Abstract

Background: Low-income senior housing offers opportunities for improving the health and well-being of vulnerable older adults. The Right Care, Right Place, Right Time (R3) initiative was an affordable housing-based supportive services intervention which partnered with health and community-services providers to enable seniors to remain in the community as long as possible and to reduce health care costs through improved service support and coordination among providers.

Objective: The purpose of this study is to determine the effect of the R3 program on health services use among residents.

Methods: A pre/post difference in difference quasi-experimental design was used, including building-level comparisons using Medicare fee-for-service claims data (based on an intent-to-treat design) across two stages of the R3 intervention. The 18-month pre-intervention period was January 2016-March 2017, while the 18-month Phase 1 intervention period was July 2017-December 2018. The 21-month Phase 2 implementation period, which introduced a targeting strategy to identify residents at high risk of poor health outcomes, was January 2019-June 2020. Key health service outcomes included hospital admissions, 30-day hospital readmission, and emergency department admissions and payments. Intervention sites included seven Boston-area buildings with roughly 1,200 individuals. Three distinct categories of comparison groups acted as controls including five baseline comparison buildings (1,100 residents); 14 additional buildings with service coordinators (5,012 residents); and six additional buildings with no service coordinators (3,100 residents). Building-level Medicare claims data was provided by the local Quality Improvement Organization.

Results: Analyses for Phase 1 found that hospital admission rates, emergency department admissions and payments, and hospital readmission rates grew more slowly for intervention sites than comparison sites. These findings strengthened after the introduction of risk-targeting in Phase 2, showing intervention group declines in inpatient hospitalization rates (16%), hospital admission days per beneficiary (25%), average hospital days (12%), hospital admission payments per beneficiary (22%), and 30-day hospital readmission rates (22%) compared to a 6%, 29%, 14%, 33% and 60% increase in these respective rates among residents in comparison buildings. When accounting for the older age of the intervention residents, the size of decline recorded in emergency department admissions was 6.7% greater for the intervention sites than the decline in comparison sites.

Conclusions: This study found that adding the R3 intervention to low-income housing sites did indeed lead to meaningful reductions in service utilization and costs when compared to buildings where the program was not operating. The introduction of risk-targeting further strengthened this effect. Residents of affordable senior housing communities, who tend to be in poorer health than their counterparts in the community, clearly benefit from the R3 program. This is because having staff more involved and dedicated to resident care, and coordinating closely with community services as well as emergency responders, can lead to lower use of costly hospital inpatient and emergency department utilization, even as observation stay status may increase. The results from this study suggest that additional investment in this type of initiative benefits both residents and the health system.