Objectives: To examine whether cluster randomized trials (1) produce baseline imbalances between intervention and control conditions; (2) give results that are substantially different individually randomized trials; and (3) give different results when adjusted for unit of analysis error.
Study design and setting: We used 14 cluster randomized trials and 20 individualized trials of the same intervention (collaborative care for depression). We conducted a random effects meta-analysis to examine imbalance in baseline depression scores. We used meta-regression to test for differential effect size and heterogeneity between clustered and individualized studies. Unit of analysis error was corrected using a range of plausible published intraclass correlation coefficients (ICCs).
Results: There were no baseline imbalances in either cluster randomized (P=0.837) or individually randomized (P=0.737) studies. Cluster randomized studies gave almost identical estimates of effect size when compared to individually randomized studies (standardized mean difference, SMDcluster=0.25, 95% confidence interval [CI]: 0.17, 0.33; SMDindividual=0.24; 95% CI: 0.13, 0.36). Adjustment for clustering had minimal effect on clinical and statistical significance (pooled SMDICC 0.02=0.249 [95% CI: 0.174, 0.325] to SMDICC 0.05=0.258 [95% CI: 0.172, 0.345]).
Conclusion: The additional effort and expense involved in cluster randomized trials needs to be justified when individualized studies might produce robust and believable results.