In a systematic review, significant heterogeneity is detected across included studies (I² = 78%, p < 0.001). Which meta-analytic model is most appropriate to use, and what is the underlying rationale?
- A Fixed-effects model; it pools studies assuming a common true effect size
- B Fixed-effects model; heterogeneity does not affect the validity of pooled estimates
- C Random-effects model; it assumes true effect sizes vary across studies and accounts for between-study variance ✓
- D Meta-regression alone; pooling should be abandoned when I² > 50%
Correct answer: C. Random-effects model; it assumes true effect sizes vary across studies and accounts for between-study variance
Explanation
When heterogeneity is substantial (I² > 50–75% is considered high), a random-effects model is preferred because it incorporates the distribution of true effect sizes (tau²) rather than assuming all studies share one true effect. Fixed-effects pooling becomes misleading under high heterogeneity. While meta-regression can explore sources of heterogeneity, it does not replace pooling. An I² of 78% strongly indicates that variability is not due to chance alone.
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.