How to interpret randomised trials of test–treatment combinations: a critical evaluation of research on uterine Doppler test to predict, and aspirin to prevent, pre‐eclampsia

Abstract

The methodology to evaluate either a test or treatment is well developed and understood. However, the design, conduct and interpretation of a combination of a test and a treatment in a single study are fraught with complexities. There are two distinct study designs to evaluate test-treatment combinations, with randomisation occurring either before or after the test. We present critical issues that need to be considered when evaluating and appraising test-treatment combinations, using the example of research on the uterine artery Doppler test to predict, and aspirin to prevent, pre-eclampsia.