PRS70 APPLYING DIFFERENT MEASURES TO ASSESS PATIENTS' NON-ADHERENCE: RESULTS OF A LINKED DATA STUDY OF PATIENTS WITH ASTHMA IN GERMANY

Abstract

This analysis aimed to compare the impact of choosing a specific method to calculate patients’ non-adherence (NA) on results of such an analysis, based on a sample of asthma patients in Germany. Within this study, primary care data of patients with asthma were collected over 12 months, including prescriptions (Rx) and recommended dosage. All primary data were linked to patient-specific claims data (CD) from the same period (AOK PLUS database), on a patient-level using insurance numbers. Rx of all long-term (LT) asthma medications were extracted for a period of 12 months from CD. NA was measured by: (1) medication possession ratio (MPR; weighted average MPR across different agents), and (2) proportion of days covered (PDC; day covered if at least one medication was available). In either case, (a) prescription-based (between first and last Rx); and (b) interval-based (12 months from first Rx) approach was applied for NA assessment. Coverage of LT drugs was measured by either (I) defined daily dose (DDD) according to WHO, and (II) physician’s recommendation of prescribed daily dose. 406 patients were included (mean age: 55.48 years, female: 73.40%). (1) Prescription-based approach: The mean weighted MPR (prescription-based approach) was 64.8/80.6 using DDD/physician recommendation for measuring drug coverage. Accordingly, average PDC was 68.9/83.2. (2) Interval-based approach: Mean MPR was 54.7/69.3 (DDD/physician recommendation), and mean PDC was 60.5/74.5. Assuming a cut-off of <80 to identify NA, percentage of patients who were classified as NA in (1) prescription-based analysis was 67.49%/38.42% (MPR-DDD/physician recommendation); 58.50%/32.52% (PDC - DDD/physician recommendation). In (2) the interval-based methodology, it was 78.85%/55.51% (MPR), and 68.94%/45.15% (PDC). The impact of the methodology to measure adherence based on prescription data has a substantial impact on study results. We therefore recommend to choose the methodology carefully, and, above all, to report it transparently in every prescription-based adherence analysis.