Chicago Housing for Health Partnership: Findings of a Randomized Controlled Trial of Supportive Housing and Case Management for Homeless Adults with Chronic.

Chicago Housing for Health Partnership: Findings of a Randomized Controlled Trial of Supportive Housing and Case Management for Homeless Adults with Chronic Medical Illness

Laura Sadowski and Romina Kee,Collaborative Research Unit Cook County Hospital (Stroger) andDavid Buchanan, Erie Family Health, Chicago

Primary Research questions

Today’s presentation are the final outcomes that answer the research questions…

Among homeless adults with chronic medical illness …

1) Will CHHP lower the use of costly medical services?

2) Will CHHP increase housing stability at 18 months?

Selected secondary findings presented: Nursing home, HIV

What did we want to learn?

Among the homeless with chronic medical illness…1) Will CHHP lower the use of costly medical services?

and if so, by how much? and do so without any harmful health effects…

2) Will they achieve stable housing at 18 months?

How did we get the information (data)?

• Interviews at baseline (enrolled in hospital)• Follow-up interviews at 1,3,6,9,12, and 18 m • Hospital electronic records (study hospitals)• Medical records (outside hospitals, 66 hospitals)• Internet data bases (incarceration, deaths)

Interviews

• Face-to-face • Baseline 2 hrs, Follow-up 35 mins• Content: Tracking and Housing

Quality of Life (ACTG-SF21)Alcohol and Drug use (ASI)Mental Health Symptoms (Prime MD)Health Service Use (HIV cost study modules)

• Compensation: $20/interview

Who could participate in this study?

• Inpatients referred to Social work for discharge planning at two hospitals.

• Social workers referred those who were homeless to research team Monday-Friday

Who could participate in this study?

• Inpatients referred to Social work for discharge planning at two hospitals. • Social workers referred those who were homeless to research team

Monday-Friday

• Who was eligible? (verified by research staff)– 18+ yrs of age, English or spanish speaking– Expected to be in hospital 24+ hours – 1 or more chronic medical illness (medical record)– Able to self care upon discharge (physician assess)– Lack of stable housing 30d prior to hospitalization– Informed consent

Who was in the study?

• 604 referrals from hospital social workers• 455 eligible referrals• 48 refused• 407 participated (89% of eligible)

2 withdrew (Usual Care)

baseline sample of 407, analysis sample of 405

Who ended up in the study?

• 76% men• 78% AA, 9% Caucasian, 8% Latino • 54% graduated from High school• Age 21-82 years of age, (median=46 yrs)• 54% never married• 9% veterans

Who ended up in the study?

• 76% men• 78% AA, 9% Caucasian, 8% Latino • 54% graduated from High school• Age 21-82 years of age, (median=46 yrs)• 54% never married• 9% veterans• 55% no insurance, 37% Medicaid, 8% Medicare, 1% HMO/PPO

• 36% HIV + • 32% High Blood Pressure• 13% Diabetes

Who ended up in this study?

Housing before study (past 30 d):

Family/Friends: 50%Shelters: 43%Streets: 27%

Spectrum of Homeless Duration

Duration homeless:26% less than 1 y58% more than 2 yMedian: 30 mMedian for HIV+: 28 mMaximum: 440 m

Hospitalizations and ER visits at Study Hospitals during 12 months before study

Usual Care CHHPHospitalizations 0.89 1.24 p<.05ER visits 2.52 2.23

Intervention group had .35 more prior hospitalizations than usual care at the study hospitals in the year prior to the study

Data Source: Electronic Medical Records, JAMA Table 1

Hospitalization at Study Hospital in prior 12 m: 53% NoneER visits at Study hospital in prior 12 m: 41% None

Spectrum of mental health and substance use

• Mental health symptoms at enrolling hospitalization (past 4 w):40% major depression23% other depression28% anxiety

• Substance use (past 30 d): 41% alcohol (to intoxication)45% cocaine

28% heroin59% any drug

Follow-up at 18 months

Interviews at 18 months:90% CHHP program73% Usual care

• Supplemental tracking: “real time” tracking in study hospitals/ERs using electronic medical records and social workers, tracked appointments and presence in ER

or hospitals

Unadjusted and Adjusted Results

• Example: Multiple factors, including CHHP, influences an outcome – e.g., hospitalization

• Unadjusted: # of hospitalizations / # of people who could have experienced an event

• Adjusting for other factors: Remove the influence of other factors on hospitalizations, so that the independent effect of CHHP can be determined. Requires collecting data on all the factors. Examples: Age, HIV, Substance use, prior hospital use

• Presentation today – unadjusted – simpler. Note: After adjusting for, the differences between the 2 groups became statistically significant (Table 3 in JAMA article)

Study Results: Hospitalizations

CHHP Usual Care Avg # hospitalizations, annual: 1.9 2.4

Avg length of stay, annual: 8.7 d 11.4 d

Not statistically significant unadjusted (JAMA, 2009, Table 2).

Study Results: ER Visits

CHHPUsual Care

Avg # ER visits annual: 2.61 3.77Not statistically significant, JAMA, 2009; Table 2Costs: Data not ready for use in terms of cost. Adjust for those who did not receive ER care (LWBS).

Preliminary Study Results: Mortality, 12%+

CHHP Usual Care

Mortality: 25 23

Mortality in HIV+: 13 9

No statistical difference, Detection Bias, Data Collection continues

Nursing Homes: Housing or Health Service? • 94 participants (23%) stayed in 62

Nursing Homes• Duration: 48% temporary, short term

– 23 stayed < 30d– 13 stayed 1-2 m– 9 stayed 2-3m

• Limitations: Self report, Differential Follow-up bias, Necessity of stay unknown,

Added late – at 9 m

Study Results: Nursing HomesAn unanticipated, important outcome

18-month Follow-up: CHHPUsual Care

Nursing home (at least once): 19% 27%

Nursing home days: 5,909 d 10,044 d

Difference is NOT statistically significant, yet important

Such a large difference in NH days, why isn’t it statistically significant?

• Majority of BOTH groups had no nursing home stays

CHHP Usual Care No NH stays 81% 73% NH stays 19% 27%

Graphs of Nursing Home days by group

Quality of Life: Physical and Mental Health

• Measures mental health and physical health functioning• Health change: 18 month score – baseline score.

CHHP Usual CareMental Health +14.7 +14.4Physical Health + 7.7 + 6.5

Both groups improved similarly (even after adjusting) No statistically important differenceScores transformed to a 100 point scale

Additional Findings for HIV: preliminary

• ER visits and Hospitals• Stable Housing at 18 months

HIV subgroup findings: ER visits and Hospital days

CHHP Usual Care n=75 n=71

Emergency Room Visits, annualAvg # hospitalizations, annual:Length of stay, annual:

Unadjusted results

HIV subgroup: Where are they at 18 m?

Ineligible for Stable Housing

Dead 15%Nursing Home 7%Incarcerated 12%Residential CD treatment 3%

Lost to follow-up 15%

Eligible for Stable Housing

48% of sample was

available for stable

housing Intervention 39

Usual Care 31

HIV subgroup: Where did you stay last night?

CHHP Usual Care n=39 n=31

Among eligible….Stable housing (rental/public): 66% 13%Less stable (friend/relative): 34% 74%

Unstable (street, vehicle, shelter): 0 13%Avg # housing changes: 2.1 3.4

Limitations

• Generalizability – to whom do our results apply?to men and women who lack stable housing (>30d),

have a chronic medical illness, and≥ 1 hospitalization in an urban setting

• Appropriateness of ER visits & NH days not assessed – use cautiously for any cost, health or housing implication….

• Nursing Home and Housing Stability data relies on self report (unverified)

Conclusions At 18 months, offering the CHHP program to homeless men and women with chronic medical illness:

•Reduced ER visits by about 1.2 annually*

•Reduced hospital days by about 2.7 annually*Although findings were in the right direction and important magnitude, we did not achieve a statistically significant difference in…

Quality of Life Incarceration (data not shown) Nursing home daysDetection of mortality difference (pending)

Other Findings: Nursing Homes, HIV

• Nursing home stays were frequent (23%) and an important outcome with cost implications (>15K d). 48% of NH stays were short term <90d

Among those with HIV, offering the CHHP program….• Add urgent health service use at 18m• Add stable housing outcomes at 18m