In a meta-analysis of randomized trials of a new antidiabetic agent, the funnel plot shows marked asymmetry, with smaller trials clustered on the side showing large treatment benefit. What does this pattern most strongly suggest?

• A Genuine dose-response relationship across trials
• B Random misclassification of the exposure
• C Reverse causation between drug use and outcomes
• D Publication bias or small-study effects ✓

Explanation

In the absence of bias, trial estimates scatter symmetrically around the pooled effect, with spread narrowing as sample size increases. Asymmetry, especially when small studies cluster on the beneficial side, suggests that small trials with negative results remain unpublished, producing publication bias. This inflates the pooled estimate. Egger's regression test or trim-and-fill methods quantify this asymmetry. Dose-response and reverse causation are causal inference concepts unrelated to funnel plot shape.

Reference: Park's Textbook of Preventive and Social Medicine, 27th ed.

High-yield for: NEET PGINI-CETNExTFMGEUSMLEPLABMRCP

Written and medically reviewed by the StethoPrep medical team.