Three Essays in Health Care.

TitleThree Essays in Health Care.
Publication TypeThesis
Year of Publication2018
AuthorsDas, A
Academic DepartmentHealth Services Organization and Policy
DegreeDoctor of Philosophy
Number of Pages213
Date Published2018
UniversityUniversity of Michigan
CityAnn Arbor, MI
Keywordsdecision modeling, depression, Diabetes, Health Care, Health Economics, health policy, hospital readmissions

This dissertation examines two common sources of increased health care costs – readmissions and the co-occurrence of depression among patients with diabetes. The first paper examines hospital performance in the Hospital Readmissions Reduction Program to determine whether sources of incentive heterogeneity are associated with differences in improvements over multiple years. I find that hospitals seem to be responding to the main incentive in the program, as those that performed poorly in previous years improve significantly more than hospitals that have avoided penalties. Hospitals also are making improvements in conditions that have the highest marginal benefit from better performance. Payer mix does not seem to be correlated with hospital performance over time even though the financial incentives of the program only apply to future Medicare reimbursements. In the second paper I develop a model to predict the onset of depression among individuals with diabetes. Using data from the Health and Retirement Study and the National Health and Nutrition Examination Survey, I find that gender, body-mass index, hypertension, history of stroke, history of heart disease, and duration of diabetes are significant predictors of annual depression status. I then build this depression prediction algorithm into the Michigan Model for Diabetes, an existing microsimulation model that allows users to evaluate the progression of diabetes. In the final paper, I use the modified diabetes simulation model to evaluate the cost-effectiveness of the collaborative care intervention to treat depression among patients with diabetes. Trials suggest that the collaborative care intervention, a multidisciplinary approach to address the depressive symptoms of patients, can be cost-effective in the short-term when used to treat patients with diabetes and comorbid depression. Using simulation models allows us to evaluate the long-term cost-effectiveness as well as the influence of a variety of inputs on the value of the program. Only when the utility loss associated with depression is small or the intervention effectiveness is substantially decreased does the intervention require a higher willingness-to-pay to be considered cost-effective. Otherwise, our base-case analysis and other one-way sensitivity analyses support the conclusion that this intervention is cost-effective.

Citation Key10245