Cochrane Evidence Synthesis and Methods最新文献

Cochrane's scientific strategy for 2025 to 2030 has four research priorities, including improving the lives of people living with multiple chronic conditions. The purpose of this article written by the Cochrane Thematic Group in Heart, Stroke and Circulation is to explore considerations around multiple chronic conditions (also referred to as ‘multiple long-term conditions’ i.e. two or more long-term conditions) in systematic reviews. Rather than using the term ‘comorbidity’, we introduce a new term ‘co-long-term conditions’. We also explore how to define ‘polyvascular disease’. We suggest that review authors consider co-long-term conditions and multiple long-term conditions in their reviews e.g. extract data about how primary studies address co-long-term conditions, perform subgroup analyses according to presence or not of co-long-term conditions, and include a section in the discussion about how well participants with co-long-term conditions were represented in the primary studies. This is especially pertinent for reviews addressing heart, circulatory or stroke disease, and polyvascular disease.

Reflecting a broader movement toward knowledge democratization, coproducing Cochrane evidence with interest holders outside universities is increasingly encouraged. However, only limited research exists on the approaches used to coproduce Cochrane reviews. Furthermore, the outcomes of coproduction are rarely described. In this commentary, we aim to address these gaps by describing the process and outcomes of coproduction used in a recently published Cochrane qualitative evidence synthesis (QES). We also reflect on power imbalances in our coproduction approach and how these could be minimized in future review coproduction activities.

This tutorial focuses on trials that assess outcomes by counting events that can occur zero, one, or more than one time in each participant. Trials and meta-analyses can estimate treatment effects for count outcomes using rate differences or rate ratios. We explain why it may be appropriate to meta-analyze count data to estimate rate ratios rather than odds ratios, risk ratios, or risk differences. We explain what count data are, how trials may estimate treatment effects, how to interpret such estimates, and how to extract data from trials that use count outcomes for meta-analysis. Finally, we discuss some common misunderstandings and subtleties. Supplementary materials include an Excel file for performing calculations, mathematical background, and additional advice.

The aim of this article is to present the most common statistical errors in meta-analyses included in systematic reviews; these are confusing standard deviation and standard error, using heterogeneity estimators for choosing between a common-effect and random-effects model, improper handling of multiarm trials, and unnecessary and misinterpreted subgroup analyses. We introduce some useful terminology and explain what authors can do to avoid these errors and how peer reviewers can spot them. We have also developed a micro-learning module to provide practical hands-on tutorial.

The common practice in meta-analysis and clinical practice guidelines is to derive the absolute treatment effect (also called risk difference, RD) from a combination of a pooled relative risk (RR) that resulted from a meta-analysis, and a user-provided baseline risk (BR). However, this method does not address the uncertainty in BR. We developed a web-based R Shiny tool to perform simple microsimulation and incorporate uncertainty in BR into the precision of RD. We empirically evaluated this approach by estimating the impact of incorporating this uncertainty when BR is derived from the control group rates in 3,128 meta-analyses curated from the Cochrane Library (26,964 individual studies). When BR was derived from the largest study in each meta-analysis, the median width of the CI of BR was 11.6% (interquartile range (IQR), 6.30%–18.5%). Incorporating this uncertainty in BR led to expansion of the RD CI by a median of 8 per 1,000 persons (IQR 2–24). This expansion increased in a linear fashion with BR imprecision and was more prominent in meta-analyses with low BR. This study provides a web-based tool to perform simple microsimulation and incorporate uncertainty in BR into the CI of RD.