The aim of this study was to evaluate self-reporting of adherence to cardiovascular medication using electronic pillboxes (medical event monitoring system [MEMS]) as the gold standard comparator. In total, 78 individuals (52% hypertensives, 21% diabetics, 27% with dyslipidemia) were recruited prospectively from an outpatient clinic setting in Switzerland. Participants completed two self-report measures (visual analogue scale [VAS] and a validated self-reporting questionnaire) at baseline and were asked to use MEMS as their pillbox for the subsequent 10 weeks. Patients expressed their medication adherence behaviour on a VAS (0 mm="I never take any tablets"; 100 mm="I take all tablets as prescribed") and entered one of six numbers (from 1: perfect adherence to 6: non-adherence) on the questionnaire. Medication compliance was monitored for 75 d on average. Mean (+/-SD, range) scores for MEMS with respect to timing adherence, correct dosing, and self-administration adherence were 79+/-25% (8-100%), 83+/-20% (24-100%), and 92+/-17% (54-118%), respectively. A majority of participants (78.8%) over-reported their adherence to the VAS (93+/-7 mm, 73-100), and VAS scores correlated poorly with MEMS recordings (Spearman's rho for timing adherence, correct dosing, self-administration adherence 0.29 [p=0.018], 0.24 [p=0.051], 0.26 [p=0.036], respectively). Similarly, we found no correlation between adherence as expressed in the questionnaire and MEMS (regression coefficients <0.1). We conclude that a majority of patients over-report adherence to cardiovascular medication if asked to complete a visual analogue scale and a validated questionnaire. Therefore, using self-reporting as the sole means of assessing medication compliance is insufficiently accurate to detect poor adherence. (Hypertens Res 2008; 31: 2037-2043).