Abstract:
|
A very general and flexible statistical approach to the analysis of randomized clinical trials uses all pairwise comparisons between two patients, one taken from the treatment group and one from the control group. Each pair favors treatment (“win”), control (“loss”), or neither. The “net treatment benefit” is the difference between the proportion of wins and the proportion of losses. We will discuss some advantages of the net treatment benefit over the win ratio or the win odds using examples from oncology and rare diseases. When a delayed treatment effect is anticipated, e.g. when testing the effect of immunotherapies on survival time, the net benefit can incorporate thresholds of clinical relevance, expressed on the time scale, to increase power. When testing new treatments in orphan indications, the power of the test can also be increased by analyzing several outcomes, either by prioritizing them from most important to least important, or by considering them simultaneously when categorizing pairs as wins or losses. When testing new treatments that have substantial toxicities, generalized pairwise comparisons can be used to assess the benefit-risk balance of these treatments.
|