Pharmacy Interventions to Improve Chronic Disease Medication Refill

Overview

Compare the effectiveness of 3 strategies by pharmacists to decrease the time to refill of prescriptions for common chronic diseases (diabetes, hypertension, hyperlipidemia, heart failure, depression, psychoses).

The design was a multi-site 8-month randomized, controlled clinical trial with the patient as the unit of randomization. The Institutional Review Board at the Medical University of South Carolina approved the study. The setting included 9 pharmacies within a medium-sized grocery store chain in South Carolina. The 9 pharmacies were selected to ensure an adequate sample size and sufficient representation by patients with respect to geographic region (urban, suburban, and rural), race/ethnicity, and socioeconomic background. Individual pharmacy staff were informed by their district manager (P.P.) that they were required to participate in the project. Beginning January 16, 2006 and continuing through August 29, 2006, all patients who were at least 7 days overdue for a qualified prescription were identified on a weekly basis (i.e. every Monday morning) using the centralized computer database. Patients identified as being overdue for a qualified medication were assigned a unique study identification number and randomized using a simple randomization technique operationalized within a Microsoft® Access® database. The randomization was based on a permuted block design with a block size of 3, with each patient thus having a one-third chance of being placed in any one of the 3 treatment arms.After eligible patients were identified and randomized each week, there were 3 different protocols followed, depending on the assigned treatment arm. The 3 treatment arms included 1) pharmacist contact with the patient via telephone, 2) pharmacist contact with the prescribing physician via facsimile, and 3) usual care.There were a total of 3,048 patients into arms 1 (n=1018), 2 (n=1016), and 3 (n= 1014). The study was designed to detect relatively small differences in the outcomes of interest. Baseline comparisons across treatment arms in the characteristics of patients and medications were made using t-tests and chi-square tests, as appropriate. For the analyses of study outcomes, an intent-to-treat approach was used. Statistical models were constructed to compare outcomes among the 3 treatment arms, after first examining whether significant interaction between treatment arm and medication disease classification was present. For the analyses of secondary outcomes, all of which were binary in nature, multivariable logistic regression models were used, with treatment arm as the key independent variable of interest, adjusting for the same covariates used in the primary analyses.

Conditions

Interventions

Eligibility

Locations (1)

Outcomes