Comparison of meta-analytic results of indirect, direct, and combined comparisons of drugs for chronic insomnia in adults: a case study.

BACKGROUND Our Center recently conducted a systematic review of the manifestations and management of chronic insomnia in adults. The efficacy and safety of benzodiazepines and nonbenzodiazepines, relative to placebo, were compared indirectly. OBJECTIVES Determine how the results of indirect comparisons made in the review compare with the results of direct comparisons, as well as with estimates derived from Bayesian mixed treatment comparisons. Establish general appropriateness of the use of results of indirect or mixed treatment comparisons. METHODS Treatments were compared using frequentist direct, indirect, and combined methods, as well as Bayesian direct and mixed methods. RESULTS Estimates for comparisons tended to be clinically and statistically similar across methods. Estimates obtained through indirect comparisons were not biased and were similar to those obtained through direct analysis. CONCLUSIONS Results of indirect comparisons made in the review, accurately reflected the current evidence. Frequentist and Bayesian methods of analysis of indirect comparisons should be considered when performing meta-analyses.