Children from marginalized racial and ethnic groups are underrepresented in health research. To improve external validity and routinize race and ethnicity reporting, a specific and standardized methodology for quantifying representativeness of participant populations is needed. To develop a standardized method for quantifying the racial and ethnic representativeness of study samples. In this cross-sectional study, data from 7 US community-based health studies (conducted between 2003 and 2017) were retrospectively pooled to assess the school-level representativeness of enrolled samples by race and ethnicity. The sampling frame for the study was constructed using the National Center of Education Statistics Common Core of Data, which provides year-specific racial and ethnic counts by grade. Representativeness was quantified by aggregating children's data at the school level, reported individually for Asian, Black, Hispanic or Latino, Native Hawaiian or other Pacific Islander, White, or multiple races. In this analysis, the Asian and Native Hawaiian or other Pacific Islander subgroups were combined. Data were analyzed from April 1 to June 15, 2022. Community-based nutritional health studies conducted with children in grades 1 to 8. Visual comparisons of percentage expected and percentage observed of the pooled sample by race and ethnicity were performed using scatterplots and Bland-Altman plots. Spearman rank-order correlation was used to assess associations. This study included 104 study schools (N = 5807 children) located in California, Kentucky, Massachusetts, Mississippi, and South Carolina. Bland-Altman analysis revealed notable patterns and variability in the representativeness of racial and ethnic groups. Differences in the overall representativeness of Asian or Native Hawaiian or other Pacific Islander children (0.45 percentage points [95% CI, -7.76 to 8.66]), Black children (0.12 percentage points [95% CI, -15.73 to 15.96]), and White children (-0.72 percentage points [95% CI, -23.60 to 22.16]) were negligible, but measures of spread suggested that target population demographics affected representativeness differently across groups. The results of this cross-sectional study suggest that replicating, testing, and scaling the proposed method for quantifying racial and ethnic representativeness, which uses measures of spread, could improve the transparency of race and ethnicity reporting during publication and lead to a more externally valid health evidence base. During implementation, investigators should adopt community-based research methods and allocate appropriate resources during recruitment, including a priori assessment of population demographics, as these conditions may affect racial and ethnic study enrollment differently. Prioritizing these methodological decisions could alleviate rising inequities.
Read full abstract