Measurement of ethnicity in clinical trials: Delphi survey and consensus statement

Ethnicity is a dynamic, multidimensional social construct shaped by cultural identity, language, and customs. It influences health outcomes and contributes to health inequities, and is therefore important to measure and report on. With no current consensus suggesting consistent collection of specific ethnicity related variables, ethnicity is often omitted from studies, including randomised clinical trials. Therefore, a core group was assembled to oversee the developments of evidence informed, consensus based recommendations to guide trialists on planning for and reporting a set of variables related to ethnic diversity. Delphi respondents considered 35 ethnicity related variables and identified self-identified ethnicity as important and critical to all trials (>81% endorsement) and 23 other variables as important to all trials. These variables were synthesised by the consensus panel into an essential minimum list of 18 variables, based on three categories: critical variables, which should be prioritised to capture ethnic diversity; context dependent variables that may be important but are not critical to every trial; and variables that are important only for select trials. Critical variables include country of birth, self-identified ethnicity and/or race, and language. Caveats on using some variables (eg, race) are described. Our guidance offers a standard for how ethnicity related variables are collected and reported in trials. These recommendations may also be useful for other types of study designs.