Assessing a missing hazard ratio and confidence interval from a result of a published subgroup analysis

Abstract

Results of subgroup analysis provide researchers valuable information in establishing research hypothesis for new clinical studies. However, it is often the case that the subgroups employed in a published paper miss the subgroup that researchers are interested in. We develop in this paper a method for assessing hazard ratio and its confidence interval in any subgroup that researchers are interested in based on the results of subgroup analysis given in a published paper, by focusing on a comparative clinical study of treatment group vs. control group having survival time as the primary endpoint, where subgroups are made by dividing the range of continuous bio-marker. The method is applied to Phase III clinical trial of comparing Nivolumab vs. Docetaxel for lung cancer where PD-L1 expression level is used to establish subgroups.

Keywords: biomarker, confidence interval, Cox proportional hazards model, hazard ratio, subgroup analysis

線上視訊請點選連結

附件下載