Integrated Behavioral-Physical Care Management in the Primary Care Setting

Abstract

Background: Care management programs for high-cost patients with complex medical and social needs are seen as an opportunity to improve quality of care, promote health, and reduce costs. However, mixed evidence of intervention effectiveness and a lack of supportive analytics tools limit the ability of programs to properly target patients with services that significantly impact health service use. The primary aim of this study is to evaluate a unified primary care and behavioral health intervention targeting high utilizer Medicaid enrollees by examining the program’s effect on subsequent health service utilization patterns. The secondary aim is to develop tools that improve the identification of patients for enrollment in high utilizer care management programs. Methods: This study employs a case-control study design that compares program participants to a carefully matched control group. Patients engaged in primary care were identified via high numbers of emergency department visits and referrals from physicians and the state Medicaid agency. Thirty-five Medicaid beneficiaries were enrolled between August 2013 and February 2014 and included in the final analysis. Priority was given to patients with comorbid mental health or substance abuse disorders, but a number of patients with significant chronic medical illness were also enrolled. Baseline rates of emergency department visits, hospital admissions, and primary care visits were calculated for the one-year pre-enrollment. Post-intervention rates were calculated as twice the number of encounters during a 6-month follow up period. Results: Enrollment of high utilizer Medicaid beneficiaries in integrated behavioral-physical care management did not result in a greater reduction of emergency department visits or hospital admissions compared to patients receiving standard primary care. Patients enrolled in the program showed an average decrease of 3.46 ED visits per year and an average increase of 0.17 hospital admissions per year. Similar high utilizer patients receiving standard primary care showed an average decrease of 4.43 ED visits per year and average decrease of 0.20 hospital admissions per year. However, program participants were significantly more engaged in primary care over the follow up period compared to the control group. Patients enrolled in the program had almost no change in primary care visit rate (increase of 0.14 visits per year), whereas control patients had an average decrease of 9.71 visits per year. Conclusion: This case-control study reveals the tough road ahead for programs attempting to show a return on investment—changes in outpatient healthcare use were not accompanied by a drop in more costly hospital healthcare use. A data application was developed to improve patient selection, which we hope reduces patient heterogeneity moving forward. More must be done to rigorously evaluate high utilizer care management programs on a larger study population.