Background: High health care costs and poor outcomes are associated with high hospitalization rates and are concentrated in a small fraction of the population. Fragmented patient care and socioeconomic disadvantage may further increase hospitalization rates. To defragment medical care for a socioeconomically diverse population of Medicare patients at increased risk of hospitalization, the University of Chicago Medicine developed the Comprehensive Care Physician (CCP) model that offers these patients inpatient and outpatient care from the same physician. This is hypothesized to improve outcomes, including reducing the likelihood of hospitalization. Patient reported data from a 2,000-person randomized clinical trial (RCT) found that CCP substantially improved patient experience and mental health status and reduces hospitalization by 15-20% over 1 year. Here we report effects on hospitalization rates as measured by Medicare claims over 1 year after enrollment and examine effects on hospitalizations in the overall study sample as well as subgroups by age, dual-eligibility status and prior hospitalization.

Methods: The CCP study is a 2,000-patient longitudinal RCT comparing outcomes for patients assigned to CCP physicians who care for them in and out of the hospital to outcomes of patients assigned to standard care (SC) with different doctors in and out of the hospital. Medicare patients with at least one hospitalization in the year before recruitment or in the emergency department at the time of recruitment were randomly assigned in equal proportions to CCP or SC between November 2012 and June 2016. We analyzed traditional Medicare claims for 1916 randomized subjects to construct a claim-based hospitalization rate. Hospitalization rates were analyzed using negative binomial models adjusting for age, gender, dual eligibility status, general health status, and ADL and IADL status at baseline.

Results: Hospitalization was significantly lower for CCP patients compared to SC patients in the overall intervention group (-15%, p< 0.05), with larger reductions for CCP vs. SC in patients age ≥65 (-21%, p< 0.03), in patients with ≥1 hospitalization in the year before enrollment (-17%, p< 0.04), and in patients who were not dual eligible (-30%, p< 0.02). Among the 87%(1669/1916) of the analytic sample with any of these favorable attributes with respect to response to CCP (age≥65, non-dual, or ≥1 hospitalization past year), hospitalization was reduced by 19% (p< 0.01), while in the 13% of patients in the analytic sample with none of those attributes (age< 65, dual and no hospitalizations in past year), the risk of hospitalization was 22% higher for CCP vs SC, though this was not statistically significant.

Conclusions: The CCP program significantly reduced hospitalization rates compared to SC in Medicare patients with ≥1 hospitalization in the year before enrollment, in patients age ≥65 and in patients who were non-dual. Further efforts are warranted to implement CCP and rigorously evaluate its effects on outcomes and costs for these patient populations, including in value-based environments. Additional interventions, such as ones that address unmet social needs, may be required to reduce hospitalization for patients who are younger and/or dual eligible.

IMAGE 1: Table 1: Effects of CCP on Hospitalization in Medicare Claims over 1 Year