Private study finds lower Medicaid costs for residents of Bud Clark Commons

From the Oregonian, April 10, 2014

Monthly health care costs for Medicaid recipients who live at the Bud Clark Commons in Portland plummeted 55 percent after moving into the alcohol-and-drug-tolerant public housing project, according to a new study.

But those savings were inflated by 9 percentage points because two high-needs residents apparently died almost immediately after moving in.

Home Forward, the housing authority that operates the commons, funded the $50,000 study. The report touts its findings as an indication that supportive housing had a “profound and ongoing impact on health care costs” for the people who live there.

READ – Integrated Housing & Health; a Health-Focused Evaluation of The Apartments at Bud Clark Commons (PDF)

The study found across-the-board cost reductions for people who moved into the 130-unit studio apartment project, said Bill Wright, associate director for research of the group that conducted the study.

But Wright cautioned there is not a cause-and-effect relationship between living in the commons and lower health care costs.

“Moving in was associated with a reduction,” he said. “That doesn’t mean it’s a direct cause.”

Read the rest of this article here.

READ – Health care study explores the impact of housing on health care use, costs and outcomes, press release from Health Share Oregon

Report Conclusions


Residents with Medicaid coverage saw significant reductions in medical costs after moving into BCC: the average resident saw a reduction of over $13,000 in annual claims, an amount greater than the estimated $11,600 it costs annually to house a resident at BCC.

Importantly, this reduction in claims was maintained into and beyond the second year of residency, suggesting that supportive housing had a profound and ongoing impact on health care costs for those living at BCC. We examined historical, pre-BCC claims data for residents to determine whether some reduction in costs might have been expected in this population even in the absence of housing.

We did not find evidence of a natural “regression to the mean” in costs for the population BCC serves; indeed, their health care costs steadily rose for the 2.5 years prior to moving into BCC, peaked just prior to move-in, and then immediately fell to a much lower level after move-in. In the absence of a formal experimental “control group” to compare out-comes, this represents the best available evidence that cost reductions are likely attributable to the acquisition of housing and would not have been expected to happen in its absence.


We examined utilization data in order to understand the mechanism by which costs were reduced. We found evidence that residents maintained connections to outpatient behavioral health, primary care, and pharmacy after moving in, but saw significant declines in inpatient and ED utilization. This suggests that cost savings among the BCC residents came from efficiently managing health care in appropriate settings, helping to reduce acute health crises and avoid more expensive types of utilization.

We also examined self-reported utilization data in order to determine if similar patterns held true for non-Medicaid residents. We found patterns in the self-report data that matched those in the claims: continued engagement in outpatient care accompanied by a reduction in acute events.

Hospitals absorb significant uncompensated care costs for such events. Given these costs, the “true” savings associated with housing at BCC are likely considerably higher than our Medicaid-only estimate.


Residents saw significant declines in unmet health care needs, and significant improvements in self-reported physical and mental health, after moving into BCC. There was also a significant increase in overall happiness.

Trauma histories were very common among BCC residents; even after moving in many residents still face traumatic events in their lives. Understanding the link between trauma survivorship and health care utilization/costs will be a key component of caring effectively for this population.


Our interviews with residents also revealed some challenges of the supportive housing model. Some residents told us that getting clean and sober was actually more difficult than they expected in an environment where others are still actively using. Others mentioned feeling unsafe or threatened by others living in the building, which sometimes hampered their involvement in social activities or use of other services. New strategies to overcome these challenges will help residents fully engage in the BCC model.


These results suggest that health care reformers would be well served to think carefully about the relationship between housing and health, particularly in vulnerable populations such as those served by BCC. Among those in our study, getting into stable housing resulted in a significant reduction in total health care costs; these savings were greater than the estimated annual cost of housing someone at BCC, do not appear likely to have reflected natural regression to the mean, and were maintained over time. Housing also im- proved self reported health outcomes. In this acutely ill and vulnerable population, supportive housing was effectively a health care intervention, and it appears to have worked.

Additional research can help replicate and substantiate these findings. For now, however, these results suggest that Oregon’s commitment toward a broader view of health care — one that thinks beyond service delivery and encompasses the social determinants of health — may have real potential to help bend the cost curve. Policy and funding pathways to support and expand such models should be strongly considered as part of Oregon’s ongoing transformation effort.