Rheumatoid arthritis affects nearly 6.5 million Americans, causing significant disability and suffering. Improvement of treatment efficacy is of considerable importance. Because adherence moderates treatment effects, efficacy studies need to account for the extent to which the patient complied. This paper reviews the methods of assessing adherence. These methods fall into four categories: (1) clinical measures (therapeutic outcome and clinician judgement); (2) self-report (interviews, questionnaires, and daily records); (3) direct measures (direct observation and biologic assays); and (4) indirect measures (pill counts, pharmacy refills, and electronic monitors). The most commonly used measure is some form of self-report. The clinical measures are problematic as compliance indicators and should not be used in this regard. Selections among the other categories should be made based upon the behavior being studied, the cost and resources available, the capabilities of the patient, the questions being asked, and the level of accuracy and detail desired. There are no perfect measures.