This study was designed to compare sulfonylurea adherence assessment b
y providers, patients' self-report, pill counts, and a medication even
t monitoring system (MEMS-3(R)) device, and correlate the estimates of
metabolic control by provider, patient, and laboratory. Forty-seven o
utpatient veterans with fair to poor metabolic control of non-insulin-
dependent diabetes mellitus were enrolled and received monthly refills
of sulfonylurea in vials with a cap containing an electronic medicati
on monitoring microprocessor. Pill counts and fasting plasma glucoses
were measured monthly, and glycohemoglobin and a 24-hour diet recall w
ere obtained at 0 and 60 days. Investigators then asked providers and
patients to assess adherence and metabolic control. Forty-seven percen
t were nonadherent to medication using MEMS-3(R), 29% using pill count
s, 29% using provider assessment, and 31% using self-report. Thirty-on
e percent of providers and 53% of patients assessed metabolic control
differently than laboratory values. Assessment of medication adherence
by provider, patient, and pill counts did not explain metabolic contr
ol as closely as assessment by MEMS-3(R).