OBJECTIVES:

A randomized phase III clinical trial was conducted to show that a drug could prevent severe adverse events (AEs) from cancer therapy. However, the efficacy analysis of the primary endpoint (incidence of severe AEs) was far less impressive in this trial than in earlier phase trials. Yet, the drug exhibited other signals of efficacy: it reduced the duration of severe AEs, and delayed their onset. Hence it was suggested to analyze three prioritized outcomes in a single analysis: incidence of severe AEs, duration of AEs in days, and time to onset of AEs in days.

METHODS:

We applied the novel statistical methodology of Generalized Pairwise Comparisons (GPC) to rigorously capture the multidimensional benefit of the drug. Using GPC and its quantitative multi-component metric, the Net Treatment Benefit, we provided statistical evidence that the overall effect of the drug consists in a composite of reduced occurrence, decreased duration and delayed onset of severe adverse events.

RESULTS:

While the univariate GPC on the occurrence of the toxicity results in a Net Treatment Benefit of 6.1% (p= .072, 95% CI: -4.9%, 17.1%), integrating the overall effect of the drug in a multivariate GPC leads to a significant Net Treatment Benefit of 18.5% (p= .001, 95% CI: 10.4%, 26.6%).

CONCLUSIONS:

GPC can highlight the multidimensional efficacy of a drug, enhancing the clinical relevance of the analysis for clinicians and patients, and making optimal use of the whole set of available data. By analyzing several outcomes in a single analysis, GPC has the potential of increasing the statistical power as well as the clinical relevance of the results.