Background: Community pharmacies vary widely in terms of ownership structures, location, and dispensing policies. It is unknown if an association exists between the type of community pharmacy and the degree of medication adherence exhibited by patrons-patients.
Objective: To describe adherence to statin therapy among subjects patronizing different types of community pharmacy categories (department- mass merchandise, chain-franchise, and independent-banner) in Saskatchewan, Canada, between 2000 and 2005.
Methods: Study data were obtained from the Saskatchewan Drug Plan and Extended Benefits database, which is maintained by the government of Saskatchewan, Canada. The study included all subjects who (a) filled a statin prescription within selected community pharmacies between January 1, 2000, and December 31, 2005; (b) had no record of statin prescriptions during the year prior to the first statin prescription, according to the records of the Saskatchewan Drug Plan and Extended Benefits; and (c) demonstrated active utilization in the drug plan database for at least 1 year after the first statin prescription. The proxy criterion for activity was any dispensing record for statin or nonstatin medications at least 1 year following the index claim. Statin adherence level was estimated as tablets per day, defined as the total number of tablets dispensed divided by the total number of days of observation. Each subject's observation period began on the index date and ended on the earlier of (a) 30 days after the last recorded fill for any type of prescription medication (statin or nonstatin), or (b) December 31, 2005. The primary end point was the proportion of subjects within each pharmacy category who maintained an adherence level of 80% or greater during their individual observation period. Additional adherence calculations were performed for each of 3 time periods, beginning on the index date and ending on days 365, 729, and 1094 (i.e., 1, 2, and 3 years). Patients were included in the analysis for each time period if they met a proxy criterion for availability for observation, defined as the dispensing of any drug at least 1 day after the end date of each period. Pearson chi square tests were used to assess the significance of differences in baseline characteristics and adherence proportions, comparing pharmacy categories. Logistic regression analysis estimated the odds of an adherence level of at least 80% during the individual observation period, adjusting for pharmacy category, sex, age 65 years or older, known low-income drug coverage, number of distinct drug classes filled concurrently during the first year of observation, loyalty to index pharmacy, and length of observation. Using similar methods, we also estimated "pharmacy loyalty" by calculating the proportion of subjects who refilled 75% or more of their statin prescriptions at the pharmacy that dispensed their first statin prescription.
Results: From an initial sample of 12,818 subjects who had at least 1 pharmacy claim for a statin in the period from January 1, 2000, through December 31, 2005, 8699 subjects met the inclusion criteria. Subjects were observed for a mean (SD, range) of 3.7 (1.7, 1.0-7.0) years after the index statin prescription. During the first year following the index claim, statin adherence rates were at least 80% for 1799 of 3761 (47.8%) patrons of department-mass merchandise, 1778 of 3235 (55.0%) patrons of chain-franchise, and 921 of 1703 (54.1%) patrons of independent-banner stores (P < 0.001). Measured from the index date through day 1094, 869 of 2292 (37.9%), 874 of 1887 (46.3%), and 457 of 975 (46.9%) subjects in the department-mass merchandise, chain-franchise, and independent banner categories, respectively, had a statin adherence level of at least 80% (P < 0.001). In logistic regression analysis, pharmacy category type was significantly associated with statin adherence; subjects in the chain franchise and independent-banner categories were more likely to be adherent to their statin medications during their observation periods than were those in the department-mass merchandise category (adjusted odds ratio [OR] = 1.36, 95% CI = 1.23-1.50, P < 0.001 and OR = 1.39, 95% CI = 1.24-1.57, P < 0.001, respectively). From the index date through day 1094, 1752 of 2292 (76.4%), 1475 of 1887 (78.2%), and 795 of 975 (81.5%) subjects remained pharmacy-loyal in the department-mass merchandise, chain franchise, and independent-banner categories, respectively (P = 0.006). Controlling for several potential confounders using logistic regression, independent-banner pharmacy patrons were more likely to remain pharmacy- loyal during their observation periods than were those patronizing department-mass merchandise (adjusted OR = 1.34, 95% CI = 1.16-1.54, P < 0.001) or chain-franchise stores (adjusted OR = 1.22, 95% CI = 1.06-1.42, P = 0.009).
Conclusion: One year after their first statin fill, subjects demonstrated low rates of adherence, ranging from 48% to 55%, regardless of the type of pharmacy they patronized. Although the differences by type of pharmacy reached statistical significance, their clinical importance is not evident, reinforcing the fact that the problem of nonadherence appears to exist among all types of community pharmacies, regardless of their categorization.