Home

Welcome to the Clinical Trial Equity wiki!

This is a supplement of the paper "Quantifying Representativeness in Randomized Clinical Trials using Machine Learning Fairness Metrics".

Objective

We formulate representativeness of randomized clinical trials (RCTs) as a machine learning (ML) fairness problem, derive new representation metrics, and deploy them in visualization tools which help users identify subpopulations that are underrepresented in RCT cohorts with respect to national or community-based target populations.

Materials and Methods

We represent RCT cohort assignments as random binary classification fairness problems, and then prove how ML fairness metrics based on RCT enrollment fraction can be efficiently calculated using easily computed rates of subpopulations in RCT cohorts and target populations. We propose standardized version of these metrics and deploy them in an interactive tool to analyze three RCTs with respect to type-2 diabetes and hypertension target populations in the National Health and Nutrition Examination Survey.

Results

We demonstrate how the proposed metrics and associated statistics enables users to rapidly examine representativeness of all subpopulations in the RCT defined by a set of categorical traits (e.g. sex, race, ethnicity, smokers, and blood pressure) with respect to target populations.

Discussion

The normalized metrics provide an intuitive common scale for evaluating representation across subgroups with vastly different enrollment fractions and rates in RCT cohorts. The metrics are beneficial complements to existing approaches (e.g. enrollment fractions and GIST) which are used to identify generalizability and health equity of RCTs.

Conclusion

By quantifying the gaps between RCTs and target populations, the proposed methods can support generalizability evaluation of existing RCTs, design of new RCTs, and the monitoring of RCT recruitment, ultimately contributing to more equitable public health outcomes.