Critical appraisal of systematic reviews and meta-analyses in English

A systematic review is a structured summary of all available evidence on a clearly defined question, using explicit and reproducible methods. A meta-analysis is the statistical technique that combines the results of several independent studies to produce a single pooled estimate. In medical English, you will often read: 'We conducted a systematic review and meta-analysis of randomised controlled trials.' The distinction is crucial: you can have a systematic review without a meta-analysis, but a meta-analysis should always be based on a systematic review. For example, a Cochrane review on a treatment for hypertension summarises dozens of trials, then may pool their results in a forest plot.

When you read a systematic review, first ask: Were the inclusion and exclusion criteria explicit? Was the literature search comprehensive and reproducible? Did the authors assess the risk of bias of each included study? In English, 'risk of bias' means the extent to which the design and conduct of a study may lead to systematic error. Tools like the Cochrane Risk of Bias tool help evaluate domains such as random sequence generation, allocation concealment, and blinding. A review with high risk of bias in most studies should be interpreted with caution, even if the meta-analysis looks precise.

A forest plot is the main graph in a meta-analysis. Each horizontal line represents one study: the square is the point estimate, and the line is its confidence interval. The diamond at the bottom shows the pooled effect and its confidence interval. If the diamond does not cross the line of no effect (1 for ratios, 0 for differences), the result is statistically significant. Heterogeneity is measured by the I² statistic: I² = 0% means no observed heterogeneity, I² > 50% suggests substantial variability between studies. In English, you will read: 'There was significant heterogeneity (I² = 72%, p < 0.01), so we used a random-effects model.'

Before applying a meta-analysis to patient care, check for publication bias: studies with negative or non-significant results are less likely to be published, which can exaggerate the pooled effect. A funnel plot can reveal asymmetry suggesting missing small studies. Reporting guidelines like PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) help you verify that the review was conducted transparently. Finally, the GRADE approach rates the certainty of evidence from high to very low. In English, you might say: 'This meta-analysis of low-risk-of-bias trials provides moderate-certainty evidence that the drug reduces mortality.'