Genevois E, Lelouer V, Vercken J B, Caillon R
Rhône-Poulenc Rorer Central Research, Biostatistics Department, Collegeville, PA 19426, USA.
J Antimicrob Chemother. 1996 May;37 Suppl A:65-72. doi: 10.1093/jac/37.suppl_a.65.
Many publications in the past 10 years have emphasised the difficulties of evaluating anti-infective drugs and the need for well-designed clinical trials in this therapeutic field. The clinical development of sparfloxacin in Europe, involving more than 4000 patients in ten countries, provided the opportunity to implement a methodology for evaluation and statistical analyses which would take into account actual requirements and past insufficiencies. This methodology focused on a rigorous and accurate patient classification for evaluability, subgroups of particular interest, efficacy assessment based on automation (algorithm) and individual case review by expert panel committees. In addition, the statistical analyses did not use significance testing but rather confidence intervals to determine whether sparfloxacin was therapeutically equivalent to the reference comparator antibacterial agents.
