Diabetes mellitus (DM) imposes a significant burden of early mortality and decreased quality of life on the 6% of Canadians affected by it. There is a growing body of research evidence showing that the use of certain medications (ACE inhibitors, blood pressure medications and cholesterol lowering agents) can reduce the major complications of diabetes such as heart, blood vessel and kidney disease and amputations. Unfortunately this research knowledge does not appear to be routinely translated into clinical practice. The proposed study examines the effect of simple mailed interventions to patients or physicians on improving the use of these therapies. The strategies to be tested are direct patient education with an emphasis on blood pressure and lipid control rather than use of particular drugs and, for physicians, the provision of confidential prescribing feedback together with targeted educational bulletins. With patient and physician consent, health care administrative data will be used to examine the impact of the interventions. If successful, the study will identify low cost reproducible interventions to promote the use of proven preventive therapies in clinical practice.

Preventing major complications is a cornerstone in the outpatient management of diabetes mellitus (DM). While glycaemic control remains important, recent evidence supports the role of blood pressure and lipids and the use of ACE inhibitors in reducing complications. Effective interventions are required to ensure implementation of these findings. Since optimal DM management involves the participation of both patients and providers, strategies to improve care can reasonably be directed toward either group.

The proposed study explores the impact of simple mailed interventions directed toward patients or providers on the rates of use of ACE inhibitors, Blood pressure lowering drugs and Cholesterol lowering therapies - the ABCs - by older patients with DM. We hypothesize that: 1) mailed interventions including prescriber feedback can increase the ABCs relative to controls; 2) that interventions directed toward patients and providers are equally effective in promoting the ABCs among patients selected to receive the intervention and 3) that interventions directed towards clusters of patients within a physician's practice will also change the care provided by those physicians to their other patients, in the absence of any direct intervention to the physician.

Practices of 300 randomly selected primary care physicians in Ontario and, through their physicians, 10 patients over age 65 with known DM in each practice will be recruited to participate. Practices (physicians and their index patients) will be randomized to one of 3 arms: 1) physician intervention, 2) patient intervention or 3) control. In arm 1, participating physicians will receive confidential prescribing feedback on the ABCs and, as a masking intervention, patients will receive educational materials regarding DM eye care. In arm 2, patients will receive a mailed educational intervention regarding blood pressure and lipid control while physicians will receive education and feedback regarding eye care as a masking intervention. In arm 3, no educational intervention regarding hypertension and lipids will be provided and both patients and physicians will receive the eye care intervention. Data sources will be claims data from the Ontario Drug Benefit Program and the Ontario Health Insurance Plan and the Ontario Diabetes Database (a validated, linkable administrative data diabetes registry).

The primary outcomes will be the proportion of patients receiving ACE inhibitors, the proportion receiving 2 or more antihypertensive drugs and the proportion receiving lipid lowering drugs. All prescriptions received by the patient will be included whether or not written by the study physician. The unit of analysis will be the physician. Data will be analyzed in a multilevel logistic model with covariates measured at the physician level and outcomes measured at the patient level. The primary analysis will include only the data on index patients in each of the three arms. A secondary analysis will include all patients with diabetes over age 65 within the practices of the participating physicians. This project will further elucidate the role of prescribing feedback in guideline implementation and will explore a novel patient-based approach to improving the quality of diabetes care. The relative simplicity and the central administration of the proposed strategy suggest that, if effective, it could readily be widely implemented.

Inclusion Criteria:

• age >65 years
• consent
• within the practices of participating primary care physicians

Exclusion Criteria:

-