Supplementary data: Evaluation of the treatment patterns among commercially insured patients with nonvalvular atrial fibrillation prescribed an oral anticoagulant by race/ethnicity

These are peer-reviewed supplementary materials for the article 'Evaluation of the treatment patterns among commercially insured patients with nonvalvular atrial fibrillation prescribed an oral anticoagulant by race/ethnicity' published in the Journal of Comparative Effectiveness Research.Table S1: Pre-IPTW demographic and baseline clinical characteristics in apixaban-warfarin cohortTable S2: Pre-IPTW demographic and baseline clinical characteristics in DOAC-warfarin cohortTable S3: Pre-IPTW demographic and baseline clinical characteristics among Asian/Pacific Islander and Other/Unknown race/ethnicity subgroups in apixaban-warfarin cohortTable S4: Pre-IPTW demographic and baseline clinical characteristics among Asian/Pacific Islander and Other/Unknown race/ethnicity subgroups in DOAC-warfarin cohortTable S5: Proportion of discontinuers and TTD in apixaban-warfarin and DOAC-warfarin cohorts (sensitivity analysis)Aim: Despite evidence of racial/ethnicity health inequities in oral anticoagulant initiation and clinicaloutcomes for nonvalvular atrial fibrillation, little is known about disparities in treatment patterns,particularly discontinuation and switching. Materials & methods: This retrospective study utilizedKomodo Healthcare Map data (1 July 2018 to 30 June 2023) to examine treatment patterns in patientswith nonvalvular atrial fibrillation initiating direct oral anticoagulants (DOACs: apixaban, rivaroxaban,dabigatran) versus warfarin, stratified by self-reported race/ethnicity (White, Black and Hispanic) in theUSA. Inverse probability of treatment weighting balanced baseline covariates and Cox model computedadjusted hazard ratio (aHR) and 95% confidence interval (CI) for apixaban-warfarin and DOAC-warfarincomparisons in the overall cohort and by race/ethnicity. Results: In the apixaban-warfarin cohort,apixaban patients were 24% less likely to discontinue treatment (aHR: 0.76, 95% CI: 0.75–0.77) and 79%less likely to switch to another oral anticoagulant (aHR: 0.21, 95% CI: 0.20–0.22) versus warfarin. Byrace/ethnicity, a higher proportion of Black patients discontinued than White and Hispanic (68, 60 and63%, respectively). Almost 5%of apixaban patients switched to another oral anticoagulant versus 31% onwarfarin. Among warfarin switchers, 77% initiated apixaban (80% in White; 76% Black; 76% in Hispanicpatients). The DOAC-warfarin cohort followed similar trends. Conclusion: Overall, apixaban/DOACs wereassociated with more favorable treatment patterns than warfarin. However, racial/ethnicity differenceswere observed. Given the potential impact of oral anticoagulant discontinuation on clinical outcomes,further research is needed to better understand factors contributing to higher discontinuation and switchrates, particularly among Black and Hispanic patients.