All Times EDT
Keywords: Mann-Whitney test, Patient Preferences, Composite Endpoint, Heterogeneity
Regulatory approval of a medical product is usually based on the balance of benefits and risks that patients may experience while taking the product. For many diseases, a medical product can benefit patients in multiple relevant endpoints. The importance of these endpoints may vary among patients. Current approaches of composite endpoints integrate multidimensional outcomes based on one hierarchical structure and don’t reflect heterogeneity of patient preferences. In this paper, we proposed a stratified randomization and composite winning probability approach to evaluate the treatment benefits. In such trials, randomized is stratified according to the baseline preference ranking of the outcome variables. Treatment benefit is evaluated using a Mann-Whitney U-statistics of pairwise comparison between patients from the treatment and the control arms. For patients from each arm in the same strata, the highest ranked outcomes that can separate the benefits between patients determined the success or failure of the treatment or tie if outcomes are identical. If a pair from different strata, the smallest commonly outcomes among the common highest rank will be compared. In that case, if the treated patient has uniformly better (or worse) outcomes than that of the controlled patient, the treatment succeed (or failed). Otherwise, the pair has a tie. We can estimate the wining probability of treatment over control arms and make statistics inferences. In this talk, we will discuss mathematical properties and its difference from other common approaches and demonstrate the benefits via simulations. We will analyze a trial of TCM for MFH diseases where physicians pre-specified the individualized treatment goals. We will also discuss our experience using this approach in the development of a shared decision making trial for AFib stroke prevention.