Purpose: Applying the comprehensive six-element Chronic Care Model (CCM) for diabetes care can improve patient and system outcomes, but its relative cost-effectiveness is not known.
Methods: We constructed a Markov decision model to estimate the cost-effectiveness of implementing the CCM compared to a provider continuing medical education (PROV) and usual care (UC) as part of a randomized controlled trial (RCT) in an underserved community. Outcomes included costs in 2000 US dollars, quality-adjusted life-expectancy, cumulative incidence of diabetes with either microvascular or macrovascular complications, costs per quality-adjusted life-year (QALY) gained, and costs per diabetes complication averted. Intervention costs and outcomes, and disease progression data were directly obtained from the RCT. Other costs, mortality rates, and utilities were drawn from published literature. Future costs and effectiveness were discounted at 3% annually. The model examined 68-year-old cohorts and assumed equal end-stage renal disease incidence and mortality among the interventions, potentially biasing against the CCM.
Results: Compared to PROV and UC over a 3-year period, the CCM reduced the absolute incidence of microvascular or macrovascular complications due to diabetes by 41.3% and 3.9% respectively. From a health care system perspective, the costs over 3 years for the CCM compared to UC were $70,317 per QALY gained and $29,573 per diabetes complication averted; the CCM dominated PROV in both analyses. In 1-way sensitivity analyses, results favoring the CCM at a threshold of $100,000/QALY were relatively insensitive to variation of costs, but were sensitive to variation of diabetes complication utilities and to strategy-specific probabilities of diabetes complication onset and treatment intensification. Over 10 years, the costs per QALY gained from a health care system and a societal perspective for the CCM compared to UC were $42,179 and $113,280 respectively. Probabilistic sensitivity analysis showed that, using a willingness-to-pay threshold of $50,000/QALY, the CCM was favored in 45% of model iterations (compared to 50% for UC and 5% for PROV); with a $100,000/QALY threshold, the CCM was favored in 51%.
Conclusion: The application of the full Chronic Care Model for diabetes care in underserved community clinical practices is a sound and cost-saving investment compared to the provider continuing medical education, and is economically reasonable compared to usual care.
Candidate for the Lee B. Lusted Student Prize Competition