SimCare: Physician Intervention to Improve Diabetes Care

• Glycosylated hemoglobin (A1c) values and A1c test rates were used to assess glycemic control.
• LDL-cholesterol levels and test rates were used to assess lipid control.
• All A1c and LDL values and test dates were recorded for a defined 12-month period before and after the date of intervention for all diabetes patients under the care of study physicians

• Age
• Gender
• Charlson comorbidity
• Years of practice experience
• Number of patients with diabetes

This randomized trial will test the hypotheses that (a) a physician opinion leader intervention, (b) a "narrative process trace" physician feedback intervention, or (c) the combined intervention (a + b), are no more effective than (d) usual care in improving the clinical care of adult patients with type 2 diabetes.

This randomized trial will test the hypothesis that (a) an influential physician feedback intervention, (b) a "narrative process trace" feedback intervention, or (c) the combined intervention (a + b), are no more effective than (d) usual care in improving the clinical care of adult patients with diabetes mellitus.

The unit of randomization and the unit of analysis will be primary care physicians who have more than 10 adult patients with diabetes mellitus. The 162 physicians eligible for this study provide care to about 6,804 adults with diagnosed diabetes at 24 primary care clinics. The narrative process trace intervention uses AMR-mounted clinical cases to assess physicians' diabetes care decisions, with results provided as feedback to physicians. The influential physician feedback intervention emphasizes individualized education of physicians, which will be guided by the narrative process trace in the combined intervention group. Usual care includes the use of a diabetes clinical guideline, use of a diabetes patient registry that provides key clinical data for each adult patient with diabetes, and access to clinic-based diabetes education nurses. We have shown in previous controlled studies that the guideline, diabetes registry, and clinic-based diabetes education nurses have failed to significantly improve diabetes care on a population basis, thus justifying the additional, stronger interventions that now will be tested.

Dependent variables include glycemic control and cardiovascular risk reduction of all 3,360 patients cared for by the 80 study physicians. Secondary analysis will assess rates of screening for microvascular complications. Physicians will be randomized in blocks based on specialty, number of diabetes patients, and years of practice experience. Hierarchical data analysis will be used to accommodate the nested data and propensity scores will be used to correct for selection effects and missing data. Results of this experiment will advance our theoretical understanding of physician behavior change and quantify cost and impact of three specific intervention strategies to improve chronic disease care in the primary care setting. The study will have substantial impact on clinical practice and policy whether the results are positive or negative. If successful, the interventions will be easily disseminated to other primary care practice settings.

• Behavioral: Case-based customized learning intervention
• Behavioral: Case-based learning + physician opinion leader feedback

• Gilmer TP, O'Connor PJ, Rush WA, Crain AL, Whitebird RR, Hanson AM, Solberg LI. Predictors of health care costs in adults with diabetes. Diabetes Care. 2005 Jan;28(1):59-64.
• O'Connor PJ. Overcome clinical inertia to control systolic blood pressure. Arch Intern Med. 2003 Dec 8-22;163(22):2677-8. No abstract available.
• O'Connor PJ. Setting evidence-based priorities for diabetes care improvement. Int J Qual Health Care. 2003 Aug;15(4):283-5. No abstract available.
• O'Connor PJ. Electronic medical records and diabetes care improvement: are we waiting for Godot? Diabetes Care. 2003 Mar;26(3):942-3. Review. No abstract available.
• Johnson PE, Veazie PJ, Kochevar L, O'Connor PJ, Potthoff SJ, Verma D, Dutta P. Understanding variation in chronic disease outcomes. Health Care Manag Sci. 2002 Aug;5(3):175-89.
• O'Connor PJ. Patient archetypes, physician archetypes, and tailored diabetes care. J Am Board Fam Pract. 2002 Jul-Aug;15(4):334-7. No abstract available.
• O'Connor PJ, Sperl-Hillen JM, Johnson PE, Rush WA, Asche SE, Dutta P, Biltz GR. Simulated physician learning intervention to improve safety and quality of diabetes care: a randomized trial. Diabetes Care. 2009 Apr;32(4):585-90. Epub 2009 Jan 26.

Inclusion Criteria:

• Patients- Diabetes patients associated with the n=57 physicians were selected according to the following criteria: (1) enrolled in HealthPartners on January 1, 1998, (2) be at least 19 years of age on January 1, 1998, (3) have an established diagnosis of diabetes mellitus in 1998 based on meeting at least one of the following two criteria: (a) have one or more inpatient, or two or more outpatient ICD-9 diagnostic codes 250.xx for diabetes, or (b) have a filled prescription for a diabetes-specific drug including insulins, sulphonylureas, biguanides, alpha-glucosidase inhibitors, thiazolidenediones, or meglitamides.
• Physicians-All primary care adult care physicians (family physicians or primary care internists) employed full-time at HealthPartners medical groups as of 2000 were eligible for this study.

Exclusion Criteria:

• None