Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement
BMJ 2009; 339 doi: https://doi.org/10.1136/bmj.b2535 (Published 21 July 2009) Cite this as: BMJ 2009;339:b2535
All rapid responses
Rapid responses are electronic comments to the editor. They enable our users to debate issues raised in articles published on bmj.com. A rapid response is first posted online. If you need the URL (web address) of an individual response, simply click on the response headline and copy the URL from the browser window. A proportion of responses will, after editing, be published online and in the print journal as letters, which are indexed in PubMed. Rapid responses are not indexed in PubMed and they are not journal articles. The BMJ reserves the right to remove responses which are being wilfully misrepresented as published articles or when it is brought to our attention that a response spreads misinformation.
From March 2022, the word limit for rapid responses will be 600 words not including references and author details. We will no longer post responses that exceed this limit.
The word limit for letters selected from posted responses remains 300 words.
David J Preiss raises an important point about the word length of abstracts and Trish Groves provides an excellent response including a historical context. We encourage authors reporting abstracts for journal articles or conference proceedings to consider using the CONSORT for Abstracts extension (1). We have shown that the CONSORT for Abstracts minimally increases the word count of abstracts and is well within the Medline 600 words mentioned by Groves. Examples of written abstracts prior to and after using the CONSORT for abstracts guidance can be seen at the CONSORT website (www.consort-statements/abstracts).
The same issues apply to reports of systematic reviews, although the PRISMA Group has not as yet developed specific guidance for abstracts of such articles. It may be that these articles require more words than reports of randomised trials, reflecting the larger number of sections recommended for structured abstracts of such articles (REF).
References
1. Hopewell S, Clarke M, Moher D, Wager E, Middleton P, Altman DG, Schulz KF and the CONSORT Group. CONSORT for reporting randomized controlled trials in journal and conference abstracts: explanation and elaboration. PLoS Med 2008;5(1):e20.
Competing interests: None declared
Competing interests: No competing interests
David J Preiss suggests that structured abstracts should have more than 250 words. The BMJ allows up to 400 words for a CONSORT-style abstract (this has proved enough to include all the important information), and I agree that we should do the same for a PRISMA abstract.
That 250 limit, used by many journals, probably harks back to the days when Medline cut abstracts off in midstream. But for the past 12 years Medline's been able to handle abstracts of up to about 600 words, as the US National Library of Medicine confirmed to us:
"There have been many changes to the NLM policy on abstracts and the length over the years, which will explain why some of the records you retrieve will have various messages (e.g., ABSTRACT TRUNCATED AT 250 WORDS or ABSTRACT TRUNCATED).
Effective in January 1996, author abstracts in MEDLINE are no longer truncated based on length of journal article. The only exception will be the occasional abstract that exceeds the maximum number of characters (4,096 [which is about 600 words]) permitted by the NLM online indexing system. And if this occurs, the following message will be at the end of the abstract: (ABSTRACT TRUNCATED). For additional information on this subject, there is a good (but brief) article, Full Abstracts in MEDLINE in the NLM Technical Bulletin (Sep-Oct 1995 issue) at http://www.nlm.nih.gov/pubs/techbull/tb.html".
Competing interests: I'm the BMJ's senior research editor and I maintain our advice to authors
Competing interests: No competing interests
The PRISMA Statement collaborators should be commended for their efforts to improve the quality of systematic reviews and meta-analyses, and their suggestions appear clear and logical [1].
The Statement suggests that the Structured Summary (or abstract) includes information relating to background, objectives, data sources used, study eligibility criteria, participants included, interventions, study appraisal, synthesis methods, results, limitations, conclusions and implications of the findings [2]. As an example they include an abstract with these headings. Importantly, however, the word count of this abstract is over 360 words, far more than the 250 words generally allowed during manuscript submission.
To support the useful recommendations of the PRISMA Statement, medical journals should therefore significantly increase the allowable word count for abstracts of systematic reviews and meta-analyses. Otherwise it will prove difficult to follow the guidance provided.
References:
1. Moher D, Liberati A, Tetzlaff J, Altman DG; PRISMA Group. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. BMJ 2009; 339:b2535.
2. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gøtzsche PC, Ioannidis JP, Clarke M, Devereaux PJ, Kleijnen J, Moher D. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare interventions: explanation and elaboration. BMJ 2009; 339:b2700
Competing interests: None declared
Competing interests: No competing interests
PRISMA Statement
RE: Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement
We congratulate the authors and The PRISMA Group on thoughtful and helpful work to improve the quality of systematic reviews. This is very needed, as they rightly point out.
One of the aspects of the checklist that we especially appreciate is the “’outcome-level’ assessment.” For example, in our own critical appraisal work, we have recognized a need to move more toward grading of individual study conclusions for validity and clinical usefulness as compared to simply grading an individual study as a whole. We have even found instances (albeit rare) in which a study is at high risk of bias, but a careful examination of evidence within the study, or sometimes within a body of evidence, may result in the discovery of an outcome that is reasonable to conclude as likely to be valid. An example is described at http://www.delfini.org/delfiniClick_PrimaryStudies.htm#westr ich. The outcome-level assessment concept is one which will help increase opportunities for determining which evidence is most likely to be reliable, and we strongly support this idea being advanced.
We would like to offer a few suggestions.
Participants The PRISMA checklist recommends, in Item 6, Eligibility Criteria, that reviewers “Specify study characteristics (e.g., PICOS…)”, the “P” referring to participants. Eligibility is important, but it is only part of the picture in providing information needed for informed decision- making. Frequently, we see participants described by inclusion and exclusion criteria only, which is what this suggests to us. While these are important, this is like publishing a guest list and omitting to report on who actually attended the party. Clinical relevance may be very different for a study for which men and women over the age of 21 were eligible, but which resulted in a range of participants, largely female of ages 60 and above. We feel it is important to create a “population-as-studied” summary statement, crafted from a review of baseline characteristics of participants. We think it would be helpful to reviewers and to those critiquing reviews to have this explicated in the PRISMA checklist.
Conflicts of Interest There may be other areas of conflict or potential conflict of interest of reviewers in addition to funding. We recommend this be explicitly called out in Item 27.
Again, we congratulate the authors and the group on a job incredibly well done. We attempt to be quite thorough in our own work and were pleased when we found ourselves in such agreement with the PRISMA Statement.
Competing interests: None declared
Competing interests: No competing interests