A systematic review finds five RCTs on a new drug for hypertension. The Cochrane I² statistic is 78%. The most appropriate next step in the meta-analysis is:
- A Use a fixed-effects model because I² > 50% means more precise estimates
- B Use a random-effects model and explore sources of heterogeneity through subgroup analysis ✓
- C Exclude all studies from the meta-analysis due to unacceptable heterogeneity
- D Convert the meta-analysis into a narrative review without pooling
Correct answer: B. Use a random-effects model and explore sources of heterogeneity through subgroup analysis
Explanation
I² > 75% indicates substantial heterogeneity, but this does not automatically preclude pooling; it mandates use of a random-effects model which accounts for between-study variance (τ²) and exploration of heterogeneity sources via subgroup or meta-regression analysis. The fixed-effects model assumes all studies estimate the same true effect and is inappropriate with substantial heterogeneity. Excluding all studies or converting entirely to narrative review discards valuable quantitative information unnecessarily.
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.