PhD candidate University of Colorado, Aurora, Colorado Denver, United States
Background: Prevalence of exposure to pharmacogenomic (PGx) medications is well established but less is known about exposure duration. Quantifying PGx medication exposure provides insight into patients who may benefit from PGx testing. Previous studies assessed frequency of PGx medication use with prescribing data from electronic health records, but prescription fill data can better estimate medication exposure using adherence measures, like proportion of days covered (PDC).
Objectives: Quantify cumulative exposure to actionable PGx medications among incident users using real-world data from a national insurance claims database.
Methods: This was a retrospective cohort study of adults ≥18 years with incident fills of CPIC level A, A/B and B PGx medications from 2012-2018. Patients were identified from the IQVIA PharMetrics® Plus for Academics claims database using generic product identifier codes. The primary outcome was PDC, measured as the total days covered (i.e., total days supplied) divided by the days in a period. Patients were followed for 1 year from the first fill. Student’s t-tests compared PDC across sex and the presence of comorbidity. ANOVA compared PDC across age groups. Pearson correlation coefficient (r) assessed correlation between PDC and incident fills of a medication.
Results: There were 1,035,918 incident PGx medication fills among 605,355 patients. The study included 72 CPIC PGx medications: 48 (67%) level A, 7 (10%) level A/B and 17 (24%) level B. Mean PDC for all medications was 0.21 (SD 0.3), indicating the average patient was exposed to their medication 21% of the year. Mean PDC was highest for medications in CPIC level A/B (0.44), level A (0.24) then level B (0.10). The highest PDCs were ivacaftor, tamoxifen, clopidogrel, HIV medications, tetrabenazine, immunosuppressants and statins (mean 0.55-0.89). Analgesic and antiemetic medications (e.g., hydrocodone, ibuprofen, ondansetron) were among the most frequently filled medications but had the lowest PDCs (mean 0.03-0.06). There was a negative correlation between mean PDC and incident fills (r=-0.343, p=0.003). Mean PDC increased with age (p < 0.001). Men had a higher mean PDC compared to women overall (0.23 vs 0.18, p< 0.001) and across age groups (p < 0.001 to p=0.003). Patients with ≥1 comorbidity had a higher mean PDC compared to patients with none (0.27 vs 0.17, p< 0.001).
Conclusions: Measures of adherence like PDC can quantify exposure to PGx medications to inform PGx testing opportunities. Medications with high PDCs could be targets for PGx testing after the decision to prescribe (i.e., reactive testing). Populations with more medication exposure, like older adults and those with comorbidities, may benefit from empiric PGx testing to inform future prescribing (i.e., preemptive testing).
