A meta-analysis of randomised controlled trials evaluating a new antihypertensive drug finds high heterogeneity (I² = 78%). The most appropriate statistical model and the most appropriate interpretation are:

• A Fixed-effects model; results represent the single true effect across all populations
• B Random-effects model; results represent the average of a distribution of true effects ✓
• C Fixed-effects model; heterogeneity is ignored if p-value for Cochran Q is >0.10
• D Random-effects model; pooling should not be attempted when I² exceeds 50%

Explanation

An I² value of 78% indicates substantial statistical heterogeneity (>75% is considered considerable by Cochrane criteria), meaning the variability in results cannot be attributed to chance alone. The random-effects model accounts for this by assuming each study estimates a different but related true effect, yielding a summary that represents the mean of a distribution of effects. The fixed-effects model assumes one common true effect, which is inappropriate with such high heterogeneity. While very high heterogeneity warrants subgroup analysis and clinical appraisal, it does not automatically prohibit pooling, making option D incorrect.

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.