Date of Award

Fall 2017

Document Type


Degree Name

Doctor of Philosophy (PhD)


Health Services Research

Committee Director

Harry Q. Zhang

Committee Member

Martha T. Early

Committee Member

Carolyn M. Rutledge


Depression is the leading cause of disability worldwide, disproportionately affecting the Medicaid population. Collaborative care programs for depression are transforming primary care to increase access and coordinate physical and behavioral health services. Understanding the relationships between components of collaborative care programs, characteristics of participants, and their effect on outcomes can maximize program effectiveness. A pilot collaborative care for depression program within a Medicaid managed care organization was evaluated using administrative claims and case management data. Participants (n=444) included adults with prior Medicaid coverage and a comparison group identified using one-to-one propensity score-matching. Multivariate logistic regression models estimated the odds of decreases in depression severity and acute care, and increases in outpatient services. Ttests and chi-squares were used to identify factors influencing clinical improvement in depression. After controlling for covariates, group status was not a significant predictor of the odds of increased health services use. Increased comorbidities was a significant predictor of increased outpatient physical health visits (OR=1.32, 95% CI [0.57,1.06]). Among intervention participants (n=234), significant individual and social determinants of health leading to higher odds of decreased depression and changes in health services use were identified. Additionally, lower illness severity was associated with clinical improvement in depression, t(45.47)=2.60, p<.05, d = 0.46, 95% CI [.40, 3.18]. Increased follow up contacts were associated with lower depression severity, OR=1.42, 95% CI [1.17, 1.71]. Comparing primarily face-to-face (FTF), primarily telephonic, and equal telephone/FTF contacts, telephonic participants were more likely to have lower depression severity and to decrease/maintain their inpatient stays compared to those with equal telephone/FTF, OR=0.28, 95% CI [1.34, 9.90]; OR=4.64, 95% CI [1.35, 15.94], respectively. Using an ecological framework for vulnerable populations, individual and social determinants associated with changes in health services use and depression outcomes were identified. Findings support adapting case management services to address complex needs, increasing follow up contacts, and utilizing telephonic along with FTF contacts. Lower contacts resulted in worse outcomes. Managed care organizations can play a bigger role with health service researchers in supplying data for evaluation of innovative programs. Additional research with collaborative care depression programs addressing Medicaid populations is needed.