METHODS OF MEASURING MEDICATION USE/COMPLIANCE IN THE COMMUNITY DWELLING POPULATION OF OLDER PEOPLE
Studies have shown that physicians’ clinical judgment of compliance are no more accurate than predicting compliance by chance. APhA summarizes a study by Gilbert et al. (1980) who found that estimates of compliance by 10 physicians for 74 patients (58 of whom the physicians had known for at least 5 years) were accurate only for 10 percent of the patients, when compared to information obtained from a pill count and blood concentration measurements.
More recently, Bikowski, Ripsin, and Lorraine (2001) conducted a study with 50 patient-physician pairs to determine the degree of disparity between physicians’ perceptions of older patients’ medication regimen and patients’ perceptions of their regimen. Patients consisted of individuals 65 and older who visited their physician on an index day, who had visited that physician at least three times in the past year, and were taking at least four prescription medications. Physicians were family medicine faculty and second- and third-year residents. The average number of prescription medications per patient was 7 (range= 3 to 7, sd=2.89). Congruence was defined as agreement between physician and patient regarding all prescription medications, dosages, and frequency. The focus of the study was on patient understanding of the medication regime, as opposed to medication-taking behavior (or adherence). First-year medical students interviewed patients in their homes and were asked about their understanding of what their medication regimen should be, including all prescription and nonprescription medications and their doses and schedules of administration.
Bikowski et al. (2001) found that only 14 percent of the patient-physician pairs demonstrated complete congruence. Seventy-four percent of the pairs had at least one medication that either the physician was unaware the patient was taking or the physician thought the patient was taking, but was not part of the patient’s regime. Twelve percent of the pairs showed dose and/or frequency discrepancies. Overall, complete congruence between physician and patient was achieved for 60 percent (225 of 375) of the individual medications. The highest congruence was found for diabetic and other endocrine medications, with 73 percent and 74 percent congruence, respectively. Pain medications (including nonsteroidal anti-inflammatory medications) and gastrointestinal medications showed the lowest congruence (29 percent and 43 percent, respectively). Antihypertensives, the most frequently prescribed medications (36 percent of the sample) showed a congruence of 66 percent. Other cardiac medications showed a congruence of 57 percent. Patient-physician congruence for psychiatric medications (including sleep medications) was 60 percent. Congruence for pulmonary medications was 55 percent. Study authors reported that it was not clear from the study why the physicians underreported medications; however, other researchers have demonstrated underreporting of drugs when physicians were surveyed about what medications their patients were taking (Spagnoli et al., 1989; Morrow, Leiter, and Sheikh, 1988).