literature review bmj

Log in using your username and password

Search More Search for this keyword Advanced search
Latest content
Current issue
BMJ Journals More You are viewing from: Google Indexer

http://orcid.org/0000-0001-6644-9845 Irma Klerings 1 ,
Shannon Robalino 2 ,
http://orcid.org/0000-0003-4808-3880 Andrew Booth 3 ,
http://orcid.org/0000-0002-2903-6870 Camila Micaela Escobar-Liquitay 4 ,
Isolde Sommer 1 ,
http://orcid.org/0000-0001-5531-3678 Gerald Gartlehner 1 , 5 ,
Declan Devane 6 , 7 ,
Siw Waffenschmidt 8
On behalf of the Cochrane Rapid Reviews Methods Group
1 Department for Evidence-Based Medicine and Evaluation , University of Krems (Danube University Krems) , Krems , Niederösterreich , Austria
2 Center for Evidence-based Policy , Oregon Health & Science University , Portland , Oregon , USA
3 School of Health and Related Research (ScHARR) , The University of Sheffield , Sheffield , UK
4 Research Department, Associate Cochrane Centre , Instituto Universitario Escuela de Medicina del Hospital Italiano de Buenos Aires , Buenos Aires , Argentina
5 RTI-UNC Evidence-based Practice Center , RTI International , Research Triangle Park , North Carolina , USA
6 School of Nursing & Midwifery, HRB TMRN , National University of Ireland Galway , Galway , Ireland
7 Evidence Synthesis Ireland & Cochrane Ireland , University of Galway , Galway , Ireland
8 Information Management Department , Institute for Quality and Efficiency in Healthcare , Cologne , Germany
Correspondence to Irma Klerings, Department for Evidence-based Medicine and Evaluation, Danube University Krems, Krems, Niederösterreich, Austria; irma.klerings{at}donau-uni.ac.at

This paper is part of a series of methodological guidance from the Cochrane Rapid Reviews Methods Group. Rapid reviews (RR) use modified systematic review methods to accelerate the review process while maintaining systematic, transparent and reproducible methods. In this paper, we address considerations for RR searches. We cover the main areas relevant to the search process: preparation and planning, information sources and search methods, search strategy development, quality assurance, reporting, and record management. Two options exist for abbreviating the search process: (1) reducing time spent on conducting searches and (2) reducing the size of the search result. Because screening search results is usually more resource-intensive than conducting the search, we suggest investing time upfront in planning and optimising the search to save time by reducing the literature screening workload. To achieve this goal, RR teams should work with an information specialist. They should select a small number of relevant information sources (eg, databases) and use search methods that are highly likely to identify relevant literature for their topic. Database search strategies should aim to optimise both precision and sensitivity, and quality assurance measures (peer review and validation of search strategies) should be applied to minimise errors.

Evidence-Based Practice
Systematic Reviews as Topic
Information Science

Data availability statement

No data are available.

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/ .

https://doi.org/10.1136/bmjebm-2022-112079

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.

WHAT IS ALREADY KNOWN ON THIS TOPIC

Compared with systematic reviews, rapid reviews (RR) often abbreviate or limit the literature search in some way to accelerate review production. However, RR guidance rarely specifies how to select topic-appropriate search approaches.

WHAT THIS STUDY ADDS

This paper presents an overview of considerations and recommendations for RR searching, covering the complete search process from the planning stage to record management. We also provide extensive appendices with practical examples, useful sources and a glossary of terms.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

There is no one-size-fits-all solution for RR literature searching: review teams should consider what search approaches best fit their RR project.

Introduction

This paper is part of a series from the Cochrane Rapid Reviews Methods Group (RRMG) providing methodological guidance for rapid reviews (RRs). 1–3 While the RRMG’s guidance 4 5 on Cochrane RR production includes brief advice on literature searching, we aim to provide in-depth recommendations for the entire search process.

Literature searching is the foundation for all reviews; therefore, it is important to understand the goals of a specific RR. The scope of RRs varies considerably (from focused questions to overviews of broad topics). 6 As with conventional systematic reviews (SRs), there is not a one-size-fits-all approach for RR literature searches. We aim to support RR teams in choosing methods that best fit their project while understanding the limitations of modified search methods. Our recommendations derive from current systematic search guidance, evidence on modified search methods and practical experience conducting RRs.

This paper presents considerations and recommendations, described briefly in table 1 . The table also includes a comparison to the SR search process based on common recommendations. 7–10 We provide supplemental materials, including a list of additional resources, further details of our recommendations, practical examples, and a glossary (explaining the terms written in italics) in online supplemental appendices A–C .

Supplemental material

View inline

Recommendations for rapid review literature searching

Preparation and planning

Given that the results of systematic literature searches underpin a review, planning the searches is integral to the overall RR preparation. The RR search process follows the same steps as an SR search; therefore, RR teams must be familiar with the general standards of systematic searching . Templates (see online supplemental appendix B ) and reporting guidance 11 for SR searches can also be adapted to structure the RR search process.

Developing a plan for the literature search forms part of protocol development and should involve an information specialist (eg, librarian). Information specialists can assist in refining the research question, selecting appropriate search methods and resources, designing and executing search strategies, and reporting the search methods. At minimum, specialist input should include assessing information sources and methods and providing feedback on the primary database search strategy.

Two options exist for abbreviating the search process: (1) reducing time spent on conducting searches (eg, using automation tools, reusing existing search strategies, omitting planning or quality assurance steps) and (2) reducing the size of the search result (eg, limiting the number of information sources, increasing the precision of search strategies, using study design filters). Study selection (ie, screening search results) is usually more resource-intensive than searching, 12 particularly for topics with complex or broad concepts or diffuse terminology; thus, the second option may be more efficient for the entire RR. Investing time upfront in optimising search sensitivity (ie, completeness) and precision (ie, positive predictive value) can save time in the long run by reducing the screening and selection workload.

Preliminary or scoping searches are critical to this process. They inform the choice of search methods and identify potentially relevant literature. Texts identified through preliminary searching serve as known relevant records that can be used throughout the search development process (see sections on database selection, development and validation of search strategies).

In addition to planning the search itself, the review team should factor in time for quality assurance steps (eg, search strategy peer review) and the management of search results (eg, deduplication, full-text retrieval).

Information sources and methods

To optimise the balance of search sensitivity and precision, RR teams should prioritise the most relevant information sources for the topic and the type of evidence required. These can include bibliographic databases (eg, MEDLINE/PubMed), grey literature sources and targeted supplementary search methods. Note that this approach differs from the Methodological Expectations of Cochrane Intervention Reviews Standards 9 where the same core set of information sources is required for every review and further supplemented by additional topic-specific and evidence-specific sources.

Choosing bibliographic databases

For many review topics, most evidence is found in peer-reviewed journal articles, making bibliographic databases the main resource of systematic searching. Limiting the number of databases searched can be a viable option in RRs, but it is important to prioritise topic-appropriate databases.

MEDLINE has been found to have high coverage for studies included in SRs 13 14 and is an appealing database choice because access is free via PubMed. However, coverage varies depending on topics and relevant study designs. 15 16 Additionally, even if all eligible studies for a topic were available in MEDLINE, search strategies will usually miss some eligible studies because search sensitivity is lower than database coverage. 13 17 This means searching MEDLINE alone is not a viable option, and additional information sources or search methods are required. Known relevant records can be used to help assess the coverage of selected databases (see also online supplemental appendix C ).

Further information sources and search techniques

Supplementary systematic search methods have three main goals, to identify (1) grey literature, (2) published literature not covered by the selected bibliographic databases and (3) database-covered literature that was not retrieved by the database searches.

When RRs search only a small number of databases, supplementary searches can be particularly important to pick up eligible studies not identified via database searching. While supplementary methods might increase the time spent on searching, they sometimes better optimise search sensitivity and precision, saving time in the long run. 18 Depending on the topic and relevant evidence, such methods can offer an alternative to adding additional specialised database searches. To decide if and what supplementary searches are helpful, it is important to evaluate what literature might be missed by the database searches and how this might affect the specific RR.

Study registries and other grey literature

Some studies indicate that the omission of grey literature searches rarely affects review conclusions. 17 19 However, the relevance of study registries and other grey literature sources is topic-dependent. 16 19–21 For example, randomised controlled trials (RCTs) on newly approved drugs are typically identified in ClinicalTrials.gov. 20 For rapidly evolving topics such as COVID-19, preprints are an important source. 21 For public health interventions, various types of grey literature may be important (eg, evaluations conducted by local public health agencies). 22

Further supplementary search methods

Other supplementary techniques (eg, checking reference lists, reviewing specific websites or electronic table of contents, contacting experts) may identify additional studies not retrieved by database searches. 23 One of the most common approaches involves checking reference lists of included studies and relevant reviews. This method may identify studies missed by limited database searches. 12 Another promising citation-based approach is using the ‘similar articles’ option in PubMed, although research has focused on updating existing SRs. 24 25

Considerations for RRs of RCTs

Databases and search methods to identify RCTs have been particularly well researched. 17 20 24 26 27 For this reason, it is possible to give more precise recommendations for RRs based on RCTs than for other types of review. Table 2 provides an overview of the most important considerations; additional information can be found in online supplemental appendix C .

Information sources for identification of randomised controlled trials (RCTs)

Search strategies

We define ‘search strategy’ as a Boolean search query in a specific database (eg, MEDLINE) using a specific interface (eg, Ovid). When several databases are searched, this query is usually developed in a primary database and interface (eg, Ovid MEDLINE) and translated to other databases.

Developing search strategies

Optimising search strategy precision while aiming for high sensitivity is critical in reducing the number of records retrieved. Preliminary searches provide crucial information to aid efficient search strategy development. Reviewing the abstracts and subject headings used in known relevant records will assist in identifying appropriate search terms. Text analysis tools can also be used to support this process, 28 29 for example, to develop ‘objectively derived’ search strategies. 30

Reusing or adapting complete search strategies (eg, from SRs identified by the preliminary searches) or selecting elements of search strategies for reuse can accelerate search strategy development. Additionally, validated search filters (eg, for study design) can be used to reduce the size of the search result without compromising the sensitivity of a search strategy. 31 However, quality assurance measures are necessary whether the search strategy is purpose-built, reused or adapted (see the ‘Quality assurance’ section.)

Database-specific and interface-specific functionalities can also be used to improve searches’ precision and reduce the search result’s size. Some options are: restricting to records where subject terms have been assigned as the major focus of an article (eg, major descriptors in MeSH), using proximity operators (ie, terms adjacent or within a set number of words), frequency operators (ie, terms have to appear a minimum number of times in an abstract) or restricting search terms to the article title. 32–34

Automated syntax translation can save time and reduce errors when translating a primary search strategy to different databases. 35 36 However, manual adjustments will usually be necessary.

The time taken to learn how to use supporting technologies (eg, text analysis, syntax translation) proficiently should not be underestimated. The time investment is most likely to pay off for frequent searchers. A later paper in this series will address supporting software for the entire review process.

Limits and restrictions

Limits and restrictions (eg, publication dates, language) are another way to reduce the number of records retrieved but should be tailored to the topic and applied with caution. For example, if most studies about an intervention were published 10 years ago, then an arbitrary cut-off of ‘the last 5 years’ will miss many relevant studies. 37 Similarly, limiting to ‘English only’ is acceptable for most cases, but early in the COVID-19 pandemic, a quarter of available research articles were written in Chinese. 38 Depending on the RR topic, certain document types (eg, conference abstracts, dissertations) might be excluded if not considered relevant to the research question.

Note also that preset limiting functions in search interfaces (eg, limit to humans) often rely on subject headings (eg, MeSH) alone. They will miss eligible studies that lack or have incomplete subject indexing. Using (validated) search filters 31 is preferable.

Updating existing reviews

One approach to RR production involves updating an existing SR. In this case, preliminary searches should be used to check if new evidence is available. If the review team decide to update the review, they should assess the original search methods and adapt these as necessary.

One option is to identify the minimum set of databases required to retrieve all the original included studies. 39 Any reused search strategies should be validated and peer-reviewed (see below) and optimised for precision and/or sensitivity.

Additionally, it is important to assess whether the topic terminology or the relevant databases have changed since the original SR search.

In some cases, designing a new search process may be more efficient than reproducing the original search.

Quality assurance and search strategy peer review

Errors in search strategies are common and can impact the sensitivity and comprehensiveness of the search result. 40 If an RR search uses a small number of information sources, such errors could affect the identification of relevant studies.

Validation of search strategies

The primary database search strategy should be validated using known relevant records (if available). This means testing if the primary search strategy retrieves eligible studies found through preliminary searching. If some known studies are not identified, the searcher assesses the reasons and decides if revisions are necessary. Even a precision-focused systematic search should identify the majority—we suggest at least 80%–90%—of known studies. This is based on benchmarks for sensitivity-precision-maximising search filters 41 and assumes that the set of known studies is representative of the whole of relevant studies.

Peer review of search strategies

Ideally, an information specialist should review the planned information sources and search methods and use the PRESS (Peer Review of Electronic Search Strategies) checklist 42 to assess the primary search strategy. Turnaround time has to be factored into the process from the outset (eg, waiting for feedback, revising the search strategy). PRESS recommends a maximum turnaround time of five working days for feedback, but in-house peer review often takes only a few hours.

If the overall RR time plan does not allow for a full peer review of the search strategy, a review team member with search experience should check the search strategy for spelling errors and correct use of Boolean operators (AND, OR, NOT) at a minimum.

Reporting and record management

Record management requirements of RRs are largely identical to SRs and have to be factored into the time plan. Teams should develop a data management plan and review the relevant reporting standards at the project’s outset. PRISMA-S (Preferred Reporting Items for Systematic Reviews and Meta-Analyses literature search extension) 11 is a reporting standard for SR searches that can be adapted for RRs.

Reference management software (eg, EndNote, 43 Zotero 44 ) should be used to track search results, including deduplication. Note that record management for database searches is less time-consuming than for many supplementary or grey literature searches, which often require manual entry into reference management software. 12

Additionally, software platforms for SR production (eg, Covidence, 45 EPPI-Reviewer, 46 Systematic Review Data Repository Plus 47 ) can provide a unified way to keep track of records throughout the whole review process, which can improve management and save time. These platforms and other dedicated tools (eg, SRA Deduplicator) 48 also offer automated deduplication. However, the time and cost investment necessary to appropriately use these tools have to be considered.

Decisions about search methods for an RR need to consider where time can be most usefully invested and processes accelerated. The literature search should be considered in the context of the entire review process, for example, protocol development and literature screening: Findings of preliminary searches often affect the development and refinement of the research question and the review’s eligibility criteria . In turn, they affect the number of records retrieved by the searches and therefore the time needed for literature selection.

For this reason, focusing only on reducing time spent on designing and conducting searches can be a false economy when seeking to speed up review production. While some approaches (eg, text analysis, automated syntax translation) may save time without negatively affecting search validity, others (eg, skipping quality assurance steps, using convenient information sources without considering their topic appropriateness) may harm the entire review. Information specialists can provide crucial aid concerning the appropriate design of search strategies, choice of methods and information sources.

For this reason, we consider that investing time at the outset of the review to carefully choose a small number of highly appropriate search methods and optimise search sensitivity and precision likely leads to better and more manageable results.

Ethics statements

Patient consent for publication.

Not applicable.

Gartlehner G ,
Nussbaumer-Streit B ,
Nussbaumer Streit B ,
Garritty C ,
Tricco AC ,
Nussbaumer-Streit B , et al
Trivella M ,
Hamel C , et al
Hartling L ,
Guise J-M ,
Kato E , et al
Lefebvre C ,
Glanville J ,
Briscoe S , et al
Higgins JPT ,
Lasserson T ,
Chandler J , et al
European network for Health Technology Assessment (EUnetHTA)
Rethlefsen ML ,
Kirtley S ,
Waffenschmidt S , et al
Klerings I , et al
Bramer WM ,
Giustini D ,
Halladay CW ,
Trikalinos TA ,
Schmid IT , et al
Frandsen TF ,
Eriksen MB ,
Hammer DMG , et al
Klerings I ,
Wagner G , et al
Husk K , et al
Featherstone R ,
Nuspl M , et al
Knelangen M ,
Hausner E ,
Metzendorf M-I , et al
Gianola S ,
Bargeri S , et al
Hillier-Brown FC ,
Moore HJ , et al
Varley-Campbell J , et al
Sampson M ,
de Bruijn B ,
Urquhart C , et al
Fitzpatrick-Lewis D , et al
Affengruber L ,
Waffenschmidt S ,
Kaiser T , et al
The InterTASC Information Specialists’ Sub-Group
Kleijnen J , et al
Jacob C , et al
Kaunelis D ,
Mensinkai S , et al
Mast F , et al
Sanders S ,
Carter M , et al
Marshall IJ ,
Marshall R ,
Wallace BC , et al
Fidahic M ,
Runjic R , et al
Hopewell S ,
Salvador-Oliván JA ,
Marco-Cuenca G ,
Arquero-Avilés R
Navarro-Ruan T ,
Hobson N , et al
McGowan J ,
Salzwedel DM , et al
Clarivate Analytics
Corporation for Digital Scholarship
Veritas Health Innovation Ltd
Graziosi S ,
Brunton J , et al
Agency for Healthcare Research and Quality
Institute for Evidence-Based Healthcare

Supplementary materials

Supplementary data.

This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

Data supplement 1

Twitter @micaelaescb

Collaborators On behalf of the Cochrane Rapid Reviews Methods Group: Declan Devane, Gerald Gartlehner, Isolde Sommer.

Contributors IK, SR, AB, CME-L and SW contributed to the conceptualisation of this paper. IK, AB and CME-L wrote the first draft of the manuscript. All authors critically reviewed and revised the manuscript. IK is responsible for the overall content.

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 AB is co-convenor of the Cochrane Qualitative and Implementation Methods Group. In the last 36 months, he received royalties from Systematic Approaches To a Successful Literature Review (Sage 3rd edn), payment or honoraria form the Agency for Healthcare Research and Quality, and travel support from the WHO. DD works part time for Cochrane Ireland and Evidence Synthesis Ireland, which are funded within the University of Ireland Galway (Ireland) by the Health Research Board (HRB) and the Health and Social Care, Research and Development (HSC R&D) Division of the Public Health Agency in Northern Ireland.

Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

Provenance and peer review Not commissioned; externally peer reviewed.

Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.

Linked Articles

Research methods and reporting Rapid reviews methods series: Guidance on team considerations, study selection, data extraction and risk of bias assessment Barbara Nussbaumer-Streit Isolde Sommer Candyce Hamel Declan Devane Anna Noel-Storr Livia Puljak Marialena Trivella Gerald Gartlehner BMJ Evidence-Based Medicine 2023; 28 418-423 Published Online First: 19 Apr 2023. doi: 10.1136/bmjebm-2022-112185
Research methods and reporting Rapid reviews methods series: Guidance on assessing the certainty of evidence Gerald Gartlehner Barbara Nussbaumer-Streit Declan Devane Leila Kahwati Meera Viswanathan Valerie J King Amir Qaseem Elie Akl Holger J Schuenemann BMJ Evidence-Based Medicine 2023; 29 50-54 Published Online First: 19 Apr 2023. doi: 10.1136/bmjebm-2022-112111
Research methods and reporting Rapid Reviews Methods Series: Involving patient and public partners, healthcare providers and policymakers as knowledge users Chantelle Garritty Andrea C Tricco Maureen Smith Danielle Pollock Chris Kamel Valerie J King BMJ Evidence-Based Medicine 2023; 29 55-61 Published Online First: 19 Apr 2023. doi: 10.1136/bmjebm-2022-112070

Read the full text or download the PDF:

Appraising systematic reviews

Systematic reviews may or may not include a meta-analysis of the primary RCTs identified. Although systematic reviews of RCTs with meta-analysis are often said to provide the most compelling evidence of effectiveness and causality, not all systematic reviews are of the highest methodological quality.

There are numerous checklists available. The SR toolbox is an online catalogue providing summaries and links to the available guidance and software for each stage of the systematic review process including critical appraisal. Examples for systematic reviews include:

AMSTAR 2 (A MeaSurement Tool to Assess systematic Reviews)
CASP systematic review checklist.

The following framework forms an initial memory aid for assessing the quality of a systematic review.

Framework for assessing systematic reviews

Content created by BMJ Knowledge Centre

What is EBM?
How to clarify a clinical question
Searching for evidence
Search strategies
Case study of a search
Appraise the evidence
Evidence synthesis
What is GRADE?
Understanding statistics: BMJ Learning modules
Understanding statistics: risk

EBM Toolkit home

Learn, Practise, Discuss, Tools

- Google Chrome

Intended for healthcare professionals

Access provided by Google Indexer
My email alerts
BMA member login
Username * Password * Forgot your log in details? Need to activate BMA Member Log In Log in via OpenAthens Log in via your institution

Search form

Advanced search
Search responses
Search blogs

About The BMJ

Resources for authors
Article submission

At The BMJ , we use a system called ScholarOne to manage our submission processes . Essentially, ScholarOne will convert your manuscript to a PDF for the review process. Most common word processing formats are accepted for text and tables, although the system prefers Microsoft Word, and images should be submitted as GIF, TIFF, EPS, or JPEG files.

The system can also accept supplementary files (for example: videos, datasets, research protocols, and checklists or statements), related articles published or available elsewhere, articles in press elsewhere, permission letters, etc. These are files that normally do not appear with the print article, although they might accompany the final version of the paper online. Supplementary files are not converted to PDF but will be available to reviewers or editors exactly as you upload them.

Coronavirus notice: If your article contains research findings relevant to the Coronavirus outbreak, please be aware that BMJ has agreed to share those findings (which will include your contact details) with WHO for the purposes set out here ( Sharing research data and findings relevant to the novel coronavirus (COVID-19) outbreak ). This applies to all new and revised submissions from 20th April 2020. If you have any concerns about this please contact us at [email protected] . A copy of BMJ’s privacy policy can be found here .

Registration

To access the system for the first time you will need to register here .

Please follow the link "Register here," which is in the right hand grey column. You will be asked to complete three steps:

Name and email information: Enter your email address. This is your unique user identification, which you will need to log into the system on all occasions.

Address information: You’ll need to fill in the name and address of the institution to which you are affiliated; you may provide secondary details here also.

User ID and password: You may change your user ID to be something more meaningful than your email address (you will, however, need this whenever you sign in). This is also where you may set your password.

Submitting a manuscript

Before you submit your manuscript please check that it is in the correct style for the appropriate section of The BMJ . Does it need a structured abstract, for example, and is the main text about the right length? Please check our advice section

Before you can use this submission website you have to register, creating your own account. When you have created an account please go to your author area and select "Click here to submit a new manuscript."

Alternative Journals

Should your article be rejected from The BMJ you have the option to have it automatically submitted to an alternative BMJ journal. You will be able to select a preferred alternate journal during the submission process, but this does not guarantee acceptance in the alternative journal.

Post-peer review pilot

The BMJ are also piloting a premier service for all authors whose articles have been through peer review. If your article is not suitable for The BMJ, an experienced Transfer Editor will discuss the suitability of your manuscript with relevant Editors from other high-quality journals in the BMJ portfolio.

The Transfer Editor will aim to find an appropriate outlet for your work that takes into account the work and reviews already received by The BMJ, potentially fast-tracking a decision for you. You will receive this feedback with your decision letter from The BMJ.

This process does not guarantee acceptance or a decision from another journal. While part of this pilot is to determine how much extra time this service will add, our aim is that it will not add more than 48 hours to your decision from The BMJ.

Both of the above services are optional for our authors.

Metadata page

On this page you can enter important information ("metadata") about the manuscript - for example, the number of authors, title, abstract, summary points, covering letter, any companion articles and other files such as questionnaires, whether this is a resubmission or a fast track article, etc.

When you are happy with the metadata page click "Save and continue" at the bottom of the page. This will take you to the next screen.

Author details

Please fill out the appropriate fields, making sure that the corresponding author box is ticked with the correct email address. You can insert and delete authors on this page. If a field has purple "req" text next to it this means that you have to fill it in: it is a required field.

When you are happy with the author details please click "Save and continue."

Please also provide key information about all of the authors in the manuscript's title page. The title page should list all authors' names, positions, institutions and full postal addresses including postal codes, as well as the corresponding author's email address.

You can also select a contributor role for each author as defined by CRediT (Contributor Roles Taxonomy). The purpose of CASRAI's CRediT taxonomy is to provide transparency in contributions to scholarly published work, to enable improved systems of attribution, credit, and accountability.

File upload

You are only able to upload one "Manuscript file," which should contain your main document but not any images or supplementary files.If you are submitting an article with no figures and no supplementary files then all you have to do is insert the word count and click "Upload files." You may then click "Save and continue" at the bottom of the screen as normal.

We would like you to upload any figures and images separately from the word file, preferably as PDF or JPEG files. Images may also be submitted as TIFF, GIF, EPS, MPEG, AVI, MOV, and WAV files. All images should be submitted on ScholarOne as separate files and not embedded into the text. Ensure that you use the colour/mono image designations when uploading these. If you need to upload tables please submit them as separate Word files. Please do not use Excel.

You can also upload supplementary files – for example, copies of questionnaires used in original research; study protocols, and checklists for the CONSORT, QUOROM, and STARD statements; video clips. There are appropriate designations for these files "Research checklists" and "Video files".

File formats

Papers can be submitted using any standard word processing software, although we prefer Word. Please ensure that you include the heading "References" at the top of the citations (see below).

Acceptable standard graphic formats include JPEG, TIFF, GIF, EPS, MPEG, AVI, MOV, and WAV.

Please note: we cannot accept PowerPoint or Excel files for tables or figures intended to accompany the main manuscript; however, they may be made available as supplementary files.

Saving your work

We are sorry that the submission process takes time. If you cannot wait, you can interrupt your submission and continue the process on another occasion. Clicking on "Save and continue" at the bottom of the screen when entering your metadata enables you to save that page or continue to the next page. If you choose not to complete the submission on that occasion, the submission will appear in your author area at the bottom – you may follow the link "Continue submission" when you return to it. If you decide not to complete the submission at all please choose "Delete."

After uploading

Everything will go off to be converted by the website host. Your figures and tables (if you have any) will be merged with the text as a single PDF, so everything will be in one file: text, tables, and figures. This makes it much easier for editors and peer reviewers to read and appraise your manuscript.

You will need to "proof" your article in its PDF form before you submit – on the last page "Review and submit" you’re able to "view PDF" and "View HTML" versions of your review file. If you upload an EndNote file the system will convert your references for you and provide Web of Science links.

Please do not worry if your reference hyperlinks appear incorrectly - it may be that the journals are not indexed or the system does not recognise the way you have formatted them. Occasionally the hyperlinking just fails for no apparent reason. Please don’t be concerned about this; it won’t stop us considering your paper.

Approving your submission

When you have viewed your PDF proofs you may return to the "File upload" step (or any other stage) to amend information you have provided. If you are satisfied with everything then click the small approval box at the end of the last stage entitled "Submit".

Keeping track of your manuscript

You can keep track of your manuscript throughout the process by going to your author area - here it will list for you manuscripts you have submitted under "Submitted manuscripts," "Manuscripts with decisions" and "Manuscripts I have co-authored". If your manuscript is sent for external peer review you will find reviewers comments in the decision letter. Please click on "Manuscripts with decision" and click on "View decision letter".

Submitting a revision/appeal

If we invite you to revise your manuscript please enter your author area and select "Manuscripts with decision." You may then select the appropriate manuscript and select "Create a revision." You will then be presented with your original manuscript metadata (the first page you filled in). Please edit this and the author page if necessary. Then please enter the number of images and supplementary files and please upload your revised manuscript.

Do not forget to enter your covering letter and your response to reviewers in the correct field on the metadata page. It is very important that you provide a new covering letter at this stage, explaining in detail how you have responded to all the peer review and editorial comments and the revision checklist.

If you believe that we have rejected your article wrongly, perhaps because we have misunderstood its scientific content, please submit an appeal (rebuttal) letter from your author area at our online editorial office. Do not try to submit a revised version of your article at this stage. Information on appeals can be found in our publishing model guide .

Providing authors with an easy and smooth submission and peer-review service is one of the most important aspects across all our journals.

We want to be able to find a home for all high-quality and sound science, clinical research in the most suited BMJ journal. We know that receiving a rejection can be disheartening. To prevent this, and the duplication of efforts of authors, editors and reviewers, BMJ Journals have improved the process of transferring manuscripts.

Read more about the BMJ transfer service .

Commissioning

If we commission you to write for The BMJ we will email you – please use the links in the email provided to respond to our request. The paper will then appear in your author area – and you may "click here to submit a commissioned manuscript" via your author area. The procedure will then be the same as for an unsolicited article. We appreciate your help with this: submitting the commissioned article in this way will help us to provide an efficient service to you.

Detailed help files are available throughout our online editorial office and can be used without stopping the submission process. If you experience serious problems please email The BMJ 's editorial office .

Publishing model
Editorial staff
Advisory panels
Explore The BMJ
BMJ Student
How green is The BMJ?
Sources of revenue
Article types and preparation
Forms, policies, and ethics
Guidance for new authors
BMJ Christmas issue
Resources for advertisers and sponsors
Resources for BMA members
Resources for media
Resources for subscribers
Resources for readers
Resources for reviewers
About The BMJ app
Poll archive
International jobs

This week's poll

Read related article

See previous polls

Log in using your username and password

Search More Search for this keyword Advanced search
Latest content
Current issue
Topic Collection
BMJ Journals More You are viewing from: Google Indexer

http://orcid.org/0000-0001-5809-7173 Oscar Lyons 1 ,
Robynne George 2 ,
Joao R Galante 3 , 4 ,
Alexander Mafi 5 ,
Thomas Fordwoh 5 ,
http://orcid.org/0000-0001-9079-7508 Jan Frich 6 ,
http://orcid.org/0000-0001-6672-3859 Jaason Matthew Geerts 7 , 8
1 Nuffield Department of Surgical Sciences , University of Oxford , Oxford , UK
2 Royal United Hospital Bath NHS Trust , Bath , UK
3 Department of Medical Education , Oxford University Hospitals NHS Foundation Trust , Oxford , UK
4 Cardiology Department , Buckinghamshire Healthcare NHS Trust , Amersham , UK
5 University of Oxford Medical School , University of Oxford , Oxford , UK
6 Department of Health Management and Health Economics , University of Oslo , Oslo , Norway
7 Research and Leadership Development , Canadian College of Health Leaders , Ottawa , Ontario , Canada
8 The Business School (formerly Cass) , University of London , London , UK
Correspondence to Dr Oscar Lyons, Nuffield Department of Surgical Sciences, University of Oxford, Oxford OX3 9DU, UK; oscar.lyons{at}nds.ox.ac.uk

Health systems invest significant resources in leadership development for physicians and other health professionals. Competent leadership is considered vital for maintaining and improving quality and patient safety. We carried out this systematic review to synthesise new empirical evidence regarding medical leadership development programme factors which are associated with outcomes at the clinical and organisational levels. Using Ovid MEDLINE, we conducted a database search using both free text and Medical Subject Headings. We then conducted an extensive hand-search of references and of citations in known healthcare leadership development reviews. We applied the Medical Education Research Study Quality Indicator (MERSQI) and the Joanna Briggs Institute (JBI) Critical Appraisal Tool to determine study reliability, and synthesised results using a meta-aggregation approach. 117 studies were included in this systematic review. 28 studies met criteria for higher reliability studies. The median critical appraisal score according to the MERSQI was 8.5/18 and the median critical appraisal score according to the JBI was 3/10. There were recurring causes of low study quality scores related to study design, data analysis and reporting. There was considerable heterogeneity in intervention design and evaluation design. Programmes with internal or mixed faculty were significantly more likely to report organisational outcomes than programmes with external faculty only (p=0.049). Project work and mentoring increased the likelihood of organisational outcomes. No leadership development content area was particularly associated with organisational outcomes. In leadership development programmes in healthcare, external faculty should be used to supplement in-house faculty and not be a replacement for in-house expertise. To facilitate organisational outcomes, interventions should include project work and mentoring. Educational methods appear to be more important for organisational outcomes than specific curriculum content. Improving evaluation design will allow educators and evaluators to more effectively understand factors which are reliably associated with organisational outcomes of leadership development.

medical leadership
development

Data availability statement

Data are available upon reasonable request.

https://doi.org/10.1136/leader-2020-000360

Statistics from Altmetric.com

Request permissions.

Introduction

Health systems invest significant resources in leadership development for physicians and other health professionals. 1 Competent leadership is considered vital for team effectiveness, for clinical and financial performance and for maintaining and improving quality and patient safety. 1–5 Clinical leadership development involves activities to promote leadership competencies among clinicians, while medical leadership development refers to activities centred on doctors.

Research suggests that medical leadership development can improve outcomes at individual, organisational and clinical levels. 6–11 Evidence backing medical leadership development activities has, however, been variable in quality. 1 7–10 12–15 There has been a particular lack of research and evaluation that goes beyond individual learner feedback and subjective outcomes. 6–9 One systematic review of 45 studies evaluating leadership development interventions for doctors found that effective interventions were characterised by the use of multiple learning methods, including seminars and group work, alongside action learning projects in multidisciplinary teams. 8 These findings were echoed in a recent study by Geerts et al , 9 who emphasised that plans need to be in place for transferring learning from the intervention into the working environment.

We undertook this systematic review to synthesise recent empirical evidence regarding medical leadership development programme factors associated with outcomes at the clinical and organisational levels. We specifically investigated links between aspects of programme design, delivery and evaluation and improved outcomes. Given the variable quality of studies highlighted in previous reviews, 7–9 we applied two validated critical appraisal instruments 16 17 to isolate higher reliability findings. This review is the first to apply both instruments in order to identify and synthesise the highest quality empirical evidence in medical leadership development.

The design of this review was guided by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses 18 and the Best Evidence in Medical Education (BEME) guide for systematic reviews. 19 Our methods were based on the review conducted by Frich et al , 8 with methodological changes drawn from other reviews. 7 9 10 14 15 20 Following the BEME recommendations for systematic reviews, 19 we hand-searched references and citations of known reviews extensively to supplement our database search. In line with recommendations from Geerts et al 9 and Rosenman et al , 7 we assessed study quality using the Medical Education Research Study Quality Indicator (MERSQI), which is designed to measure the methodological quality of quantitative medical education research studies. 16 We added the Joanna Briggs Institute (JBI) Critical Appraisal Checklist, 17 which is designed for meta-aggregation of qualitative research and is well-established in healthcare research. 21

Search strategy

We began this review by re-examining the data set identified in the review of leadership development for physicians by Frich et al . 8 With assistance from a specialist librarian at the University of Oxford, we then based our search strategy on Frich et al ’s review. 8 Using the Ovid MEDLINE database, we conducted a search using both free text and Medical Subject Headings. The full search terms are listed in the online supplemental material . This search identified 501 unique publications. We then conducted an extensive hand-search of references and of citations in known healthcare leadership development reviews using Web of Science and Google Scholar. This identified an additional 107 studies for possible inclusion, for a total of 608 records for screening ( figure 1 ).

Supplemental material

Download figure
Open in new tab
Download powerpoint

PRISMA diagram. PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses.

Inclusion criteria

We included any peer-reviewed study published in English between January 2000 and January 2020 which:

Describes a leadership development intervention (programme, workshop, course and so on).

Includes physicians as learners (defined here as any practising doctor post-qualification).

Evaluates the leadership development intervention.

Qualitative, quantitative and mixed evaluations were included. We excluded studies where leadership development was a minor focus or where the proportion of physicians was lower than 10% of intervention participants.

Screening process

Two members of the review team (OL and TF) independently screened all study titles and abstracts for eligibility. Articles that were approved by either reviewer progressed to full-text review. Two members of the review team independently reviewed for inclusion the full text of all 207 articles that passed the title and abstract screen (TF and RG reviewed half each, OL reviewed all). Where there was disagreement about inclusion, all three reviewers (OL, TF, RG) reached consensus by discussion, with the third reviewer (TF or RG) arbitrating where required.

Data abstraction

After screening and reviewing for eligibility, 117 unique studies were included for abstraction and analysis. Data were abstracted and coded for educational setting, methods, content, evaluation methods and outcomes. Outcome data were categorised according to an adapted version of Kirkpatrick’s Framework for evaluation of training programmes (see table 1 ). 19 22 One reviewer abstracted and coded all 117 included studies (OL). The second reviewers (RG/JRG/AM/TF) each abstracted and coded at least five studies in full to ensure consistency between reviewers. Data abstraction and coding for all 117 studies was then cross-checked by the second reviewers. Any differences were resolved by consensus, with a third reviewer arbitrating where required. Where possible, statistical tests performed in studies were replicated and checked for accuracy.

View inline

Kirkpatrick’s Framework for evaluation of training programmes, with adaptations from Frich et al 8

Study quality appraisal

Previous reviews have shown marked variation in the quality of studies of medical leadership development. 7 9 10 14 15 20 To isolate the most reliable evidence linking medical leadership programmes to improved outcomes, two researchers independently critically appraised each included study using the MERSQI and JBI Instruments. 16 21 Differences in MERSQI and JBI quality score were resolved by consensus, and a third researcher arbitrated where needed.

The MERSQI was applied to all 117 studies. The MERSQI is a validated appraisal tool consisting of 10 items in six domains which relate to design, sampling, type of data collected, validity of evaluation methods, analysis and outcomes. 16 Each domain is scored to a maximum of 3, for a total score of 5–18. In line with Geerts et al , 9 studies with scores of 12 or higher were categorised as higher reliability studies (see the Data analysis section).

The JBI Checklist for Qualitative Studies was also applied where a study used mixed methods (k=53) or qualitative methods (k=10). Fundamental differences in study design, sampling, evaluation instruments and analysis preclude summative comparison of mixed-methods or qualitative studies to quantitative studies using the MERSQI. 16 21 23 24 The JBI Checklist is considered the most appropriate qualitative critical appraisal tool for use in pragmatic meta-aggregation of qualitative research. 24 It includes 10 items which regard the study’s research questions, methods, analysis and reporting, for a total score of 0–10. Following recommendations from the JBI Reviewers’ Manual, 17 a cut-off score for higher reliability studies was predetermined at 6/10. This score was chosen as studies obtaining six or more points included most key elements of high-quality design.

Data analysis

MERSQI and JBI Scores were used to establish which studies presented more reliable evidence of outcomes. Summary statistics were calculated for all 117 studies. In line with Geerts et al , 9 studies with a final MERSQI Score of 12/18 or higher were also analysed separately to isolate the most reliable evidence, as were qualitative and mixed-methods studies which achieved the pre-determined JBI Score of 6/10 or higher. As there was substantial methodological heterogeneity, study characteristics and outcomes were synthesised using a meta-aggregation approach. 25 All study quality appraisal scores are presented in the Online supplemental table 1 , and full data extraction tables are available on request.

Study reliability (MERSQI and JBI)

Twenty-eight of 117 studies (25%) were categorised as higher reliability. Two studies were categorised as higher reliability by both the MERSQI and the JBI tool, 26 27 14 studies (12%) by the MERSQI only and 12 studies (10%) by the JBI tool only. The median critical appraisal score according to the MERSQI was 8.5 (range 5–16 from possible range of 5–18) and the median critical appraisal score according to the JBI was 3 (range 0–9 from possible range of 0–10). Online supplemental table 1 includes the MERSQI and JBI Scores for all included studies.

Study design showed considerable room for improvement, as shown in online supplemental tables 2 and 3 . Nearly half the of studies (46%) relied on post-programme evaluations only, and 92% did not include a control group. Of the nine studies that did include control groups, most had substantial methodological flaws in their selection of control groups. One common method for control group recruitment was to use unsuccessful course applicants. 28–30 In terms of evaluation design, the median evaluation instrument score was 0 (range 0–3). The majority of studies (59%) did not fulfil any of the MERSQI requirements for evaluation instruments, including reporting questionnaire design, wording and content. Objective outcome measures were used in only a minority of studies, with 60% relying solely on self-reported measures.

Data analysis and reporting likewise showed considerable limitations. Only one in five studies (20%) met criteria for comprehensive analysis and reporting of data. Few studies analysed their data beyond descriptive statistics to consider the generalisability and implications (13%). In many cases, studies omitted basic statistical significance tests.

Many studies did not contain key reporting elements for qualitative research as outlined in the JBI tool (see online supplemental table 3 ). There was clear congruity between research methodologies chosen and the research objectives and methods employed in 60% of studies. A minority of studies adequately reported their analysis (28%) and interpretation of data (25%), the potential for the researcher to have influenced data collection and interpretation (23%) and the researcher’s cultural or theoretical orientation (15%). Participant voices were clearly represented through quotes in only 16/53 (30%) of mixed-methods studies and 5/10 (50%) of qualitative studies. There was a statement of ethical approval or ethics exemption in only 26 of 63 studies (40%) which used qualitative methods. No study included a statement of philosophical perspective (normally expected for qualitative research). 17

Programme design

There was considerable heterogeneity in leadership development intervention design. It was often unclear whether established good practice for development of medical education interventions was followed, as shown in figure 2 . 9 31 Only 52 studies (44%) reporting having conducted a needs assessment before their intervention, and only 20 studies (17%) explicitly reported using an established capability or competency framework to inform leadership programme goals and objectives. There was, however, a plan for training transfer reported or built into 68 of 117 interventions (59%).

Educational design components: studies which reported Kirkpatrick level 4 outcomes (k=34) compared with studies that did not report Kirkpatrick level 4 outcomes (k=83).

The majority of interventions were carried out in a single hospital department (27%), single hospital (22%) or a single university (12%). Just under a quarter (23%) of interventions were conducted in multiple healthcare centres. A further 15% of studies were conducted within a specialty training programme outside healthcare centres.

Most of the studies took place in the USA (67%) or the UK (16%). The remainder of studies were in other European countries (7%), Canada (4%) or Australia (3%), with a single study each from Africa, 32 India, 33 Israel 34 and Qatar. 35

Programmes ranged in length from 2 hours to 4 years. The median intervention length was 6 months, and the most common length was 1 year (19%). Only 18 interventions (15%) lasted longer than 1 year. Five interventions (4%) were shorter than 1 day.

Programme faculty

Programmes were predominately delivered by either in-house faculty (36%) or a mix of in-house and external faculty (32%). Programmes delivered by mixed faculty were most likely to show organisational outcomes, as shown in figure 3 . The professional backgrounds, qualifications and experience of faculty were generally not reported.

Relationship between faculty source and programme outcomes. Higher reliability studies were those with Medical Education Research Study Quality Indicator Score of at least 12/18 or Joanna Briggs Institute Score of at least 6/10. NR, not reported.

Participants

In terms of participant selection criteria, the majority of interventions included participants who volunteered (27%), were nominated (19%) or who applied to the programme (16%). In some cases the application process was highly competitive. Interventions were mandatory in one-fifth of studies (20%). A considerable proportion of all studies (23%) did not report the selection process for their learners, including one quarter (25%) of the studies categorised as higher reliability by MERSQI criteria.

Educational methods

A wide range of educational methods were employed in various combinations across the reviewed studies, as shown in figure 4 . Most interventions included lectures (68%) and small group work (61%). Project work was included in the majority of studies with organisational outcomes (68%), but only in a minority of studies which did not report organisational outcomes (33%). Individual or team mentoring was also more prevalent where organisational outcomes were reported (47% vs 23%).

Educational methods: studies which reported Kirkpatrick level 4 outcomes (k=34) compared with studies that did not report Kirkpatrick level 4 outcomes (k=83).

Educational content

Educational content varied considerably among interventions. The most consistent content area was leadership theory (reported in 65% of interventions). The other common content areas were performance management (44%), self-management (41%), change management (39%), communication (36%), teamwork (33%), quality improvement (30%), healthcare policy (27%), healthcare finance (26%) and leadership behaviours (20%). There were no notable educational content differences in higher reliability studies or in studies which reported organisational outcomes (Kirkpatrick level 4).

Evaluation methods

A wide range of evaluation methods were employed across the included studies. Nearly half used quantitative methods only for their evaluation (46%). Of the remainder, most studies used mixed methods (45%), with 10 studies (9%) using purely qualitative methods. These proportions were similar in the higher reliability studies (41% quantitative, 48% mixed methods, 10% qualitative).

Four out of every five studies (82%) used questionnaires in their evaluation. Almost all of these employed Likert Scale items (92%) and one-third included open questions (34%). Only 8% used content or construct validated questionnaires. The proportion of higher reliability studies using validated questionnaires was slightly higher at 20% (MERSQI) and 18% (JBI). An additional six studies (6%) had conducted an expert review of their questionnaire for content validity only.

More than two-thirds of the included studies relied solely on self-ratings (69%). A minority of studies included ratings from subordinates (3%), peers (7%), superiors (12%) or experts (20%). The proportion of higher reliability studies which relied on self-ratings was lower (39%), with increased use of ratings from peers (14%), superiors (25%) or experts (39%).

The majority of studies (72%) included the collection of outcome data regarding behavioural changes (Kirkpatrick level 3, 57%) or organisational outcomes (Kirkpatrick level 4, 24%). Only three studies relied solely on Kirkpatrick level 1 outcomes (reaction). 36–38

Nearly half of the studies used single group post-programme only designs (46%), with most of the other half using single group pre-programme and post-programme designs (46%). Most studies included a post-programme evaluation completed immediately at the end of the programme (90%). Only 18 studies (15%) included a longer-term evaluation. In higher reliability studies, longer-term evaluations were associated with increased reporting of organisational outcomes (56%) when compared with immediately-post designs (31%). All 16 higher reliability studies as assessed by the MERSQI used pre and post designs. Six of these included a non-randomised control group (38%), and one study included a randomised control group (6%). This was the only randomised control group used in any of the 117 studies.

Behavioural and organisational outcomes in higher reliability studies

A full summary of outcomes from all 117 studies is provided in online supplemental table 1 .

There was a range of behavioural (Kirkpatrick level 3) and organisational (Kirkpatrick level 4) outcomes demonstrated in higher reliability studies.

Behavioural changes were objectively demonstrated in higher reliability studies through observed changes in behaviour, 26 27 39–43 promotions, 44 45 increased responsibilities or titles 28 46–49 and project completion. 50–52 Subjective changes in behaviour included improved communication, 39 influence, 50 delegation, 27 collaboration, 53 involvement in service improvement 47 and application of skills learnt or improved leadership in general. 39 40 54–57 These changes were indicated through interviews, free text questionnaire responses and behavioural self-assessments.

Organisational outcomes in higher reliability studies (Kirkpatrick level 4) were defined prospectively and in most cases were objectively demonstrated through leadership project impact evaluations. Projects achieved a range of outcomes, including reduced waiting times, 50 improved patient care 46 50 and cost savings. 27 46 47 50 By assessing the financial impact of projects completed during the intervention and relating this to programme costs, one higher reliability study reported a 364% financial return-on-investment (ROI). 27 Other objective outcomes included reduced organisational turnover of participants, 28 improved departmental working climate, 39 reduced sick leave 44 and increased promotion of women. 45 Organisational outcomes were subjectively indicated through reports of increased staff retention 56 and improvement in organisational effectiveness. 27 One study reported that ‘intangible benefits’ resulted in a 106% financial ROI. 51

Organisational outcomes in higher reliability studies were reported more frequently from programmes delivered by a mix of internal and external faculty than from programmes delivered by only external faculty (83% vs 11%), as shown in figure 2 . Organisational outcomes were also more frequently reported from interventions conducted in a whole hospital (57%) or multiple hospitals (40%), compared with interventions conducted in a single specialty (conference or outside-hospital training programme) (33%), single university (25%) or in a single department (0%). There were no notable differences in outcomes related to specific educational content.

Higher reliability studies that reported organisational outcomes were more likely have included project work (70% vs 44%), mentoring (50% vs 22%), coaching (22% vs 11%) and reflective instruments such as personality type assessments (22% vs 6%) than higher reliability studies that did not report organisational outcomes. Organisational outcomes were reported less frequently in higher reliability studies that included simulation or role play (10% vs 33%).

The aim of this review was to synthesise recent empirical evidence and explore factors associated with higher level outcomes in physician leadership development.

We found a substantial increase in the number of studies which evaluate medical leadership development interventions compared with previous reviews. 6–10 14 15 In many studies, it is still not clear whether best practices for design, delivery and evaluation are being followed. 31 It is also not clear whether there are sufficient behavioural and organisational outcomes to justify the considerable and increasing investments in medical leadership development.

Compared with previous reviews, we found an increase in the proportion of studies which report the use of active learning methods such as project work, simulation, discussions and reflections, which are widely accepted to be a vital component of leadership development 58 and which were associated in our review with increased Kirkpatrick level 4 outcomes.

No single leadership development content area was particularly associated with improved outcomes. With respect to educational methods, however, there was an association between the inclusion of individual or group project work and of mentoring with organisational outcomes. This may support the established position that educational methods are more important than specific curriculum content for leadership development. 1 58 Simulation and role play were less common in higher reliability studies which reported organisational outcomes that those that did not report organisational outcomes. This unexpected finding could result from these studies being situated in a training environment rather than a working environment. Alternatively, it could result from the evaluation process and study designs rather than from a lack of organisational impact. Studies which included simulation and role play tended to focus their evaluations on objective changes in behaviour at the expense of evaluating organisational outcomes (see online supplemental table 1 ). Interestingly, lacking a leadership development framework did not seem to impede programmes from reporting organisational outcomes. This may indicate that programmes which are designed as bespoke solutions to local needs are more likely to achieve organisational impact than pre-packaged approaches to leadership development.

There was an additional association of more senior participant level with organisational outcomes. This may be related to the wider scope of influence or practice of senior physicians compared with resident physicians. It could also indicate that there is a longer post-programme development period before residents are able to have an impact on organisational outcomes. This would align with the finding that programmes which evaluated longer-term outcomes were more likely to report organisational outcomes.

Importantly, our findings indicated that leadership development interventions which used a combination of internal and external faculty were most likely to report organisational outcomes, and those interventions which used external faculty only were least likely. This could have significant implications for procurement and design of leadership development interventions across healthcare, particularly as courses run internally are associated with significantly reduced costs. 59 60

As in previous physician leadership development reviews that used critical appraisal instruments, 7 9 we found that studies frequently did not meet criteria for high reliability. Many studies failed to report important methodological features, which restricts readers’ ability to appraise studies and learn from their findings. This was particularly notable in terms of questionnaire design, with fewer than one in 10 studies using validated questionnaires or reporting their questionnaire content in detail. Most studies also did not report or analyse outcome evaluation data comprehensively. Many study designs were biased towards obtaining positive results, particularly in terms of the absence of control groups, having stringent or undisclosed selection criteria, including leading questions on questionnaires and relying solely on self-ratings. This is likely to have resulted in improved reported outcomes. The lack of evaluation quality seems to indicate perfunctory attention paid to evaluation design and precludes confident conclusions from these studies. Future studies could benefit from consulting study quality appraisal checklists such as the MERSQI and JBI in advance, in order to effectively design their evaluations.

This review does indicate that certain recommendations for improved programme evaluation are beginning to be applied into research. Whereas only 29% of the studies reviewed by Frich et al 8 included qualitative components, 63 (54%) of the 117 studies included in our review used mixed or qualitative methods. In a nascent and complex field such as medical leadership development research, 1 8 9 61 qualitative methods can have value in terms of establishing effective programme design features to achieve desired outcomes, 21 25 31 as well as helpful nuances of how, for whom, to what extent or in what circumstances interventions are effective or not. 9 10 62

Additionally, many studies in this systematic review evaluated outcomes at Kirkpatrick level 3 behavioural change (57%) or level 4 organisational outcomes (24%). This is a significant improvement from previous reviews. 7 8 14 Changes in behaviour (level 3) and organisational outcomes (level 4) are more closely associated with transfer of learning to the working environment than participant reaction (level 1) and learning (level 2). 63–65

Limitations and strengths

This review was limited by the reliability of the studies included. We attempted to control for study reliability using critical appraisal tools with cut-off scores for higher reliability studies. To the best of our knowledge, this is the first systematic review of healthcare leadership development interventions to use the JBI critical appraisal tool to critically appraise qualitative studies. The JBI tool enabled us to identify 12 additional higher reliability qualitative and mixed-methods studies which were not identified using the MERSQI. Marked heterogeneity of studies and evaluations precluded a formal meta-analysis, therefore, we adopted a meta-aggregation approach. This enabled us to highlight design components that are correlated with behavioural and organisational outcomes in higher reliability studies.

A substantial majority of studies reported only positive outcomes, which could represent a publication bias, and we limited our review to English language peer-reviewed studies. In line with Frich et al , 8 our database search was limited to MEDLINE, however, we augmented our database search with an extensive hand-search of reference lists and citations using Web of Science and Google Scholar. The hand-search revealed that many relevant empirical studies were absent from recent reviews despite some of those reviews searching a greater range of research databases. This could indicate flaws in healthcare leadership development literature tagging and filing procedures within medical and educational databases.

Our review has practical implications for those commissioning, designing and evaluating medical leadership development programmes in healthcare. No specific area of curriculum content and no particular leadership development framework were clearly associated with behavioural or organisational outcomes. While relevance and appropriateness of educational content is important, 31 this systematic review has more clear implications for leadership development methods than for specific content. Where possible, interventions should include projects and individual or group mentoring. Transfer of learning from the programme into learners’ daily work and their organisations should be planned into the programme and where possible active learning educational designs should be employed, including opportunities for learners to set their own goals for development. External faculty should be judiciously used to supplement in-house faculty, not as a replacement for in-house expertise.

In terms of evaluation design, efforts should be made to ensure that evaluations are cost-effective and produce data that is useful for both practitioners and researchers. 66 67 Effective mixed-methods evaluation strategies should be integrated into evaluation designs. Study quality checklists such as the MERSQI and JBI could be consulted in the programme design phase to help build high quality quantitative and qualitative evaluation methods into programmes. At the minimum, evaluation design should include consideration of assessment at multiple time points, inclusion of control groups and collection of objective data, as well as collection of qualitative data from interviews, focus groups, questionnaires or observations. Programme goals and intended organisational outcomes should be explicitly considered during evaluation design 67 so that measures of organisational outcomes (including project outcomes) can be incorporated into the evaluation design. Improving study design and building robust evaluation methods into programmes will allow evaluators and educators to more effectively understand factors which are reliably associated with high level programme outcomes. This could both inform the improvement of individual programmes and contribute to the medical leadership literature as a whole. It is only through more considered and thorough evaluation of physician leadership development programmes that we will be able to justify the investment they represent.

Ethics statements

Patient consent for publication.

Not required.

Acknowledgments

We would like to thank Tatjana Petrinic, University of Oxford Health CareHealthcare Librarian, for her invaluable assistance and advice in the search process.

Loewenthal L , et al
Falcone RE ,
Spurgeon P ,
Clark J , et al
Keswani A ,
Husebø SE ,
Akerjordet K
Rosenman ED ,
Shandro JR ,
Ilgen JS , et al
Brewster AL ,
Cherlin EJ , et al
Geerts JM ,
Goodall AH ,
Steinert Y ,
Naismith L ,
Leslie LK ,
Miotto MB ,
Liu GC , et al
Straus SE ,
Soobiah C ,
Sadowski B ,
Cantrell S ,
Barelski A , et al
Beckman TJ , et al
Aromataris E ,
Liberati A ,
Altman DG ,
Tetzlaff J , et al
Hammick M ,
Olsen Øystein Evjen ,
Lockwood C ,
Kirkpatrick DL
Porritt K ,
Lockwood C , et al
Fassiotto M ,
Maldonado Y ,
Berkenbosch L ,
Muijtjens AMM ,
Zimmermann LJI , et al
Sin H-P , et al
Thomas PA ,
Nakanjako D ,
Namagala E ,
Semeere A , et al
Kumar S , et al
Shechter E ,
Pur Eizenberg N , et al
Al-Mutawa N ,
Elmahdi H ,
Ringdahl EN ,
Tarwater KD ,
Lindbloom EJ
Johnson JM ,
Morris CA ,
Wright SC , et al
Giordano CR ,
Vasilopoulos T , et al
Ten Have ECM ,
Tulleken JE
Gilfoyle E ,
Gottesman R ,
von Vultée PJ ,
Dannels SA ,
Yamagata H ,
McDade SA , et al
Brockbank A ,
Baron R , et al
Hickford D ,
Lees P , et al
Barr JW , et al
Haftel HM ,
Anderson MS , et al
Hopkins J ,
Ku MC , et al
Throgmorton C ,
Mitchell T ,
Morley T , et al
Levine SA ,
Brett B , et al
Pradarelli JC ,
Lemak CH , et al
Wurster AB ,
Pearson K ,
Sonnad SS , et al
Bergman D ,
Wahlstrom R , et al
Monkhouse A ,
Boyd A , et al
McMahon L , et al
MacPhail A ,
Gagliano NJ ,
Colton D , et al
van Dijk H ,
Agyepong IA , et al
Kozlowski SWJ
Kennedy PE ,
Chyung SY ,
Winiecki DJ , et al
Edmonstone J

Supplementary materials

Supplementary data.

This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

Data supplement 1

Twitter @oscarlyonsnz, @J_Frich, @jaasongeerts

Contributors OL, RG and JRG planned the review. OL, RG and TF screened studies for inclusion. OL, RG, JRG, AM and TF abstracted and coded studies. OL, RG, JRG, AM, TF, JF and JMG contributed to analysis, writing and editing the manuscript.

Funding Oscar Lyons was supported during this work by a Rhodes Scholarship, a Goodger and Schorstein Research Scholarship (University of Oxford) and the Shirtcliffe Fellowship (Universities New Zealand)

Competing interests None declared.

Provenance and peer review Not commissioned; externally peer reviewed.

Read the full text or download the PDF:

Log in using your username and password

Search More Search for this keyword Advanced search
Latest content
Current issue
Write for Us
BMJ Journals More You are viewing from: Google Indexer

Helen Noble 1 ,
Joanna Smith 2
1 School of Nursing and Midwifery , Queen’s University Belfast , Belfast , UK
2 School of Healthcare , University of Leeds , Leeds , UK
Correspondence to Dr Helen Noble, School of Nursing and Midwifery, Queen’s University Belfast, Belfast BT9 7BL, UK; helen.noble{at}qub.ac.uk

https://doi.org/10.1136/eb-2018-102895

Statistics from Altmetric.com

Request permissions.

Many health professionals, students and academics including health researchers will have grappled with the challenges of undertaking a review of the literature and choosing a suitable design or framework to structure the review. For many undergraduate and master’s healthcare students their final year dissertation involves undertaking a review of the literature as a way of assessing their understanding and ability to critique and apply research findings to practice. For PhD and Master’s by Research students, a rigorous summary of research is usually expected to identify the state of knowledge and gaps in the evidence related to their topic focus and to provide justification for the empirical work they subsequently undertake. From discussions with students and colleagues, there appears to be much confusion about review designs and in particular the use and perhaps misuse of the term ‘systematic review’. For example, some quantitatively focused researchers subscribe to a ‘Cochrane’ approach as the only method to undertake a ‘systematic review’, with other researchers having a more pragmatic view, recognising the different purposes of a review and ways of applying systematic methods to undertake a review of the literature. Traditionally, systematic reviews have included only quantitative, experimental studies, usually randomised controlled trials. 1 More recently, systematic reviews of qualitative studies have emerged, 2 and integrative reviews which include both quantitative and qualitative studies. 3

In this article, we will build on a previous Research Made Simple article that outlined the key principles of undertaking a review of the literature in a structured and systemic way 4 by further exploring review designs and their key features to assist you in choosing an appropriate design. A reference to an example of each review outlined will be provided.

What is the purpose of undertaking a review of the evidence?

The purpose of a review of healthcare literature is primarily to summarise the knowledge around a specific question or topic, or to make recommendations that can support health professionals and organisations make decisions about a specific intervention or care issue. 5 In addition, reviews can highlight gaps in knowledge to guide future research. The most common approach to summarising, interpreting and making recommendations from synthesising the evidence in healthcare is a traditional systematic review of the literature to answer a specific clinical question. These reviews follow explicit, prespecified and reproducible methods in order to identify, evaluate and summarise the findings of all relevant individual studies. 6 Systematic reviews are typically associated with evaluating interventions, and therefore where appropriate, combine the results of several empirical studies to give a more reliable estimate of an intervention’s effectiveness than a single study. 6 However, over the past decade the range of approaches to reviewing the literature has expanded to reflect broader types of evidence/research designs and questions reflecting the increased complexity of healthcare. While this should be welcomed, this adds to the challenges in choosing the best review approach/design that meets the purpose of the review.

What approaches can be adopted to review the evidence?

View inline

Key features of the common types of healthcare review

In summary, we have identified and described a variety of review designs and offered reasons for choosing a specific approach. Reviews are vital research methodology and help make sense of a body of research. They offer a succinct analysis which avoids the need for accessing individual research reports included in the review, increasingly vital for health professionals in light of the increasing vast amount of literature available. The field of reviews of the literature continues to change and while new approaches are emerging, ensuring methods are robust and remain paramount. This paper offers guidance to help direct choices when deciding on a review and provides an example of each approach.

5. ↵ Canadian Institutes of Health Research . Knowledge translation. Canadian Institutes of Health Research . 2008 . http://www.cihr.ca/e/29418.html ( accessed Jan 2018 ).
6. ↵ Centre for Reviews and Dissemination . Guidance for undertaking reviews in heathcare . 3rd ed . York University, York : CRD , 2009 .
Buchwald H ,
Braunwald E , et al
Horvath M ,
Massey K , et al
Sheehan KJ ,
Sobolev B ,
Villán Villán YF , et al
Christmals CD ,
Whittemore R ,
McInnes S ,
Bonney A , et al
Greenhalgh T ,
Harvey G , et al
Rycroft-Malone J ,
McCormack B ,
DeCorby K , et al
Mitchison D ,
19. Joanna Briggs Institute Umbrella reviews . 2014 . http://joannabriggs.org/assets/docs/sumari/ReviewersManual-Methodology-JBI_Umbrella_Reviews-2014.pdf ( accessed Jan 2018 )
van der Linde R , et al

Competing interests None declared.

Provenance and peer review Commissioned; internally peer reviewed.

Read the full text or download the PDF:

Log in using your username and password

Search More Search for this keyword Advanced search
Latest content
Supplements
BMJ Journals More You are viewing from: Google Indexer

http://orcid.org/0000-0002-0261-807X Lisa-Marie Thomas 1 ,
http://orcid.org/0000-0002-8505-3368 Lucia D’Ambruoso 1 , 2 , 3 ,
Dina Balabanova 4
1 Centre for Global Development and Institute of Applied Health Sciences , University of Aberdeen , Aberdeen , UK
2 Umeå Centre for Global Health Research , Umeå University , Umeå , Sweden
3 MRC/Wits Rural Public Health and Health Transitions Research Unit (Agincourt) , School of Public Health, Faculty of Health Sciences, University of the Witwatersrand , Johannesburg , South Africa
4 Department of Global Health and Development , London School of Hygiene & Tropical Medicine (LSHTM) , London , UK
Correspondence to Lisa-Marie Thomas; lisamarie.thomas.16{at}aberdeen.ac.uk

Introduction Estimates suggest that one in two deaths go unrecorded globally every year in terms of medical causes, with the majority occurring in low and middle-income countries (LMICs). This can be related to low investment in civil registration and vital statistics (CRVS) systems. Verbal autopsy (VA) is a method that enables identification of cause of death where no other routine systems are in place and where many people die at home. Considering the utility of VA as a pragmatic, interim solution to the lack of functional CRVS, this review aimed to examine the use of VA to inform health policy and systems improvements.

Methods A literature review was conducted including papers published between 2010 and 2017 according to a systematic search strategy. Inclusion of papers and data extraction were assessed by three reviewers. Thereafter, thematic analysis and narrative synthesis were conducted in which evidence was critically examined and key themes were identified.

Results Twenty-six papers applying VA to inform health policy and systems developments were selected, including studies in 15 LMICs in Africa, Asia, the Middle East and South America. The majority of studies applied VA in surveillance sites or programmes actively engaging with decision makers and governments in different ways and to different degrees. In the papers reviewed, the value of continuous collection of cause of death data, supplemented by social and community-based investigations and underpinned by electronic data innovations, to establish a robust and reliable evidence base for health policies and programmes was clearly recognised.

Conclusion VA has considerable potential to inform policy, planning and measurement of progress towards goals and targets. Working collaboratively at sub-national, national and international levels facilitates data collection, aggregation and dissemination linked to routine information systems. When used in partnerships between researchers and authorities, VA can help to close critical information gaps and guide policy development, implementation, evaluation and investment in health systems.

verbal autopsy
cause of death
health policy
systems research
population surveillance

This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/

https://doi.org/10.1136/bmjgh-2017-000639

Statistics from Altmetric.com

Request permissions.

Key questions

What is already known.

Verbal autopsy (VA) is an established health surveillance method that provides information on levels and causes of death in populations where medical death certification is weak or absent.

From 1998 to 2010, methodological and technical developments and high level advocacy have transitioned VA from use in research environments towards pragmatic applications on a wider scale, including in routine registration.

An extensive literature exists on the validity of VA and applications in research settings. Information on its use in health policy and systems, ie on working with health authorities to provide evidence for action, is less common by comparison.

What are the new findings?

Since 2010, applications of VA in health policy and systems were identified in 15 low and middle-income countries. There were various forms and extents of engagement with decision makers and governments.

Advancements in electronic data collection, automated data interpretation and electronic databases have supported collaboration in terms of data sharing, and monitoring and evaluation of disease burdens.

When combined with complementary methods, such as, social autopsy, the explanatory potential of VA is extended to identify contributory factors in conjunction with information on levels and causes of deaths in populations.

Active engagement with decision makers and communities was seen to help establish a relevant evidence-base that directly informs the means for action.

What do the new findings imply?

VA has significant potential in health policy and systems to deliver robust and reliable evidence, help close gaps in statistics, and guide implementation, evaluation and investment.

In conjunction with complementary data collection efforts, VA can contribute to more holistic views of health systems performance combining evidence on burden of disease with social determinants and local knowledge.

The active collaboration of various stakeholders: governments, health authorities, communities and research groups, can foster engagement in, and coverage of, data collection and enhance the validity and utility of the process.

Introduction

It is vitally important that data on cause of death are available, reliable, timely, and collected and aggregated at low cost for a robust evidence base to inform strategic health policy and evaluation. 1 Collecting data on births and deaths including cause of death in high-income countries relies on well-established civil registration and vital statistics (CRVS) systems. However, in many low and middle-income countries (LMICs), there is a pervasive lack of death registration inclusive of notification of medical cause of death due to low investment over decades. 1 2 This situation highlights an important gap, as one in two deaths go unreported globally, the majority occurring in LMICs. 1 3

The importance of recording vital events through improved CRVS systems is well recognised. The Lancet series ‘ Who counts ?’ in 2007 4 and ‘ Counting births and deaths’ in 2015 5 emphasised the significance of robust data on vital events. 6 There is also increasing recognition among the global health community that progress towards universal health coverage (UHC) depends on effective monitoring of equitable coverage and utilisation. 7 8 To achieve goals such as UHC and the Sustainable Development Goals (SDGs), countries need functional CRVS systems, and the collection of data on deaths of people excluded from access to health systems is an important step for addressing health inequalities and saving lives. 9

Verbal autopsy (VA) is a method that can be used to collect and analyse data on cause of death. VA is defined by the WHO as ‘a method used to ascertain the cause of a death based on an interview with next of kin or other caregivers’ that can be applied for deaths without certification of medical causes. 2 Its main objective is to deliver a simple identification of cause of death at community or population level in countries where no other functional registration system is in place and/or where many people die at home without contact with the health system. 2 Over the past 25 years, VA has become a primary source of information about cause of death in several LMICs. 2

The development of VA can be traced back to the 1950s and 1960s, where researchers in Asia and Africa used physician interviews with relatives and carers of deceased persons to assess cause of death and generate cause of death statistics. 10 Workers from the Narangwal project in India named this technique VA. 11 In the 1970s, the method gained attention when WHO encouraged the use of ‘lay reporting’ by people with no medical training. 12 Subsequently, Reproductive Age Mortality Studies, Matlab (Bangladesh) and Niakhar (Senegal) developed VA questionnaires to ascertain possible cause of death diagnosis in the late 1970s and early 1980s, which are still used in research settings and national/regional surveys. 2 13

Concerns about the validity of instruments and comparability of data arose in the early 1990s and led to the convening of expert committees to develop standardised VA tools for childhood and maternal deaths, 14 15 leading to the development of VA standards for maternal deaths in 1994. 16 In 2007, the WHO published three standard VA questionnaires, ‘death of a child aged under four weeks’, ‘death of a child aged four weeks to 14 years’ and ‘death of a person aged 15 years and above’. 2 These tools sought to permit certification and coding commensurate with International Statistical Classification of Diseases and Related Health Problems, 10th Revision (ICD-10) and ascertain all causes of death with reasonable accuracy drawing on well-administered VA interviews. 2 The standards were updated in 2012, 2014 and 2016 to improve cause-specific mortality data and to ensure consistency and comparability between countries. 10 Appendix 1 contains a description of the development of the WHO VA standards from 2012-2016 (online supplementary appendix 1 ).

Supplementary file 2

In its present form, VA typically consists of two main stages. First, information is collected via structured interviews with family members and caregivers of the deceased on their signs, symptoms, medical history and circumstances at and around the time of death. 2 Second, interview data are interpreted by physicians (Physician-Certified VA, [PCVA] 1 17 ) or using automated methods (‘Computerised Coding of VA’ [CCVA]) for example, InterVA and SmartVA, using algorithms and probability theory, 1 18 to obtain probable cause(s) of death.

The method is applied in a range of study designs and research settings: clinical trials and large-scale epidemiological studies, health and demographic surveillance systems (HDSS), national sample surveillance systems, and household surveys. 1 2 The resulting data can help to determine gaps in vital statistics and help to establish population-level disease burden estimates. 19 Furthermore, it can assist in monitoring and evaluation of health policy, planning and instituting programmes.

Due to the lack of effective national CRVS systems in many LMICs, statistics on disease burdens are often calculated by researchers using prediction models and estimation procedures. 20 Alternative approaches to provide reliable sources of vital data are HDSS or Sample Vital Registration with verbal autopsy (SAVVY) systems. While surveys collect information from a representative population sample, 21 HDSS and SAVVY systems aim to regularly monitor ‘demographic and health characteristics of a population living in a well-defined geographical area’ and collect prospective, longitudinal data. 20

VA has been used to investigate population health in selected contexts by examining sequences of events in detail, particularly rare events, for example, maternal deaths. 21 HDSSs are commonly situated in deprived rural, semiurban or urban areas where generalisability is conditional on a geographically, well-identified context. 21 VA is frequently used in HDSS contexts to determine levels and causes of deaths for surveillance populations. Rather than precise descriptions of events at a given time, HDSS focus on the relationships between events at the community, household and individual levels over time. 21 The information generated can help to evaluate the effects of, and give evidentiary support to, the scaling up of interventions. 20 Furthermore, HDSS data collected and analysed with standardised tools, compared and cross-validated with national and international data, has shown extensive conformity, for example, recently with data from the International Network for the Demographic Evaluation of Populations and their Health (INDEPTH) and Global Burden of Disease estimates, which supports generalisability beyond specific surveillance sites. 21

VA is a well established surveillance method used in over 45 LMICs for over 25 years mainly in research settings, and/or as part of large household surveys to calculate disease burdens in populations. 9 Developments in VA tools and analysis, combined with wider recognition of the dearth of mortality data globally, have led to VA being recognised as a valuable, interim approach for use outside research environments, and towards use in CRVS systems. 22 On this basis, this review is concerned with how VA has been used to inform health systems, policy and management in LMICs.

Major collaborative networks also support the use of VA, eg initiatives such as ‘Bloomberg Data for Health’, a network of local and national authorities in 20 LMICs, seek to strengthen the collection and use of critical public health information using the method. One of the key objectives of the initiative is to provide resources to understand broad systems issues, as well as promote the integration of VA data into CRVS systems, enhancing the use of data for maximum impact in policymaking and priority setting. 23 24

Furthermore, the Million Death Study (MDS) in India (1998–2014), one of the largest studies of premature mortality in the world, involves the use of VA on a large scale, to influence policy and decision making. The MDS worked in collaboration with the Registrar General of India, which since 2002 has integrated an enhanced form of VA into its large-scale, nationally representative Sample Registration System. 25 Monitoring births and deaths that occur outside healthcare facilities with VA in approximately one million randomly selected households has provided invaluable information to governments, research agencies and media, leading to action against preventable deaths. 26

The wider use of VA is often hampered by concerns about validity. While there have been several studies investigating this issue, their findings need careful consideration. Medical autopsies and hospital records in LMICs are often of poor quality. 27 However, these are frequently used as a ‘gold standard’ in testing the validity of VA. Validation studies are therefore a comparison of two imperfect cause of death assignments. 27 To this end, the Population Health Metrics Consortium (PHMRC) conducted research in four countries documenting deaths in high-level hospitals and followed these up with VA, developing a data set labelled as ‘gold standard’ VA data. 28 However, the authors acknowledged that even though the VA data might have achieved internal validity, the external validity was likely to have been limited. 28 While VA may be seen as an imperfect tool for ascertaining cause of death, it is often the only alternative in the absence of medical certification. 22

A further stream of methodological development aims to combine VA with social autopsy (SA). SA is a method aiming to ‘collect the data needed to connect the fatal illness or the act of diagnosing or recognizing that illness to a set of socio-demographic, economic, cultural conditions or factors’ representing a social diagnosis of the death. 29 Relative to VA, SA is a novel approach, which has not been widely practised and still lacks standardisation for data collection and analysis to the extent seen with VA. 30 Current work by the Child Health Epidemiology Reference Group (CHERG) and INDEPTH seeks to arrive at standardised SA tools. 31 Methodologists posit that VA combined with SA provides enhanced data that can generate more holistic information on causes and determinants of deaths and situate cause of death in a richer account of context. 32 In addition, ‘VASA’ (verbal autopsy and social autopsy) studies have been conducted in Bangladesh (2007–2011), Malawi (2011–2012), Niger (2012–2013) and Nigeria (2013), using retrospective surveys for data collection. 32 VASA data have also been incorporated into updates of modelled data as part of global and regional mortality estimates, supplemented national mortality estimates, and informed policy and programme development. 32

Aims and objectives

Considering the methodological developments and global recognition of the broader use of VA both for research and policy development, the contribution of this paper is twofold: first, it aims to examine how VA is used to inform the operation of health systems (ie, beyond surveillance or condition-specific purposes); and second, to identify and review how VA is used in an ‘embedded’ manner in health policy and systems development. That is, with and for health authorities, aligning to and reflecting the realities of implementation 33 and informing action. It updates a previous review by Fottrell and Byass, charting the methodological transition of VA from research environment to routine application in CRVS systems 13 with a focus on the use of VA to inform health system and policy research and development.

The review was conducted using a systematic search strategy. Information was gathered on implementation features, the use and purpose of VA to strengthen health policy and systems, and strengths and limitations of the approaches used.

Search strategy

We examined online databases, including PubMed, POPLINE, Web of Science and Scopus using combinations of keywords and phrases ( figure 1 ). Further sources and especially grey literature were identified through manual searches of references quoted in original publications. Additional hand searches and web searches were conducted on organisational portals, for example, Centres for Disease Control and Prevention, INDEPTH, CHERG and PHMRC, to identify further papers. The time frame 2010–2017 was defined as a key paper on VA considering its methodological transition was published in 2010. 13

Download figure
Open in new tab
Download powerpoint

Search strategy used for PubMed, POPLINE and Scopus.

Inclusion and exclusion criteria

Inclusion criteria were as follows: (1) papers published after 2010 in a journal or reports accessible through web search; (2) papers that described engagement in health systems to employ VA findings and (3) papers written in English. Studies that did not describe a contribution to health systems and policy strengthening and/or health policy and systems research by identifying needs or changes in policies and practice (eg, reforms to secure healthier communities) 34 were excluded.

Data extraction and analysis

A form was developed to extract data from retrieved articles on (1) study setting; (2) stated objective(s); (3) VA/SA method used; (4) remarks/limitations on the use of VA in health policy and system research; (5) applicability for health systems and policy strengthening and/or research; and (6) key findings. Articles were selected and reviewed, and data were extracted from the retrieved studies by one reviewer (LMT). Subsequently, data extraction was performed independently by two reviewers (LD and DB). Thereafter, data extractions were collated from all reviewers, and divergences were identified and resolved via discussion. Thematic analysis and narrative synthesis were subsequently conducted to analyse the data extracted. Themes and subthemes that emerged from the literature were identified, compared, discussed and associated with each other by the reviewers to ensure plausibility, validity and consistency. In accordance to the aims and objectives of the review, a framework was developed to demonstrate the extent of use of VA/SA in health policy and systems. The framework delineated a range of uses comprising: (1) disease-specific or vertical donor-driven policies and programmes, seeking to measure specified programme outcomes; (2) analysis and interpretation of information at the health system level, seeking to capture a broader range of (intended and unintended) outcomes and considerations for health policies; and (3) ‘embedded in policy’ processes at the system level, that is, active engagement with the Department of Health or Ministry of Health (MOH) ( figure 2 ).

Framework: continuum of the use of verbal autopsy in health policy and systems.

The methodological quality of included studies was assessed by the reviewers using the relevant Critical Appraisal Skills Programme (CASP) checklists 35 for qualitative studies and an adapted version of the Effective Public Health Practice Project (EPHPP) 36 assessment tool for quantitative studies. The tool usually involves rating articles (strong, moderate, weak) on six components: selection bias, study design, confounders, blinding, data collection methods, withdrawals and dropouts. In the modified version developed, the blinding section was disregarded as studies containing VA do not usually involve blinding, and so inclusion of this component would have had a misleading adverse impact on the quality assessment. The quality appraisal process sought to identify common strengths and weaknesses, rather than to exclude studies. All studies were analysed and used in the analysis with strengths and weaknesses taken into account.

Search results

The search identified 634 articles. 362 publications were identified through PubMed, while 132, 117 and 23 publications were identified through Scopus, Web of Science and POPLINE respectively. An additional eight papers were identified through hand searches. After title and abstract review, 26 publications were considered eligible for the review ( figure 3 ).

Literature search and review process.

Study characteristics

From the selected papers, 14 were research papers. 22 25 37–49 Seven were quantitative, from which the rating according to the EPHPP tool was strong for two articles, moderate for four, and weak for one. 38–40 43 45–47 According to the CASP tool, the one qualitative study identified had a low risk of bias. 49 Three papers were considerations of contextual factors for systems integrating VA, 22 25 44 and two investigated feasibility issues (one of community informants and one, the cost of VA in a mortality surveillance system). 42 48 One was a pilot study that identified weaknesses in an existing maternal death review system and introduced a community-based process to improve the recording of maternal deaths. 37 From the remaining papers, four were HDSS profiles, 50–53 four were correspondences/commentaries, 54–57 two were information sheets, 41 58 one was a summary of study results 59 and one was a research proposal. 60 The studies were conducted in 15 LMICs: 10 in Africa, 37 40 41 43 47 51–53 59 60 8 in Asia, 25 39 42 44 46 49 50 55 2 in the Middle East 45 48 and 1 in South America 38 ( figure 4 ). In five studies, the country was not specified. 22 54 56–58 The finding that the majority of studies took place in LMICs was not surprising as VA is predominantly used in countries where CRVS systems are incomplete or absent. Ten studies used a WHO VA tool, 38 39 43 45–48 56 58 60 two an adapted version developed by INDEPTH, 52 53 two used the PHMRC tool 49 59 and one applied a study-specific tool (Vital Registration and VA tool). 40 Eleven studies mentioned the use of a VA tool but gave no specifics on which. 22 25 37 41 42 44 50 51 54 55 57 For data collection with VA, five stated the use of electronic devices, 40 49 52 56 60 two used paper-based questionnaires, 42 50 while the remaining studies did not state how data were collected. Eight mentioned the additional use of questionnaires investigating social determinants. 37 40 43 44 47 48 53 59 Specifics of how governments were involved in the process varied considerably. In eight studies, government involvement was in more than one element of the process: five studies described involvement in planning and research programme implementation, 37–39 49 52 and in three studies, in implementation and scale up. 45 48 59 In a further nine studies, government involvement was described in a single element of the process: five in implementation, 25 41 43 46 47 two in scale up, 44 60 one in funding support, 50 and one through a working relationship which was stated but not described, 53 respectively (online supplementary table 1 ). Twelve studies were conducted within established routine health information systems, 25 38 40 41 46 49–53 59 60 and eight introduced or established new methods of data collection on cause of death. 37 39 42–45 47 48

Supplementary file 1

Countries in which verbal autopsy (VA) studies were conducted and reported in scientific literature 2010-2017.

Implementation features

VA was used to ascertain cause of death predominantly in studies where no other form of certification was available 22 54 58 60 or for deaths outside facilities, 25 40 45–49 investigating mainly maternal, child and neonatal deaths. 37 41 43–45 47 48 53 59 Other studies used VA in demographic surveillance sites in conjunction with epidemiological and demographic research, 50–53 and for recording and investigating deaths in HDSS for the purposes of informing health policy and planning. 38 40 42 46 50–52 55 60

Through methodological advances, and often with reference to the use of CCVA, VA was stated to have become a valued and feasible approach for use outside research settings, for wider application in health systems and policy as well as in national CRVS systems. 22 25 38 40 42 46 48 49 54–56 58 60 This was supported by the study from Joshi et al , 42 which investigated the cost of a VA-based mortality surveillance system in rural India. The study concluded that the approach was cost-effective (US$0.10 per capita per year), and that with data collection on electronic devices the cost could be further reduced. 42 The authors demonstrated the use of VA in a mortality surveillance system as feasible, and as capable of providing valuable, timely evidence on cause of death for policymakers and health professionals. 42

Other studies did not directly investigate costs but stated that cost savings could be achieved by conducting VA on a representative sample of all deaths or in a selection of registration administrative areas. 22 As noted by Sankoh and Byass, 56 the time and resource implications of establishing and maintaining CRVS systems inclusive of VA are inevitable, however the benefits that functional CRVS systems confer on people and societies are considerable. These include enabling access to public goods, facilitating development across all sectors and delivering valuable information for health policy and planning. 56 In this sense, the cost of not collecting vital information is acknowledged as far higher. Government funding and commitment, as well as support and governance through an independent body, were also recommended to ensure longevity, continued evolution and provision of academic freedom to achieve this. 43 44 50

Several studies described VA as an applicable tool that adds value across intersectoral domains and beyond public health. 38 40–42 44 50 59 60 These included ministries of health, local governments, specialised agencies (eg, national and international organisations), national health insurance, local health authorities, health facilities, community workers as well as departments of justice and interior. 22 43 54 Several authors recommended an interdisciplinary and intersectoral approach that combines empirical and analytical efforts with different organisations, stakeholders and the community to enable functional registration in poorly resourced health systems. 22 37 44 47 53 59 In the paper by Bensaïd et al , 59 an intersectoral collaboration is described to demonstrate the power of implementing surveillance and response interventions in settings with high levels of mortality. The collaborative effort was initiated by the government of Niger and implemented by the National Statistics Institute and a VASA working group including the MOH, the Niger country office of UNICEF and other partners in partnership with the John Hopkins School of Public Health, which provided technical assistance on behalf of WHO/UNICEF and supported by CHERG. 59

Furthermore, communities were described as a critical foundation on which to build CRVS systems and ensure sustainability. 37 44 52 54 So-called hybrid approaches, using a combination of reporting through community-based systems, for example, health facilities, burial authorities, key informants as well as official registration, were detailed. 22 38 58 Several authors pointed out, however, that communities should not be ‘used’ as passive data collectors but instead adopt roles as active partners in the surveillance and response processes. 37 44 48 54 Authors also noted that this approach has the potential to generate self-sustaining sources of information on mortality, improve trust in health systems, and counteract the potential for ’community fatigue' from repeated data collection cycles. 37 52

In this sense, active community involvement can enhance understandings of, and ultimately transform social exclusion from, access to health systems by shifting control over the production, use and exchange of knowledge on health towards those most directly affected. 54 One example of such an approach was a pilot study by Bayley et al , 37 which investigated a community-linked maternal death review (CLMDR) to measure and prevent maternal mortality. 37 The study demonstrated that CLMDR improved identification and review of deaths, providing opportunities for education and stimulating action in facilities. Due to the success of this approach in a sample site in Malawi, the MOH initiated a nationwide roll-out of the programme. 37

Studies employing VA to examine facility-based deaths were also identified. VA was described to become particularly important where postmortem pathological examinations are not available, not mandatory, unreliable, or death certificates are issued for non-specific causes, which are of little value for public health decision making. 22 38 55 58 De Savigny et al 22 and Bayley et al 37 note the importance of investigations of this nature when recording maternal deaths that were not attended by trained providers.

VA in health systems and policy

Of the papers reviewed, 18 explicitly acknowledged a close intersection between the use of VA methods, and the functions of the health system to improve population health. This was achieved via: (1) active engagement of the government in the process, 39 43 46 47 49 55 59 60 for example, high-level engagement of the Niger government and stakeholders for dissemination of VASA study results to support policy development, 59 (2) demands from, or working relationship with, governments/health authorities, 38 45 48 50 52 53 for example, the request from the Ministry of Public Health in Afghanistan to expand an intervention programme to reduce maternal deaths, 45 and (3) purposeful dissemination of information to governmental, national and international agencies, 37 40 41 51 58 for example, VA within the Millennium Global Village-Network provided feedback to local providers. 40

In combination with questionnaires about social determinants, including as part of SA processes, VA was also used to identify social factors contributing to deaths and to determine gaps in health systems. 44 52–54 SA has been reported as effective in supporting a more comprehensive analysis of the extent to which interventions achieve intended effects or help to explain why actions to improve people’s health are not taken by community members or health authorities. 44 47 53

The studies demonstrated that when other routinely collected data, for example, surveys on demographic characteristics, are supplemented with VA data on a regular basis, they can provide information that can help to enhance health planning, inform policies and contribute to health system strengthening. However, to do so, data need to be shared and interpreted at all levels, from communities to systems. 38 40–42 44 46 47 52 54 59 60

Strengths and limitations of VA for health systems and policy

According to the WHO, the gold standard for cause of death reporting is certification by a medical practitioner trained using the ICD system. 61 However, in LMICs there is often insufficient supply of appropriate medical equipment, trained forensic experts and pathologists. Cultural norms and traditions can also interfere with timing and performance of autopsies. 38 46 As de Savigny et al 22 state, the absence of locally relevant gold standard data presents challenges in conducting locally applicable validation studies to determine the external validity of VA results. As stated above, VA has become increasingly standardised in recent years and more frequently applied globally, allowing for comparison with other standardised methods, for example, hospital records. 56 58 The literature also acknowledged, however, the value of adaptations to context, and that standardisation may inadvertently compromise accuracy, completeness and validity, limiting relevance to local policy and planning. 22 49 54 58

Use of mobile devices and electronic databases, ‘mHealth’, was a relatively new development affecting how VA is used. Electronic platforms were frequently described as able to enhance data sharing and link data of a range of types and sources: demographic, epidemiological, mortality, morbidity, clinical, laboratory, household, environmental, health system for a more holistic view. 57 National electronic records systems, ‘SmartCare’ in Zambia, 62 for example, have been reported to create new opportunities to enhance the timeliness of data collection and analysis, among others with VASA, enabling real-time data capture. 22 40 47 52 53 60 These advances were reported to assist in the detection of rapid changes in health systems and disease burdens, as well as track key indicators related to local, regional and global goals and associated targets, 48 59 aiding in efficient decision making and delivery of care. 25 40

Encouraging the use of VA in countries lacking adequate or representative data via the expansion of HDSS, and with VA complementing hospital data were also seen as promising approaches. 22 38 41 De Savigny et al 22 recommend that data gathered with VA in the community should be analysed with physician-certified hospital deaths. To realise potential, it was further recognised that expansion of infrastructure and collaboration between research and service organisation and delivery are necessary. 54 In South Africa, for example, the Department of Science and Technology has recently launched an initiative to consolidate three existing HDSS sites and expand the platform with three further urban and rural nodes, to cover over 1% of the national population. 63

Discussion and conclusion

This review sought to examine how VA is used in a manner consistent with the emerging health policy and system research paradigm. 33 The review fills an important gap in the literature on how VA is (and can be) used to support the routine operation of health systems, not just in terms of surveillance but in policy and programme development. Further, it explores its use in an ‘embedded’ manner in health policy and systems research, working with planners and implementers.

A number of limitations of this review need to be acknowledged. Findings may have been influenced by publication and reporting bias. The search may have missed studies in which VA was used in health policy and systems due to the lack of uniform reporting or indexing methods. Many key studies are also reported in the grey literature, with uneven quality and variable availability. On this basis, grey literature was included to increase the breadth, relevance and utility of the review.

The studies identified were conducted in 15 LMICs in Africa, Asia and South America using mostly standardised VA tools to investigate generally maternal and child mortality. Of the studies selected, eight used VA ‘embedded’ in projects investigating the effectiveness of single or multiple interventions implemented in the study area. Four used VA in established HDSS sites to ascertain cause of death in the population, while the remaining studies looked at the feasibility of VA in national health and vital registration systems.

The studies either disseminated data to health authorities and government representatives or stated the direct involvement of a government representative in the process (ie, in planning, implementation and/or scale-up). A notable shift in the time period covered by this review was the application and use of VA beyond research settings. The method has been developed to enhance appropriateness and feasibility of implementation in vital registration where information gathered is actively and routinely used by health authorities. The review revealed a continued transition of the VA method beyond research environments into health policy and planning, contributing to achievements in, and demonstrating of progress towards UHC and other targets. With methodological advancements and technical developments, it is regarded as a valuable and feasible approach to supplement and strengthen national sample and CRVS systems in LMICs. 22 38 49 58 The increasing use of electronic means to collect, compute and share real-time data, for example, using tablets, CCVA, e-health and mHealth recording systems, is also contributing to development and use on a wider scale and routine basis. 22 49 62 64

An intersectoral focus and the involvement of communities were reported as further key developments to the method, which were seen to facilitate sustainability. Here an important tension was identified whereby despite efforts to standardise and generate comparable data across settings, it was noted that VA is not a one-size-fits-all approach, 20 and adaptations to local conditions were described as critical for maintaining relevance to national and local health contexts. 10 13 Standardised and validated VA were also described as important for evidence-based resource allocation when used by regional, national, cross-national and international actors. 65 Moreover, when VA is used in conjunction with other data collection methods, for example, SA on demographic and health system characteristics, a more holistic view of health systems performance was achieved.

Improving CRVS systems in LMICs was described to involve national policymakers and governments, and to require commitments from the global health community to help fund and build these systems. This was seen to require an interdisciplinary and intersectoral effort involving a variety of organisations and stakeholders such as MOHs, ministries of local government, national and international agencies, as well as health facilities and researchers. There was also evidence of merit in the involvement of communities and families living in registration sites. 37 44 52 54 Several partnerships models were seen to help extend CRVS systems in settings lacking resources and capacity, fostering improved health systems performance and, ultimately, reductions in avoidable mortality. 22 38 58

There is a growing demand for evidence-based and data-driven interventions in LMICs. With new technologies facilitating data collection and analysis, and in combination with supplementary data collection methods, VA is used in a versatile manner beyond research settings to inform policy, programming and planning. Where governments are involved in the planning, implementation and/or scale-up of VA activities, information tended to be used more directly in planning and policy processes. Increasing coverage and use of VA on a routine and continuous basis may offer information gains on many fronts: to inform local health institutions and policy development, and in planning and research at national and global levels. Data that are developed collaboratively and communicated effectively, recognising strengths and limitations, were further promising developments. The literature reviewed suggests that VA has potential to help close the gaps in the availability and use of intelligence, and guide policy implementation, evaluation and investment in health systems.

Chandramohan D ,
Byass P , et al
2. ↵ World Health Organization . Verbal autopsy standards: the 2016 WHO verbal autopsy instrument . Geneva : World Health Organization , 2016 .
3. ↵ World Health Organization . Almost half of all deaths now have a recorded cause, WHO data show . Genava : World Health Organization , 2017 .
Macfarlane SB ,
Szreter S , et al
Mikkelsen L ,
Phillips DE ,
AbouZahr C , et al
AbouZahr C ,
de Savigny D ,
Mikkelsen L , et al
Shibuya K ,
Eozenou P ,
Evans D , et al
D’Ambruoso L ,
Wager RG , et al
10. ↵ World Health Organization . Verbal autopsy standards: ascertaining and attributing causes of death: The 2016 WHO verbal autopsy instrument . 2017 http://www.who.int/healthinfo/statistics/verbalautopsystandards/en/ ( accessed 8 Mar 2017 ).
12. ↵ World Health Organization . Verbal autopsy standards: ascertaining and attributing causes of death: The 2014 WHO verbal autopsy . 2014 http://www.who.int/healthinfo/statistics/verbalautopsystandards/en/index1.html ( accessed 18 Jan 2017 ).
Fottrell E ,
14. ↵ World Health Organization . International standard verbal autopsy questionnaires . Verbal autopsy standards: ascertaining and attributing cause of death . Geneva : World Health Organization , 2007 : 5 – 51 .
15. ↵ World Health Organization . WHO technical consultation on verbal autopsy tools: review of the literature and currently-used verbal autopsy tools . Geneva , 2005 . http://www.who.int/healthinfo/statistics/mort_verbalautopsy.pdf
Campbell O ,
18. ↵ World Health Organization . Verbal autopsy standards: ascertaining and attributing causes of death: The 2012 WHO verbal autopsy instrument . 2012 http://www.who.int/healthinfo/statistics/verbalautopsystandards/en/index2.html ( accessed 18 Jan 2017 ).
19. ↵ World Health Organization . Verbal autopsy standards: The 2012 WHO verbal autopsy instrument . Geneva , 2012 .
Wamukoya M ,
Ezeh A , et al
Bocquier P ,
Chandramohan D , et al
23. ↵ Bloomberg Philanthropies . Data for Health . 2017 https://www.bloomberg.org/program/public-health/data-health/#problem ( accessed 5 Jan 2018 ).
24. ↵ Centers for Disease Control and Prevention (CDC) . Data for health initiative . 2017 https://www.cdc.gov/globalhealth/healthprotection/data.html ( accessed 5 Jan 2018 ).
Aleksandrowicz L ,
Malhotra V ,
Dikshit R , et al
26. ↵ Centre for Global Health Research . Million Death Study (MDS) . 2017 http://www.cghr.org/projects/million-death-study-project/ ( accessed 5 Jan 2018 ).
Murray CJ ,
Black R , et al
Libite PR ,
Moluh S , et al
Kalter HD ,
Jakob R , et al
32. ↵ John Hopkins Bloomberg School of Public Health . Verbal Autopsy and Social Autopsy Studies (VASA) . 2017 http://www.jhsph.edu/research/centers-and-institutes/institute-for-international-programs/current-projects/verbal-autopsy-and-social-autopsy-studies-vasa/ ( accessed 18 Jan 2017 ).
34. ↵ World Health Organization . Universal Health Coverage . Geneva , 2012 . http://www.who.int/universal_health_coverage/en/
35. ↵ CASP . Critical appraisal skills. Crit. Apprais. Ski. Program . 2013 http://www.casp-uk.net/checklists ( accessed 4 Jan 2018 ).
36. ↵ Effective Public Health Practice Project . Quality assessment tool for quantitative studies. EPHPP . 2009 http://www.ephpp.ca/tools.html ( accessed 13 Jan 2018 ).
Chapota H ,
Kainja E , et al
Guimarães MD , et al
Alwi Q , et al
Ohemeng-Dapaah S ,
Akosa E , et al
Konopka S ,
Praveen D ,
Jan S , et al
Orobaton N ,
Abdulazeez J ,
Abegunde D , et al
Mishra A , et al
Zainullah P , et al
Soemantri S ,
Djaja S , et al
Serbanescu F ,
Goldberg HI ,
Danel I , et al
Shaikh MS ,
Qomariyah SN , et al
Sun J , et al
Majumder S , et al
Elung’ata P ,
Mberu B , et al
Levira F , et al
Gyapong M ,
Sarpong D ,
Awini E , et al
Dharmaratne SD ,
Jayasuriya RL ,
Perera BY , et al
58. ↵ United Nations Economic Commission for Africa . Cause of death fact sheet . 2015 http://www.uneca.org/sites/default/files/uploaded-documents/Statistics/CRVS/cause_of_death_fact_sheet_en.pdf ( accessed 14 Mar 2017 ).
Bensaïd K ,
Kalter HD , et al
60. ↵ Centers for Disease Control and Prevention (CDC) . Community-based verbal autopsy linked to Kenya’s civil registration and vital statistics system . 2017 https://www.cdc.gov/ophss/chiic/projects/2016/ideas/community-based-verbal-autopsy-linked-to-kenyas-civil-registration-and-vital-statistics-system.html ( accessed 7 Mar 2017 ).
Rampatige R ,
Hernandez B , et al
Mudenda SS ,
Kamocha S ,
Mswia R , et al
63. ↵ Department of Science and Technology . Research Infrastructure Roadmap . 1st edn . Pretoria : Department of Science and Technology , 2016 . https://www.gov.za/sites/default/files/sa research infrastructure road mapa.pdf . ( accessed 7 Sep 2017 ).
Moshabela M ,
Nanne I , et al

Handling editor Stephanie M Topp

Contributors LMT conducted the initial literature review, analysis and drafted the manuscript, with significant input from LD and DB. LD and DB conducted the double data extraction. All authors have given final approval of the version to be published.

Funding The research presented in this paper is funded by the Health Systems Research Initiative from the Department for International Development (DFID)/Medical Research Council (MRC)/Wellcome Trust/Economic and Social Research Council (ESRC) (MR/P014844/1).

Competing interests None declared.

Patient consent Not required.

Provenance and peer review Not commissioned; externally peer reviewed.

Data sharing statement No additional data are available.

Read the full text or download the PDF:

Log in using your username and password

Search More Search for this keyword Advanced search
Latest content
Current issue
BMJ Journals More You are viewing from: Google Indexer

http://orcid.org/0000-0002-8498-7329 Manjekah Dunn 1 , 2 ,
Iva Strnadová 3 , 4 , 5 ,
Jackie Leach Scully 4 ,
Jennifer Hansen 3 ,
Julie Loblinzk 3 , 5 ,
Skie Sarfaraz 5 ,
Chloe Molnar 1 ,
Elizabeth Emma Palmer 1 , 2
1 Faculty of Medicine & Health , University of New South Wales , Sydney , New South Wales , Australia
2 The Sydney Children's Hospitals Network , Sydney , New South Wales , Australia
3 School of Education , University of New South Wales , Sydney , New South Wales , Australia
4 Disability Innovation Institute , University of New South Wales , Sydney , New South Wales , Australia
5 Self Advocacy Sydney , Sydney , New South Wales , Australia
Correspondence to Dr Manjekah Dunn, Paediatrics & Child Health, University of New South Wales Medicine & Health, Sydney, New South Wales, Australia; manjekah.dunn{at}unsw.edu.au

Objective To identify factors acting as barriers or enablers to the process of healthcare consent for people with intellectual disability and to understand how to make this process equitable and accessible.

Data sources Databases: Embase, MEDLINE, PsychINFO, PubMed, SCOPUS, Web of Science and CINAHL. Additional articles were obtained from an ancestral search and hand-searching three journals.

Eligibility criteria Peer-reviewed original research about the consent process for healthcare interventions, published after 1990, involving adult participants with intellectual disability.

Synthesis of results Inductive thematic analysis was used to identify factors affecting informed consent. The findings were reviewed by co-researchers with intellectual disability to ensure they reflected lived experiences, and an easy read summary was created.

Results Twenty-three studies were included (1999 to 2020), with a mix of qualitative (n=14), quantitative (n=6) and mixed-methods (n=3) studies. Participant numbers ranged from 9 to 604 people (median 21) and included people with intellectual disability, health professionals, carers and support people, and others working with people with intellectual disability. Six themes were identified: (1) health professionals’ attitudes and lack of education, (2) inadequate accessible health information, (3) involvement of support people, (4) systemic constraints, (5) person-centred informed consent and (6) effective communication between health professionals and patients. Themes were barriers (themes 1, 2 and 4), enablers (themes 5 and 6) or both (theme 3).

Conclusions Multiple reasons contribute to poor consent practices for people with intellectual disability in current health systems. Recommendations include addressing health professionals’ attitudes and lack of education in informed consent with clinician training, the co-production of accessible information resources and further inclusive research into informed consent for people with intellectual disability.

PROSPERO registration CRD42021290548.

Decision making
Healthcare quality improvement
Patient-centred care
Quality improvement
Standards of care

Data availability statement

Data are available upon reasonable request. Additional data and materials such as data collection forms, data extraction and analysis templates and QualSyst assessment data can be obtained by contacting the corresponding author.

https://doi.org/10.1136/bmjqs-2023-016113

Statistics from Altmetric.com

Request permissions.

What is already known on this topic

People with intellectual disability are frequently excluded from decision-making processes and not provided equal opportunity for informed consent, despite protections outlined in the United Nations Convention on the Rights of Persons with Disabilities.

People with intellectual disability have the capacity and desire to make informed medical decisions, which can improve their well-being, health satisfaction and health outcomes.

What this review study adds

Health professionals lack adequate training in valid informed consent and making reasonable adjustments for people with intellectual disability, and continue to perpetuate assumptions of incapacity.

Health information provided to people with intellectual disability is often inaccessible and insufficient for them to make informed decisions about healthcare.

The role of support people, systemic constraints, a person-centred approach and ineffective healthcare communication also affect informed consent.

How this review might affect research, practice or policy

Health professionals need additional training on how to provide a valid informed consent process for people with intellectual disability, specifically in using accessible health information, making reasonable adjustments (e.g., longer/multiple appointments, options of a support person attending or not, using plain English), involving the individual in discussions, and communicating effectively with them.

Inclusive research is needed to hear the voices and opinions of people with intellectual disability about healthcare decision-making and about informed consent practices in specific healthcare settings.

Introduction

Approximately 1% of the world’s population have intellectual disability. 1 Intellectual disability is medically defined as a group of neurodevelopmental conditions beginning in childhood, with below average cognitive functioning and adaptive behaviour, including limitations in conceptual, social and practical skills. 2 People with intellectual disability prefer an alternative strength-based definition, reflected in the comment by Robert Strike OAM (Order of Australia Medal): ‘We can learn if the way of teaching matches how the person learns’, 3 reinforcing the importance of providing information tailored to the needs of a person with intellectual disability. A diagnosis of intellectual disability is associated with significant disparities in health outcomes. 4–7 Person-centred decision-making and better communication have been shown to improve patient satisfaction, 8 9 the physician–patient relationship 10 and overall health outcomes 11 for the wider population. Ensuring people with intellectual disability experience informed decision-making and accessible healthcare can help address the ongoing health disparities and facilitate equal access to healthcare.

Bodily autonomy is an individual’s power and agency to make decisions about their own body. 12 Informed consent for healthcare enables a person to practice bodily autonomy and is protected, for example, by the National Safety and Quality Health Service Standards (Australia), 13 Mental Capacity Act (UK) 14 and the Joint Commission Standards (USA). 15 In this article, we define informed consent according to three requirements: (1) the person is provided with information they understand, (2) the decision is free of coercion and (3) the person must have capacity. 16 For informed consent to be valid, this process must be suited to the individual’s needs so that they can understand and communicate effectively. Capacity is the ability to give informed consent for a medical intervention, 17 18 and the Mental Capacity Act outlines that ‘a person must be assumed to have capacity unless it is established that he lacks capacity’ and that incapacity can only be established if ‘all practicable steps’ to support capacity have been attempted without success. 14 These assumptions of capacity are also decision-specific, meaning an individual’s ability to consent can change depending on the situation, the choice itself and other factors. 17

Systemic issues with healthcare delivery systems have resulted in access barriers for people with intellectual disability, 19 despite the disability discrimination legislation in many countries who are signatories to the United Nations (UN) Convention on the Rights of Persons with Disabilities. 20 Patients with intellectual disability are not provided the reasonable adjustments that would enable them to give informed consent for medical procedures or interventions, 21 22 despite evidence that many people with intellectual disability have both the capacity and the desire to make their own healthcare decisions. 21 23

To support people with intellectual disability to make independent health decisions, an equitable and accessible informed consent process is needed. 24 However, current health systems have consistently failed to provide this. 21 25 To address this gap, we must first understand the factors that contribute to inequitable and inaccessible consent. To the best of our knowledge, the only current review of informed consent for people with intellectual disability is an integrative review by Goldsmith et al . 26 Many of the included articles focused on assessment of capacity 27–29 and research consent. 30–32 The review’s conclusion supported the functional approach to assess capacity, with minimal focus on how the informed consent processes can be improved. More recently, there has been a move towards ensuring that the consent process is accessible for all individuals, including elderly patients 33 and people with aphasia. 34 However, there remains a paucity of literature about the informed consent process for people with intellectual disability, with no systematic reviews summarising the factors influencing the healthcare consent process for people with intellectual disability.

To identify barriers to and enablers of the informed healthcare consent process for people with intellectual disability, and to understand how this can be made equitable and accessible.

A systematic literature review was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-Analysis Protocols (PRISMA-P) systematic literature review protocol. 35 The PRISMA 2020 checklist 36 and ENhancing Transparency in REporting the synthesis of Qualitative research (ENTREQ) reporting guidelines were also followed. 37 The full study protocol is included in online supplemental appendix 1 .

Supplemental material

No patients or members of the public were involved in this research for this manuscript.

Search strategy

A search strategy was developed to identify articles about intellectual disability, consent and healthcare interventions, described in online supplemental appendix 2 . Multiple databases were searched for articles published between January 1990 to January 2022 (Embase, MEDLINE, PsychINFO, PubMed, SCOPUS, Web of Science and CINAHL). These databases include healthcare and psychology databases that best capture relevant literature on this topic, including medical, nursing, social sciences and bioethical literature. The search was limited to studies published from 1990 as understandings of consent have changed since then. 38 39 This yielded 4853 unique papers which were imported into Covidence, a specialised programme for conducting systematic reviews. 40

Study selection

Citation screening by abstract and titles was completed by two independent researchers (MD and EEP). Included articles had to:

Examine the informed consent process for a healthcare intervention for people with intellectual disability.

Have collected more than 50% of its data from relevant stakeholders, including adults with intellectual disability, families or carers of a person with intellectual disability, and professionals who engage with people with intellectual disability.

Report empirical data from primary research methodology.

Be published in a peer-reviewed journal after January 1990.

Be available in English.

Full text screening was completed by two independent researchers (MD and EEP). Articles were excluded if consent was only briefly discussed or if it focused on consent for research, capacity assessment, or participant knowledge or comprehension. Any conflicts were resolved through discussion with an independent third researcher (IS).

Additional studies were identified through an ancestral search and by hand-searching three major journals relevant to intellectual disability research. Journals were selected if they had published more than one included article for this review or in previous literature reviews conducted by the research team.

Quality assessment

Two independent researchers (MD and IS) assessed study quality with the QualSyst tool, 41 which can assess both qualitative and quantitative research papers. After evaluating the distribution of scores, a threshold value of 55% was used, as suggested by QualSyst 41 to exclude poor-quality studies but capture enough studies overall. Any conflicts between the quality assessment scores were resolved by a third researcher (EEP). For mixed-method studies, both qualitative and quantitative quality scores were calculated, and the higher value used.

Data collection

Two independent researchers (MD and JH) reviewed each study and extracted relevant details, including study size, participant demographics, year, country of publication, study design, data analysis and major outcomes reported. Researchers used standardised data collection forms designed, with input from senior researchers with expertise in qualitative research (IS and EEP), to extract data relevant to the review’s research aims. The form was piloted on one study, and a second iteration made based on feedback. These forms captured data on study design, methods, participants, any factors affecting the process of informed consent and study limitations. Data included descriptions and paragraphs outlining key findings, the healthcare context, verbatim participant quotes and any quantitative analyses or statistics. Missing or unclear data were noted.

Data analysis

A pilot literature search showed significant heterogeneity in methodology of studies, limiting the applicability of traditional quantitative analysis (ie, meta-analysis). Instead, inductive thematic analysis was chosen as an alternative methodology 42 43 that has been used in recent systematic reviews examining barriers and enablers of other health processes. 44 45 The six-phase approach described by Braun and Clarke was used. 46 47 A researcher (MD) independently coded the extracted data of each study line-by-line, with subsequent data grouped into pre-existing codes or new concepts when necessary. Codes were reviewed iteratively and grouped into categories, subthemes and themes framed around the research question. Another independent researcher (JH) collated and analysed the data on study demographics, methods and limitations. The themes were reviewed by two senior researchers (EEP and IS).

Qualitative methods of effect size calculations have been described in the literature, 48 49 which was captured in this review by the number of studies that identified each subtheme, with an assigned frequency rating to compare their relative significance. Subthemes were given a frequency rating of A, B, C or D if they were identified by >10, 7–9, 4–6 or <3 articles, respectively. The overall significance of each theme was estimated by the number of studies that mentioned it and the GRADE framework, a stepwise approach to quality assessment using a four-tier rating system. Each study was evaluated for risk of bias, inconsistency, indirectness, imprecision and publication bias. 50 51 Study sensitivity was assessed by counting the number of distinct subthemes included. 52 The quality of findings was designated high, moderate or low depending on the frequency ratings, the QualSyst score and the GRADE scores of studies supporting the finding. Finally, the relative contributions of each study were evaluated by the number of subthemes described, guided by previously reported methods for qualitative reviews. 52

Co-research

The findings were reviewed by two co-researchers with intellectual disability (JL and SS), with over 30 years combined experience as members and employees of a self-advocacy organisation. Guidance on the findings and an easy read summary was produced in line with best-practice inclusive research 53 54 over multiple discussions. Input from two health professional researchers (MD and EEP) provided data triangulation and sense-checking of findings.

Twenty-three articles were identified ( figure 1 ): 14 qualitative, 6 quantitative and 3 mixed-methods. Two papers included the same population of study participants: McCarthy 55 and McCarthy, 56 but had different research questions. Fovargue et al 57 was excluded due to a quality score of 35%. Common quality limitations were a lack of verification procedures to establish credibility and limited researcher reflexivity. No studies were excluded due to language requirements (as all were in English) or age restrictions (all studies had majority adult participants).

Download figure
Open in new tab
Download powerpoint

PRISMA 2020 flowchart for the systematic review. 36

Studies were published from 1999 to 2020 and involved participant populations from the UK (n=18), USA (n=3), Sweden (n=1) and Ireland (n=1). Participant numbers ranged from 9 to 604 (median 21), and participants included people with intellectual disability (n=817), health professionals (n=272), carers and support people (n=48), and other professionals that work with people with intellectual disability (n=137, community service agency directors, social workers, administrative staff and care home staff). Ages of participants ranged from 8 to 84 years, though only Aman et al 58 included participants <18 years of age. This study was included as the article states very few children were included. Studies examined consent in different contexts, including contraception and sexual health (6/23 articles), 58–60 medications (5/23 articles), 58–62 emergency healthcare, 63 cervical screening, 64 community referrals, 58–61 65 mental health, 66 hydrotherapy, 64 blood collection 67 and broad decision-making consent without a specific context. 65 68–71 A detailed breakdown of each study is included in online supplemental appendix 3 .

Six major themes were identified from the studies, summarised in figure 2 . An overview of included studies showing study sensitivity, effect size, QualSyst and GRADE scores is given in online supplemental appendix 4 . Studies with higher QualSyst and GRADE scores contributed more to this review’s findings and tended to include more subthemes; specifically, Rogers et al , 66 Sowney and Barr, 63 Höglund and Larsson, 72 and McCarthy 55 and McCarthy. 56 Figure 3 gives the easy read version of theme 1, with the full easy read summary in online supplemental appendix 5 .

Summary of the identified six themes and subthemes.

Theme 1 of the easy read summary.

Theme 1—Health professionals’ attitudes and lack of education about informed consent

Health professionals’ attitudes and practices were frequently (18/21) identified as factors affecting the informed consent process, with substantial evidence supporting this theme. Studies noted the lack of training for health professionals in supporting informed consent for people with intellectual disability, their desire for further education, and stereotypes and discrimination perpetuated by health professionals.

Lack of health professional education on informed consent and disability discrimination legislation

Multiple studies reported inconsistent informed consent practices, for various reasons: some reported that health professionals ‘forgot’ to or ‘did not realise consent was necessary’, 63 73 but inconsistent consent practices were also attributed to healthcare providers’ unfamiliarity with consent guidelines and poor education on this topic. Carlson et al 73 reported that only 44% of general practitioners (GPs) were aware of consent guidelines, and there was the misconception that consent was unnecessary for people with intellectual disability. Similarly, studies of psychologists 66 and nurses 63 found that many were unfamiliar with their obligations to obtain consent, despite the existence of anti-discrimination legislation. People with intellectual disability describe feeling discriminated against by health professionals, reflected in comments such as ‘I can tell, my doctor just thinks I’m stupid – I'm nothing to him’. 74 Poor consent practices by health professionals were observed in Goldsmith et al , 67 while health professionals surveyed by McCarthy 56 were unaware of their responsibility to provide accessible health information to women with intellectual disability. Improving health professional education and training was suggested by multiple studies as a way to remove this barrier. 63 65–67 69 73

Lack of training on best practices for health professions caring for people with intellectual disability

A lack of training in caring for and communicating with people with intellectual disability was also described by midwives, 72 psychologists, 66 nurses, 63 pharmacists 61 and GPs. 56 72 75 Health professionals lacked knowledge about best practice approaches to providing equitable healthcare consent processes through reasonable adjustments such as accessible health information, 56 60 66 longer appointments times, 60 72 simple English 62 67 and flexible approaches to patient needs. 63 72

Health professionals’ stereotyping and assumptions of incapacity

Underlying stereotypes contributed to some health professionals’ (including nurses, 63 GPs 56 and physiotherapists 64 ) belief that people with intellectual disability lack capacity and therefore, do not require opportunities for informed consent. 56 64 In a survey of professionals referring people with intellectual disability to a disability service, the second most common reason for not obtaining consent was ‘patient unable to understand’. 73

Proxy consent as an inappropriate alternative

People with intellectual disability are rarely the final decision-maker in their medical choices, with many health providers seeking proxy consent from carers, support workers and family members, despite its legal invalidity. In McCarthy’s study (2010), 18/23 women with intellectual disability said the decision to start contraception was made by someone else. Many GPs appeared unaware that proxy consent is invalid in the UK. 56 Similar reports came from people with intellectual disability, 55 56 60 64 69 76 health professionals (nurses, doctors, allied health, psychologists), 56 63 64 66 77 support people 64 77 and non-medical professionals, 65 73 and capacity was rarely documented. 56 62 77

Exclusion of people with intellectual disability from decision-making discussions

Studies described instances where health professionals made decisions for their patients with intellectual disability or coerced patients into a choice. 55 72 74 76 77 In Ledger et al 77 , only 62% of women with intellectual disability were involved in the discussion about contraception, and only 38% made the final decision, and others stated in Wiseman and Ferrie 74 : ‘I was not given the opportunity to explore the different options. I was told what one I should take’. Three papers outlined instances where the choices of people with intellectual disability were ignored despite possessing capacity 65 66 69 and when a procedure continued despite them withdrawing consent. 69

Theme 2—Inadequate accessible health information

Lack of accessible health information.

The lack of accessible health information was the most frequently identified subtheme (16/23 studies). Some studies reported that health professionals provided information to carers instead, 60 avoided providing easy read information due to concerns about ‘offending’ patients 75 or only provided verbal information. 56 67 Informed consent was supported when health professionals recognised the importance of providing medical information 64 and when it was provided in an accessible format. 60 Alternative approaches to health information were explored, including virtual reality 68 and in-person education sessions, 59 with varying results. Overall, the need to provide information in different formats tailored to an individual’s communication needs, rather than a ‘one size fits all’ approach, was emphasised by both people with intellectual disability 60 and health professionals. 66

Insufficient information provided

Studies described situations where insufficient information was provided to people with intellectual disability to make informed decisions. For example, some people felt the information from their GP was often too basic to be helpful (Fish et al 60 ) and wanted additional information on consent forms (Rose et al 78 ).

Theme 3—The involvement of support people

Support people (including carers, family members and group home staff) were identified in 11 articles as both enablers of and barriers to informed consent. The antagonistic nature of these findings and lower frequency of subthemes are reflected in the lower quality assessments of evidence.

Support people facilitated communication with health professionals

Some studies reported carers bridging communication barriers with health to support informed consent. 63 64 McCarthy 56 found 21/23 of women with intellectual disability preferred to see doctors with a support person due to perceived benefits: ‘Sometimes I don’t understand it, so they have to explain it to my carer, so they can explain it to me easier’. Most GPs in this study (93%) also agreed that support people aided communication.

Support people helped people with intellectual disability make decisions

By advocating for people with intellectual disability, carers encouraged decision-making, 64 74 provided health information, 74 77 emotional support 76 and assisted with reading or remembering health information. 55 58 76 Some people with intellectual disability explicitly appreciated their support person’s involvement, 60 such as in McCarthy’s 55 study where 18/23 participants felt supported and safer when a support person was involved.

Support people impeded individual autonomy

The study by Wiseman and Ferrie 74 found that while younger participants with intellectual disability felt family members empowered their decision-making, older women felt family members impaired their ability to give informed consent. This was reflected in interviews with carers who questioned the capacity of the person with intellectual disability they supported and stated they would guide them to pick the ‘best choice’ or even over-ride their choices. 64 Studies of psychologists and community service directors described instances where the decision of family or carers was prioritised over the wishes of the person with intellectual disability. 65 66 Some women with intellectual disability in McCarthy’s studies (2010, 2009) 55 56 appeared to have been coerced into using contraception by parental pressures or fear of losing group home support.

Theme 4—Systemic constraints within healthcare systems

Time restraints affect informed consent and accessible healthcare.

Resource limitations create time constraints that impair the consent process and have been identified as a barrier by psychologists, 66 GPs, 56 hospital nurses 63 and community disability workers. 73 Rogers et al 66 highlighted that a personalised approach that could improve informed decision-making is restricted by inflexible medical models. Only two studies described flexible patient-centred approaches to consent. 60 72 A survey of primary care practices in 2007 reported that most did not modify their cervical screening information for patients with intellectual disability because it was not practical. 75

Inflexible models of consent

Both people with intellectual disability 76 and health professionals 66 recognised that consent is traditionally obtained through one-off interactions prior to an intervention. Yet, for people with intellectual disability, consent should ideally be an ongoing process that begins before an appointment and continues between subsequent ones. Other studies have tended to describe one-off interactions where decision-making was not revisited at subsequent appointments. 56 60 72 76

Lack of systemic supports

In one survey, self-advocates highlighted a lack of information on medication for people with intellectual disability and suggested a telephone helpline and a centralised source of information to support consent. 60 Health professionals also want greater systemic support, such as a health professional specialised in intellectual disability care to support other staff, 72 or a pharmacist specifically to help patients with intellectual disability. 61 Studies highlighted a lack of guidelines about healthcare needs of people with intellectual disabilities such as contraceptive counselling 72 or primary care. 75

Theme 5—Person-centred informed consent

Ten studies identified factors related to a person-centred approach to informed consent, grouped below into three subthemes. Health professionals should tailor their practice when obtaining informed consent from people with intellectual disability by considering how these subthemes relate to the individual. Each subtheme was described five times in the literature with a relative frequency rating of ‘C’, contributing to overall lower quality scores.

Previous experience with decision-making

Arscott et al 71 found that the ability of people with intellectual disability to consent changed with their verbal and memory skills and in different clinical vignettes, supporting the view of ‘functional’ capacity specific to the context of the medical decision. Although previous experiences with decision-making did not influence informed consent in this paper, other studies suggest that people with intellectual disability accustomed to independent decision-making were more able to make informed medical decisions, 66 70 and those who live independently were more likely to make independent healthcare decisions. 56 Health professionals should be aware that their patients with intellectual disability will have variable experience with decision-making and provide individualised support to meet their needs.

Variable awareness about healthcare rights

Consent processes should be tailored to the health literacy of patients, including emphasising available choices and the option to refuse treatment. In some studies, medical decisions were not presented to people with intellectual disability as a choice, 64 and people with intellectual disability were not informed of their legal right to accessible health information. 56

Power differences and acquiescence

Acquiescence by people with intellectual disability due to common and repeated experiences of trauma—that is, their tendency to agree with suggestions made by carers and health professionals, often to avoid upsetting others—was identified as an ongoing barrier. In McCarthy’s (2009) interviews with women with intellectual disability, some participants implicitly rejected the idea that they might make their own healthcare decisions: ‘They’re the carers, they have responsibility for me’. Others appeared to have made decisions to appease their carers: ‘I have the jab (contraceptive injection) so I can’t be blamed for getting pregnant’. 55 Two studies highlighted that health professionals need to be mindful of power imbalances when discussing consent with people with intellectual disability to ensure the choices are truly autonomous. 61 66

Theme 6—Effective communication between health professionals and patients

Implementation of reasonable adjustments for verbal and written information.

Simple language was always preferred by people with intellectual disability. 60 67 Other communication aids used in decision-making included repetition, short sentences, models, pictures and easy read brochures. 72 Another reasonable adjustment is providing the opportunity to ask questions, which women with intellectual disability in McCarthy’s (2009) study reported did not occur. 55

Tailored communication methods including non-verbal communication

Midwives noted that continuity of care allows them to develop rapport and understand the communication preferences of people with intellectual disability. 72 This is not always possible; for emergency nurses, the lack of background information about patients with intellectual disability made it challenging to understand their communication preferences. 63 The use of non-verbal communication, such as body language, was noted as underutilised 62 66 and people with intellectual disability supported the use of hearing loops, braille and sign language. 60

To the best of our knowledge, this is the first systematic review investigating the barriers and enablers of the informed consent process for healthcare procedures for people with intellectual disability. The integrative review by Goldsmith et al 26 examined capacity assessment and shares only three articles with this systematic review. 69 71 73 Since the 2000s, there has been a paradigm shift in which capacity is no longer considered a fixed ability that only some individuals possess 38 39 but instead as ‘functional’: a flexible ability that changes over time and in different contexts, 79 reflected in Goldsmith’s review. An individual’s capacity can be supported through various measures, including how information is communicated and how the decision-making process is approached. 18 80 By recognising the barriers and enablers identified in this review, physicians can help ensure the consent process for their patients with intellectual disability is both valid and truly informed. This review has highlighted the problems of inaccessible health information, insufficient clinical education on how to make reasonable adjustments and lack of person-centred trauma-informed care.

Recommendations

Health professionals require training in the informed consent process for people with intellectual disability, particularly in effective and respectful communication, reasonable adjustments and trauma-informed care. Reasonable adjustments include offering longer or multiple appointments, using accessible resources (such as easy read information or shared decision-making tools) and allowing patient choices (such as to record a consultation or involve a support person). Co-researchers reported that many people with intellectual disability prefer to go without a support person because they find it difficult to challenge their decisions and feel ignored if the health professional only talks to the support person. People with intellectual disability also feel they cannot seek second opinions before making medical decisions or feel pressured to provide consent, raising the possibility of coercion. These experiences contribute to healthcare trauma. Co-researchers raised the importance of building rapport with the person with intellectual disability and of making reasonable adjustments, such as actively advocating for the person’s autonomy, clearly stating all options including the choice to refuse treatment, providing opportunities to contribute to discussions and multiple appointments to ask questions and understand information. They felt that without these efforts to support consent, health professionals can reinforce traumatic healthcare experiences for people with intellectual disability. Co-researchers noted instances where choices were made by doctors without discussion and where they were only given a choice after requesting one and expressed concern that these barriers are greater for those with higher support needs.

Co-researchers showed how these experiences contributed to mistrust of health professionals and poorer health outcomes. In one situation, a co-researcher was not informed of a medication’s withdrawal effects, resulting in significant side-effects when it was ceased. Many people with intellectual disability describe a poor relationship with their health professionals, finding it difficult to trust health information provided due to previous traumatic experiences of disrespect, coercion, lack of choice and inadequate support. Many feel they cannot speak up due to the power imbalance and fear of retaliation. Poor consent practices and lack of reasonable adjustments directly harm therapeutic alliances by reducing trust, contribute to healthcare trauma and lead to poorer health outcomes for people with intellectual disability.

Additional education and training for health professionals is urgently needed in the areas of informed consent, reasonable adjustments and effective communication with people with intellectual disability. The experiences of health professionals within the research team confirmed that there is limited training in providing high-quality healthcare for people with intellectual disability, including reasonable adjustments and accessible health information. Co-researchers also suggested that education should be provided to carers and support people to help them better advocate for people with intellectual disability.

Health information should be provided in a multimodal format, including written easy read information. Many countries have regulation protecting the right to accessible health information and communication support to make an informed choice, such as UK’s Accessible Information Standard, 81 and Australia’s Charter of Health Care Rights, 24 yet these are rarely observed. Steps to facilitate this include routinely asking patients about information requirements, system alerts for an individual’s needs or routinely providing reasonable adjustments. 82 Co-researchers agreed that there is a lack of accessible health information, particularly about medications, and that diagrams and illustrations are underutilised. There is a critical need for more inclusive and accessible resources to help health professionals support informed consent in a safe and high-quality health system. These resources should be created through methods of inclusive research, such as co-production, actively involving people with intellectual disability in the planning, creation, and feedback process. 53

Strengths and limitations

This systematic review involved two co-researchers with intellectual disability in sense-checking findings and co-creating the easy read summary. Two co-authors who are health professionals provided additional sense-checking of findings from a different stakeholder perspective. In future research, this could be extended by involving people with intellectual disability in the design and planning of the study as per recommendations for best-practice inclusive research. 53 83

The current literature is limited by low use of inclusive research practices in research involving people with intellectual disability, increasing vulnerability to external biases (eg, inaccessible questionnaires, involvement of carers in data collection, overcompliance or acquiescence and absence of researcher reflexivity). Advisory groups or co-research with people with intellectual disability were only used in five studies. 58 60 68 74 76 Other limitations include unclear selection criteria, low sample sizes, missing data, using gatekeepers in patient selection and predominance of UK-based studies—increasing the risk of bias and reducing transferability. Nine studies (out of 15 involving people with intellectual disability) explicitly excluded those with severe or profound intellectual disability, reflecting a selection bias; only one study specifically focused on people with intellectual disability with higher support needs. Studies were limited to a few healthcare contexts, with a focus on consent about sexual health, contraception and medications.

The heterogeneity and qualitative nature of studies made it challenging to apply traditional meta-analysis. However, to promote consistency in qualitative research, the PRISMA and ENTREQ guidelines were followed. 36 37 Although no meta-analyses occurred, the duplication of study populations in McCarthy 2009 and 2010 likely contributed to increased significance of findings reported in both studies. Most included studies (13/23) were published over 10 years ago, reducing the current relevance of this review’s findings. Nonetheless, the major findings reflect underlying systemic issues within the health system, which are unlikely to have been resolved since the articles were published, as the just-released final report of the Australian Royal Commission into Violence, Abuse, Neglect and Exploitation of People with Disability highlights. 84 There is an urgent need for more inclusive studies to explore the recommendations and preferences of people with intellectual disability about healthcare choices.

Informed consent processes for people with intellectual disability should include accessible information and reasonable adjustments, be tailored to individuals’ needs and comply with consent and disability legislation. Resources, guidelines and healthcare education are needed and should cover how to involve carers and support people, address systemic healthcare problems, promote a person-centred approach and ensure effective communication. These resources and future research must use principles of inclusive co-production—involving people with intellectual disability at all stages. Additionally, research is needed on people with higher support needs and in specific contexts where informed consent is vital but under-researched, such as cancer screening, palliative care, prenatal and newborn screening, surgical procedures, genetic medicine and advanced therapeutics such as gene-based therapies.

Ethics statements

Patient consent for publication.

Not applicable.

Ethics approval

Maulik PK ,
Mascarenhas MN ,
Mathers CD , et al
World Health Organisation
Council for Intellectual Disability
Emerson E ,
Shogren KA ,
Wehmeyer ML ,
Reese RM , et al
Cordasco KM
Hallock JL ,
Jordens CFC ,
McGrath C , et al
Brenner LH ,
Brenner AT ,
United Nations Population Fund
Australian Commission on Safety and Quality in Health Care
The Joint Commission
Beauchamp TL ,
Childress JF
New South Wales Attorney General
United Nations General Assembly
Strnadová I ,
Loblinzk J ,
Scully JL , et al
MacPhail C ,
McKay K , et al
Keywood K ,
Fovargue S ,
Goldsmith L ,
Skirton H ,
Cash J , et al
Morris CD ,
Niederbuhl JM ,
Arscott K ,
Fisher CB ,
Davidson PW , et al
Giampieri M
Shamseer L ,
Clarke M , et al
McKenzie JE ,
Bossuyt PM , et al
Flemming K ,
McInnes E , et al
Appelbaum PS
↵ Covidence systematic review software . Melbourne, Australia ,
Proudfoot K
Papadopoulos I ,
Koulouglioti C ,
Lazzarino R , et al
Onwuegbuzie AJ
BMJ Best Practice
Guyatt GH ,
Vist GE , et al
Garcia-Lee B
Brimblecombe J , et al
Benson BA ,
Farmer CA , et al
Ferguson L ,
Graham YNH ,
Gerrard D ,
Laight S , et al
Huneke NTM ,
Halder N , et al
Ferguson M ,
Jarrett D ,
McGuire BE , et al
Woodward V ,
Jackson L , et al
Conboy-Hill S ,
Leafman J ,
Nehrenz GM , et al
Höglund B ,
Carlson T ,
English S , et al
Wiseman P ,
Walmsley J ,
Tilley E , et al
Khatkar HS , et al
Holland AJ , et al
Beauchamp TL
England National Health Service
National Health Service England
Royal Commission into Violence, Abuse, Neglect and Exploitation of People with Disability

Supplementary materials

Supplementary data.

This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

Data supplement 1
Data supplement 2
Data supplement 3
Data supplement 4
Data supplement 5

Contributors MD, EEP and IS conceived the idea for the systematic review. MD drafted the search strategy which was refined by EEP and IS. MD and EEP completed article screening. MD and IS completed quality assessments of included articles. MD and JH completed data extraction. MD drafted the original manuscript. JL and SS were co-researchers who sense-checked findings and were consulted to formulate dissemination plans. JL and SS co-produced the easy read summary with MD, CM, JH, EEP and IS. MD, JLS, EEP and IS reviewed manuscript wording. All authors critically reviewed the manuscript and approved it for publication. The corresponding author attests that all listed authors meet authorship criteria and that no others meeting the criteria have been omitted. MD is the guarantor responsible for the overall content of this manuscript.

Funding This systematic literature review was funded by the National Health & Medical Research Council (NHMRC), Targeted Call for Research (TCR) into Improving health of people with intellectual disability. Research grant title "GeneEQUAL: equitable and accessible genomic healthcare for people with intellectual disability". NHMRC application ID: 2022/GNT2015753.

Competing interests None declared.

Provenance and peer review Not commissioned; externally peer reviewed.

Linked Articles

Editorial It is up to healthcare professionals to talk to us in a way that we can understand: informed consent processes in people with an intellectual disability Jonathon Ding Richard Keagan-Bull Irene Tuffrey-Wijne BMJ Quality & Safety 2024; 33 277-279 Published Online First: 30 Jan 2024. doi: 10.1136/bmjqs-2023-016830

Read the full text or download the PDF:

Log in using your username and password

Search More Search for this keyword Advanced search
Latest Content
BMJ Journals More You are viewing from: Google Indexer

http://orcid.org/0000-0002-1463-9206 Andi Orlowski 1 , 2 ,
http://orcid.org/0000-0001-7044-4780 Jack Ettinger 1 ,
Alex Bottle 2 ,
Sally Snow 1 ,
http://orcid.org/0000-0003-1755-7975 Rachel Ashton 1 and
Jennifer K Quint 2
1 Health Economics Unit , London , UK
2 Department of Primary Care and Public Health , Imperial College London , London , UK
Correspondence to Andi Orlowski; andi.orlowski{at}nhs.net

Chronic obstructive pulmonary disease (COPD) is a multisystem disease, and many patients have multiple conditions. We explored multimorbidity patterns that might inform intervention planning to reduce health-care costs while preserving quality of life for patients. Literature searches up to February 2022 revealed 4419 clinical observational and comparative studies of risk factors for multimorbidity in people with COPD, pulmonary emphysema, or chronic bronchitis at baseline. Of these, 29 met the inclusion criteria for this review. Eight studies were cluster and network analyses, five were regression analyses, and 17 (in 16 papers) were other studies of specific conditions, physical activity and treatment. People with COPD more frequently had multimorbidity and had up to ten times the number of disorders of those without COPD. Disease combinations prominently featured cardiovascular and metabolic diseases, asthma, musculoskeletal and psychiatric disorders. An important risk factor for multimorbidity was low socioeconomic status. One study showed that many patients were receiving multiple drugs and had increased risk of adverse events, and that 10% of medications prescribed were inappropriate. Many patients with COPD have mainly preventable or modifiable multimorbidity. A proactive multidisciplinary approach to prevention and management could reduce the burden of care.

COPD epidemiology

https://doi.org/10.1136/bmjresp-2023-002272

Statistics from Altmetric.com

Request permissions.

Introduction

Multimorbidity is the coexistence of at least two chronic diseases where one is not more central to a person’s health than another. 1 By contrast, the term comorbidity is used to indicate an illness that is seen as secondary to an index disorder. It is generally in research but also in secondary and tertiary care systems that have traditionally been structured around diseases and/or body systems. 2 These concepts are important in healthcare, as they may affect whether management of long-term conditions is more or less holistic and how well it reflects reality for patients.

Chronic obstructive pulmonary disease (COPD) is a multisystem disease characterised by pulmonary and systemic inflammation. 3 Many people with COPD have more than one long-term chronic condition, particularly diseases that have strong associations with smoking, ageing and anxiety and/or depression. 4 Furthermore, age-related disorders frequently occur earlier in life in people with COPD than in those without. 5 In a large UK study, among 51 928 patients aged 25 years or older (mean age 65 years) with COPD, 86% had multimorbidity compared with 51% of 1 220 757 people without COPD, and 22% vs 5% had five or more conditions. 6 Frequently occurring conditions in patients with COPD are cardiovascular, musculoskeletal, psychiatric and metabolic disorders, gastro-oesophageal reflux disease, chronic kidney disease and cancer. 4 7

Multimorbidity in people with COPD increases the risk of exacerbation and decreases quality of life and exercise tolerance. 8 Together, these effects increase the use of healthcare resources 9 and the likelihood of hospital admission 10 11 and mortality. 12 Iheanacho et al 13 performed a systematic literature review of studies published worldwide and found that the average number of primary care visits per person per year ranged from 2.3 to 13.0 for mild-to-moderate COPD and from 2.8 to 15.1 for severe COPD. The annual number of COPD hospital admissions per year for patients with moderate and severe disease was between 0 and 0.57 in patients with moderate disease and between 0 and 0.88 for patients with very severe disease. Direct costs rose with worsening disease severity, increasing by 1.3 times for annual per-person costs from mild-to-severe disease in two UK studies, and more than doubling in a study from Italy. The review also found that while the main drivers for hospital admissions in patients with COPD were increasing disease severity, restricted lung function and higher baseline concentrations of C-reactive protein, those for longer hospital stays were many and varied and included multiple factors associated with other conditions.

A further important issue associated with multimorbidity is polypharmacy, which can reduce quality of life, increase burden on patients and raise risks of adverse drug events, morbidity and mortality. Potential fragmentation and reduced coordination of care can increase healthcare use, including unplanned and emergency care. 14 Schnell et al 15 found that more than half (52%) of 995 adults aged 45 years or older with COPD in their study were receiving more than four medications compared with 32% of 14 828 without COPD. Hanlon et al 16 found that 52% of 8317 patients with COPD reported polypharmacy with five or more medications compared with 18% of 494 323 of those without COPD.

Predicting which people with COPD will go on to develop or experience extension of multimorbidity and planning how to intervene to address modifiable risk factors could help to reduce costs and preserve quality of life for patients. This systematic literature review was performed as part of the groundwork for development of an impactibility model 17 and had the objective of exploring the breadth of relevant information in the literature. In this paper, we discuss which modifiable risk factors are most reported in people with COPD and might be possible to address in routine care to prevent progression to or extension of multimorbidity in people with an existing COPD diagnosis. We also discuss potential management strategies.

Literature search and selection of reports

For the purposes of this review, multimorbidity was defined as progression from COPD to one or more of the long-term conditions used by Barnett et al 18 that could be treated, prevented or changed by routine care ( table 1 ).

View inline

Disorders of interest included in exploratory analysis of multimorbid conditions

A systematic literature review was carried out to identify all papers published from the start of each database's collection up to February 2022. The Ovid search platform was used to search Embase Classic and Embase, Ovid MEDLINE and the Healthcare Management Information Consortium (HMIC). Web of Science was also searched ( table 2 ).

Databases searched with dates

Search strategies were built iteratively, with relevant keywords and subject headings for each database being added after review of retrieved publications. The final set of search terms (see online supplemental appendix ) included synonyms of multimorbidity, including “polymorbidity” and “polypathology”, and terms relating to COPD, including “chronic bronchitis”, “emphysema” and “obstructive airway disease”. Terms associated with the Charlson Comorbidity Index (CCI) were included in the final set of search terms for Embase, Ovid MEDLINE and HMIC (see online supplemental appendix ).

Supplemental material

An exploratory search for papers discussing progression from COPD to the specific chronic conditions of interest was conducted. This search was performed using Embase Classic and Embase for all papers published prior to 6 August 2021 ( table 2 ). The Boolean operator ‘AND’ was used to link words relating to specific chronic disease diagnoses with "COPD". Conditions relating to infectious diseases were excluded.

Database search results were exported to Covidence. Two reviewers (JE and SS) independently screened titles and abstracts for relevance and reviewed the full texts of reports that included a population diagnosed with COPD and mentioned risk factors for progression to multimorbidity.

Papers that reported observational or comparative studies of risk factors for multimorbidity in people with COPD, pulmonary emphysema or chronic bronchitis were included. We excluded reports focusing on diagnosis of COPD, COPD exacerbation, hospital admission for COPD or COPD-related mortality. Papers were also excluded if participants did not have a diagnosis of COPD at baseline, the report did not discuss risk factors for progression to multimorbidity (even if they reported the prevalence of multimorbidity) or were preclinical, pathophysiology, genetic and/or biomarker studies. Additional exclusions were made for non-English-language publications, editorials, opinion pieces, case reports, narrative reviews and predictive models. To gain the widest possible evidence base in this field, studies were assessed based on methodological quality (completeness of outcome data, selective reporting and other sources of bias). However, because the aim of this study was to explore what is discussed in the literature and not make clinical recommendations, no formal quality assessment process was applied and no papers were excluded on this basis. For reports that were not unanimously included or excluded by the two reviewers, they were discussed with a third reviewer (AO) until a decision was reached.

Data extraction

Data extraction was performed by SS and JE. Risk factors for progression to multimorbidity in patients with COPD were extracted into data tables. Results were considered by type of study performed and/or by disease being discussed. Thus, some studies might be discussed more than once in the results section depending on the data they contained.

Statistical analysis

Due to the range of study designs, methodologies and participant samples, it was judged that meta-analysis would not be statistically meaningful. Therefore, we provide a descriptive analysis of progression to or extension of multimorbidity in patients with COPD.

Selection of studies

Of 4418 studies (in 4419 papers) initially identified, 1335 were deemed unsuitable for review at initial screening and 3053 were excluded at later assessment stages ( figure 1 ). Therefore, 29 studies were deemed to be relevant to this report.

Download figure
Open in new tab
Download powerpoint

Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow diagram. COPD, chronic obstructive pulmonary disease.

Multimorbidity prevalence and characteristics

Shen et al 19 performed a cluster analysis to investigate disease combinations in 91 453 US patients with COPD. They reviewed electronic health records and identified four distinct profiles: low morbidity (based on CCI; 61%; mean CCI score 1.9 SD±1.4), metabolic–renal (21%; 4.7±1.8), cardiovascular (12%; 4.6±1.9) and multimorbidity (7%; 7.5±1.7).

A cluster analysis study in Lithuania assessed records of 321 297 patients for 32 chronic diseases. 20 In this cohort, 4834 had COPD, which was associated with significantly increased prevalence of cardiovascular diseases, arrhythmia, heart failure, kidney diseases and lung cancer (all p<0.0001). They further assessed the 19 conditions seen in at least 5% of patients (n=3338) and revealed six disease clusters in men with COPD: cardiovascular diseases (99% had at least one), asthma–musculoskeletal (49%), stroke–cancer–sensory (44%), endocrine–metabolic (42%), gout–renal (12%) and mental disorders (8%). Five clusters were identified in women with COPD. Again, the most common contained cardiovascular diseases (100%), followed by glaucoma–mental disorders–osteoarthritis–back pain–asthma–obesity–dyslipidaemia–diabetes (87%), cancer–osteoporosis–hypothyroidism–hearing loss–cancer (38%), dementia–stroke (14%) and anaemia (5%).

Using network visualisation software, Divo et al 21 compared comorbidity networks for 79 disorders in 1969 patients with COPD and 316 individuals without COPD. From the 79 disease nodes, the COPD network showed 428 links with significance of p≤0.01. By comparison, the non-COPD network contained only 56 of the disease nodes and had 149 links. Thus, the prevalence, diversity and degrees of association of comorbidities seem much greater in people with COPD. Nevertheless, the sample size for people without COPD was much smaller. The authors further identified four distinct clusters of anthropometric and clinical characteristics in which nodes were highly interlinked with COPD. One cluster that included 50% of patients had a cardiovascular disease ‘theme’. Although cardiovascular disease clustering was also seen in the non-COPD group, the prevalence was 30% and the number of links was considerably fewer. Another node with 50% prevalence in the COPD group centred around individuals with less obstruction, higher body mass index (BMI) and comorbidities mainly associated with metabolic syndrome. Metabolic syndrome components were present in the non-COPD controls but they were seen mostly in older individuals. Two clusters had 30% prevalence in the COPD group. The first included mainly younger current smokers with psychiatric disorders and diseases associated with high-risk bevaviours (eg, schizophrenia, anxiety, hepatitis, liver cirrhosis, pancreatitis and HIV), whereas for individuals without COPD, the cluster contained only anxiety, asthma and depression and had prevalence of 5%. The theme of the second was gastrointestinal diseases, musculoskeletal diseases and cancer. While the prevalence in the non-COPD group was also 30%, the cluster contained fewer nodes and links. The authors concluded that patients with COPD are affected by larger numbers of interlinked morbidities, the clustering patterns of which may suggest common pathobiological processes that might be useful for screening and/or therapeutic interventions. In another study, Divo et al 5 also showed that COPD affects the timing of multimorbidity. They extracted data from the EpiChron Cohort in Aragón, Spain, for 27 617 people with COPD and 27 617 controls without COPD matched for age, sex and site. Patients were separated into incremental age groups (40–55, 56–65, 66–75, 76–85 and ≥85 years) and the prevalence of chronic disorders seen mainly in elderly people was compared between the two groups. The number of comorbidities increased with age in both but occurred earlier in the COPD group. For instance, in the youngest age group of 40‒55 years, 50% of controls had any disease of interest compared with 82% of people with COPD (p<0.001). Furthermore, multimorbidity with two or more disorders were seen 10–15 years earlier among patients with COPD at baseline than among controls.

Carmona-Pírez et al 22 also performed a network analysis based on data from the EpiChron Cohort to assess multimorbidity in 28 608 patients with COPD. The findings revealed several clusters of diseases that were specifically associated with COPD, such as behavioural risk disorders. These clusters featured psychiatric multimorbidity in women and cancer multimorbidity in men with COPD. Risks of gastro-oesophageal reflux disease and obstructive sleep apnoea were also raised in men.

In an exploration of multimorbidity patterns in a population of 12 032 men aged over 50 years, Zacarías-Pons et al 23 identified that the strongest risk factors for multimorbidy among men with COPD were social deprivation (OR 3.3, 95% CI 2.3 to 4.7), current smoking (OR 3.2, 95% CI 2.4 to 4.3), former smoking (OR 1.9, 95% CI 1.4 to 2.4), obesity (OR 1.7, 95% CI 1.3 to 2.4), being separated or single (OR 1.7, 95% CI 1.3 to 2.4) and age (OR 1.16, 95% CI 1.15 to 1.18). They found also that COPD was clustered with osteoarthritis, rheumatoid arthritis, ulcer and cataracts.

Knorst et al 24 assessed multimorbidity in 470 patients with COPD for whom they could obtain Global Initiative for Chronic Obstructive Lung Disease stage and BMI data. The mean number of comorbidities per patient was 3.1 (SD 1.9), rising to 4.1 in patients with obesity, indicating a significant correlation with BMI ( r =0.32, p<0.001). Five or more multimorbid disorders were present in 105 (22%). No correlation was found between the number of comorbidities and severity of COPD. The most common comorbidities were hypertension (45%), cardiac disease (20%), diabetes (15%), osteoporosis (14%) and dyslipidaemia (13%).

In an observational longitudinal study of Medicaid data for 37 151 people with COPD in the USA, Ajmera et al 25 investigated multimorbidity with inflammatory diseases (arthritis, cardiovascular diseases, diabetes, hypertension, hyperlipidaemia and osteoporosis) and psychiatric disorders (anxiety, bipolar disorder, depression and schizophrenia). The overall prevalence of multimorbidity was 79%. Multinomial logistic regression revealed that the risk of coexisting disorders was increased in women versus men (adjusted OR 1.88, 95% CI 1.75 to 2.01) and older adults (age 55‒64 years) versus younger adults (age 18‒24 years; adjusted OR 6.14, 95% CI 5.05 to 7.04).

Cardiovascular and metabolic diseases

Heart diseases are seen in a substantial proportion of people with COPD. In an analysis by Hansen et al 26 of 70 274 people with a diagnosis of COPD in Danish health registries, hypertension was reported in 48% and heart disease in 16%. In an investigation of disease combinations in 11 734 long-term residents in 1174 nursing homes in the USA, COPD was seen in combination with hypertension in 7% of residents (7% of women and >10% of men). 27 COPD was noted in a three-disease combination with hypertension and composite vascular diseases in 4% of residents (3.5% of women and 5.5% of men). Jurevičienė et al 20 found that among 4834 patients with COPD, rates of cardiovascular diseases, arrhythmia and heart failure were significantly increased (all p<0.0001). At least one cardiovascular disease was seen in nearly all men (98.7% of 3338) and women (99.5% of 1496) with COPD.

Nesterovska et al 28 investigated whether risk of atrial fibrillation was increased by the presence of COPD. The study included 86 patients with coexisting asthma and COPD but no cardiovascular disease or thyroid dysfunction. Around half (42 (49%)) had paroxysmal atrial fibrillation.

Asker et al 29 found that among 95 people with COPD and pulmonary hypertension, 68 (72%) had coronary artery disease. The presence of coronary artery disease correlated positively with male sex ( rs =0.224, p=0.029) and hypertension ( rs =0.227, p=0.07) but negatively with the ration of forced expiratory volume in 1 s forced vital capacity (FEV 1 :FVC) ( rs =‒0.253, p=0.013) and systolic pulmonary artery pressure ( rs =‒0.215, p=0.037). No correlation was found between the severity of coronary artery disease and pulmonary hypertension. In a large population study in Copenhagen, Denmark, Ingebrigtsen et al 30 compared the risks of coronary heart disease and heart failure in people with varying phenotypes of COPD and/or coexisting asthma and COPD (n=11 988) or no respiratory disease (n=42 058). Risks of coronary heart disease and heart failure were greatest in the presence of COPD with FEV 1 <50% of predicted (HRs 1.8, 95% CI 1.4 to 2.3 p<0.001 and 2.9, 2.2 to 3.9, p<0.001, respectively) and COPD with late-onset asthma with FEV 1 <50% of predicted (HRs 2.2, 95% CI 1.6 to 3.0, p<0.001 and 2.9, 2.0 to 4.3, p<0.001, respectively).

Miller et al 31 found that heart disease in people with COPD increases the risk of further disorders. They used US household survey data for 968 people with self-reported COPD and heart disease and 767 with COPD and no heart disease. Those with heart disease were most likely to be men and were nearly five times as likely to have diabetes (OR 4.8, 95% CI 3.5 to 6.5; p<0.001) and more than twice as likely to have arthritis (OR 2.2, 95% CI 1.7 to 2.8; p<0.001) as people without heart disease. Risks of other conditions (arthritis, sleep apnoea, chronic pain, depression, gastro-oesophageal reflux disease, osteoporosis and overactive bladder) were also significantly increased, with ORs ranging from 1.5 to 1.8.

Triest et al 32 identified a cachetic disease cluster in 39 (19%) of 208 patients with COPD that was characterised particularly by low muscle mass and underweight (both 80%), hyperglycaemia (54%), arterial stiffness (29%) and hypertension (28%). This cluster of disorders was not relevant in the control group without COPD. A metabolic diseases cluster was also identified among patients with COPD, which featured obesity, hyperglycaemia, insulin resistance, dyslipidaemia and cardiovascular disorders. However, it was also present in the control group and did not have specificity for COPD.

Respiratory disease

O’Kelly et al 33 performed a multivariable analysis to assess risk of multimorbidity in patients with chronic respiratory conditions in Dublin, Ireland. The authors separated respiratory disease into asthma (n=432), COPD (n=206) or other (n=15). Among all 653 patients, 393 (60%) had multimorbidity. While asthma was the most frequently reported disease, the rate of multimorbidity was lower than that for COPD (213 (54%) of 432 vs 169 (82%) of 206). In multivariable analysis, only a diagnosis of COPD was associated with increased rates of multimorbidity. The most common comorbidities in the whole cohort were depression or anxiety, hypertension and cardiovascular diseases (all 28%), musculoskeletal disorders (23%) and endocrine disorders (20%).

Nesterovska et al 28 assessed 86 patients with asthma and COPD overlap syndrome (ACOS). The respiratory effects of this condition in people who developed paroxysmal atrial fibrillation were a reduction in FEV 1 by 36.5% and an increase in hypoxaemia by 12.3%. These changes correlated significantly with the development of atrial fibrillation when seen alongside increased blood pressure and systemic inflammation.

Ganga et al 34 performed a single-centre study to investigate new-onset atrial fibrillation in 2873 patients older than 65 years with obstructive sleep apnoea alone (n=60), COPD alone (n=416), both (n=28) or neither (n=2369). Non-adjusted incidence of atrial fibrillation was 7%, 11%, 21% and 5%, respectively. The risk of multimorbid atrial fibrillation was significantly increased in the COPD subgroup (OR 1.79, 95% CI 0.190 to 0.962, p=0.003) and notably more so in the subgroup of patients with both obstructive sleep apnoea and COPD (3.66, 95% CI 1.06 to 6.9, p=0.007). By contrast, obstructive sleep apnoea alone and neither disease were not predictive.

Spicuzza et al 35 reported a retrospective observational study of people with obstructive sleep apnoea with and without COPD. The risk of cardiovascular diseases, metabolic disorders and gastro-oesophageal reflux disease was substantially increased by COPD (OR 7.8, 95% CI 4.86 to 11.39; p<0.001). Lacedonia et al 36 found similar trends in a retrospective analysis where they compared people with coexisting asthma and COPD or obstructive sleep apnoea alone.

Psychological disorders

Anxiety and depression are common comorbidities in COPD. Phan et al 37 reported in a cross-sectional study that among 242 people with COPD, 124 (51%) had symptoms of depression and/or anxiety, and 81 (34%) had symptoms of both. Multiple regression revealed associations with younger age, having a carer, psychological medical history, comorbidities and reduced quality of life. Silva Júnior et al 38 hypothesised that the presence of COPD would increase the risk of major depression even in people with mild hypoxaemia. They assessed 30 patients with major depression and 30 controls without depression. A significant association was seen between COPD Assessment Test scores greater than 20 and major depression (OR 7.88; 95% CI 1.96 to 31.7; p=0.004), making COPD a predictive factor.

A UK study used questionnaires in 44 patients with COPD and lung cancer who were attending an outpatient clinic to assess the risk of depression. 39 The relative risk and ORs of patients with COPD developing depression were 1.4 and 1.6, respectively. The number of coexisting comorbidities significantly raised the OR to 2.13 (95% CI 1.02 to 4.49).

Triest et al 32 performed a validation study of comorbidity clusters previously identified in patients with COPD. They compared the clusters in 208 patients with COPD and a control group of 200 elderly patients without COPD. The psychological cluster included 40 patients with COPD, of whom 95% had anxiety and 59% had depression. By contrast, very few controls had anxiety and/or depression and, therefore, this cluster was deemed relevant only to COPD.

Physical activity

Associations between COPD and 31 comorbidities and how these were affected by level of physical activity were assessed in 601 adults in Spain. 40 94% of participants had comorbidities. Low levels of physical activity were significantly associated with increased risk of urinary incontinence (OR 2.12, 95% CI 1.21 to 3.69), chronic constipation (OR 1.97, 95% CI 1.12 to 3.46), cataracts (OR 1.84, 95% CI 1.07 to 3.15), chronic anxiety (OR 1.51, 95% CI 1.00 to 2.27) and chronic lumbar back pain (OR 1.49, 95% CI 1.04 to 2.13). The authors concluded that recommending increased physical activity could improve the quality of life for patients with COPD.

Yu et al 41 performed a longitudinal study to assess the relationship between physical activity and multimorbidity risk in 409 patients with COPD selected from primary care in the Netherlands and Switzerland. Patients were followed up for 5 years and self-reported physical activity, occurrences of cardiovascular, neurological, endocrine, musculoskeletal, malignant and infectious diseases and mental health. Physical activity showed significant associations with reduced anxiety (adjusted HR 0.89, 95% CI 0.79 to 1.00; p=0.045) and depression (adjusted HR 0.85, 0.75 to 0.95; p=0.005). For other disorders, likelihood of occurrence was reduced with physical activity, but not significantly so.

Polypharmacy

One study addressed polypharmacy (defined by the authors as taking five or more drugs per day) along with multimorbidity. Among 245 patients with COPD in Crete, Greece, Ierodiakonou et al 42 found that 77% of patients had multimorbidity, which increased to 84% in those with age 65 years or older. More than half (55%) of patients were receiving multiple drugs, but 10% of medications were found to be inappropriate. Polypharmacy was associated with COPD Assessment Test scores of 10 or greater, multimorbidity, several cardiometabolic diseases, cancer, depression and anxiety and prostate disorders. Co-administration of medications increased the cumulative risk of falls in 22%, constipation in 17% and cardiovascular events in 13% of patients. The authors concluded that polypharmacy increases the risk of worse health outcomes in patients with COPD.

Garneau-Picard et al 43 investigated the influence of smoking on multimorbidity in people with ACOS. In a group of 154 patients with irreversible airway obstruction (47% men, 53% women; 100 smokers with ACOS and 54 with irreversible airway obstruction not related to smoking), a sex-specific symptom pattern was found that suggested greater susceptibility of women to smoke. Woman had lower prevalence of severe asthma (44.9% vs 64.7%, p=0.0264) but had higher FEV 1 :FVC values (67.2, 95% CI 66.7 to 67.7 vs 65.5, 95% CI 64.8 to 66.2, p=0.0002) and more comorbidities (4.6, 95% CI 3.9 to 5.4 vs 3.2, 95% CI 2.7 to 3.7, p=0.0012) than men, despite having lower tobacco exposure (36.0 pack-years, 95% CI 32.5 to 39.5 vs 41.9 pack-years, 95% CI 38.1 to 45.7, p=0.0278). Furthermore, compared with female non-smokers with airway obstruction, more women with ACOS had severe asthma, used more medication and had a worse FEV 1 :FVC, whereas men showed no significant differences from other men with ACOS or men or women with non-smoking-related disease.

Cunningham et al 44 explored the relationships between smoking, COPD and 10 other conditions (arthritis, asthma, cancer, coronary heart disease, depression, diabetes, high blood pressure, high cholesterol, kidney disease and stroke). In a cross-sectional study of 405 856 adults in the US general population who had responded to surveys from the national Behavioral Risk Factor Surveillance System, 33 088 (7%) had COPD. The prevalence of COPD was 14% among current smokers, 7% among former smokers and 3% among never smokers. Only a quarter of people with COPD (24%) were never smokers, compared with 57% of those without COPD (39% vs 27% were former smokers and 37% vs 16% were current smokers). 95% of those with COPD had any of the 10 comorbidities of interest, compared with 69% of those without COPD, and the prevalence of all conditions was higher in the COPD group than in the non-COPD group. Significant interactions (p<0.001) were seen between smoking status and COPD and each of the other chronic conditions. Zacarías-Pons et al 23 calculated ORs of 3.2 (95% CI 2.4 to 4.3) for current smoking and 1.92 (95% CI 1.4 to 2.4) for former smoking in a disease cluster that included COPD.

In a retrospective study, Le et al 45 investigated the multimorbidity burden in 739 118 Medicare beneficiaries with COPD in the USA, aged 65 years or older. The authors calculated the prevalence of multimorbidity at COPD diagnosis and 1 year after diagnosis and estimated the rates of onset per 100 person-years 1 year before versus 1 year after diagnosis. The findings were compared with the same number of Medicare beneficiaries without COPD, matched for age, sex and race. In the COPD group, the average number of comorbidities was 10 (SD 4.7) compared with only 1 (SD 3.3) in the non-COPD group. The most frequent comorbidities seen at COPD diagnosis had all increased in prevalence 1 year later: hypertension changed from 70.8% to 80.2%; hyperlipidaemia from 52.2% to 64.8%; anaemia from 42.1% to 52.0%; arthritis from 39.8% to 47.7%; and congestive heart failure from 31.3% to 38.8%. The rates of new onset 1 year before and after COPD diagnosis were hyperlipidaemia (22.8 and 27.6 cases per 100 person-years, respectively), anaemia (17.8 and 20.3), arthritis (12.9 and 13.2), hypertension (39.8 and 32.3) and congestive heart failure (16.2 and 13.2). The ORs for all diseases assessed were increased in the COPD group compared with those in the non-COPD group.

This systematic literature review indicated high prevalence of multimorbid cardiovascular disorders, metabolic disorders and anxiety and depression in people with COPD, and disease profiles were frequently complex. Multimorbid age-related diseases occur earlier in people with COPD than those in the general population. Therefore, there is an accompanying risk of long-term increased healthcare needs and increased risks of polypharmacy, adverse drug-related events, morbidity and mortality. Differences between the sexes in preventable or modifiable diseases that should be noted are the increased likelihood of gastro-oesophageal disorders in men and psychiatric and musculoskeletal disorders in women.

Most patients with COPD have complex disease profiles, and have a much-heightened risk of developing or extending multimorbidity. Our findings highlighted the associations between COPD, multimorbidity and high-risk conditions and behaviours, such as smoking, obesity, low physical activity and low socioeconomic status. Additionally, Le et al reported increased prevalence of many chronic conditions after the diagnosis of COPD in elderly participants. That study was not designed to assess causality, but Alter et al suggest that the increased prevalence of some multimorbid diseases are due to worsening of COPD whereas some are age related. 46 However, this area is not well researched, and other reasons, such as increased investigations for associated disorders and/or lack of testing for or modification of other risk factors, should be explored further. Finally, compared with knowledge of other disorders, patients’ understanding of COPD symptoms, treatment and long-term disease course, including multimorbidity, is poor and can lead to undertreatment. 47 Care for COPD should, therefore, consider preventive measures at population, national, regional and individual levels to address risk of frequently seen multimorbid diseases . Such measures might be holistic approaches that involve primary and secondary care teams; creation of educational materials and opportunities (eg, to address and maintain lifestyle changes); development of individualised care plans that underscore the importance of COPD in relation to other diseases and support self-management (eg, exercise, diet, etc), as recommended by the Global Initiative for Chronic Obstructive Lung Disease 48 and national guidelines, such as those of the National Institute for Health and Care Excellence 49 ; and multidisciplinary management with healthcare professionals in other specialties. 50

Jassem et al 51 proposed an integrated care model for patients with advanced COPD in Poland. Long-term maintenance care involved planned visits to healthcare providers in primary care (general practitioners and community and specialty nurses) and secondary care (pneumonologists) and other professional or volunteer carers and social workers. As the disease or the patient’s health worsened, they recommended spiritual, psychological and palliative care support. Information on health status after exacerbation, including multimorbidity, would be shared and managed as appropriate during or in addition to planned visits. This model could be well adapted for earlier-stage disease with the addition of other specialties as appropriate.

Only one study assessing polypharmacy met our selection criteria, but the increased risk of adverse drug reactions, overtreatment with redundant drugs and drug–drug and/or drug–disease interactions 16 make this an important factor in the risk of multimorbidity among patients with COPD. Furthermore, use of high numbers of drugs can contribute to reduced adherence across therapies, and, in COPD, increased disease severity and risk of exacerbation. Regular medicine reviews that include clinical and pharmacy healthcare professionals should be considered as part of an integrated care pathway for people with COPD.

This review has some limitations. The papers included showed substantial heterogeneity. Furthermore, we chose to include abstracts providing clear data where studies were relevant but full reports were not available. This decision was made to ensure wide exploration of the literature and include poorly represented areas of research. However, these factors meant that it was not possible to perform a meta-analysis or account for bias in studies and that descriptive analysis of the findings was more appropriate. Due to the nature of what is being investigated, many of the studies published are observational. We cannot, therefore, cite causal links between COPD and multimorbidity. However, our findings reveal positive associations between COPD and the development or worsening of multimorbidity that occurs in reasonably consistent, predictable patterns. Only one study was found that dealt with polypharmacy. This area warrants more investigation to maximise the effectiveness of treatment without overly high pill burden on patients and to reduce the risk of drug-related unplanned and emergency care. Study designs that focus on these aspects and/or more long-term epidemiological data would be beneficial in this area.

Conclusions

Multimorbidity is an extremely common and important feature of COPD. People experience a wide range of disorders, but the most common are generally considered preventable and/or modifiable. Patients seen in general practice with cardiovascular, metabolic and musculoskeletal disorders, particularly arthritis or osteoporosis and who are current or former smokers should be considered for education about the risks of COPD and new or worsening multimorbidity. A proactive holistic approach to management involving primary and secondary care healthcare professionals that includes regular review of all aspects of health, treatment and lifestyle factors could reduce the burden of care even for patients with several severe long-term conditions. Important areas for future research are to assess changes in multimorbidity over time, as diagnosis of other diseases seems to increase quickly around the time of COPD diagnosis, and the risks associated with polypharmacy.

Ethics statements

Patient consent for publication.

Not applicable.

Ethics approval

Harrison C ,
van den Akker M , et al
Wilkinson TMA
Poblador-Plou B , et al
Morrison D , et al
Santos NC dos ,
Miravitlles M ,
Camelier AA , et al
Edwards LD ,
Agustí A , et al
Soley-Bori M ,
Ashworth M ,
Bisquera A , et al
Mannino DM ,
Swensen A , et al
Williams NP ,
Coombs NA ,
Johnson MJ , et al
Anthonisen NR ,
Soriano JB , et al
Iheanacho I ,
King D , et al
Schnell K ,
Lee T , et al
Nicholl BI ,
Jani BD , et al
Orlowski A ,
Humphreys H , et al
Barnett K ,
Mercer SW ,
Norbury M , et al
Mularski RA , et al
Jurevičienė E ,
Burneikaitė G ,
Dambrauskas L , et al
Casanova C ,
Marin JM , et al
Carmona-Pírez J ,
Poblador-Plou B ,
Díez-Manglano J , et al
Zacarías-Pons L ,
Vilalta-Franch J ,
Turró-Garriga O , et al
Ries L , et al
Ajmera MR ,
Sambamoorthi U
Hansen NS ,
Ängquist L ,
Lange P , et al
Boscardin WJ ,
Steinman MA , et al
Nesterovska O ,
Stupnytska G ,
Fediv O , et al
Kucuk U , et al
Ingebrigtsen TS ,
Marott JL ,
Vestbo J , et al
Miller DP ,
Triest FJJ ,
Franssen FME ,
Reynaert N , et al
O’Kelly S ,
Lane S , et al
Puppala VK , et al
Spicuzza L ,
Campisi R ,
Crimi C , et al
Lacedonia D ,
Carpagnano GE ,
Patricelli G , et al
Waterer G , et al
Silva Júnior JLR ,
de Sousa Corrêa K , et al
Md Yusof MY ,
Sánchez Castillo S ,
Díaz Suárez A , et al
ter Riet G ,
Puhan MA , et al
Ierodiakonou D ,
Theodorou E ,
Sifaki-Pistolla D , et al
Garneau-Picard FA ,
Boulay ME ,
Cunningham TJ ,
Rolle IV , et al
Magder L , et al
Kahnert K ,
Trudzinski FC , et al
Mannino DM , et al
↵ Global initiative for chronic obstructive lung disease. global initiative for chronic obstructive lung disease: GOLD pocket guide to COPD diagnosis, management, and prevention. A guide for health professionals. 2021 report . 2020 . Available : https://goldcopd.org/wp-content/uploads/2020/11/GOLD-2021-POCKET-GUIDE-v1.0-16Nov20_WMV.pdf [Accessed 3 Jan 2023 ].
Ambrosino N ,
Kozielski J ,
Górecka D , et al

Supplementary materials

Supplementary data.

This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

Data supplement 1

Contributors AO devised the study, AO, JE, SS and JKQ performed the literature search and reviewed the papers for inclusion. AO, SS, JE, RA, AB and JKQ interpreted the data. AO and RA drafted the article. All authors reviewed the article and approved the final version for publication.

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.

Provenance and peer review Not commissioned; externally peer reviewed.

Read the full text or download the PDF:

This website is intended for healthcare professionals

{ $refs.search.focus(); })" aria-controls="searchpanel" :aria-expanded="open" class="hidden lg:inline-flex justify-end text-gray-800 hover:text-primary py-2 px-4 lg:px-0 items-center text-base font-medium"> Search

Search menu

Literature review.

Autistic women's experiences of the antenatal, intrapartum and early postnatal periods

The PICO mnemonic (Stern et al, 2014) was used to identify key words and develop the research question: what can midwives in England learn from studies exploring the experiences of autistic women in...

Perinatal outcomes in persistent occiput posterior fetal position: a systematic review and meta-analysis

Meta-analysis is a quantitative, formal, epidemiological study design used to systematically assess the results of previous research to derive conclusions about that body of research (Haidich, 2010)....

Carbetocin vs oxytocin in third stage labour: a quantitative review of low- and middle-income countries

This review was carried out to determine if the use of carbetocin in low- and middle-income countries would reduce the risk of postpartum haemorrhage, and associated morbidity and mortality, in...

mHealth interventions to improve self efficacy and exclusive breastfeeding: a scoping review

The electronic search was carried out in September 2022, using the population/problem/patient, intervention, comparison, outcome and study design strategy. The primary source of literature was online...

Routine examinations such as palpation can help to recognise adverse events

The use of gender-neutral language in maternity settings: a narrative literature review

A preliminary search of the Cochrane Library, CINAHL, and MEDLINE databases was undertaken to identify articles relating to the topic. Search terms or text words contained in titles, abstracts and...

Midwives’ practice of maternal positions throughout active second stage labour: an integrative review

An integrative review was considered suitable for this study, as this methodology allows inclusion of data from all types of literature to fully answer review questions (Whittemore and Knafl, 2005;...

Health-seeking behaviours of pregnant adolescents: a scoping review

This study was developed based on Arskey and O'Malley's (2005) scoping review methodology. According to this framework, there are six stages: (1) identifying the research question, (2) identifying...

Maternal intrapartum fluids and neonatal weight loss in the breastfed infant

Searches of key databases (CINAHL, MEDLINE, EMBASE, EMCARE) were conducted using a search strategy developed in collaboration with the local NHS library service (Table 1). Known researchers in this...

Midwives' and women's understanding of cytomegalovirus infection during pregnancy

The electronic databases CINAHL, Maternity and Infant Care, EMBASE and PubMed were systematically searched using key words and terms, which included: midwife, midwives, woman, women,...

Figure 8a. Roll over onto all fours position

Simulation and midwifery education 2011–2021: a systematic review

The focus of this review was the evaluation of articles describing the use of simulation in midwifery education. A search was carried out using the online database PUBMED for articles published...

How does social media influence expectations, decision making and experiences of childbirth?

This literature review focused on qualitative research, to gather insight from rich data about women's experiences (Aveyard, 2019). Initial searches were undertaken in April 2022 in CINAHL Plus,...

Midwives’ experiences using personal protective equipment during COVID-19: a scoping review

A scoping review method was used to review the literature. This method is recommended when the topic has not yet been broadly reviewed or is of a complex nature (Mays et al, 2001). The methodological...

Showing 1 to 12 of 59 results

Why choose British Journal of Midwifery?

BJM supports midwives by sharing expertise and advice to help you build confidence, grow professionally and improve care.

What's included

Evidence-based best practice

Peer-reviewed research

Practical guidance

CPD support

Subscriptions start:

Log in using your username and password

Search More Search for this keyword Advanced search
Latest content
Current issue
For authors
BMJ Journals More You are viewing from: Google Indexer

Jeremy Josef Heit 1 ,
http://orcid.org/0000-0003-0630-2547 Neeraj Chaudhary 2 ,
http://orcid.org/0000-0001-9409-5810 Justin R Mascitelli 3 ,
http://orcid.org/0000-0002-0978-8012 Mais Al-Kawaz 4 ,
http://orcid.org/0000-0001-9973-5792 Amanda Baker 5 ,
Ketan R Bulsara 6 ,
Jan Karl Burkhardt 7 ,
Franklin A Marden 8 ,
Daniel Raper 9 ,
http://orcid.org/0000-0002-1054-9414 Stavropoula I Tjoumakaris 10 ,
http://orcid.org/0000-0003-1743-8781 Clemens M Schirmer 11 ,
http://orcid.org/0000-0001-5885-7259 Steven W Hetts 12 ,
SNIS Standards and Guidelines Committee
SNIS Board of Directors
1 Departments of Radiology and Neurosurgery , Stanford University Medical Center , Stanford , California , USA
2 Departments of Radiology, Neurosurgery, and Otorhinolaryngology , University of Michigan Health System , Ann Arbor , Michigan , USA
3 Department of Neurosurgery , University of Texas Health Science Center at San Antonio , San Antonio , Texas , USA
4 Departments of Neurology, Neurosurgery, and Radiology , University of Kentucky , Lexington , KY , USA
5 Department of Radiology , Montefiore Medical Center , New York , NY , USA
6 Division of Neurosurgery , University of Connecticut , Farmington , Connecticut , USA
7 Department of Neurosurgery , Hospital of the University of Pennsylvania , Philadelphia , Pennsylvania , USA
8 Alexian Brothers Medical Center, Brain and Spine Institute , Elk Grove Village , Illinois , USA
9 Departments of Neurological Surgery, Radiology, and Biomedical Imaging , University of California , San Francisco , California , USA
10 Department of Neurosurgery , Thomas Jefferson University Hospital , Philadelphia , Pennsylvania , USA
11 Department of Neurosurgery and Neuroscience Institute , Geisinger and Geisinger Commonwealth School of Medicine , Wilkes-Barre , Pennsylvania , USA
12 Departments of Radiology, Biomedical Imaging and Neurological Surgery , University of California , San Francisco , California , USA
Correspondence to Dr Steven W Hetts, Radiology, UCSF, San Francisco, California, USA; steven.hetts{at}ucsf.edu

Background Endovascular therapy (EVT) dramatically improves clinical outcomes for patients with anterior circulation emergent large vessel occlusion (ELVO) strokes. With recent publication of two randomized controlled trials in favor of EVT for basilar artery occlusions, the Society of NeuroInterventional Surgery (SNIS) Standards and Guidelines Committee provides this focused update for the existing SNIS guideline, ‘Current endovascular strategies for posterior circulation large vessel occlusion stroke.’

Methods A structured literature review and analysis of studies related to posterior circulation large vessel occlusion (basilar or vertebral artery) strokes treated by EVT was performed. Based on the strength and quality of the evidence, recommendations were made by consensus of the writing committee, with additional input from the full SNIS Standards and Guidelines Committee and the SNIS Board of Directors.

Results Based on the results of the most recent randomized, controlled trials on EVT for basilar or vertebral artery occlusion, the expert panel agreed on the following recommendations. For patients presenting with an acute ischemic stroke due to an acute basilar or vertebral artery occlusion confirmed on CT angiography, National Institutes of Health Stroke Scale (NIHSS) score of ≥6, posterior circulation Alberta Stroke Program Early CT Score (PC-ASPECTS) ≥6, and age 18–89 years: (1) thrombectomy is indicated within 12 hours since last known well (class I, level B-R); (2) thrombectomy is reasonable within 12–24 hours from the last known well (class IIa, level B-R); (3) thrombectomy may be considered on a case by case basis for patients presenting beyond 24 hours since last known well (class IIb, level C-EO). In addition, thrombectomy may be considered on a case by case basis for patients aged <18 years or >89 years on a case by case basis (class IIb, level C-EO).

Conclusions The indications for EVT of ELVO strokes continue to expand and now include patients with basilar artery occlusion. Further prospective, randomized controlled trials are warranted to elucidate the efficacy and safety of EVT in populations not included in this set of recommendations, and to confirm long term outcomes.

Thrombectomy

https://doi.org/10.1136/jnis-2024-021705

Statistics from Altmetric.com

Request permissions.

Introduction

Endovascular therapy (EVT) dramatically improves clinical outcomes for patients with emergent large vessel occlusion (ELVO). The Society of NeuroInterventional Surgery (SNIS) Standards and Guidelines Committee updates and supplements the existing SNIS guideline ‘Current endovascular strategies for posterior circulation large vessel occlusion stroke’. 1–3 Acute ischemic stroke within the posterior circulation accounts for up to 20% of ischemic stroke diagnoses. Patients with a large vessel occlusion of the vertebral or basilar artery are now often treated with endovascular thrombectomy. This document provides recommendations for the treatment of these patients.

Posterior circulation large vessel occlusion strokes

Basilar artery occlusion (BAO) accounts for approximately 1% of all ischemic strokes and about 10% of strokes due to an intracranial large vessel occlusion. 4 Although this condition may be relatively uncommon, it captures our attention due to its potentially devastating consequences, which may include death or locked-in syndrome. Despite reports of high morbidity and mortality, the natural history can be quite variable, and clinical outcomes likely depend on a variety of patient factors, such as age, baseline health, severity of presenting symptoms, location of the occlusion along the basilar artery, presence of collaterals, and territory of any associated ischemic injury. 5–8 Nonetheless, BAO remains one of the most dreaded types of stroke.

Similar to the transition from IMS-III (Interventional Management of Stroke III) to the post-MR CLEAN (Multicenter Randomized Clinical trial of Endovascular treatment for Acute ischemic stroke in the Netherlands) era for anterior circulation ELVO stroke treatment, new data in BAO strokes provide encouraging results regarding improved clinical outcomes. The previously reported randomized BEST (Basilar Artery Occlusion Endovascular Intervention vs Standard Medical Treatment) 9 and BASICS (Basilar Artery International Cooperation Study) 10 trials did not show a definite benefit of EVT over medical therapy. However, a trend towards better outcomes with thrombectomy was noted in these trials. Concerns regarding clinical equipoise and treatment crossover may have undermined these studies. In contrast, the more recent ATTENTION (Endovascular Treatment for Acute Basilar Artery Occlusion) 11 and BAOCHE (Basilar Artery Occlusion Chinese Endovascular Trial) 12 studies demonstrated that EVT within 0–12 and 6–24 hours of symptom onset, respectively, is superior to medical therapy.

New trial summary

The ATTENTION trial randomized acute ischemic stroke patients with BAO to best medical therapy or thrombectomy plus best medical therapy. Enrolled patients had an National Institutes of Health Stroke Scale (NIHSS) score of ≥10. The study was performed in China, and patients in ATTENTION were randomized between 0 and 12 hours from symptom onset in a 2:1 manner to EVT versus best medical therapy. In total, 226 patients were assigned to the EVT arm and 114 to the medical therapy arm. 11 Good functional outcome was defined as a modified Rankin Scale (mRS) score of 0–3, 90 days after symptom onset; this was achieved in 46% of the EVT group versus 23% in the medical therapy group (absolute risk reduction (ARR) 2.06, 95% CI 1.46 to 2.91, P<0.001). Although the symptomatic intracranial hemorrhage rate was 5% in the thrombectomy group and 0% in the medical group, the difference did not reach statistical significance. Mortality at 90 days was lower in the EVT arm (37%) compared with the medical arm (55%, ARR 0.66, 95% CI 0.52 to 0.82). Of note, 14% of EVT cases were noted to have a procedural complication, including one death due to arterial perforation.

The BAOCHE trial was also performed in China, and this study randomized patients with acute ischemic stroke due to BAO to medical therapy or thrombectomy plus medical therapy. 12 Patients in BAOCHE were enrolled 6–24 hours after symptom onset and had an NIHSS score of ≥6. The primary endpoint of a favorable clinical outcome was initially defined as mRS 0–4, but was later changed to mRS 0–3 after the study started. The trial was stopped early due to an interim analysis that demonstrated the superiority of the thrombectomy arm. When the study was halted, 110 patients had received thrombectomy and 107 had received medical therapy. Favorable outcomes (mRS 0–3) were achieved in 46% of thrombectomy patients versus 24% of medical therapy patients (ARR 1.81, 95% CI 1.26 to 2.60, P<0.001). Like ATTENTION, symptomatic intracranial hemorrhage was uncommon and non-significantly higher in the thrombectomy group (6% vs 1%, relative risk (RR) 5.18, 95% CI 0.46 to 42.18). Mortality at 90 days was also non-significantly lower in the thrombectomy group (31% vs 42%, RR 0.75, 95% CI 0.54 to 1.04). Procedural complications in thrombectomy patients occurred in 11%, which was similar to the rate in ATTENTION.

Pre-thrombectomy imaging considerations

The previously reported Basilar Artery International Cooperation Study (BASICS 10 and the Basilar Artery Occlusion Endovascular Intervention vs Stanford Medical Treatment (BEST 9 trials—both of which used simple imaging criteria for enrollment—failed to demonstrate the superiority of thrombectomy compared to medical therapy in patients with acute ischemic stroke due to a basilar artery occlusion. Patients had BAO on CT angiography (CTA) and a non-contrast head CT free of bleeding and free of extensive brainstem infarction, cerebellar mass effect, or acute hydrocephalus. However, there was no other quantitative assessment of ischemic injury applied. ATTENTION and BAOCHE also required demonstration of BAO on CTA, and evidence of limited ischemic injury, defined as a posterior circulation Alberta Stroke Program Early CT Score (PC-ASPECTS) of ≥6 ( figure 1 ).

Download figure
Open in new tab
Download powerpoint

Posterior circulation Alberta Stroke Program Early CT Score (PC-ASPECTS) on MRI and CT. PC-ASPECTS regions are noted on MRI (top row) and non-contrast CT (bottom row). PC-ASPECTS are assigned in the following manner: pons, 2 points; cerebellum, 1 point per hemisphere; midbrain, 2 points; thalamus, 1 point per hemisphere; and posterior cerebral artery territory in occipital lobes, 1 point per hemisphere. Note that MRI images are shown on T2 weighted images to better delineate brain anatomy, but PC-ASPECTS is assigned based on restricted diffusion on diffusion weighted images. Cer, cerebellum; T, thalamus; PCA, posterior cerebral artery territory.

Further research is required to determine the optimal imaging selection for BAO patients before thrombectomy, but at this time, the evidence suggests that eligible patients should have a PC-ASPECTS ≥6 on non-contrast head CT and BAO on CTA. The use of MRI is not required before basilar artery thrombectomy. However, this imaging modality remains helpful in the evaluation of these patients, given the greater sensitivity of the modality for ischemia within the posterior fossa, which is notoriously difficult to assess on non-contrast head CT. The role of CT perfusion techniques has shown promise in identifying patients who do and do not benefit from thrombectomy, 13 and ongoing prospective studies will provide additional guidance as to the use of CT perfusion in patients with posterior circulation large vessel occlusion (LVO).

Thrombectomy technical considerations

Endovascular thrombectomy techniques generally include aspiration, mechanical thrombectomy with stentretriever devices, cervical or cerebral artery angioplasty, and cervical or cerebral artery stent placement. Adjunctive medical therapies may also be performed, which include intravenous antiplatelet medication administration and intra-arterial thrombolysis. Moreover, thrombectomy may be performed with patients undergoing moderate sedation or general anesthesia. The recently reported ATTENTION and BAOCHE trials used varied endovascular thrombectomy techniques, briefly summarized below.

In the ATTENTION trial, the thrombectomy technique was left to the treating physician’s discretion, and general anesthesia was used in 56% of treated patients. Treatment strategies included stent retrievers alone in 5%, aspiration alone in 35%, and combined aspiration and stent retriever use in 50%. Notably, intracranial angioplasty and stent placement were performed in 40%, extracranial angioplasty and stent placement in 8%, intra-arterial thrombolysis in 5%, and intravenous tirofiban administration in 40%. The specific devices and intracranial stents used in the ATTENTION trial have not been reported.

In the BAOCHE trial, the thrombectomy treatment protocol was more restricted. General anesthesia was used in 65% of patients. All patients were treated with the Solitaire SR (Medtronic), and other rescue aspiration, additional stent retrievers, and pharmacological treatment options were not allowed except in cases of cervical or cerebral artery angioplasty and stenting. Intracranial angioplasty or stenting after a failed thrombectomy was performed in 55%, the Solitaire SR was detached in 26%, and other stents were used in 12% of cases. Tirofiban was administered during the procedure in 54% of patients in the BAOCHE trial. Hemorrhagic rates were low, even in the setting of angioplasty and stenting.

In summary, all of the commonly used thrombectomy techniques (aspiration, stent retriever, and combined techniques) were performed with high frequency in the ATTENTION and BAOCHE trials. They were shown to be effective for the treatment of posterior circulation LVO. The high rates of cervical and intracranial angioplasty and stent placement, and intravenous GP2b3a inhibitor administration in the ATTENTION and BAOCHE trials likely reflects the frequency of intracranial atherosclerosis in the posterior circulation 14 as well as within the Chinese population 15 that comprised these studies. The use of angioplasty and stent placement with associated medical therapy was an effective treatment for posterior circulation LVO in these studies. It may be a reasonable rescue technique to consider for the emergent treatment of intracranial atherosclerosis after failed aspiration based or stent retriever based thrombectomy, despite the results of SAMMPRIS (Stenting vs Aggressive Medical Management for Preventing Recurrent Stroke in Intracranial Stenosis) 16 in symptomatic intracranial atherosclerosis a decade ago. However, nearly all commonly used angioplasty balloons and stents are not approved by the US Food and Drug Administration for use in the cerebral circulation.

Areas of uncertainty

Posterior circulation large vessel occlusion and low nihss.

Patients with a basilar or vertebral artery occlusion may present with relatively mild symptoms (NIHSS <6), and whether thrombectomy is indicated in these patients remains uncertain. Some proportion of patients who present with initially mild symptoms may have clinical deterioration, but clinical and imaging factors that predict subsequent neurological decline have not yet been defined.

The ATTENTION and BAOCHE trials enrolled patients with NIHSS ≥6 (BAOCHE) and ≥10 (ATTENTION), and there is currently a lack of guidance from prospective and randomized studies to guide the treatment of patients with a posterior circulation LVO and mild symptoms. 17 There is a need for future prospective and randomized trials to address these limitations in our knowledge and, at this time, thrombectomy treatment of patients with a vertebral artery occlusion or BAO and NIHSS ≤6 should be based on individual situations that balance the risks and benefits of thrombectomy compared with medical therapy.

Thrombectomy for posterior circulation large vessel occlusion beyond 24 hours

Thrombectomy treatment of patients with BAO was superior to medical therapy in patients treated within 0–12 hours of the last known well (ATTENTION) and 6–24 hours of the last known well (BAOCHE). However, many patients with BAO have a delayed diagnosis due to the non-specific symptoms that are frequently associated with posterior circulation ischemic stroke, and these patients may only be evaluated for possible thrombectomy after >24 hours from the last known well.

There have been isolated case reports of BAO, which have been treated with mechanical thrombectomy after a long duration beyond even 24 hours. Pandhi et al 18 reported their experience treating 10 patients who underwent thrombectomy beyond 24 hours from symptom onset. Successful recanalization (Thrombolysis in Cerebral Infarction 2b–3) was achieved in five patients (50%), which indicates that thrombectomy is feasible and technically successful in this population. 19 There are also isolated case reports of BAO patients who underwent thrombectomy >48 hours from the last known well. However, there is tremendous uncertainty as to the benefit of treatment in these extended time windows. Data from the BASICS registry found that functional outcomes were universally poor when recanalization was achieved >9 hours after stroke onset or last known well. 20 We agree that there needs to be more clarity to guide thrombectomy treatment of posterior circulation LVO patients >24 hours, and additional studies are needed to address these gaps in our knowledge.

Posterior circulation distal or medium vessel Occlusions

Thrombectomy treatment of patients with occlusion of the P1, P2, or more distal segments of the posterior cerebral artery (posterior circulation distal or medium vessel occlusions) remains unproven. There are anecdotal and retrospective studies that describe the benefits of thrombectomy in these patients, but there are no prospective or randomized data to guide treatment of these patients, which may be reasonable in some circumstances. Future studies are needed to address these gaps in our knowledge.

Updated recommendations

In patients with acute BAO who meet criteria from the BAOCHE or ATTENTION trials (NIHSS score ≥6, PC-ASPECTS ≥6, CTA with BAO, age 18–89 years), thrombectomy is indicated within 12 hours of last known well (class I, Level B-R).

In patients with acute BAO who present between 12 and 24 hours from the last known well, NIHSS score ≥6, or PC-ASPECTS ≥6, thrombectomy is reasonable (class IIa, level B-R).

In patients with acute BAO who present beyond 24 hours from last known well, with NIHSS score ≥6 or PC-ASPECTS≥6, it may be reasonable to consider thrombectomy on a case by case basis (class IIb, level C-EO).

In aged <18 years or >89 years presenting with acute BAO, it is reasonable to consider thrombectomy on a case by case basis (class IIb, level C-EO).

In summary, multiple randomized trials now exist that show a benefit of thrombectomy in populations where there was a lack of data previously. These results justify changes in local patient selection criteria for thrombectomy. Hub and spoke hospital systems should consider incorporating these results in determining which patients are appropriate for transfer to a center capable of performing thrombectomy.

Ethics statements

Patient consent for publication.

Not applicable.

Ethics approval

Acknowledgments.

The authors acknowledge Teri Moore and Eddie Woods for logistical support in preparing this document and Justin Fraser for critical review.

Jayaraman MV ,
Hussain MS ,
Abruzzo T , et al
Ansari SA ,
McTaggart RA , et al
Meyers PM ,
Prestigiacomo CJ , et al
Mattle HP ,
Lindsberg PJ , et al
Higashida R ,
Wechsler L , et al
Schonewille WJ ,
Wijman CAC ,
Michel P , et al
Khomenko A ,
Hill MD , et al
Ye R , et al
Langezaal LCM ,
van der Hoeven EJRJ ,
Mont’Alverne FJA , et al
Nogueira RG ,
Zhu Y , et al
Wu L , et al
Cereda CW ,
Mlynash M , et al
Park SM , et al
Hurford R ,
Feng X , et al
Derdeyn CP ,
Chimowitz MI ,
Lynn MJ , et al
Alemseged F ,
Van der Hoeven E ,
Di Giuliano F , et al
Chandra R ,
Abdulrazzak MA , et al
Vergouwen MDI ,
Pfefferkorn T , et al
Nguyen TN ,
Coutts SB , et al

X @NIR_ChaudharyMD, @jmascite, @Mais_Alk, @amandaebaker, @stavtjouamakris

Collaborators Society of NeuroInterventional Surgery (SNIS) Standards and Guidelines Committee: Steven Hetts, Clemens Schirmer, Jeremy Heit, Matthew Amans, Charles Prestigiacomo, Matthew Tenser, Neil Haranhalli, Ketan Bulsara, Stavropoula Tjoumakaris, Fawaz Al-Mufti, Shivani Pahwa, Kazim Narsinh, Neeraj Chaudhary, Jan-Karl Burkhardt, Franklin Marden.

SNIS Board of Directors: Mahesh Jayaraman, Johanna Fifi, Guilherme Dabus, Justin Fraser, Peter Kan, Jeremy Heit, Matthew Amans, Vitor Pereira, Edgar Samaniego, J Mocco, Michael Chen, Jonathan Grossberg, Jenny Tsai, Michael Froehler, Felipe Albuquerque, Blaise Baxter, Sameer Ansari, Steven Hetts.

Contributors All named authors had substantial contributions to the conception (JJH, NC, JRM, MA-K, AB, KRB, JKB, FAM, DR, SIT, CMS, and SWH), design (JJH, NC, JRM, MA-K, AB, KRB, JKB, FAM, DR, SIT, CMS, and SWH), or analysis (JJH, NC, JRM, MA-K, AB, KRB, JKB, FAM, DR, SIT, CMS, and SWH) of the work presented. All named authors took part in drafting or revising the manuscript critically for intellectual content (JJH, NC, JRM, MA-K, AB, KRB, JKB, FAM, DR, SIT, CMS, and SWH). All named authors reviewed the final version of the manuscript submitted for publication (JJH, NC, JRM, MA-K, AB, KRB, JKB, FAM, DR, SIT, CMS, and SWH). SWH agrees to be accountable for the work.

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. The research was funded solely from internal resources of the Society of NeuroInterventional Surgery.

Disclaimer This guideline update is provided for informational and educational purposes only. Adherence to any recommendations included in this review will not ensure successful treatment in every situation. Furthermore, the recommendations contained in this review should not be interpreted as setting a standard of care, or deemed inclusive of all proper methods of care, nor exclusive of other forms of care reasonably directed to obtaining the same results. The ultimate judgment regarding the propriety of any specific therapy must be made by the physician and the patient, considering all the circumstances presented by the individual patient and the known variability and biological behavior of the medical condition. This review and its conclusions and recommendations reflect the best available information at the time the review was prepared. The results of future studies may require revisions to the recommendations in this review to reflect new data. The Society of NeuroInterventional Surgery does not warrant the accuracy or completeness of the review and assumes no responsibility for any injury or damage to persons or property arising out of or related to any use of this review or for any errors or omissions.

Competing interests The contents of this manuscript are solely the responsibility of the authors and do not necessarily represent the official views of the National Institutes of Health (NIH). JJH is a consultant for Medtronic and MicroVention, and a member of the medical and scientific advisory board for iSchemaView. JJH’s research is supported by the NIH (R01 NS121720-01, R01 NS075209, R01 NS11351704, R01 NS130172-01, and R01 EB032417-01A1). JRM is a consultant for Stryker. CMS has contract or research support from Penumbra, Medtronic, MIVI, Cerenovus, Stryker, and Balt, and ownership interest in NTI. JKB serves as a consultant for Stryker, Microvention, Cerenovus, Balt, Medtronic, Q’Apel Medical, Longeviti Neuro Solutions, and Siemens Healthineers. DR consults for Q’Apel, Penumbra, and Phenox. SWH is an ex officio member of the Society of NeuroInterventional Surgery Board of Directors. SWH’s research is supported by NIH (R01CA194533, R42CA265316, and R01EB012031). SWH has consulting agreements with Medtronic, Imperative, and Cerenovus and an ownership interest in Filtro. SWH’s institution has contract and grant support from Siemens, Stryker Neurovascular, and Route 92.

Provenance and peer review Not commissioned; externally peer reviewed.

Read the full text or download the PDF:

Log in using your username and password

Search More Search for this keyword Advanced search
Latest content
Current issue
Hosted content
BMJ Journals More You are viewing from: Google Indexer

http://orcid.org/0000-0002-5111-7861 Amit K Dutta 1 ,
http://orcid.org/0000-0003-2472-3409 Vishal Sharma 2 ,
Abhinav Jain 3 ,
Anshuman Elhence 4 ,
Manas K Panigrahi 5 ,
Srikant Mohta 6 ,
Richard Kirubakaran 7 ,
Mathew Philip 8 ,
http://orcid.org/0000-0003-1700-7543 Mahesh Goenka 9 ,
Shobna Bhatia 10 ,
http://orcid.org/0000-0002-9435-3557 Usha Dutta 2 ,
D Nageshwar Reddy 11 ,
Rakesh Kochhar 12 ,
http://orcid.org/0000-0002-1305-189X Govind K Makharia 4
1 Gastroenterology , Christian Medical College and Hospital Vellore , Vellore , India
2 Gastroenterology , Post Graduate Institute of Medical Education and Research , Chandigarh , India
3 Gastroenterology , Gastro 1 Hospital , Ahmedabad , India
4 Gastroenterology and Human Nutrition , All India Institute of Medical Sciences , New Delhi , India
5 Gastroenterology , All India Institute of Medical Sciences - Bhubaneswar , Bhubaneswar , India
6 Department of Gastroenterology , Narayana Superspeciality Hospital , Kolkata , India
7 Center of Biostatistics and Evidence Based Medicine , Vellore , India
8 Lisie Hospital , Cochin , India
9 Apollo Gleneagles Hospital , Kolkata , India
10 Gastroenterology , National Institute of Medical Science , Jaipur , India
11 Asian Institute of Gastroenterology , Hyderabad , India
12 Gastroenterology , Paras Hospitals, Panchkula , Chandigarh , India
Correspondence to Dr Amit K Dutta, Gastroenterology, Christian Medical College and Hospital Vellore, Vellore, Tamil Nadu, India; akdutta1995{at}gmail.com

https://doi.org/10.1136/gutjnl-2024-332154

Statistics from Altmetric.com

Request permissions.

PROTON PUMP INHIBITION
META-ANALYSIS

We read with interest the population-based cohort studies by Abrahami et al on proton pump inhibitors (PPI) and the risk of gastric and colon cancers. 1 2 PPI are used at all levels of healthcare and across different subspecialties for various indications. 3 4 A recent systematic review on the global trends and practices of PPI recognised 28 million PPI users from 23 countries, suggesting that 23.4% of the adults were using PPI. 5 Inappropriate use of PPI appears to be frequent, although there is a lack of compiled information on the prevalence of inappropriate overuse of PPI. Hence, we conducted a systematic review and meta-analysis on the inappropriate overuse of PPI globally.

Supplemental material

Overall, 79 studies, including 20 050 patients, reported on the inappropriate overuse of PPI and were included in this meta-analysis. The pooled proportion of inappropriate overuse of PPI was 0.60 (95% CI 0.55 to 0.65, I 2 97%, figure 1 ). The proportion of inappropriate overuse by dose was 0.17 (0.08 to 0.33) and by duration of use was 0.17 (0.07 to 0.35). Subgroup analysis was done to assess for heterogeneity ( figure 2A ). No significant differences in the pooled proportion of inappropriate overuse were noted based on the study design, setting (inpatient or outpatient), data source, human development index of the country, indication for use, sample size estimation, year of publication and study quality. However, regional differences were noted (p<0.01): Australia—40%, North America—56%, Europe—61%, Asia—62% and Africa—91% ( figure 2B ). The quality of studies was good in 27.8%, fair in 62.03% and low in 10.12%. 6

Download figure
Open in new tab
Download powerpoint

Forest plot showing inappropriate overuse of proton pump inhibitors.

(A) Subgroup analysis of inappropriate overuse of proton pump inhibitors (PPI). (B) Prevalence of inappropriate overuse of PPI across different countries of the world. NA, data not available.

This is the first systematic review and meta-analysis on global prescribing inappropriateness of PPI. The results of this meta-analysis are concerning and suggest that about 60% of PPI prescriptions in clinical practice do not have a valid indication. The overuse of PPI appears to be a global problem and across all age groups including geriatric subjects (63%). Overprescription increases the patient’s cost, pill burden and risk of adverse effects. 7–9 The heterogeneity in the outcome data persisted after subgroup analysis. Hence, this may be inherent to the practice of PPI use rather than related to factors such as study design, setting or study quality.

Several factors (both physician and patient-related) may contribute to the high magnitude of PPI overuse. These include a long list of indications for use, availability of the drug ‘over the counter’, an exaggerated sense of safety, and lack of awareness about the correct indications, dose and duration of therapy. A recently published guideline makes detailed recommendations on the accepted indications for the use of PPI, including the dose and duration, and further such documents may help to promote its rational use. 3 Overall, there is a need for urgent adoption of PPI stewardship practices, as is done for antibiotics. Apart from avoiding prescription when there is no indication, effective deprescription strategies are also required. 10 We hope the result of the present systematic review and meta-analysis will create awareness about the current situation and translate into a change in clinical practice globally.

Ethics statements

Patient consent for publication.

Not applicable.

Ethics approval

Abrahami D ,
McDonald EG ,
Schnitzer ME , et al
Jearth V , et al
Malfertheiner P ,
Megraud F ,
Rokkas T , et al
Shanika LGT ,
Reynolds A ,
Pattison S , et al
O’Connell D , et al
Choudhury A ,
Gillis KA ,
Lees JS , et al
Paynter S , et al
Targownik LE ,
Fisher DA ,

Supplementary materials

Supplementary data.

This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

Data supplement 1

X @drvishal82

Contributors AKD: concept, study design, data acquisition and interpretation, drafting the manuscript and approval of the manuscript. VS: study design, data acquisition, analysis and interpretation, drafting the manuscript and approval of the manuscript. AJ, AE, MKP, SM: data acquisition and interpretation, critical revision of the manuscript, and approval of the manuscript. RK: study design, data analysis and interpretation, critical revision of the manuscript and approval of the manuscript. MP, MG, SB, UD, DNR, RK: data interpretation, critical revision of the manuscript and approval of the manuscript. GKM: concept, study design, data interpretation, drafting the manuscript, critical revision and approval of the manuscript.

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.

Provenance and peer review Not commissioned; internally peer reviewed.

Read the full text or download the PDF:

IMAGES

The PRISMA 2020 statement: an updated guideline for reporting
Integrating qualitative research with trials in systematic reviews
Performing a literature review
Diagnosing dying: an integrative literature review
Diabetic retinopathy screening in persons with mental illness: a
What is the impact of clinically assisted hydration in the last days of

VIDEO

2014 Alfa Romeo MiTo Distinctive 1.4T 0-100km/h & engine sound
JFK Assassination Theories That Will Make Your Eyes Roll
Organize Your Life with the PRINLUCK Wood Phone Docking Station!
|| LIVE
RoboCop Police Cruiser from Jada Toys
They Chose the Wrong Operators to Mess With 😳

COMMENTS

Performing a literature review
Literature reviews are most commonly performed to help answer a particular question. While you are at medical school, there will usually be some choice regarding the area you are going to review. Once you have identified a subject area for review, the next step is to formulate a specific research question. This is arguably the most important ...
Five tips for developing useful literature summary tables for writing
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 ...
PDF The PRISMA 2020 statement: an updated guideline for reporting ...
the bmj | BMJ 2021;372:n71 | doi: 10.1136/bmj.n71 1 ... of the literature were used to inform the content of a ... review conduct, for which comprehensive resources are available.43-46 However, familiarity with PRISMA 2020 is useful when planning and conducting systematic
Reviewing the literature
Implementing evidence into practice requires nurses to identify, critically appraise and synthesise research. This may require a comprehensive literature review: this article aims to outline the approaches and stages required and provides a working example of a published review. Literature reviews aim to answer focused questions to: inform professionals and patients of the best available ...
PDF Reviewing the literature: choosing a review design
The purpose of a review of healthcare literature is primarily to summarise the knowledge around a specific question or topic, or to make recommendations that can support health professionals and organisations make decisions about a specific intervention or care issue.5 In addition, reviews can highlight gaps in knowledge to guide future research.
Rapid reviews methods series: Guidance on literature search
This paper is part of a series of methodological guidance from the Cochrane Rapid Reviews Methods Group. Rapid reviews (RR) use modified systematic review methods to accelerate the review process while maintaining systematic, transparent and reproducible methods. In this paper, we address considerations for RR searches. We cover the main areas relevant to the search process: preparation and ...
PDF Reviewing the literature
Reviewing the literature Joanna Smith,1 Helen Noble2 Implementing evidence into practice requires nurses to identify, critically appraise and synthesise research. This may require a comprehensive literature review: this article aims to outline the approaches and stages required and provides a working example of a published review.
Systematic review study protocol of literature from 2018 to ...
The search strategy was based on the strategy of a previous scoping review of the literature with a similar methodology conducted in 2017 but is slightly ... peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any ...
Appraising systematic reviews
Appraising systematic reviews. Systematic reviews may or may not include a meta-analysis of the primary RCTs identified. Although systematic reviews of RCTs with meta-analysis are often said to provide the most compelling evidence of effectiveness and causality, not all systematic reviews are of the highest methodological quality.
Article submission
At The BMJ, we use a system called ScholarOne to manage our submission processes.Essentially, ScholarOne will convert your manuscript to a PDF for the review process. Most common word processing formats are accepted for text and tables, although the system prefers Microsoft Word, and images should be submitted as GIF, TIFF, EPS, or JPEG files.
Identifying and understanding the factors that influence the ...
This review considered literature published in the years 1997-2020, as 'The New NHS', published on the 8 December 1997, represents one of the earliest examples of literature calling to 'replace the internal market with integrated care', thereby serving as the starting point for research activity on the topic of integration in the NHS.15
How to Write a Literature Review
Examples of literature reviews. Step 1 - Search for relevant literature. Step 2 - Evaluate and select sources. Step 3 - Identify themes, debates, and gaps. Step 4 - Outline your literature review's structure. Step 5 - Write your literature review.
Evidence-based medical leadership development: a systematic review
Methods. The design of this review was guided by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses18 and the Best Evidence in Medical Education (BEME) guide for systematic reviews.19 Our methods were based on the review conducted by Frich et al,8 with methodological changes drawn from other reviews.7 9 10 14 15 20 Following the BEME recommendations for systematic reviews ...
Reviewing the literature: choosing a review design
The purpose of a review of healthcare literature is primarily to summarise the knowledge around a specific question or topic, or to make recommendations that can support health professionals and organisations make decisions about a specific intervention or care issue. 5 In addition, reviews can highlight gaps in knowledge to guide future research.
Defining global health: findings from a systematic review and thematic
Method A systematic review was conducted with qualitative synthesis of findings using peer-reviewed literature from key databases. Publications were identified by the keywords of 'global health' and 'define' or 'definition' or 'defining'. Coding methods were used for qualitative analysis to identify recurring themes in definitions of global health published between 2009 and 2019.
PDF Defining global health: findings from a systematic review and thematic
SalmfiM, etflal BMJ Global Health 20216:e005292 doi:101136bmjgh-2021-005292 1 Defining global health: findings from a systematic review and thematic analysis of the literature Melissa Salm , 1 Mahima Ali,2 Mairead Minihane,2 Patricia Conrad3 Original research To cite: Salm M, Ali M, Minihane M, et al.Defining
Verbal autopsy in health policy and systems: a literature review
Methods A literature review was conducted including papers published between 2010 and 2017 according to a systematic search strategy. Inclusion of papers and data extraction were assessed by three reviewers. Thereafter, thematic analysis and narrative synthesis were conducted in which evidence was critically examined and key themes were identified.
Doctors' handovers in hospitals: a literature review
Aim To review studies on hospital doctors' handovers to identify the methods and main findings. Method A literature search of electronic databases Medline and Embase (via Ovid) was conducted against a set of inclusion criteria. Results A total of 32 papers were identified. The most common methods of studying handovers were observations and interviews, which typically focused on the sign-out ...
Equitable and accessible informed healthcare consent process for people
Qualitative methods of effect size calculations have been described in the literature,48 49 which was captured in this review by the number of studies that identified each subtheme, with an assigned frequency rating to compare their relative significance. Subthemes were given a frequency rating of A, B, C or D if they were identified by >10, 7 ...
Modifiable risk factors that may be addressed in routine care to
Chronic obstructive pulmonary disease (COPD) is a multisystem disease, and many patients have multiple conditions. We explored multimorbidity patterns that might inform intervention planning to reduce health-care costs while preserving quality of life for patients. Literature searches up to February 2022 revealed 4419 clinical observational and comparative studies of risk factors for ...
British Journal Of Midwifery
Midwives' practice of maternal positions throughout active second stage labour: an integrative review. An integrative review was considered suitable for this study, as this methodology allows inclusion of data from all types of literature to fully answer review questions (Whittemore and Knafl, 2005;...
PDF Identifying and understanding the factors that influence the ...
NHS: a systematic literature review Karthik Bhat , Rokshan Easwarathasan , Milan Jacob , William Poole , Vithullan Sapaetharan , Manu Sidhu , Ashvin Thomas ... BMJ Open: first published as 10.1136/bmjopen-2021-049296 on 5 April 2022. Downloaded from . BhatK etal Open 202212:e049296 doi:101136bmjopen2021049296 3
Focused update to guidelines for endovascular therapy for emergent
Methods A structured literature review and analysis of studies related to posterior circulation large vessel occlusion (basilar or vertebral artery) strokes treated by EVT was performed. Based on the strength and quality of the evidence, recommendations were made by consensus of the writing committee, with additional input from the full SNIS Standards and Guidelines Committee and the SNIS ...
Clinical and genetic spectrum of
Background Cases of RNF216 -related disorder have been reported sporadically. However, the clinical and genetic spectrum of this disorder has not been fully studied. Methods We identified an individual with a novel causative RNF216 variant in our institution and reviewed all individuals with causative RNF216 variants in previous reports. The clinical and genetic features of all the described ...
Inappropriate use of proton pump inhibitors in clinical practice
We read with interest the population-based cohort studies by Abrahami et al on proton pump inhibitors (PPI) and the risk of gastric and colon cancers.1 2 PPI are used at all levels of healthcare and across different subspecialties for various indications.3 4 A recent systematic review on the global trends and practices of PPI recognised 28 million PPI users from 23 countries, suggesting that ...

Log in using your username and password

You are here

Data availability statement

Statistics from Altmetric.com

WHAT IS ALREADY KNOWN ON THIS TOPIC

WHAT THIS STUDY ADDS

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

Introduction

Supplemental material

Preparation and planning

Information sources and methods

Choosing bibliographic databases

Further information sources and search techniques

Study registries and other grey literature

Further supplementary search methods

Considerations for RRs of RCTs

Search strategies

Developing search strategies

Limits and restrictions

Updating existing reviews

Quality assurance and search strategy peer review

Validation of search strategies

Peer review of search strategies

Reporting and record management

Ethics statements

Supplementary materials

Linked Articles

Read the full text or download the PDF:

Appraising systematic reviews

EBM Toolkit home

Search form

About The BMJ

Registration

Submitting a manuscript

Metadata page

Author details

File upload

Saving your work

After uploading

Keeping track of your manuscript

Submitting a revision/appeal

Commissioning

This week's poll

Log in using your username and password

You are here

Data availability statement

Statistics from Altmetric.com

Introduction

Search strategy

Supplemental material

Inclusion criteria

Screening process

Data abstraction

Study quality appraisal

Data analysis

Study reliability (MERSQI and JBI)

Programme design

Programme faculty

Participants

Educational methods

Educational content

Evaluation methods

Behavioural and organisational outcomes in higher reliability studies

Limitations and strengths

Ethics statements

Acknowledgments

Supplementary materials

Read the full text or download the PDF:

Log in using your username and password

You are here

Statistics from Altmetric.com

What is the purpose of undertaking a review of the evidence?

What approaches can be adopted to review the evidence?

Read the full text or download the PDF:

Log in using your username and password

You are here

Statistics from Altmetric.com

Key questions

What are the new findings?

What do the new findings imply?