The assessment of compliance is critical in the evaluation of the effectiveness of a new therapeutic agent. Fifteen patients with transfusion-dependent β-thalassemia, many of whom had previously demonstrated erratic compliance with deferoxamine, were enrolled in a clinical trial of a new oral iron chelator, 1,2-dimethyl-3-hydroxypyrid-4-one (L1). Their compliance with this medication was estimated by several existing methods and the novel Medication Event Monitoring System (MEMS). Overall compliance as assessed by the MEMS was 78.5 ± 13.0% of prescribed doses taken, significantly lower than the corresponding rates calculated by pill counts and diaries (91.5 ± 9.2% and 94.1 ± 4.3%, respectively). However, several serious problems were encountered with the MEMS, mostly in the form of incorrect use of the device by the patients. Disclosure of the nature of the MEMS and the compliance monitoring process did not alter the rate of adherence with L1 therapy. Compliance as determined by pill counts did not differ between the 1st and 2nd 6-month periods. Although not reaching statistical significance, a trend towards better L1 compliance occurred in those patients in whom serum ferritin levels decreased. Patients who filled at least 50% of their diaries had significantly better compliance by pill counts than those who completed less than 50% of their diaries (95.9 ± 4.1% and 86.5 ± 11.1%, respectively). Steady-state L1 trough concentrations and 24-hour urinary iron excretion did not correlate with L1 compliance. Given the limitations of the available techniques for monitoring compliance, including the new MEMS device, a combination of methods should be used in the evaluation of medication compliance during the investigation of a new drug.