Published Online:June 22, 2013
Assumptions about cost do matter—whether or not those assumptions are realistic, considerations for treatment are often affected. The analysis of long-term treatment is an important tool in modeling the long-term cost-effectiveness of treatment in diabetes. The sensitivity of the analysis to variations in how costs are measured also plays a role in evaluating the validity of cost-effectiveness analyses (CEAs).
In long-term CEAs, economic outcomes from one treatment option are compared to those of another. In these long-term studies, overall costs are evaluated with the goal of cost reduction over time due improvements in patients’ health status. For example, while the costs of medications and the management team must be accounted for, intensifying glycemic control in patients with diabetes reduces the risk of micro- and macro-vascular complications over time. Using short-term changes in markers such as blood glucose, blood pressure, and cholesterol, one can calculate the likelihood of complications over time that may result from diabetes. The costs of these complications may then be quantified, and the overall effect on quality of life (QOL) may also be extrapolated.
Two examples of CEAs are the United Kingdom Prospective Diabetes Study (UKPDS) that, according to Gilmer, is probably the most "visible" trial of long-term cost-effectiveness studies, and the Diabetes Prevention Program (DPP), which is the first trial to test the effectiveness of long-term preventive strategies. Using data from these studies, it will be possible to develop simulation models. While there may be some discrepancies and variations between certain subpopulations, studies that evaluate the association between clinical risk factors and outcomes are essential for determining the long-term cost-effectiveness of pharmacologic interventions. Investigators conducted the UKPDS study in a country with fairly low healthcare costs (the UK), so the results might not be applicable to health costs in other nations where health costs are higher.
Interventions for the management of diabetes include the education of patients on comorbid diseases and instruction on self-management. These interventions and programs provide education about lifestyle changes, proper nutrition, and the importance of integrating regular exercise. Patients with low income, who have limited proficiency in English, and who may have limited access to health care tend to benefit the most from these training programs.
Project Dulce, a chronic care model for Latinos with T2DM, demonstrated that the type of insurance coverage provided to a patient may affect how care is delivered. Compared with groups with Medicaid or commercial insurance, subjects who were uninsured had received the shortest duration of care (ie, they sought or received care for less time than other groups). However, when provided, case management and diabetes education resulted in a 1.3 percentage point reduction in glycosylated hemoglobin (A1C) among the uninsured, compared with a 0.4 percentage point and a 0.5 percentage point reduction in A1C in the Medicaid and commercial insurance groups. Furthermore, the cost per quality-adjusted life year (QALY) gained was more expensive among the commercial insurance group than it was among the uninsured, with the cost of education programs increasing from $10,000 per QALY in uninsured subjects to $45,000 in commercially insured subjects.
In an effort to see how treatment quality and A1C improved among patients with diabetes, another intervention integrated guideline-based recommendations into an electronic health record system. On average, A1C levels decreased by 0.3 percentage points in hospital-treated patients as a result of the EHR system. Investigators enumerated study costs including the cost of extra time from health care professionals, the cost of incentivizing the program for physicians, and the cost of training programs for staff. Overall, the per capita cost was $5, but increased to $76 when the cost of incentivization was included in the analysis, with the total cost of incentives amounting to $15,000. Since treatment for diabetes is considered cost-effective at approximately $50,000 per QALY gained, this program, without incentivizing, would be a cost-saving intervention, since the long-term CEA estimated the cost per QALY gained to be only $3000.
It is important for investigators to develop models that apply to different settings, as they might provide significant public health benefits, especially in areas with less developed treatment systems. Electronic medical record (EMR)-based algorithms and mobile health for personalized medicine, among other interventions, will be important strategies for improving cost-effectiveness of disease management in the future. Concluding his session, Gilmer noted that, “Overall, diabetes treatment is cost effective, but it is challenging to achieve the outcomes. Future improvements in clinical care and cost effectiveness studies are on the horizon.”