OBJECTIVES: Published pharmacoepidemiological studies use MPR and PDC to calculate medication adherence and usually fail to incorporate hospitalization and coverage overlap/gap from claims data. We created a refined algorithm to capture medication exposure level(MEL) for patients with End Stage Renal Disease(ESRD) who are more frequently hospitalized.

METHODS: United States Renal Data System(USRDS) registry of ESRD claims (Medicare PartA- PartB-PartD) were used to accurately capture entire prescription filling patterns of those receiving new prescriptions for oral P2Y12inhibitors, including clopidogrel(95.19%), prasugrel(2.67%),ticagrelor (2.14%) between 2011-2015. P2Y12 naïve patients were followed until death, kidney transplantation, switching medications, or lost to follow-up. Resulted in patients with median study periods of 367days(IQR:147,1627). After flagging/censoring key variables, algorithm adjustments were made for hospital length of stay(LOS), and medication overlap. Analyses were generated with SAS software.

RESULTS:Hospitalization occurred for 78% of the cohort (n=46,514). Median LOS was 12(IQR:2,34). Median age was 64(IQR:55,73), 54% male, Caucasian(40.6%), majority on hemodialysis(92.9%) with dialysis vintage of 3.8years, modified Liu Index median was 7. Patients were on 7 medications (median:7,IQR:5,10) on the day new P2Y12 was prescribed. The four most common medications were: clopidogrel(95.20%), sevelamer carbonate(24.79%), amlodipine besylate(24.79%), carvedilol(24.49%). Calculated MPR(61.10%, 95%CI:60.79%-61.41%), PDC(59.00%, 95%CI:58.70%-59.30%) using traditional adherence methods and compared findings with refined MEL calculation. MEL adjustment affected 41% of the cohort and 52.7% of the cohort after applying modification for overlap coverage days and hospital stay, respectively. When adjustments for overlap and hospital stay were made concurrently, medication adherence improved 5.78%. Leading to overall MEL 68%(median 0.68,IQR:0.31,0.93). Sensitivity analysis considered restricted medication counts of ≥14days and first ≥30day gap.

CONCLUSIONS: Patients had higher medication exposure level after adjusting for coverage overlap, gap, hospital stays, and other data refinement. This sophisticated algorithm is applicable to other disease condition where all-cause hospitalization burden is higher. Patient insurance eligibility status and hospitalization significantly influence MEL.