This study examined a brief intervention to improve medication knowledge and compliance in family medicine patients. The study was conducted in a network of six family medicine residency programs in South Texas, funded by a grant from the Health Research Services Administration. Family medicine patients were invited to participate in the study if they took medicines for type 2 diabetes, hypertension, or hyperlipidemia and responded "yes" to the screening question: "Do you sometimes have trouble taking all your medications as prescribed?"
At enrollment, clinic staff administered a Medication Survey to patients to assess the variables: readiness to change, reported medication knowledge; reported compliance, and predictors of knowledge and compliance. Self-reports of compliance were validated with a social desirability scale, and with physician impressions. Six and twelve weeks post-enrollment, interviewers administered followup surveys by telephone to assess change over time in medication knowledge and compliance. One year post-enrollment, investigators conducted a chart review to examine changes in health outcomes: blood pressure or HbA1c or cholesterol levels, and number of hospitalizations.
Three physicians per clinic site (18 total) conducted medication interventions for the purpose of this study. This intervention was brief behavior change counseling done in the context of a routine office visit, guided by simple strategies outlined in Rollnick et al (1999). First, physicians reviewed subjects= medication regimen and addressed their concerns. We selected a simple strategy from Rollnick et al. - a brainstorming session - that built patients' confidence to change health behaviors. Afterward, physicians offered pillboxes to patients who desired them, and reviewed potential adverse reactions to medicines. One week after the intervention, physicians telephoned patients to reinforce behavior change and to answer questions about the medications. Prior to training physicians to conduct this intervention, we enrolled five subjects per physician into a control group. We attempted to blind physicians to these patients= enrollment status; usual care was provided to control subjects. After training, enrollees' charts were flagged so that physicians knew to conduct the intervention. We sought five patients per physician to receive the intervention.
