The Heart Protection Study established that statins are beneficial for most patients with diabetes regardless of baseline LDL cholesterol (1–2). Kaiser Permanente Colorado, a group-model, nonprofit health-maintenance organization in the Denver/Boulder metropolitan area, developed a letter-based population management strategy to improve statin utilization in patients with diabetes.
Kaiser Permanente Colorado maintains a disease state registry of patient-specific data for members with validated diagnoses, which serves as an ideal tool for patient tracking, quality improvement, and population management. Although letters have previously been employed for population education purposes (3–5), no published literature describes their effectiveness at initiating chronic medication(s). We designed a prospective, quasi-experimental study to demonstrate the effectiveness and safety of a letter-based strategy in increasing statin use in eligible patients with diabetes.
Patients considered statin eligible were aged 40–80 years with a confirmed diabetes diagnosis and total cholesterol ≥135 mg/dl. Statin-eligible and statin-naïve subjects were included. Subjects were excluded for using gemfibrozil, fenofibrate, or cyclosporine; for having an alanine aminotransferase >60 IU/l or serum creatinine >2.0 mg/dl; or for not having a primary care provider. We provided region-wide education describing the benefits of statin therapy to the diabetic population. Eighteen primary care clinics were stratified by size into two groups, and the letter-based strategy was implemented in two phases 3 months apart.
The primary care providers reviewed subjects and authorized statin initiation via letter notification. A letter was sent to those subjects with an explanation of statin benefits and instructions to pick up a prescription and have baseline and follow-up laboratory tests drawn. Medication and laboratory orders were entered into the electronic medical record centrally. Primary care clinical pharmacy specialists at each clinic reviewed and triaged laboratory results with the aid of monthly reports identifying subjects overdue for tests.
We identified 17,464 statin-eligible subjects, of whom 3,352 (19.2%) met inclusion criteria. Of these, 1,570 (46.8%) were assigned to phase one (intervention group) and 1,782 (53.2%) to phase two (control group). The groups were similar with respect to age; sex; and total cholesterol, LDL cholesterol, and alanine aminotransferase values. After 3 months, statin initiation was significantly higher in the intervention group (365 [23.2%] vs. the control group 161 [9.0%], P < 0.001). Overall, current or recent statin use in eligible subjects increased from 75.6 to 82.3%.
More statin starters in the intervention group had completed laboratory follow-up at 3 months (61.1 vs. 46.0%, P = 0.002). Among statin starters, statin persistence at 1 year did not differ between groups (64.7 vs. 61.1% for the control and intervention groups, respectively, P = 0.511). No serious adverse events related to statin initiation occurred.
Findings from this investigation support the concept that a letter-based strategy is safe and more effective than educational efforts alone at initiating statin therapy in patients likely to substantially benefit. Principles applied in this population management approach, including the patient identification process, the informational and interventional letter, and the central monitoring system, can feasibly be incorporated into other models of health care.