Supporting Diversity and Inclusion Through the Evaluation of Differential Item Functioning (DIF) By Race and Ethnicity in the Asthma Control Questionnaire and Asthma Quality of Life Questionnaire Using Dupilumab Phase 3 Data

OBJECTIVES: There is a need to examine the extent to which patient-reported outcome (PRO) measures administered in clinical trials support the generalization of their respective endpoint results across diverse populations.

METHODS: Differential item functioning (DIF) was used to investigate if respondents had a different probability of endorsing a response on the Asthma Control Questionnaire – 5 items (ACQ-5) and Asthma Quality of Life Questionnaire (AQLQ) due to their race and ethnicity. Global Dupilumab clinical trial data of participants with persistent asthma (NCT02414854) were used to examine the presence of DIF between Hispanic and non-Hispanic; Black/African American and White; and Asian and White. Baseline and post-treatment data (i.e., Weeks 24 and 52) were used and DIF effect size (negligible: R²<0.13; moderate: 0.13≤ R²≤0.26 [with p-value significance]; large: R²>0.26 [with p-value significance]) was examined. These analyses were repeated using baseline data from United States (US)-based sites to assess the effect of culture/language on DIF direction.

RESULTS: For the analysis, 1902 participants were enrolled (including 490 Hispanic, 1412 non-Hispanic; 223 Asian, 80 Black/African American, 1577 White, 22 other races). DIF analysis of ACQ-5 showed negligible underestimation of asthma symptom severity for non-Hispanic versus Hispanic, negligible DIF in opposite directions between Asian and White, and no evidence of DIF between Black/African American and White. DIF analysis of AQLQ showed negligible underestimation of asthma burden on quality of life for Hispanic and negligible DIF in opposite directions between Black/African American and White, and between Asian and White. Negligible DIF was also found for US only participants (n = 423).

CONCLUSIONS: Negligible DIF was observed by racial and ethnic groups, and DIF directions showed no consistent trends in the overall and US populations. These results support the use of ACQ-5 and AQLQ for assessing treatment outcomes in asthma clinical trials enrolling diverse racial and ethnic populations.