Interpreting Randomized Controlled Trials

Simple Summary We provide an extensive review of the fundamental principles of statistical science that are needed to accurately interpret randomized controlled trials (RCTs). We use these principles to explain how RCTs are motivated by the powerful but strange idea that flipping a coin to choose each patient’s treatment is the most statistically reliable way to compare treatments. Random treatment assignment ensures fair comparisons between treatments because it does away with bias and confounding from variables other than treatment. If the goal is to estimate treatment effects in a patient population, rather than compare two or more treatments, then random sampling, not random treatment assignment, is required. However, random sampling is virtually impossible to carry out in a clinical trial because patients are accrued over time as they arrive in the clinic, subject to a trial’s entry criteria. Consequently, in practice, a trial provides a nonrepresentative convenience sample. Valid treatment comparisons provided by RCT data subsequently require additional causal assumptions of transportability of between-treatment effects from the sample to the patient population of interest. This may be used as a basis for choosing treatments for future patients. The present paper discusses what this means for practicing physicians who encounter RCT data in the literature. Abstract This article describes rationales and limitations for making inferences based on data from randomized controlled trials (RCTs). We argue that obtaining a representative random sample from a patient population is impossible for a clinical trial because patients are accrued sequentially over time and thus comprise a convenience sample, subject only to protocol entry criteria. Consequently, the trial’s sample is unlikely to represent a definable patient population. We use causal diagrams to illustrate the difference between random allocation of interventions within a clinical trial sample and true simple or stratified random sampling, as executed in surveys. We argue that group-specific statistics, such as a median survival time estimate for a treatment arm in an RCT, have limited meaning as estimates of larger patient population parameters. In contrast, random allocation between interventions facilitates comparative causal inferences about between-treatment effects, such as hazard ratios or differences between probabilities of response. Comparative inferences also require the assumption of transportability from a clinical trial’s convenience sample to a targeted patient population. We focus on the consequences and limitations of randomization procedures in order to clarify the distinctions between pairs of complementary concepts of fundamental importance to data science and RCT interpretation. These include internal and external validity, generalizability and transportability, uncertainty and variability, representativeness and inclusiveness, blocking and stratification, relevance and robustness, forward and reverse causal inference, intention to treat and per protocol analyses, and potential outcomes and counterfactuals.