Compliance: the case for objective measurement.

This paper evaluates the current status and importance of methods for measuring patients' compliance with medical regimens. Studies are cited on the inaccuracy of subjective measures (patients' statements) and one study, presented in detail, demonstrates that the magnitude of error can be large. However, research has generally failed to use or develop valid quantitative measures of compliance. One review finds only 31% of compliance studies using objective measures prior to 1978; 50 more recent studies were only slightly better. Although it appears that inaccurate measurement of compliance may jeopardize research and clinical practice, major reviewers of the compliance literature have not provided clear guidelines. Investigators should develop methods appropriate to their own research question before commencing a project, or else should restrict themselves to measurable compliance variables, e.g. attendance. Improved measurement may come from blood tests, from medication monitoring (counting) and perhaps ultimately from research to enhance the validity of interview procedures.