Session number: 1143
Abstract title: Medication Adherence among Veterans: Comparing Patient, Provider Report with Electronic Monitoring
Author(s):
AM Kilbourne - Center for Health Equity Research and Promotion,
VA Pittsburgh Healthcare System
AC Justice - Center for Health Equity Research and Promotion,
VA Pittsburgh Healthcare System
CB Good - Center for Health Equity Research and Promotion,
VA Pittsburgh Healthcare System
RL Matheus - Department of Health Services Administration,
University of Pittsburgh School of Public Health
Objectives: Clinically useful yet comprehensive medication adherence measures are needed for optimal disease management. Validity of patient reported adherence has gained some recognition, but provider assessment has been generally discounted. We determined whether patient and provider report are independently associated with electronic monitoring adherence data.
Methods: Veterans with diabetes mellitus receiving primary care at the VA Pittsburgh Healthcare System in 2001 were given electronic monitoring caps (EMCs) that recorded the date and time of each opening for their oral hypoglycemic medication for 30 days. When returning the caps, patients completed an adherence assessment that measured short-term (past 4 days) and overall adherence (past 3 months). Providers rated their patients’ adherence using similar questions. Patient and provider-reported adherence were categorized as perfect (missing no pills and perfect timing) or inadequate (missing at least some pills or did not closely follow schedule). We used linear regression to determine the association between patient and provider report and EMC adherence (defined as percentage of days out of 30 with correct number of doses per day and timing of doses).
Results: Of the 200 diabetic patients enrolled, 99% returned their EMCs. All patients were male, 16% nonwhite, and the mean age was 67 years. Eighty-nine percent of patients self-reported perfect dosage adherence-past 4 days, 69% reported perfect timing-past 4 days, and 65% reported perfect overall adherence. Only 48% of providers thought their patients had perfect overall adherence. Based on EMC data, the mean percentage was 69% (sd=31%) for correct doses and 60% (sd=32%) for correct timed doses, excluding patients who used pillboxes (N=14) or did not use the cap (N=28). Patient-reported perfect adherence-past 4 days (B=27.2, P=0.002) and provider-reported perfect overall adherence (B=16.8, P=0.003) were significantly associated with EMC percent correct dose. Adding both patient-reported perfect adherence-past 4 days and provider reported perfect overall adherence into the regression model best predicted EMC data (R-square=0.10). Neither patient nor provider report were associated with EMC percent correct timing.
Conclusions: A combined provider report of overall adherence and patient reported correct dose improves prediction of medication adherence.
Impact statement: Questioning both the patient and provider regarding medication adherence may be a practical alternative in assessing adherence.