Transcript Department of Health, Behavior & Society David Holtgrave, PhD, Professor & Chair © 2005, Johns Hopkins University.

Department of Health, Behavior & Society
David Holtgrave, PhD, Professor & Chair
© 2005, Johns Hopkins University. All rights reserved.
Overview
 Examples of studies
of cost offset
due to housing provision
 Mechanisms of
prevention
 Relative
housing as HIV
cost-effectiveness of
housing as a structural intervention
Recent studies of cost offset

Larimer et al. (JAMA, 2009):


Housing first model for persons with severe alcohol
challenges created stability, reduced alcohol
consumption, & decreased health costs 53%
relative to wait-list condition
Gilmer et al. (Psych Services, 2009):

Participants in a San Diego housing first program
had increased case management and outpatient
care costs but these were nearly entirely offset by
decreases in inpatient, ER and criminal justice
system
Recent studies of cost offset (continued)

Economic Roundtable (Report: “Where We Sleep,” 2009)
 Study of 10,193 persons in LA County
 9186 were homeless while receiving General Relief
public Assistance
 1007 exited homelessness via supportive housing
 Typical public monthly cost in group experiencing
homelessness: $2897
 Typical public monthly cost in supportive housing
group: $605
Mechanisms of housing as HIV prevention

Two large-scale, randomized controlled trials examined the
impact of housing on health care utilization & outcomes
among homeless/unstably housed persons with HIV & other
chronic medical conditions

The Chicago Housing for Health Partnership
(CHHP) study followed 407 chronically ill homeless
persons over 18 months following discharge from the
hospital, including an HIV sub-study of 105 participants
who are HIV+

The Housing and Health (H&H) Study examined
the impact of housing on HIV risk behaviors, medical
care and treatment adherence among 630 HIV+ persons
who were homeless or unstably housed at baseline
CHHP Background & Methods

“Housing first” program providing supportive housing
for homeless persons with medical issues such as
HIV/AIDS, hypertension, diabetes, cancer and other
chronic illnesses

18 month random controlled trial (RCT)
 Half received CHHP supportive housing
 Half continued to rely on “usual care” - a
piecemeal system of emergency shelters, family &
recovery programs

Results published in JAMA (Sadowski et al., 2009)
and AJPH (Buchanan et al., 2009)
CHHP Findings

“Housed participants:




More likely to be stably housed at 18 months
Fewer housing changes
29% fewer hospitalizations, 29% fewer hospital days, and 24%
fewer emergency department visits than “usual care”
counterparts
Reduced nursing home days by 50%

For every 100 persons housed, this translates annually into
49 fewer hospitalizations, 270 fewer hospital days, and 116
fewer emergency department visits

CHHP cost analyses showed that reductions in avoidable
health care utilization translated into cost savings for the
housed participants, even after taking into account the cost
of the supportive housing (Cost aspects of study previously
described in Wall Street Journal)
CHHP HIV Sub-Study

HIV sub-study examined the impact of housing on disease
progression among the 105 CHHP participants who were
HIV+ (and randomized like other participants)

At 12 months, housed HIV+ CHHP had significantly better
health status:



55% of housed were alive with “intact immunity”, compared to
only 34% of HIV+ participants left to “usual care”
Housed HIV+ participants were much more likely to have
undetectable viral load (36%) as compared to who did not
receive housing (19%)
Such impact on viral load has relevance to HIV
prevention
H&H Study: Background & Methods

Conducted by CDC and HUD
HOPWA program - in
Baltimore, Chicago & Los
Angeles

Half were randomly selected to
receive an immediate HOPWA
voucher

630 HIV+ participants were
homeless (27%), doubled up
(62%) other otherwise at risk of
homelessness (11%) at baseline

Data on HIV risk and health
indicators collected at baseline
and at 3 follow up assessments
over an 18-month period

All received case management,
help finding housing, referral to
medical care and behavioral
prevention interventions

Results just published in AIDS
& Behavior, 2009
H&H Findings

At 18 months, 82.5% of voucher recipients had their “own
place,” compared to 50.6% of control group members



At 6 months, these figures were 54.2% vs. 16.0%
At 12 months, these figures were 87.0% vs. 37.2%
Compared to housed participants, and controlling for
demographics & health status, those who experienced
homelessness during follow up:
 Were significantly more likely to use an ER
 Were significantly more likely to have a
detectible viral load


an outcome with HIV prevention relevance
Reported significantly higher levels of perceived
stress

an outcome which relates to quality-adjusted life expectancy
H&H Study: Typical Annual Service
Delivery Costs Per Client (AIDS & Behavior, 2007)
City
Payor Perspective
Costs
Societal Perspective
Costs
Baltimore
$9256
$10048
Chicago
$11651
$14032
Los Angeles
$10639
$12785
How many transmissions must be averted to be
cost-saving or cost-effective? (AIDS & Behavior, 2007)
Baltimore
Chicago
LA
Average
Cost-saving
Cost-effective
threshold
threshold
.0454
.0128
(23)
(78)
.0634
.0179
(16)
(56)
.0578
.0163
(18)
(62)
.0555
.0157
(1 per 19
(1 per 64
clients)
clients)
Did H&H achieve the thresholds?





To transmit HIV, let’s make a simplifying assumption that there
needs to be some detectable viral load (and of course some risk
behavior)
Applying this to H&H as-treated analysis indicates that 8.2% of
persons with housing might possibly transmit, but 10.5% of
persons without housing could potentially transmit (a difference
of about 2.4%) [(.1332*.614) – (.1332*.791)]
Let’s also assume a mean of 3.81 sexual partnerships per year
in H&H for HIV+ persons who had any seronegative or unknown
serostatus partners, and a one-year, per-partnership
transmission probability of 17.4%
So, .024*3.81*.174 equals roughly .0157 HIV transmission
averted for each housed client in a given year
Cost-per-quality-adjust-life-year-saved is appox. $62,493
Sensitivity Analysis in H&H
Number of QALYs Saved
Per HIV Transmission
Averted
(input parameter)
5.33
{base case}
Cost-Utility Ratio
$62,493
{base case}
7.50
$48,337
11.23
$34,780
How does housing compare to other public health
interventions in terms of cost-effectiveness?
Intervention
Kidney dialysis
Approx. cost per QALY saved
(varies by study)
$52,000 to $129,000
Mammography, 50-69 y.o.
$57,500
Colon cancer screening, 50-85 y.o.
$53,600
Type 2 diabetes screening,>25 y.o.
$63,000
HIV screening every 5 years
$42,200
Syringe exchange
HIV behavioral interventions
PrEP
Cost-saving
Generally cost-saving
$298,000
HIV vaccine
$22,617 to $111,277
Early vs deferred HAART
$15,159 to $36,301
Deferred vs no HAART
$46,423
Mycobacterium avium complex
(MAC) prophylaxis
$44,500
Evaluating HIV Housing: Conclusions

Results of economic evaluation studies of
housing indicate it is either cost-saving or within
the range of interventions generally considered to
be “cost-effective” and “well accepted” by society

Each prevented HIV infection saves hundreds of
thousands of dollars in life-time medical costs,
and even more importantly, years of (qualityadjusted) life

Evidence indicates housing is an effective
and efficient HIV care and prevention
strategy