Using data to end homelessness

USING DATA TO END HOMELESSNESS

Joshua D. Bamberger, MD, [email protected]

San Francisco Department of Public HealthUniversity of California, San Francisco, Dept. of

Family and Community Medicine

Housing and Homeless Studies

• Cost• Before and after studies• Randomized controlled trials

• Mortality• Retrospective case control

• Quality of Life• Retrospective cohort studies

• Populations based homeless prevalence• Creating a Narrative

Creating a Narrative

• Housing is less expensive than homelessness• For people w/ homelessness and AIDS, ARVs are

necessary but not sufficient to improve mortality• The right treatment for the condition

Direct Access to Housing- 1600 units in 40 buildings Tailor housing to needs of individual

Initially SRO, now new buildingsPriority to people with multiple disabilities93% with Axis I mental illnessAt least 18% HIV+

SF Health Dept’s Housing

DAH Portfolio

•

253 286604 678 704 704 878

0 0

177 247 339

689

020040060080010001200140016001800

99-00 2001-2 2003-4 2005-6 2007-8 2009-10

2011-2015

NewMaster-lease

Cost: Plaza Retrospective Before and After• 106 Chronically homeless adults• Cost year before housing: $3,132,856 • Cost year after housing: $906,228• Reduction in healthcare costs: $2,226,568 • Cost of program: $1.1million/year• Reduction in public cost in first year: $1.1 million• More than 90% of reduction

among 15 tenants who cost more than $50,000/year prior to being housed

• Regression to the mean

• Brand new building with 174 units• Homeless, high users of a managed care system• Comprehensive healthcare utilization• Randomly assigned to treatment or regular care• Followed prospectively for 5 years• Outcomes included: Healthcare cost,

mortality, jail

Cost: KCC Random assignment trial

Cost- 1811 Eastlake, Seattle

• Compared to controls, housed Ps showed greater reductions in overall costs

• Cost offsets of housing > $4m for 1st year

• More time in housing associated with greater reduction in costs

• 6-mo within-subjects reductions in typical alcohol use

Figure and findings from Larimer et al. (2009)

Mortality

• Ranking of housing from worst to best housing• Private bath better than shared bath• New building better than renovated• Nursing better than no nursing• Senior better than non-senior

Quality of Housing and Outcome

Windsor Empress LeNain PBI CCR West Folsom Dore

Plaza 149 Mason

990 Polk Mission Creek

0.0

5.0

10.0

15.0

20.0

25.0

30.0

R² = 0.76418262009445

Move-out not death

Move-out not deathLinear (Move-out not death)Linear (Move-out not death)

Windsor Empress LeNain PBI CCR West Folsom Dore Plaza 149 Mason 990 Polk Mission Creek

7.6

3.5

6.8

3.9

5.3

2.7

5.0

3.5

2.5

4.0

3.1

R² = 0.388887624467414

Death by Quality of Housing%death

Death Rate/year0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

Death rate Le Nain vs. Mission Creek 2006-2011

Le Nain death %MCSC death %

The more beautiful the housing the better the outcome

POPULATION SNAPSHOT

Veteran PIT Counts, 2009-2012

* CoCs only required to conduct a new count of unsheltered homelessness in odd numbered years; in 2012, only 32% of CoCs opted not to do a new unsheltered count, providing an incomplete picture of trends in the number of unsheltered homeless VeteransSource: PIT data, 2009 - 2012

2009 2010 2011 2012 -

10,000

20,000

30,000

40,000

50,000

60,000

70,000

80,000

90,000

75,609 76,329

67,495 62,619

43,409 43,437 40,033

35,143

32,200 32,892 27,462 27,476

Total VeteransSheltered VeteransUnsheltered Veterans

Num

ber o

f Vet

eran

s

*

2010 2011 2012 2013 2014 20150

100

200

300

400

500

600

HennepinLexingtonTacomaFort WorthBirmingham

Measured

-------Projected

______

Number of Homeless Veterans in 5 Communities with Greater than 40% reduction 2010-2012

2005 2006 2007 2008 2009 2010 2011 20120

500

1,000

1,500

2,000

2,500

0%

2%

4%

6%

8%

10%

12%

14%

16%

1,932

1,9141,530

1,470 1,400

812601

542

14%14%

13%

10%

9%

5%4%

3%

Utah Annualized Chronic Homeless Count: 2005-2012

Chronic Count% Chronic of Total Homeless Persons

Source: 2012 Utah Homeless Point-In-Time Count

2009 2010 2011 20120

50

100

150

200

250

300267

224

177

126

Veterans in Minneapolis/Hennepin County 2009 - 2011

total veterans

2009 2010 2011 2012

775 779

566

351

Point-in-time count for Minneapolis/Hennepin County Con-tinuum

total chronic homeless

21.8424.26

17.59

10.36

• Common values and philosophy of practice, strong leadership, housing first

• Targeting• High level of communication (HIPPA busters)• Use of data to inform policy and measure success

Common aspects of “positive outliers”

Creating a Narrative

• Housing is less expensive than homelessness• For people w/ homelessness and AIDS, ARVs are

necessary but not sufficient to improve mortality• The right treatment for the condition

USING DATA TO END HOMELESSNESS

Joshua D. Bamberger, MD, [email protected]

San Francisco Department of Public HealthUniversity of California, San Francisco, Dept. of

Family and Community Medicine