Article type
Year
Abstract
Background: The methodology to conduct systematic reviews (SR) of prognostic issues is not yet fully developed and unanimously agreed. In particular, risk of bias assessment and quality scoring of research (both primary studies and SR) are meant to be largely inconsistent. Objectives: To measure the role for SR dealing with prognostic issues, using the following bibliometric indicators: proportion of all SR over all published articles, proportion prognostic SR over all systematic reviews, temporal trends in those proportion, adherence of prognostic SR to PRISMA and MOOSE guidelines. Methods: JAMA, BMJ, Annals of Internal Medicine, Circulation, Stroke and Blood were searched for: a) total number of articles b) number of SR c) number of prognostic SR. The figures were extracted for any year from 2000 to 2009. The search for items a) and b) were performed on PubMed, while item c was hand-searched. The SR about prognosis were analyzed with the MOOSE and PRISMA checklist. All searched and assessments were performed by two of the authors. Results: Main results are shown in Figure 1. While the overall number of original article remained the same, SR and particularly prognostic SR showed a significant increase over the last 10 years. JAMA and Stroke showed the steeper increase in the proportion of prognostic SR, Blood the flattest profile with a mean of 1 SR per year. The fulfillment of PRISMA and MOOSE requirements was generally low. Conclusions: Notwithstanding the methodological flaws inherent in prognostic research, the space allocated in core clinical journal to prognostic systematic reviews increased over the last 10 years.
PDF