This randomized trial is evaluating whether socioeconomically disadvantaged Medicare patients at increased risk of hospitalization experience fewer hospitalization if those patients are offered care in: 1) ACCT, where patients receive care from different physicians in the hospital and the clinic settings and have access to nurse and social worker care coordination services, 2) CCP where patients receive care from one physician in the inpatient and outpatient settings or 3) C4P which adds screening of unmet social needs, community health worker support and arts and culture programming to CCP. The study will determine how these programs affect patient activation and engagement in care, satisfaction with care, general health and mental health, and goal attainment.

Health disparities and the health of communities are influenced both by the health care system and by the social context in which people live. Accordingly, interventions that seek to meaningfully reduce disparities should consider patients' medical and social needs. The fragmentation of medical care is one aspect of the health care system that adversely affects health, perhaps particularly for socioeconomically disadvantaged individuals with more limited resources to bridge gaps in care. Indeed, while many care coordination programs have been developed, evidence supporting their effectiveness is quite limited and especially for vulnerable populations. This study will fill important gaps in evidence concerning the effects of 3 diverse care coordination models on hospitalization rates for a socioeconomically disadvantaged population at increased risk of hospitalization that is served by the University of Chicago Medicine (UCM). Secondary outcomes include patient activation and engagement with care, satisfaction with care, general health and mental health and personal goal attainment.

The first model is an example of a commonly implemented class of care coordination models that relies primarily on the use of care coordinators. The specific care coordinator (CC) model that is being studied - the Partners HealthCare Care Management Program (PHCMP) - was initially developed and studied by Partners HealthCare. In PHCMP, "high-risk" patients (defined primarily as ≥2 hospitalizations in the year before enrollment) have access to nurse care coordinators who seek to help manage the patient's care across the continuum. UCM's Medicare Shared Savings Plan Accountable Care Organization has recently implemented a model based on PHCMP that it calls the Ambulatory Care Coordination Team (ACCT). In ACCT, nurses and social workers provide proactive care coordination to "high-risk" patients. CC and ACCT are representative of common care coordination models implemented nationally. The fact that these models typically involve hiring additional staff and increase the number of hand-offs may explain why programs such as these have often failed to produce desired improvements in health outcomes or decreases in utilization.

The second model is a novel care delivery program called the Comprehensive Care Physician (CCP) program. The CCP model seeks to more effectively integrate inpatient and outpatient care for patients at increased risk of hospitalization by offering them care from the same physician in the inpatient and the outpatient settings so that these patients can benefit from the advantages of continuity in the doctor patient relationship. Since 2012, the investigators have developed and tested this model at UCM in a randomized trial funded by the Center for Medicare and Medicaid Innovation comparing CCP to standard care (SC) in which patients receive inpatient and outpatient care from different doctors and do not have access to care coordinators. The investigators enrolled 2,000 patients in this study, of whom ~90% are African American, with a median income of ~$20,000 per year and 1-year mortality rate of 15-20%. The results are striking: care ratings on Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS) measures correspond to the 95th percentile nationally for CCP patients vs. the 80th for SC, about 30% of CCP patients have a 1 point improvement in self-rated mental health status (e.g., from good to very good) compared to SC patients, hospitalizations are 20% lower over 1 year follow-up and 26% lower for CCP patients who experienced ≥2 hospitalizations in the year before enrollment. We think the efficacy of this model comes from the deep connection that the CCP team develops with patients, understanding thise patients as individuals and recognizing and beginning to address the deeper social determinants of their health.

The third model, the Comprehensive Care, Community and Culture Program (C4P), builds on the CCP program to better engage patients in their care. C4P was motivated by the finding that ~30% of patients who enrolled in the CCP program did not engage with it despite having expressed interest in the program and that a wide range of social factors might be barriers to their engagement. To better address social determinants of health, C4P builds on CCP by adding 1) systematic screening of 17 domains of unmet social needs, 2) access to a community health worker and 3) access to community-based arts and culture programming. Preliminary findings from a pilot of C4P indicate that unmet social needs are diverse, highly concentrated in a small number of patients and linked to each other, presenting barriers to addressing these needs. However, unmet needs were also found to cluster in ways that suggested actionable strategies to address needs more effectively. Early results suggest that C4P increases patient activation and engagement in care compared to CCP and SC.

While we have rigorously compared CCP to SC and performed a pilot study of C4P, CCP and SC at UCM, we have not compared CCP or C4P to the more commonly-used CC model, nor have we performed an adequately powered study comparing CCP and C4P. The rigorous findings comparing these models that we will generate are sorely needed by patients and health systems to inform choices about care coordination models, and particularly for socioeconomically disadvantaged individuals.

This study has 2 specific aims:

Aim 1: To compare the effect of 3 models of care on hospitalization (primary outcome) and patient activation, engagement in care, satisfaction, health outcomes and goal attainment over 1 year for patients at increased risk of hospitalization: 1) a care coordinator (CC) model in which patients receive inpatient and outpatient care from different physicians with access to care coordinator services for patients identified as at high risk of hospitalization, 2) the Comprehensive Care Physician (CCP) program in which patients at increased risk of hospitalization can receive inpatient and outpatient care from the same physician and 3) the Comprehensive Care, Community and Culture Program(C4P), which enhances CCP care with systematic screening for unmet social needs, access to a community health worker and access to programming to better engage patients and address their unmet social needs.

Aim 2: To determine whether the effects of the 3 models differ based on the number of hospitalizations that a patient has experienced in the year before entering these models of care.

Study Description: This study is a randomized clinical trial that compares 3 approaches to care for patients at increased risk of hospitalization: 1) ACCT in which patients receive inpatient and outpatient care from different doctors with the option to receive care from UCM physicians and care coordination from nurses and social workers who manage their care with the larger clinical team, 2) CCP in which patients receive care from the same UCM physician in the inpatient and outpatient settings and 3) C4P in which patients receive care from a CCP in addition to systematic screening of unmet social needs and access to a community health worker and community based arts and culture programming. Patients are insured by Medicare or Medicare and Medicaid, be at increased risk of hospitalization based on a history of prior hospitalization or emergency department use and be recruited from mixed, but predominantly low-income communities on Chicago's South Side. A total of 3,000 patients will be recruited, with 1,000 patients in the ACCT, CCP and C4P arms. The primary outcome is hospitalization because of its importance to both patients and health systems. Our primary measure of hospitalization is the number of hospitalizations over 1 year measured in Medicare claims data. Other outcomes are patient activation and engagement in care (measured using the Patient Activation Measure and the rate of completion of primary care visits), satisfaction with care (measured using HCAHPS scores), self-rated general and mental health status, and goal attainment.

Inclusion Criteria:

• Must have Medicare Part A and Part B
• Must have been hospitalized once in the past 2 years or be in emergency department at time recruitment is initiated

Exclusion Criteria:

- None

no plan at present