Abstract
Background/Aims The Medication Possession Ratio (MPR) is a common claims-based metric for describing medication adherence. The MPR can only be calculated for patients who fill at least 2 prescriptions. Thus, patients with <2 prescription fills are excluded from the MPR denominator in population-level adherence reports. Combining prescription-ordering data available in an electronic health record (EHR) with matching claims data, we characterized the population excluded from a MPR-based adherence calculation.
Methods We conducted a retrospective (2004–2008) analysis of EHR and claims data for patients treated in the Geisinger Clinic. Hypertensive patients continuously enrolled for at least two years in the Geisinger Health Plan (GHP) with GHP pharmacy benefit coverage were identified using blood pressure (BP) data in the EHR and ICD-9 codes associated with clinic encounters. For all patients with an EHR order for an angiotensin converting enzyme inhibitor (ACEI) or an angiotensin receptor blocker (ARB), we matched each EHR order with the corresponding claim if one existed.
Results We identified 89,403 patients with hypertension, of whom 14,387 (16.1%) had at least 2 years of continuous pharmacy benefits. Median age was 67 years and 51% of the cohort was female. There were 4,934 patients with = 1 order for an ACEI and 1,251 patients had = 1 order for an ARB. The MPR could not be calculated for 18.8% (n=928) of patients based on ACEI orders and for 32.3% (n=404) of patients based on ARB orders. Patients for whom the MPR could not be calculated had a higher systolic BP than patients with an MPR measure (137 compared to 134 for patients with an ACEI; 139 compared to 134 for ARBs), and a higher diastolic BP (77.91 compared to 76.92 for ACEI; 76.16 compared to 75.33 for ARBs).
Conclusion In this primary care population, a sizable proportion of patients failed to fill more than one antihypertensive prescription. These patients are likely to be excluded from population level adherence estimates. Further analyses will examine differences between the “MPR” cohort and those patients excluded from the calculation.
