Objective: The rationale for selection of a cut-point between acceptable and unacceptable medication adherence has not been previously evaluated. The objectives of this study were to describe the dichotomized measures for medication adherence and assess the rationale for selection of the cut-point. Methods: A systematic Medline review and examination of studies assessing adherence using pharmacy claims from January 1997-June 2007 was performed. Studies containing partial or incomplete methods for measuring adherence were excluded. A sub-analysis of articles was conducted to determine the rationale for selection of the cut-point. Results: The review identified 98 studies with 103 measures of adherence. These studies investigated adherence to cardiovascular drugs (39.8%), hormones and synthetic substitutes (33.0%) and others (28.2%). The following types of measures were used: medication possession ratio (MPR) (45.6%), fixed gap between refills (38.8%), proportion of days covered (7.8%) and others (7.8%). Dichotomous measures were used by 32.0% of the studies. A cut point was specified for 35.9% measures of adherence. The mean cutoff value was 76.0%±12.9%; with a median of 80.0%, and a range of 20.0%-90.0%.The sub-analysis contained 28 articles of which 35.7% refer to cut-points derived from previous clinical studies, 21.4% selected the cutpoint arbitrarily and 14.3% offered no explanation, 7.1% provided clinical evidence, and 7.1% mentioned a reference that actually indicated that dichotomization of the continuous adherence variable was inappropriate. Conclusion: MPR and gap between refills were the most commonly used measures of medication adherence. Almost one third of the studies used dichotomous measures. A medication adherence of 80% of the therapy was typically indicated as the cut-point between adherence and non-adherence. There is no accepted clinical or pharmacological rationale for medication adherence threshold selection. The use of continuous variables to measure medication adherence is recommended.