Measurement of adherence is complex and many methods, both direct and indirect are used; there is no universal gold standard. In this article, we share our experiences in a randomised controlled study, the Hypertension Adherence Program in Pharmacy trial, evaluating a community pharmacy-based intervention for improving adherence to antihypertensive medication. Several objective and subjective measures of adherence (Morisky score, TABS score, MedsIndex, Medicines Possession Ratio) were used, but produced varying results, limiting confidence in the conclusions that could be drawn. Despite using a specifically designed data mining software program to identify potentially nonadherent patients from dispensing records, many participants were found to be adherent by the self reported Morisky scale. A lesson to be learned when targeting people for interventions to improve adherence is that information from dispensing records should be supplemented by other methods in order to identify patients most in need of assistance.