Partial compliance: implications for clinical practice.

Modern therapeutics unavoidably requires integration of a patient's medication-taking behavior in assessing the clinical response to treatment. Automatic escalation of the drug regimen whenever the treatment goal is not achieved carries major risks and should be discouraged. Medication-taking behavior, when studied carefully by dynamic measures like electronic monitors, displays marked inter- and intra-subject variability over time. Most deviations from the prescription are underdosings and occur randomly, rather than consistently or systematically. Such deviations are frequent, difficult to detect by traditional measures, and hard to predict from common baseline characteristics. Compliance tends to fall as dosing frequency rises above once daily. Better measures of medication-taking behavior permit evaluation of both adverse drug reactions and secondary resistance to therapy as resulting from pill taking itself. Negative consequences of partial compliance among hypertensive patients include marked increases in rates of rehospitalization and rates of coronary events. One alternative strategy to labor-intensive interventions for improving compliance is to develop longer-acting medications, which compensate in part for lapses in dosing frequency. Such a strategy reflects a search for therapeutic sufficiency rather than a rigid concordance between the prescription and medication-taking behavior.