2016. Association between Hypertension Control and Adherence Measures Using Self-report and Pharmacy Data
K Taneda, HSR&D Center of Excellence, Puget Sound Health Care System, CL Bryson, Center of Excellence, Puget Sound Health Care System, MB McDonnell, Center of Excellence, Puget Sound Health Care System, SD Fihn, HSR&D Center of Excellence, Puget Sound Health Care System

Objectives: We assessed the relationships between adherence to hypertension medication as measured by self-report, a new pharmacy-based measure, and control of blood pressure (BP). Hypertension affects the majority of veterans over 60 and has severe sequelae if not appropriately managed. Recent studies document sub-optimal control of BP among VA patients.  This may be due in part to poor medication adherence.

Methods: As part of the Ambulatory Care Quality Improvement Project (ACQUIP), a randomized trial involving 7 VA general internal medicine clinics, we collected medication data and self-reported comorbidities. Using a novel algorithm (ReComp) that combines information from several previously reported indexes, we computed refill compliance, and assessed its association with self-reported adherence (SA) and BP measured at clinic visits within 90 days of SA among treated hypertensive patients.ReComp compensates for the limitation of prior methods for calculating medication adherence by taking into account both overstocking and days out of medication. SA during the past 4 weeks was assessed by mailed questionnaire using a 5-point Likert scale.

Results: BP records were available for 7,727 of 15,763 patients sent the hypertension questionnaire during the 2-year study period. Among these, 7,268 patients received at least one antihypertensive medication in the 6-months prior to returning the questionnaire. 80% were out of medication fewer than 80% of the days, while 24% obtained a 50% overstock for the same period. SA was modestly associated with ReComp (4% increase in days covered for each 1-point increase in SA, 95% CI 2 6%, Spearman R 0.06, p<0.01). Each 1-point increase in SA reflected a 19% increase in the risk of being at least 80% adherent (RR 1.19, 95% CI 1.1 1.3).   For each 1-unit increase in ReComp, subjects were 6% more likely to have a BP <140/90 (RR 1.06; 95% CI 1.0 1.13, p=0.05). No difference was observed using previously reported, pharmacy-based, measures of compliance.

Conclusions: ReComp, a new measure of medication adherence, is readily computed from computerized pharmacy data and is associated with both SA and BP control.

Impact: ReComp appears to be a better measure of adherence for both descriptive studies and studies of association than prior measures.