Study Documents Higher Health Costs Among HUD-Assisted Seniors

CHPS | June 11, 2014

A new study has, for the first time, successfully linked administrative data from the U.S. Department of Health and Human Services (HHS) and the U.S Department of Housing and Urban Development (HUD). An initial analysis of the data confirmed that HUD-assisted senior residents in the study areas were more likely to be dually eligible for Medicare and Medicaid. These residents were also sicker and more costly to both programs than their non-subsidized peers in the community.

A new study has, for the first time, successfully linked administrative data from the U.S. Department of Health and Human Services (HHS) and Housing and the U.S. Department of Urban Development (HUD).

The merged data provide important insights into the health and health care utilization of older residents of HUD-subsidized housing living in 12 geographic areas, according to the study’s final report.

An initial analysis of the data confirmed that HUD-assisted residents in the study areas were more likely to be dually eligible for Medicare and Medicaid. These residents were also sicker and more costly to both programs than their non-subsidized peers in the community.

The Lewin Group led the study, which was funded by HHS and HUD. The LeadingAge Center for Applied Research and The Moran Company worked as subcontractors on the project.

The pilot effort may help researchers understand how older Medicare beneficiaries who receive federal housing assistance differ from similar community-dwelling older beneficiaries who do not receive such assistance. The study compared these populations based on their demographic characteristics, insurance enrollment status, health conditions, and health care utilization and costs.

Part of a Larger Research Effort 

The study, reported in Picture of Housing and Health: Medicare and Medicaid Use Among Older Adults in HUD-Assisted Housing, explored the potential for publicly subsidized senior housing to serve as a platform for efficiently managing the population health of low-income older adults with various levels of physical and mental health risk.

  • The study was part of a larger research effort that included case studies of programs that link housing with services within several of the study’s 12 geographic areas.


“HUD-subsidized senior housing properties have great potential to reduce Medicare and Medicaid costs by linking with community-based health and supportive services and efficiently coordinating care,” says Dr. Robyn Stone, executive director of the Center for Applied Research. “Federal and state coordinated and integrated care demonstrations should consider HUD housing as a platform for testing new coordinated care models and better managing population health in low-income communities.”  

Matching HUD and CMS Data for Individuals 

The HHS-HUD study explored the feasibility of matching HUD administrative data to administrative data from the Centers for Medicare & Medicaid Services (CMS).

Researchers attempted to match HUD and CMS data for individuals living in HUD-assisted housing in the study areas in 2008 (the most recent data available to them). Using a conservative methodology, 85% of individuals age 65 and older matched to Medicare. All the data was de-identified to protect the privacy of the older adults in the study. 

Dual Eligibility 

Of the 85% of HUD-assisted older adults who were identified Medicare beneficiaries, approximately 68% were dually eligible for both Medicare and Medicaid.

Dual eligibles are known to be sicker and frailer than their Medicare-only counterparts, says Alisha Sanders, senior policy research associate with the Center for Applied Research. Consequently, they use more health care services and are more costly. According to the Kaiser Family Foundation, Medicare spends 1.8 times more for dual eligibles than it does for beneficiaries who only receive Medicare.  

Health Care Costs and Utilization 

To examine health care utilization and payments, the HHS-HUD study stratified Medicare beneficiaries into 5 subgroups to reduce variability between the HUD-assisted and the unassisted community-dwelling groups. Analysis focused primarily on Medicare-Medicaid enrollees (MME), also known as dual eligibles, who were age 65+ and had Fee-for-Service (FFS) coverage. This group was compared to MME beneficiaries in the community who did not receive assistance from HUD.

Researchers are hopeful that the merged dataset can help to track health and housing outcomes and could reliably support future research and policy analysis. Their initial comparison of individuals in the study areas yielded these findings: 

  • Chronic conditions: HUD-assisted MMEs had more chronic conditions than unassisted MMEs in the community. Fifty-five percent of HUD-assisted MMEs and 43% of unassisted MMEs in the 12 jurisdictions had 5 or more chronic conditions. The higher number of chronic conditions translated into higher health care utilization and costs.
  • Monthly Medicare costs: Average Medicare per-month per member (PMPM) costs for Medicare FFS were 16% higher for HUD-assisted MME beneficiaries ($1,222) than for unassisted MMEs in the community ($1,054). These higher costs were driven by the fact that HUD-assisted MMEs had 31% more home health visits, 45% more ambulatory surgery center visits, 26% more physician office visits and 13% more emergency department visits than unassisted MMEs in the community. 
  • Monthly Medicaid costs: PMPM costs for Medicaid FFS were 32% higher for HUD-assisted MMEs ($1,180) than for unassisted MMEs ($895). When compared with MMEs in the community, HUD-assisted MMEs used over 100% more personal care services, 80% more “other home and community-based services,” and 67% more Durable Medical Equipment services covered by Medicaid. 


“To some extent this data confirmed what we suspected,” says Sanders. “We have done case studies that show the high levels of multiple chronic conditions among residents of affordable senior housing, and their high health care needs and relatedly high costs. But now we have the hard data to confirm this anecdotal evidence.”

Preliminary analysis of the new dataset also raises additional questions about this population, says Sanders.

“The data raises the question of whether linking housing properties with services could help residents better address their chronic conditions and lower their use of health services,” she says. “We can’t tell that from this data, but we’re looking at that through a follow-up study funded by the MacArthur Foundation. After surveying properties about their available services, we’ll attempt to see whether those services had any effect on residents’ health service utilization and costs.”  

Data Sources

Researchers working on the HHS-HUD study used data from the Tenant Rental Assistance Certification System (TRACS), the Public and Indian Housing Information Center (PIC), the Medicare Beneficiary Summary File, and the Medicaid Analytic eXtract (MAX) Person Summary file.

The study included the following 12 jurisdictions: 

  • New Haven-Milford, CT.
  • Bridgeport-Stamford, CT.
  • Milwaukee-Waukeesha, WI.
  • San Francisco-Oakland, CA.
  • Boston-Cambridge-Quincy, MA.
  • Durham-Chapel Hill, NC.
  • Richmond, VA.
  • New York-Northern New Jersey-Long Island, NY.
  • Columbus, OH.
  • Akron, OH.
  • Cleveland, OH.
  • The state of Vermont.