Claims of equivalence in randomized controlled trials of the treatment of bacterial meningitis in children

Abstract

Objective: To evaluate claims of therapeutic equivalence in studies of the treatment of bacterial meningitis in children.

Methods: We performed a systematic review of randomized controlled trials of antimicrobial therapy for bacterial meningitis in children indexed in MEDLINE and published after 1980 and that claimed equivalency. The sample size of each trial was compared with the minimum sample size needed to rigorously claim equivalence. The primary endpoint was case fatality.

Results: Twenty-five studies were identified that met the inclusion criteria. Two of these were specifically designed to test equivalence, and the remaining based claims of equivalence on failed tests of superiority. The majority of these trials (24 of 25) that claimed equivalence had sufficient sample size to exclude a 20% difference in mortality between the tested therapies. Only 3 of the 25 trials could exclude a 10% difference in mortality.

Conclusion: Few of the trials in this study had sufficient sample size to claim equivalence within 10% of the expected mortality. Proving equivalency is challenging because large sample sizes are often needed to ensure adequate statistical power to rule out clinically important differences between the standard of care and new therapies.