreatment with selenomethionine (Se Met) can inhibit the inflammatory response and prevent cell aging by blocking the cGAS/STING pathway in A549 cells exposed to PM2.5. In addition, the in vivo C57BL/6J mouse model showed a decrease in cGAS expression after Se Met treatment, which can alleviate PM2.5-induced lung tissue aging in mice.

Collectively, the studies discussed here have established the implications and characteristics of STING signaling activation in COPD development. These findings suggest that known COPD-causative factors (e.g., CS, silica, and PM 2.5) can trigger the activation of the cGAS-STING signaling pathway and that targeting this pathway could help alleviate inflammation in COPD patients. Nonetheless, most of the studies discussed here were mouse model-based; therefore, there is an urgent need for human-based research to elucidate the involvement of STING in COPD.

Antagonists of the cGAS-STING pathway

There is no cure for COPD; however, the emergence of the first drug-like compounds selectively targeting cGAS or STING has opened the door for the development of clinical candidates. Several small-molecule agonists have been developed and are being tested in tumor immunotherapy [ 68 ]. Although the cGAS-STING pathway antagonists have the optimal beneficial effects on tumor diseases, prominent efforts are underway to develop novel compounds to control the severe inflammation and acute tissue damage observed from chronic stimulation by antagonizing cGAS and STING. (Table  1 )

Inhibitors of cGAS

Numerous cGAS antagonists have been identified as favorable targets for ameliorating cGAS- and STING-dependent inflammatory diseases. One example is antimalarial drugs (e.g., hydroxychloroquine, quinacrine, suramin, and oligodeoxynucleotides A151 and X6), which specifically bind to two drug sites on the 2:2 cGAS/dsDNA dimer and have the potential to suppress cGAS activity [ 69 , 71 , 72 ]. At each site, the antimalarial drug acts in the dsDNA minor groove between the cGAS/DNA interface (the interface connecting the dsDNA-binding site A/B on two neighboring monomers). This results in the interaction of the antimalarial drug with the DNA-binding sites (A and B) on the two neighboring cGAS monomers, which alters the stability of the cGAS/dsDNA complex and thus inhibits the activation of cGAS by dsDNA and its enzymatic activities. Follow-up studies have indicated that in the presence of the antimalarial drugs studied, IFN-β expression, cGAMP production, and the levels of a number of dsDNA-stimulated cytokines (IL-6 and TNF-α) are inhibited.

Furthermore, RU.521 affected cGAS activity by occupying the catalytic site of cGAS and decreasing its binding affinity for ATP and GTP without directly interfering with dsDNA binding, as identified in mouse studies [ 73 ]. It was demonstrated to reduce Ifnb1 mRNA expression in bone marrow-derived macrophages from Trex1-/- mice. PF-06928125 has also been shown to act as a cGAS inhibitor [ 74 ]. High-throughput biochemical screening of cGAS inhibitors identified PF-06928215. Although PF-06928215 was also able to bind to the cGAS active site with a high affinity value of 0.2 µmol/L and inhibit cGAS activity with an IC50 value of 4.9 µmol/L, PF-06928215 showed no activity in the cellular cGAS assay [ 80 ].

Inhibitors of STING

STING is the critical signaling molecule for the cGAS–STING pathway; therefore, developing antagonists of STING may exploit cGAS–STING inhibitors. Haag et al. reported that nitrofuran derivatives, including C-170, C-171, C-176, C-178, and H-151, can block the STING-mediated signaling pathway by covalently modifying the Cys91 residue of STING [ 77 ]. Cys91 in STING has been shown to be targeted by the covalent ligand BPK-25, which inhibits STING activation by disrupting the binding of the cyclic dinucleotide ligand cGAMP [ 81 ]. Moreover, Tetrahydroisoquinolinone acetate (Compound 18) stabilizes the open, inactive conformation of STING and binds to the cGAMP binding site in a 2:1 ratio, displacing cGAMP from its binding site on STING. Compound 18 potently inhibited in vitro cGAMP-dependent signaling and displayed slow dissociation kinetics and good oral bioavailability [ 75 ]. Astin C is a natural cyclopeptide derived from the traditional Chinese medicinal plant Aster tataricus and was identified by Li et al. as a potent bioactive compound that restricts the cGAS-STING signaling to alleviate autoinflammatory response in a Trex1−/− mouse model and in macrophages [ 76 ]. It was demonstrated that astin C binds competitively to the CDN site via pull-down experiments using biotinylated astin C and human STING. Astin C blocks the recruitment of IRF3 to the STING signalosome, thus preventing downstream signaling through this pathway.

Inhibitors of unknown mechanisms

Aside from the findings discussed above, there are also several inhibitors whose mechanisms are still unknown. For example, the small molecules CU-32 and CU-76 bind to cGAS without disrupting the binding between cGAS and dsDNA; these small molecules can inhibit the protein − protein interactions (PPIs), interfaces required for IRF3 activation and downstream IFN-I induction, in human monocyte THP-1 cells, but the exact mechanism is still unclear [ 78 ]. Additionally, the small heterocyclic compound VS-X4 has been shown to inhibit STING with no elucidated mechanism of action [ 79 ]. Further research is needed to reveal the specific mechanisms of action of these inhibitors.

cGAS-STING inhibitors in clinical trials

Currently, some of these compounds are being phased into clinical studies. María Gómez Antúnez reported a higher survival rate in COPD patients hospitalized with SARS-CoV-2 treated with hydroxychloroquine, but they only recommend its use in clinical trials [ 82 ]. Moreover, studies has shown that Quinacrine is a potential treatment for COVID-19 virus infection [ 83 ] and that Nitrofuran is a therapy for uncomplicated lower urinary tract infection in women [ 84 ], indicating their anti-inflammatory roles. However, many of these studied inhibitors primarily treat tumors rather than COPD or other inflammatory diseases. More researches should be devoted to controlling inflammation through antagonizing the cGAS-STING pathway, and we expect that more inhibitors of cGAS-STING will enter clinical trials for COPD treatment in the future.

Targeting the cGAS-STING signaling pathway alleviates COPD

A better understanding of the cGAS-STING signaling pathway has led to the identification of several potential therapies for inhibiting inflammation; these therapies have been termed “protectors of COPD patients”. Recent studies focused on cGAS and STING have provided new directions for treating COPD.

Circumventing cellular senescence attenuates COPD by targeting the cGAS-STING signaling pathway

Cellular senescence is a state of cell cycle arrest and is among the 9 hallmarks of aging proposed by López-Otín in a landmark paper [ 85 ]. Senescent cells including alveolar epithelial and endothelial cells accumulate in the lungs COPD patients [ 86 , 87 ], and CS-induced oxidative stress is likely to play an important role in the induction of senescence in COPD [ 88 ]. In a mouse model-based study, Takao demonstrated the induction of senescence in lung parenchymal cells during the progression of COPD [ 89 ].And evidence from clinical samples of primary human bronchial epithelial cells and lung homogenates from COPD patients indicates the same conclusion [ 90 ].

Studies have shown that interfering with DNA binding to the DNA sensor cGAS can induce cellular senescence [ 10 , 11 ]. Aging-accelerated factors induce the proinflammatory Senescence Associated Secretory Phenotype (SASP), leading to leakage of DNA into the cytoplasm and triggering of the cGAS-STING pathway of the innate immune response [ 12 ]. The cGAS-STING also induces the SASP phenotype by accumulating cytoplasmic DNA during senescence [ 10 , 91 ], thus aggravating the aging response. Wang et al. reported that Se-Met treatment prevents PM2.5-induced senescence via attenuating inflammatory response regulated by cGAS/STING/NF-κB pathway, and further causes a reduction in COPD [ 9 ]. Since cGAS-STING pathway plays important roles in COPD, these studies further indicate that targeting this pathway may circumvent cellular senescence and thus has therapeutic potential for mitigating COPD.

Natural products relieve COPD by targeting the cGAS-STING signaling pathway

The Tanreqing (TRQ) injection is a Chinese patent medicine. It can significantly improve the partial pressure of oxygen (PaO2), partial pressure of carbon dioxide (PaCO2) and lung function in patients with COPD combined with respiratory failure, and is commonly used to treat AECOPD. Several in vivo experimental studies have also revealed that TRQI can reduce the expression of IL-8, TNF-α and mucin 5AC (MUC5A) in alveolar lavage fluid in CS-and LPS-induced rats with COPD, thereby improving the inflammatory response of airway mucosa and inhibiting airway mucus hypersecretion in rats [ 92 ]. TRQ injection inhibited STING levels, suggesting that TRQ has therapeutic efficacy by blocking the increase in the cGAS-STING pathway in COPD patients [ 93 ]. However, the specific mechanism of action of TRQI for the treatment of COPD is still unclear. With respect to the five traditional Chinese medicines, further experiments are needed to identify the specific components of TRQ that regulate the cGAS-STING pathway and alleviate COPD.

Panax ginseng C.A Meyer (ginseng) root is another important traditional Chinese medicinal herb. Its pharmacologically active constituents have been identified, most notably ginsenosides, which are triterpenoid saponins that include protopanaxadiol (PPD) and protopanaxatriol (PPT) [ 94 ]. The PPD group contains ginsenosides Rb1, Rb2, Rb3, Rc, Rd, Rg3 and Rh2 and compound K, while the PPT group comprises ginsenosides Re, Rf, Rg1, Rg2 and Rh1. These compounds possess pharmacological effects, such as antiviral, antioxidant, and immunomodulatory activities, and have potentially relevant effects on COPD, including the inhibition of proinflammatory mediators and cytokines [ 95 ]. Recently, studies have shown that ginsenosides can alleviate COPD and reduce lung injury [ 96 , 97 ]. In a CS-induced BALV/c mouse model, X. Guan et al. reported that ginsenosides Rg3 and Rb3 could negatively regulate PI3K activation, NF-κB activity, and proinflammatory cytokines in a CS-induced BALV/c rat model, basal cells, and a coculture model of bronchial epithelial cells and neutrophils, thus reducing neutrophil migration. Moreover, ginsenosides basically inhibit various COPD-related pathogenesis processes, such as inflammatory responses (TNF-α, IL-6, IL-1β, NF-κB induction and translocation), kinase phosphorylation (MAPK and ERK1/2), and oxidative stress (ROS) [ 98 ]. Mechanistically, PPD suppressed the cGAS-SING pathway through the activation of AMPK and the inhibition of TNF, IL and NF-κB. The therapeutic effect of PPD in COPD patients awaits further clinical investigation.

Both Radix Pseudostellariae and Juglanin, which are types of tonic Chinese medicine, have been shown to reduce lung inflammation by inhibiting the cGAS-STING pathway and therefore alleviating COPD [ 99 , 100 ]. The findings above suggest that natural products, especially Chinese medical herbs, can alleviate inflammation by inhibiting the cGAS-SING pathway, thus exerting therapeutic effects on COPD. This finding provides potential avenues for future drug development and therapeutic strategies for this disease.

Discussion: summary, outstanding questions, and future directions

Airway inflammation is a consistent characteristic of COPD and is related to its pathogenesis and progression. As an innate immune pathway, cGAS-STING plays an undeniable role in immune-related diseases. The cGAS-STING pathway has been shown to alleviate inflammatory responses and lung function damage in COPD patients, indicating its potential as a therapeutic target [ 7 ]. This pathway offers a specific and selective means to modulate immune responses, particularly in DNA-induced inflammation associated with COPD. Current preclinical development efforts are focused on several cGAS and STING inhibitors, including antimalarial drugs [ 77 ], RU.521 [ 73 ], PF-06928215 [ 74 ], nitrofuran [ 77 ], compound 18 [ 75 ] and astin C [ 76 ], ect., which are expected to open up new avenues for treating COPD. Furthermore, targeting the cGAS-STING signaling pathway circumvents cellular senescence to attenuate COPD [ 10 , 11 ] and an increasing number of natural products, mainly Chinese herbs, have been discovered to alleviate COPD via a pharmacological mechanism involving cGAS-STING [ 93 , 98 , 99 , 100 ].

However, the pathogenesis of COPD is not fully understood. Most relevant studies are animal-based experiments rather than human-based ones, which makes their applicability in humans a question. Although the cGAS-STING pathway is highly involved in the inflammatory mechanism of COPD, it is merely a fraction of the overall picture, necessitating further exploration of additional mechanisms. Moreover, inhibitors of the cGAS-STING signaling pathway have not been extensively employed as COPD therapeutic drugs, and the mechanisms by which certain drugs function as inhibitors remain unclear.

In this review, we discuss the inflammatory pathogenesis of COPD and provide an overview of the current understanding of the cGAS-STING signaling pathway as well as its potential as a COPD therapeutic target. The cGAS-STING pathway is activated in COPD, and its activation further exacerbates the development of COPD. Targeting the cGAS-STING signaling pathway is highly important for curing COPD, and many studies have suggested that small molecule inhibitors are effective controlling the development of COPD. However, whether COPD is related to cGAS and STING remains an area that requires further research. In conclusion, targeting the cGAS-STING signaling pathway provides a promising direction for COPD therapy and intervention.

Data availability

No datasets were generated or analysed during the current study.

Abbreviations

Chronic Obstructive Pulmonary Disease

cyclic GMP-AMP Synthase

Stimulator of Interferon Genes

Cigarette Smoking

Pattern Recognition Receptors

cyclic GMP-AMP

Endoplasmic Reticulum

The Global Initiative of Obstructive Lung Disease

Damage-Associated Molecular Patterns

Interleukin

Thymic Stromal Lymphopoietin

Innate Lymphoid Cells

Monocyte Chemotactic Proteins

C-C motif chemokine ligand

double-stranded DNA

Pathogen Associated Molecular Patterns

Damage Associated Molecular Pattern

TANK-binding Kinase 1

Interferon Regulatory Factor 3

Immune Stimulating Genes

Interferons

ER-Golgi Intermediate Compartment

Selenomethionine

Protein − protein Interface

Senescence Associated Secretory Phenotype

Partial Pressure of Oxygen

Pressure of Carbon Dioxide

Panax Ginseng C.A Meyer

Protopanaxadiol

Protopanaxatriol

Reactive Oxygen Species

Chen S, Kuhn M, Prettner K, Yu F, Yang T, Barnighausen T, Bloom DE, Wang C. The global economic burden of chronic obstructive pulmonary disease for 204 countries and territories in 2020-50: a health-augmented macroeconomic modelling study. LANCET GLOB HEALTH. 2023;11(8):e1183–93.

Article   CAS   PubMed   PubMed Central   Google Scholar  

Agusti A, Celli BR, Criner GJ, Halpin D, Anzueto A, Barnes P, Bourbeau J, Han MK, Martinez FJ, Montes DOM, et al. Global Initiative for Chronic Obstructive Lung Disease 2023 Report: GOLD Executive Summary. AM J RESP CRIT CARE. 2023;207(7):819–37.

Article   CAS   Google Scholar  

Fletcher C, Peto R. The natural history of chronic airflow obstruction. Br Med J. 1977;1(6077):1645–8.

Civril F, Deimling T, de Oliveira MC, Ablasser A, Moldt M, Witte G, Hornung V, Hopfner KP. Structural mechanism of cytosolic DNA sensing by cGAS. Nature. 2013;498(7454):332–7.

Ishikawa H, Barber GN. STING is an endoplasmic reticulum adaptor that facilitates innate immune signalling. Nature. 2008;455(7213):674–8.

Gulen MF, Samson N, Keller A, Schwabenland M, Liu C, Gluck S, Thacker VV, Favre L, Mangeat B, Kroese LJ, et al. cGAS-STING drives ageing-related inflammation and neurodegeneration. Nature. 2023;620(7973):374–80.

Nascimento M, Gombault A, Lacerda-Queiroz N, Panek C, Savigny F, Sbeity M, Bourinet M, Le Bert M, Riteau N, Ryffel B, et al. Self-DNA release and STING-dependent sensing drives inflammation to cigarette smoke in mice. SCI REP-UK. 2019;9(1):14848.

Article   Google Scholar  

Benmerzoug S, Rose S, Bounab B, Gosset D, Duneau L, Chenuet P, Mollet L, Le Bert M, Lambers C, Geleff S, et al. STING-dependent sensing of self-DNA drives silica-induced lung inflammation. NAT COMMUN. 2018;9(1):5226.

Wang X, Lu W, Xia X, Zhu Y, Ge C, Guo X, Zhang N, Chen H, Xu S. Selenomethionine mitigate PM2.5-induced cellular senescence in the lung via attenuating inflammatory response mediated by cGAS/STING/NF-kappaB pathway. ECOTOX ENVIRON SAFE. 2022;247(12):114266.

Gluck S, Guey B, Gulen MF, Wolter K, Kang TW, Schmacke NA, Bridgeman A, Rehwinkel J, Zender L, Ablasser A. Innate immune sensing of cytosolic chromatin fragments through cGAS promotes senescence. NAT CELL BIOL. 2017;19(9):1061–70.

Yang H, Wang H, Ren J, Chen Q, Chen ZJ. cGAS is essential for cellular senescence. P NATL ACAD SCI USA. 2017;114(23):E4612–20.

Chilosi M, Carloni A, Rossi A, Poletti V. Premature lung aging and cellular senescence in the pathogenesis of idiopathic pulmonary fibrosis and COPD/emphysema. TRANSL RES. 2013;162(3):156–73.

Article   CAS   PubMed   Google Scholar  

Celli BR, Locantore N, Yates J, Tal-Singer R, Miller BE, Bakke P, Calverley P, Coxson H, Crim C, Edwards LD, et al. Inflammatory biomarkers improve clinical prediction of mortality in chronic obstructive pulmonary disease. AM J RESP CRIT CARE. 2012;185(10):1065–72.

Singh D, Edwards L, Tal-Singer R, Rennard S. Sputum neutrophils as a biomarker in COPD: findings from the ECLIPSE study. RESP RES. 2010;11(1):77.

Mendy A, Forno E, Niyonsenga T, Gasana J. Blood biomarkers as predictors of long-term mortality in COPD. CLIN RESPIR J. 2018;12(5):1891–9.

Zhang MY, Jiang YX, Yang YC, Liu JY, Huo C, Ji XL, Qu YQ. Cigarette smoke extract induces pyroptosis in human bronchial epithelial cells through the ROS/NLRP3/caspase-1 pathway. LIFE SCI. 2021;269:119090.

Shyam PSB, Chaya SK, Kumar VS, Mahendra M, Jayaraj BS, Lokesh KS, Ganguly K, Mahesh PA. Inflammatory Biomarkers Interleukin 1 Beta (IL-1beta) and Tumour Necrosis Factor Alpha (TNF-alpha) Are Differentially Elevated in Tobacco Smoke Associated COPD and Biomass Smoke Associated COPD. TOXICS 2021, 9(4).

Benjamin JT, Plosa EJ, Sucre JM, van der Meer R, Dave S, Gutor S, Nichols DS, Gulleman PM, Jetter CS, Han W et al. Neutrophilic inflammation during lung development disrupts elastin assembly and predisposes adult mice to COPD. J CLIN INVEST 2021, 131(1).

Genschmer KR, Russell DW, Lal C, Szul T, Bratcher PE, Noerager BD, Abdul RM, Xu X, Rezonzew G, Viera L, et al. Activated PMN exosomes: pathogenic entities causing Matrix Destruction and Disease in the lung. Cell. 2019;176(1–2):113–26.

Mahor D, Kumari V, Vashisht K, Galgalekar R, Samarth RM, Mishra PK, Banerjee N, Dixit R, Saluja R, De S, et al. Elevated serum matrix metalloprotease (MMP-2) as a candidate biomarker for stable COPD. BMC PULM MED. 2020;20(1):302.

Aida Y, Shibata Y, Abe S, Inoue S, Kimura T, Igarashi A, Yamauchi K, Nunomiya K, Kishi H, Nemoto T, et al. Inhibition of elastase-pulmonary emphysema in dominant-negative MafB transgenic mice. INT J BIOL SCI. 2014;10(8):882–94.

Guyot N, Wartelle J, Malleret L, Todorov AA, Devouassoux G, Pacheco Y, Jenne DE, Belaaouaj A. Unopposed cathepsin G, neutrophil elastase, and proteinase 3 cause severe lung damage and emphysema. AM J PATHOL. 2014;184(8):2197–210.

Guo Y, Ma L, Zhang F, Sun R, Li T. Neutrophil elastase ameliorates matrix metalloproteinase-9 to promote lipopolysaccharide-induced acute lung injury in mice 1. ACTA CIR BRAS. 2016;31(6):382–8.

Article   PubMed   Google Scholar  

Arai N, Kondo M, Izumo T, Tamaoki J, Nagai A. Inhibition of neutrophil elastase-induced goblet cell metaplasia by tiotropium in mice. EUR RESPIR J. 2010;35(5):1164–71.

Yun JH, Lamb A, Chase R, Singh D, Parker MM, Saferali A, Vestbo J, Tal-Singer R, Castaldi PJ, Silverman EK, et al. Blood eosinophil count thresholds and exacerbations in patients with chronic obstructive pulmonary disease. J ALLERGY CLIN IMMUN. 2018;141(6):2037–47.

Eltboli O, Bafadhel M, Hollins F, Wright A, Hargadon B, Kulkarni N, Brightling C. COPD exacerbation severity and frequency is associated with impaired macrophage efferocytosis of eosinophils. BMC PULM MED. 2014;14:112.

Article   PubMed   PubMed Central   Google Scholar  

Ying S, Robinson DS, Meng Q, Rottman J, Kennedy R, Ringler DJ, Mackay CR, Daugherty BL, Springer MS, Durham SR, et al. Enhanced expression of eotaxin and CCR3 mRNA and protein in atopic asthma. Association with airway hyperresponsiveness and predominant co-localization of eotaxin mRNA to bronchial epithelial and endothelial cells. EUR J IMMUNOL. 1997;27(12):3507–16.

Hirai H, Tanaka K, Yoshie O, Ogawa K, Kenmotsu K, Takamori Y, Ichimasa M, Sugamura K, Nakamura M, Takano S, et al. Prostaglandin D2 selectively induces chemotaxis in T helper type 2 cells, eosinophils, and basophils via seven-transmembrane receptor CRTH2. J EXP MED. 2001;193(2):255–61.

Davoine F, Lacy P. Eosinophil cytokines, chemokines, and growth factors: emerging roles in immunity. FRONT IMMUNOL. 2014;5:570.

Bafadhel M, Peterson S, De Blas MA, Calverley PM, Rennard SI, Richter K, Fageras M. Predictors of exacerbation risk and response to budesonide in patients with chronic obstructive pulmonary disease: a post-hoc analysis of three randomised trials. LANCET RESP MED. 2018;6(2):117–26.

Hoffmann A, Levchenko A, Scott ML, Baltimore D. The IkappaB-NF-kappaB signaling module: temporal control and selective gene activation. Science. 2002;298(5596):1241–5.

Lee UJ, Choung SR, Prakash KV, Lee EJ, Lee MY, Kim YJ, Han CW, Choi YC. Dual knockdown of p65 and p50 subunits of NF-kappaB by siRNA inhibits the induction of inflammatory cytokines and significantly enhance apoptosis in human primary synoviocytes treated with tumor necrosis factor-alpha. MOL BIOL REP. 2008;35(3):291–8.

Zhuan B, Yu Y, Yang Z, Zhao X, Li P. Mechanisms of oxidative stress effects of the NADPH oxidase-ROS-NF-kappaB transduction pathway and VPO1 on patients with chronic obstructive pulmonary disease combined with pulmonary hypertension. EUR REV MED PHARMACO. 2017;21(15):3459–64.

CAS   Google Scholar  

Xu H, Sun Q, Lu L, Luo F, Zhou L, Liu J, Cao L, Wang Q, Xue J, Yang Q, et al. MicroRNA-218 acts by repressing TNFR1-mediated activation of NF-kappaB, which is involved in MUC5AC hyper-production and inflammation in smoking-induced bronchiolitis of COPD. TOXICOL LETT. 2017;280:171–80.

Kaur G, Batra S. Regulation of DNA methylation signatures on NF-kappaB and STAT3 pathway genes and TET activity in cigarette smoke extract-challenged cells/COPD exacerbation model in vitro. CELL BIOL TOXICOL. 2020;36(5):459–80.

Brancho D, Tanaka N, Jaeschke A, Ventura JJ, Kelkar N, Tanaka Y, Kyuuma M, Takeshita T, Flavell RA, Davis RJ. Mechanism of p38 MAP kinase activation in vivo. GENE DEV. 2003;17(16):1969–78.

Gaffey K, Reynolds S, Plumb J, Kaur M, Singh D. Increased phosphorylated p38 mitogen-activated protein kinase in COPD lungs. EUR RESPIR J. 2013;42(1):28–41.

Knobloch J, Jungck D, Kronsbein J, Stoelben E, Ito K, Koch A. LABAs and p38MAPK inhibitors reverse the corticosteroid-insensitivity of IL-8 in Airway smooth muscle cells of COPD. J CLIN MED 2019, 8(12).

Oh SY, Kim YH, Kang MK, Lee EJ, Kim DY, Oh H, Kim SI, Na W, Kang IJ, Kang YH. Aesculetin inhibits airway thickening and mucus Overproduction Induced by Urban Particulate Matter through blocking inflammation and oxidative stress involving TLR4 and EGFR. ANTIOXIDANTS-BASEL 2021, 10(3).

Fang S, Crews AL, Chen W, Park J, Yin Q, Ren XR, Adler KB. MARCKS and HSP70 interactions regulate mucin secretion by human airway epithelial cells in vitro. AM J PHYSIOL-LUNG C. 2013;304(8):L511–8.

Lai Y, Fois G, Flores JR, Tuvim MJ, Zhou Q, Yang K, Leitz J, Peters J, Zhang Y, Pfuetzner RA, et al. Inhibition of calcium-triggered secretion by hydrocarbon-stapled peptides. Nature. 2022;603(7903):949–56.

Cloonan SM, Lam HC, Ryter SW, Choi AM. Ciliophagy: the consumption of cilia components by autophagy. AUTOPHAGY. 2014;10(3):532–4.

Sun L, Wu J, Du F, Chen X, Chen ZJ. Cyclic GMP-AMP synthase is a cytosolic DNA sensor that activates the type I interferon pathway. Science. 2013;339(6121):786–91.

Du M, Chen ZJ. DNA-induced liquid phase condensation of cGAS activates innate immune signaling. Science. 2018;361(6403):704–9.

Kranzusch PJ, Lee AS, Berger JM, Doudna JA. Structure of human cGAS reveals a conserved family of second-messenger enzymes in innate immunity. CELL REP. 2013;3(5):1362–8.

Diner EJ, Burdette DL, Wilson SC, Monroe KM, Kellenberger CA, Hyodo M, Hayakawa Y, Hammond MC, Vance RE. The innate immune DNA sensor cGAS produces a noncanonical cyclic dinucleotide that activates human STING. CELL REP. 2013;3(5):1355–61.

Blander JM, Sander LE. Beyond pattern recognition: five immune checkpoints for scaling the microbial threat. NAT REV IMMUNOL. 2012;12(3):215–25.

Ablasser A, Goldeck M, Cavlar T, Deimling T, Witte G, Rohl I, Hopfner KP, Ludwig J, Hornung V. cGAS produces a 2’-5’-linked cyclic dinucleotide second messenger that activates STING. Nature. 2013;498(7454):380–4.

Gao P, Ascano M, Wu Y, Barchet W, Gaffney BL, Zillinger T, Serganov AA, Liu Y, Jones RA, Hartmann G, et al. Cyclic [G(2’,5’)pA(3’,5’)p] is the metazoan second messenger produced by DNA-activated cyclic GMP-AMP synthase. Cell. 2013;153(5):1094–107.

Li XD, Wu J, Gao D, Wang H, Sun L, Chen ZJ. Pivotal roles of cGAS-cGAMP signaling in antiviral defense and immune adjuvant effects. Science. 2013;341(6152):1390–4.

Shang G, Zhang C, Chen ZJ, Bai XC, Zhang X. Cryo-EM structures of STING reveal its mechanism of activation by cyclic GMP-AMP. Nature. 2019;567(7748):389–93.

Zhao B, Du F, Xu P, Shu C, Sankaran B, Bell SL, Liu M, Lei Y, Gao X, Fu X et al. A conserved PLPLRT/SD motif of STING mediates the recruitment and activation of TBK1. NATURE 2019, 569(7758):718–722.

Zhang C, Shang G, Gui X, Zhang X, Bai XC, Chen ZJ. Structural basis of STING binding with and phosphorylation by TBK1. NATURE 2019, 567(7748):394–398.

Liu S, Cai X, Wu J, Cong Q, Chen X, Li T, Du F, Ren J, Wu YT, Grishin NV, et al. Phosphorylation of innate immune adaptor proteins MAVS, STING, and TRIF induces IRF3 activation. Science. 2015;347(6227):aaa2630.

Gonugunta VK, Sakai T, Pokatayev V, Yang K, Wu J, Dobbs N, Yan N. Trafficking-mediated STING degradation requires sorting to Acidified endolysosomes and can be targeted to Enhance Anti-tumor Response. CELL REP. 2017;21(11):3234–42.

Fitzgerald KA, McWhirter SM, Faia KL, Rowe DC, Latz E, Golenbock DT, Coyle AJ, Liao SM, Maniatis T. IKKepsilon and TBK1 are essential components of the IRF3 signaling pathway. NAT IMMUNOL. 2003;4(5):491–6.

Wu J, Sun L, Chen X, Du F, Shi H, Chen C, Chen ZJ. Cyclic GMP-AMP is an endogenous second messenger in innate immune signaling by cytosolic DNA. Science. 2013;339(6121):826–30.

Sun B, Sundstrom KB, Chew JJ, Bist P, Gan ES, Tan HC, Goh KC, Chawla T, Tang CK, Ooi EE. Dengue virus activates cGAS through the release of mitochondrial DNA. SCI REP-UK. 2017;7(1):3594.

Gentili M, Lahaye X, Nadalin F, Nader G, Puig LE, Herve S, De Silva NS, Rookhuizen DC, Zueva E, Goudot C, et al. The N-Terminal Domain of cGAS Determines Preferential Association with centromeric DNA and Innate Immune activation in the Nucleus. CELL REP. 2019;26(9):2377–93.

Volkman HE, Cambier S, Gray EE, Stetson DB. Tight nuclear tethering of cGAS is essential for preventing autoreactivity. ELIFE 2019, 8.

Jiang H, Xue X, Panda S, Kawale A, Hooy RM, Liang F, Sohn J, Sung P, Gekara NO. Chromatin-bound cGAS is an inhibitor of DNA repair and hence accelerates genome destabilization and cell death. EMBO J. 2019;38(21):e102718.

Gao D, Li T, Li XD, Chen X, Li QZ, Wight-Carter M, Chen ZJ. Activation of cyclic GMP-AMP synthase by self-DNA causes autoimmune diseases. P NATL ACAD SCI USA. 2015;112(42):E5699–705.

Kerur N, Fukuda S, Banerjee D, Kim Y, Fu D, Apicella I, Varshney A, Yasuma R, Fowler BJ, Baghdasaryan E, et al. cGAS drives noncanonical-inflammasome activation in age-related macular degeneration. NAT MED. 2018;24(1):50–61.

Pouwels SD, Zijlstra GJ, van der Toorn M, Hesse L, Gras R, Ten HN, Krysko DV, Vandenabeele P, de Vries M, van Oosterhout AJ, et al. Cigarette smoke-induced necroptosis and DAMP release trigger neutrophilic airway inflammation in mice. AM J PHYSIOL-LUNG C. 2016;310(4):L377–86.

Sears CR. DNA repair as an emerging target for COPD-lung cancer overlap. RESPIR INVESTIG. 2019;57(2):111–21.

You R, Lu W, Shan M, Berlin JM, Samuel EL, Marcano DC, Sun Z, Sikkema WK, Yuan X, Song L, et al. Nanoparticulate carbon black in cigarette smoke induces DNA cleavage and Th17-mediated emphysema. ELIFE. 2015;4:e9623.

Ahn J, Gutman D, Saijo S, Barber GN. STING manifests self DNA-dependent inflammatory disease. P NATL ACAD SCI USA. 2012;109(47):19386–91.

Zhu HF, Li Y. Small-molecule targets in Tumor Immunotherapy. NAT Prod BIOPROSP. 2018;8(4):297–301.

An J, Woodward JJ, Sasaki T, Minie M, Elkon KB. Cutting edge: antimalarial drugs inhibit IFN-beta production through blockade of cyclic GMP-AMP synthase-DNA interaction. J IMMUNOL. 2015;194(9):4089–93.

An J, Woodward JJ, Lai W, Minie M, Sun X, Tanaka L, Snyder JM, Sasaki T, Elkon KB. Inhibition of cyclic GMP-AMP synthase using a Novel Antimalarial Drug Derivative in Trex1-Deficient mice. ARTHRITIS RHEUMATOL. 2018;70(11):1807–19.

Steinhagen F, Zillinger T, Peukert K, Fox M, Thudium M, Barchet W, Putensen C, Klinman D, Latz E, Bode C. Suppressive oligodeoxynucleotides containing TTAGGG motifs inhibit cGAS activation in human monocytes. EUR J IMMUNOL. 2018;48(4):605–11.

Wang M, Sooreshjani MA, Mikek C, Opoku-Temeng C, Sintim HO. Suramin potently inhibits cGAMP synthase, cGAS, in THP1 cells to modulate IFN-beta levels. FUTURE MED CHEM. 2018;10(11):1301–17.

Vincent J, Adura C, Gao P, Luz A, Lama L, Asano Y, Okamoto R, Imaeda T, Aida J, Rothamel K, et al. Small molecule inhibition of cGAS reduces interferon expression in primary macrophages from autoimmune mice. NAT COMMUN. 2017;8(1):750.

Zhao W, Xiong M, Yuan X, Li M, Sun H, Xu Y. In Silico Screening-based Discovery of Novel inhibitors of human cyclic GMP-AMP synthase: a Cross-validation Study of Molecular Docking and experimental testing. J CHEM INF MODEL. 2020;60(6):3265–76.

Siu T, Altman MD, Baltus GA, Childers M, Ellis JM, Gunaydin H, Hatch H, Ho T, Jewell J, Lacey BM, et al. Discovery of a novel cGAMP competitive ligand of the inactive form of STING. ACS MED CHEM LETT. 2019;10(1):92–7.

Li S, Hong Z, Wang Z, Li F, Mei J, Huang L, Lou X, Zhao S, Song L, Chen W, et al. The Cyclopeptide Astin C specifically inhibits the Innate Immune CDN Sensor STING. CELL REP. 2018;25(12):3405–21.

Haag SM, Gulen MF, Reymond L, Gibelin A, Abrami L, Decout A, Heymann M, van der Goot FG, Turcatti G, Behrendt R, et al. Targeting STING with covalent small-molecule inhibitors. Nature. 2018;559(7713):269–73.

Padilla-Salinas R, Sun L, Anderson R, Yang X, Zhang S, Chen ZJ, Yin H. Discovery of Small-Molecule Cyclic GMP-AMP synthase inhibitors. J ORG CHEM. 2020;85(3):1579–600.

Huffman BJ, Chen S, Schwarz JL, Plata RE, Chin EN, Lairson LL, Houk KN, Shenvi RA. Electronic complementarity permits hindered butenolide heterodimerization and discovery of novel cGAS/STING pathway antagonists. NAT CHEM. 2020;12(3):310–7.

Hall J, Brault A, Vincent F, Weng S, Wang H, Dumlao D, Aulabaugh A, Aivazian D, Castro D, Chen M, et al. Discovery of PF-06928215 as a high affinity inhibitor of cGAS enabled by a novel fluorescence polarization assay. PLoS ONE. 2017;12(9):e184843.

Vinogradova EV, Zhang X, Remillard D, Lazar DC, Suciu RM, Wang Y, Bianco G, Yamashita Y, Crowley VM, Schafroth MA et al. An Activity-Guided Map of Electrophile-Cysteine Interactions in Primary Human T Cells. CELL 2020, 182(4):1009–1026.

Gomez AM, Muino MA, Bendala EA, Maestro DLCG, Monge MD, Boixeda R, Ena J, Mella PC, Anton SJ, Lumbreras BC. Clinical characteristics and prognosis of COPD patients hospitalized with SARS-CoV-2. INT J CHRONIC OBSTR. 2020;15:3433–45.

Google Scholar  

Zhou Y, Hou Y, Shen J, Huang Y, Martin W, Cheng F. Network-based drug repurposing for novel coronavirus 2019-nCoV/SARS-CoV-2. CELL DISCOV. 2020;6:14.

Huttner A, Kowalczyk A, Turjeman A, Babich T, Brossier C, Eliakim-Raz N, Kosiek K, Martinez DTB, Roux X, Shiber S, et al. Effect of 5-Day nitrofurantoin vs single-dose fosfomycin on clinical resolution of uncomplicated lower urinary tract infection in women: a Randomized Clinical Trial. JAMA-J AM MED ASSOC. 2018;319(17):1781–9.

Lopez-Otin C, Blasco MA, Partridge L, Serrano M, Kroemer G. Hallmarks of aging: an expanding universe. Cell. 2023;186(2):243–78.

Tsuji T, Aoshiba K, Nagai A. Alveolar cell senescence in patients with pulmonary emphysema. AM J RESP CRIT CARE. 2006;174(8):886–93.

Rutten EP, Gopal P, Wouters EF, Franssen FM, Hageman GJ, Vanfleteren LE, Spruit MA, Reynaert NL. Various mechanistic pathways representing the aging process are altered in COPD. Chest. 2016;149(1):53–61.

Rivas M, Gupta G, Costanzo L, Ahmed H, Wyman AE, Geraghty P. Senescence: pathogenic driver in Chronic Obstructive Pulmonary Disease. MEDICINA-LITHUANIA 2022, 58(6).

Tsuji T, Aoshiba K, Nagai A. Cigarette smoke induces senescence in alveolar epithelial cells. AM J RESP CELL MOL. 2004;31(6):643–9.

Fujii S, Hara H, Araya J, Takasaka N, Kojima J, Ito S, Minagawa S, Yumino Y, Ishikawa T, Numata T, et al. Insufficient autophagy promotes bronchial epithelial cell senescence in chronic obstructive pulmonary disease. ONCOIMMUNOLOGY. 2012;1(5):630–41.

Lan YY, Heather JM, Eisenhaure T, Garris CS, Lieb D, Raychowdhury R, Hacohen N. Extranuclear DNA accumulates in aged cells and contributes to senescence and inflammation. Aging Cell. 2019;18(2):e12901.

Liu W, Zhang X, Mao B, Jiang H. Systems pharmacology-based study of tanreqing injection in airway mucus hypersecretion. J ETHNOPHARMACOL. 2020;249:112425.

Deng J, He Y, Sun G, Yang H, Wang L, Tao X, Chen W. Tanreqing injection protects against bleomycin-induced pulmonary fibrosis via inhibiting STING-mediated endoplasmic reticulum stress signaling pathway. J ETHNOPHARMACOL. 2023;305:116071.

Choi KT. Botanical characteristics, pharmacological effects and medicinal components of Korean Panax ginseng C a Meyer. ACTA PHARMACOL SIN. 2008;29(9):1109–18.

Cho JY, Yoo ES, Baik KU, Park MH, Han BH. In vitro inhibitory effect of protopanaxadiol ginsenosides on tumor necrosis factor (TNF)-alpha production and its modulation by known TNF-alpha antagonists. PLANTA MED. 2001;67(3):213–8.

Guan X, Yuan Y, Wang G, Zheng R, Zhang J, Dong B, Ran N, Hsu AC, Wang C, Wang F. Ginsenoside Rg3 ameliorates acute exacerbation of COPD by suppressing neutrophil migration. INT IMMUNOPHARMACOL. 2020;83:106449.

Li H, Cui L, Liu Q, Dou S, Wang W, Xie M, Xu X, Zheng C, Li T, Huang S, et al. Ginsenoside Rb3 alleviates CSE-induced TROP2 upregulation through p38 MAPK and NF-kappaB pathways in basal cells. AM J RESP CELL MOL. 2021;64(6):747–59.

Shergis JL, Di YM, Zhang AL, Vlahos R, Helliwell R, Ye JM, Xue CC. Therapeutic potential of Panax ginseng and ginsenosides in the treatment of chronic obstructive pulmonary disease. COMPLEMENT THER MED. 2014;22(5):944–53.

Pang W, Lin S, Dai Q, Zhang H, Hu J. Antitussive activity of Pseudostellaria heterophylla (miq.) Pax extracts and improvement in lung function via adjustment of multi-cytokine levels. Molecules. 2011;16(4):3360–70.

Sun SC, Han R, Hou SS, Yi HQ, Chi SJ, Zhang AH. Juglanin alleviates bleomycin-induced lung injury by suppressing inflammation and fibrosis via targeting sting signaling. BIOMED PHARMACOTHER. 2020;127:110119.

Download references

Acknowledgements

We thank the library of Anhui Medical University for providing literature search service and Prof. Zhenxiu Liao of Anhui Jianzhu University for providing technical support in case of computer failure.

This work was supported by the National Natural Science Foundation of China (82104267).

Author information

Authors and affiliations.

First Clinical Medical College, Anhui Medical University, Hefei, 230022, People’s Republic of China

Kexin Liao & Wenqi Bi

Department of Respiratory and Critical Care Medicine, The First Affiliated Hospital of Anhui Medical University, Hefei, 230022, People’s Republic of China

Chenhao Xia, Ze Xu, Sen Zhong & Jingjing Ruan

College of Pharmacy, Anhui Medical University, Hefei, 230022, People’s Republic of China

Fengshuo Wang

You can also search for this author in PubMed   Google Scholar

Contributions

J.R. provided the main direction and important guidance for this manuscript. K.L., F.W. and C.X. conceived the paper. K.L., Z.X., S.Z. and W.B. wrote the original draft and illustrated the figure for the manuscript. All authors approved the final manuscript.

Corresponding author

Correspondence to Jingjing Ruan .

Ethics declarations

Ethics approval and consent to participate.

Not applicable.

Consent for publication

Competing interests.

The authors declare no competing interests.

Additional information

Publisher’s note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ . The Creative Commons Public Domain Dedication waiver ( http://creativecommons.org/publicdomain/zero/1.0/ ) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Cite this article.

Liao, K., Wang, F., Xia, C. et al. The cGAS-STING pathway in COPD: targeting its role and therapeutic potential. Respir Res 25 , 302 (2024). https://doi.org/10.1186/s12931-024-02915-x

Download citation

Received : 27 March 2024

Accepted : 12 July 2024

Published : 07 August 2024

DOI : https://doi.org/10.1186/s12931-024-02915-x

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• cGAS-STING pathway
• Therapeutic potential

Respiratory Research

ISSN: 1465-993X

• General enquiries: [email protected]