Gray B, Weng W, Lipner RS. — American Board of Internal Medicine
Schuetz A, Adler J, Peskin B. — Archimedes, Inc.
Presented: AcademyHealth Annual Research Meeting, June 2010
Research Objective: To determine the relative importance of measures of physician quality in terms of long-term health outcomes for adults with type 2 diabetes.
Study Design: We used the Archimedes Model to perform 30-year simulations that measured the long-term benefits of various interventions. The Model was a person-specific, large-scale simulation model of physiology, disease and health care systems, represented at a high level of detail. Its accuracy was checked by successfully simulating the results for more than 50 randomized controlled trials.
The Archimedes Model captured a rich variety of outcomes, including myocardial infarctions (MIs), strokes, retinopathy, amputation and renal disease. Discounted quality adjusted life years (D-QALYs) incorporated all of these measures into a single metric capturing mortality and morbidity at different points in time. We constructed policy simulations by incrementally supplementing interventions according to physician control from most to least: 1) process-of-care measures office visit procedures (retinal and foot examinations, microalbumunuria screening, aspirin therapy); 2) meeting biomarker targets (conservative or aggressive treatment of HbA1c (9% or 7%), blood pressure (140/90 or 130/80mmHg), LDL cholesterol (130 or 100mg/dl)); and 3) lifestyle interventions (smoking cessation and diabetic diet). Each intervention was performed at the status-quo level observed in NHANES or at 100% physician performance and patient compliance.
Population: 25,000 simulated 30 to 75-year-old old patients, with diagnosed type 2 diabetes derived from NHANES (1999-2006).
Results: Our findings indicated that, if all process-of-care measures were performed at 100%, there would be a 3.6% increase in D-QALYs among diabetes patients over the status-quo (0.43 difference over status-quo 10.9 D-QAYs). This improvement in outcomes grew by 81.3% when compliance with lifestyle interventions was added to the benefit of meeting all process-of-care standards (change in D-QALYs 0.78 versus 0.43). The benefits resulting from patients making these difficult life-style changes were equivalent to meeting the conservative biomarker targets (D-QALYs of 11.6 versus 11.7).
Assuming no lifestyle changes, treating to aggressive rather than conservative biomarkers increased the benefit of biomarker control by 52.4% (0.41 difference over conservative biomarkers 11.7 D-QALYs). Adding smoking cessation and compliance with a diabetic diet increased this benefit to 75.8% (0.59 difference over conservative biomarkers11.7 D-QALYs).
For event counts, our results indicated that performing all process-of-care measures at 100% resulted in a 16.8% reduction in MIs, and a 7.5% reduction in strokes over 30 years. These reductions grew to 27.3% / 43.1% for MI and 27.0% / 46.7% for stroke when all biomarker targets were met at conservative/aggressive levels. Adding compliance with lifestyle measures increased this reduction to 35.3% / 49.4% for MI and 30.9% / 50.3% for stroke.
Conclusion: Our results suggest that improvement in the process-of-care for diabetics would result in moderate improvements in long-term health outcomes as measured by D-QALYs and substantial reductions in the number of MIs and strokes. Further, by meeting conservative biomarker standards, there is an opportunity for a significant increase in the effectiveness of care. These improvements match the effects of patients making difficult lifestyle changes (smoking cessation, diabetic diet).
Implications for Practice: While process-of-care and lifestyle measures are important, clinical emphasis on meeting aggressive biomarker targets may offer the greatest opportunity for improvement in diabetes care.
For more information about this presentation, please contact Research@abim.org.