Understanding treatment response heterogeneity using crossover and n-of-1 randomised controlled trials in exercise and nutrition research: A Primer

Crossover randomised controlled trials (CRCT) are common in exercise/nutrition sciences. Because researchers randomise participants to different sequences of both the treatment/intervention and comparator/control conditions, CRCTs are powerful for detecting mean treatment effects under certain circumstances. We aim to review the information that researchers can derive from CRCTs about treatment response heterogeneity - a fundamental issue in precision medicine for tailoring treatments to individuals. After covering CRCT design issues, we describe the variance components that underlie observed data. Researchers cannot derive the crucial participant-by-treatment interaction from a typical single-cycle CRCT, but this interaction can be quantified from a “replicate” CRCT by exposing participants to multiple cycles of experimental conditions. Related to n-of-1 trials, replicate CRCTs have important design, statistical power, and analysis considerations. By synthesising findings from six published replicate CRCTs, we compared the various data analysis approaches. We found general agreement between approaches, and a logical link between within-person consistency of response and detection of a participant-by-treatment interaction. It may be that a paired “variance comparison”, e.g., the Pitman-Morgan test, can provide some initial exploratory information regarding response heterogeneity from a single-cycle CRCT. Nevertheless, underlying assumptions are critical, and researchers cannot draw inferences about responses of specific individuals without repeat cycles of experimental conditions. Replicate CRCTs or n-of-1 studies are underused but necessary for robust inferences about personalised responses to interventions. However, these trials are still only one component of the process for predicting individual magnitude of response from any personal traits, which is the “holy grail” of personalised treatment.