Meta-analyses: tests of heterogeneityBMJ 2012; 344 doi: https://doi.org/10.1136/bmj.e3971 (Published 13 June 2012) Cite this as: BMJ 2012;344:e3971
- Philip Sedgwick, senior lecturer in medical statistics
- 1Centre for Medical and Healthcare Education, St George’s, University of London, Tooting, London, UK
Researchers investigated the association between consumption of white rice and type 2 diabetes. They performed a meta-analysis of prospective cohort studies that reported the relative risk of type 2 diabetes by intake of white rice (high v low). In total, four publications were identified that included seven distinct prospective cohort analyses in Asian and Western populations. Rice intake and type 2 diabetes were identified through self report. A total of 13 284 incident cases of type 2 diabetes were ascertained among 352 384 participants with follow-up periods ranging from four to 22 years.1
For each study the researchers identified the relative risk of type 2 diabetes for high consumption of white rice compared with low intake. Statistical tests of heterogeneity were undertaken across the seven sample estimates (Cochran’s Q test, P=0.001; I2=72.2%). The overall relative risk was 1.27 (95% confidence interval 1.04 to 1.54). The researchers concluded that higher consumption of white rice was associated with a significantly higher risk of type 2 diabetes.
Which of the following statements, if any, are true for the statistical test of heterogeneity?
a) Null hypothesis: heterogeneity exists between the sample relative risks as estimates of the population parameter
b) Statistical heterogeneity existed between the seven sample estimates of the population relative risk
c) A random effects model …