PRISMA Statement

To the Editor: In their recent commentary Takkouche and Norman1 seriously misrepresent the PRISMA guideline for reporting systematic reviews and meta-analyses. They state that PRISMA “demand[s] the prior registration of the protocol of any systematic review and meta-analysis, requiring that this protocol should be made accessible before any hands-on work is done.” No such demands are made in PRISMA. Item 5 of the checklist asks authors to “Indicate if a review protocol exists, if and where it can be accessed (such as Web address), and, if available, provide registration information including registration number.”2 The rationale for this request is: “registration may reduce unnecessary duplication of the same review question, particularly in jurisdictions with limited budgets; minimize study publication bias and selective reporting bias; and provide important information for those involved in updating systematic reviews.” That request anticipated the development of SR registration; indeed an international register is now being developed.3 Likewise, the PRISMA reporting guideline makes no statements as to the need for publication of a protocol prior to the initiation of the systematic review. It is possible that, as with the advent of clinical-trial registration, more avenues will appear whereby authors can publish protocols of such reviews. The authors also misrepresent the nature of registration, suggesting that this process would involve a judgement of originality and scientific merit of the planned research; we know of no such proposal. Takkouche and Norman1 also state “We also dispute the view that subgroup analyses and similar approaches should be used only if they were prespecified.” Again PRISMA makes no such demands on authors. Rather, the PRISMA explanatory paper says: “Having a protocol can help restrict the likelihood of biased post hoc decisions in review methods, such as selective outcome reporting.”2 In our explanatory article about PRISMA,2 we also recognized that “Authors may modify protocols during the research, and readers should not automatically consider such modifications inappropriate.” PRISMA seeks transparency and clarity: authors need to report how the systematic review was conducted to enable all readers to decide upon its merits. Lastly, we are not as confident as Takkouche and Norman about the effectiveness of peer review. A recent systematic review4 indicates that the peer review process is at best marginal and at worst ineffective. The scant research into the decision-making process of editors supports the need for clear and transparent reporting.5 David Moher Ottawa Methods Centre Ottawa Hospital Research Institute Ottawa, Ontario, Canada [email protected] Douglas G. Altman Centre for Statistics in Medicine University of Oxford Oxford, United Kingdom Alesandro Liberati Università di Modena e Reggio Emilia Centro Cochrane Italiano Milan, Italy Jennifer Tetzlaff Ottawa Methods Centre Ottawa Hospital Research Institute Ottawa, Ontario, Canada