Veteran Homelessness:Risk FactorsVeteran Homelessness:Risk Factors and Effective Services1

Robert Rosenheck MDYale Medical School

VA New England MIRECCVA New England MIRECCMay 21, 2013

1. Based in part on “Rosenheck RA. Service Models for Assisting Homeless People with Mental Health Problems: Cost-Effectiveness and Policy Relevance in Ellen IG and O’Flaherty B (Eds.) How to House the Homeless. New York: Russell Sage Foundation, 2010.”

Risk factors for Homelessness Among Veterans: Highlights Alcohol and Drug Abuse increase risk 2-5 times

stronger risk risk factor than mental illness alone (NCS NESARCH Edens et al 2011)(NCS, NESARCH, Edens et al. 2011)

VA service connection is strongest measurable protective factor, reducing homelessness risk by 62% (greater than 50% SC) or 47% (lessby 62% (greater than 50% SC) or 47% (less than 50% SC)

OEF/OIF veterans were 65% less likely to have b h l th th tbeen homeless than other veterans

Female veterans 3-4 times greater risk than female non-veterans

Edens E, Kasprow W, Tsai, Rosenheck RA (2011). Association of Substance use and VA service-connected disability benefits with risk of homelessness among veterans. The American Journal on Addictions;20 (5): 412-419.

From “Rosenheck RA and Fontana AF (1994) A Model of Homelessness Among Male Veterans of the Vietnam Generation. American Journal of Psychiatry 151:421-427”

Percentage of veterans among homeless and domiciled US males in 1996: by age group

60%63%

60%ans

O.R =2.0 O.R =3.2 O.R 1.4 O.R =.89 O.R =.88

47% 46%49%

60%

nt v

eter

a

Homeless

34%

38%40%

Perc

en HomelessDomiciled

15% 14%20%

8%

0%20-34

Post VN('82 - )

35-44 PostVN ('73-

'81)

45-54 (VN) 55-64(Korea-

VN)

>64 (WWII+ Korea)

From NSHAPC, 2000 and CPS, 1996

Current Family Status Among y gHomeless Veterans (N=43,853)

Male FemaleMarried 9 6% 8 6%Married 9.6% 8.6%Children

dCustody 12.8% 33.3%Living w 8.2% 27.8%

Family Relationships (N=1,493)

One close family member 84% Number of close family members 6 6 Number of close family members 6.6 Family Support

C ld t l f $100 52 7% Could get a loan of $100 52.7% Could get help with a ride 43.4% Could get emotional support if

suicidal 55.7%

Family Relationships(2) (N=1,493)

Family of… Origin ProcreationOne close rel 76 2% 57 2%One close rel. 76.2% 57.2%Number of rels. 4.0 1.5Support

Loan 45.1% 18.7%Ride 35.4% 16.4%Emotional 48 7% 23 5%Emotional 48.7% 23.5%

Five Intervention Models

1. System Integration. 2 Supported housing 2. Supported housing. 3. Case management alone.

f h 4. Benefits outreach. 5. Supported Employment. 6. Towards Social Re-

integration/Recovery: Group Intensiveintegration/Recovery: Group Intensive

Peer Support (GIPS)

System Integration: The ACCESS Program

“System fragmentation is the problem” 18 site demonstration of efforts to integrate

homeless service systems 7,000 homeless people with serious mental

illness in 4 annual cohorts followed for 12illness in 4 annual cohorts followed for 12 months.

9 sites received $150,00 and technical 9 sites received $150,00 and technical assistance to improve system integration

Other 9 sites did notFrom “Rosenheck RA, Lam J, Morrissey JP, Calloway M, Marilyn Stolar, Randolph F, Blasinsky M, Johnsen, M, Steadman H, Cocozza J, Dennis DD, Goldman HH. Do Efforts to Improve Service Systems Integration Enhance Outcomes for Homeless Persons with Serious Mental Illness? Evidence from the ACCESS Program. Psychiatric Services. 2002:53(8):958-966.

ACCESS Outcomes By Systems Integration Site Assignment and by Annual Cohort: Exited From Homelessness for 30 Days

50

s

30

40

omel

ess

20

30

nt N

ot H

10

Perc

en

0BL 3 Months 12 Months

CohortCohort

Int 1 Int 2 Int 3 Int 4 Cont 1Cont 2 Cont 3 Cont 4

OUTCOMES BY SYSTEMS INTEGRATION ASSIGNMENT AND BY COHORT:PSYCHIATRIC SYMPTOMS*

2

dex*

ptom

Ind

1

ric

Sym

pPs

ychi

atr

0BL 3 Months 12 Months

Cohort

P

Cohort

Int 1 Int 2 Int 3 Int 4 Cont 1Cont 2 Cont 3 Cont 4

Supported Housing: HUD-VA Supported Housing

Combines case management and Section 8 vouchersSection 8 vouchers

Experimental study of three groups:HUD VASH HUD-VASH

Case management onlyStandard VA care Standard VA care

Rosenheck RA, Kasprow W, Frisman LK, Liu-Mares W. Cost-effectiveness of Supported Housing for Homeless Persons with Mental Illness. Archives of General Psychiatry 2003;60: 940-951.

Days housed in past 90: HUD VASH Experimental Evaluation

90

65% days housed

60

ast 9

0

57% days housed

hts i

n pa

HUD-VASH exp. :N=182

53% days housed

30Nig

h N 182Case mgt. N=90

Std. care: N=188

Key:0

Baseline 6 mos.*# 1 year *# 18 mos. *# 2 years*# 3 years

Key:

*Group 1>2, p<.05

#Group 1>3, p<.05

Outcomes in the HUD-VA Supported Housing Program:Nights Homeless in Past 90

45

30

ast 9

0

HUD-VASH exp. :N=182Case mgt. N=90

hts i

n pa

Std. care: N=188Both 22% days homeless

15Nig

h homeless

Key:

14% days homeless

0Baseline 6 mos. *# 1 year *# 18 mos. *# 2 years # 3 years

Key:

*Group 1<2, p<.05

#Group 1<3, p<.05

Other HUD-VASH Benefits

Superior therapeutic allianceM i l More social support

Reduced alcohol and drug use Reduced alcohol and drug use

Outcomes in the HUD-VA Supported Housing Program:Days of Alcohol Use in past 30 days

1010

VASH exp. : N=1827.5

coho

l use

pCase mgt. N=88Std. care: N=187

5

Day

s of a

loc

2.5

D

0Baseline 6 mos. 1 year 18 mos. 2 years 3 years

Three-year VA health care costs by treatment group.$40,000

$36,524$35,095

,

$28,515$30,000

Cos

ts 1. Experimental VASH (N=182)2. CM controls (N=90)

$17,267$19,257

$22,315

$18,628$20,000

VA

C

( )3. Standard care (N=188)

$12,779

$9,886$10,000

$0

Total VA VAIP VA OP

Outcomes in the CICH Supported Housing Program:Percent Nights Housed in Past 90

100%t 9

0

75%

used

in p

ast

50%

Nig

hts H

ou

CICH

25%Perc

ent CICH

Controls

0%Baseline 3 mos 6 mos 9 mos 12 mos

Mares AS, Rosenheck RA. A Comparison of Treatment Outcomes among Chronically Homelessness Adults Receiving Comprehensive Housing and Health Care Services Versus Usual Local Care. Administration and Policy in Mental Health and Mental Health Services Research 2011; 38 (6):459-475

Benefits Outreach VA-VBA initiatives: 70% pensions Joint VA-SSA Benefits Outreach Linked SSA disability specialists and VA case

managers and medical evaluatorsC d J i h i d h VA Compared Joint outreach sites and other VA homeless programs using VA and SSA administrative data over a 4 year periodadministrative data over a 4 year period before and after program implementation.

Program Cost $1,700-$3,200/ additional awardee

SSA-VA Joint Outreach: Rates of Application for Benefits (N=34,431)

2325

Intervention

14

1819

15

20

App

lied

Intervention

87 7

10

14

78

109

1012

1112

10

erce

nt A

7 77

0

5Pe

0Year -

4Year -

3Year -

2Year -

1Year +

1Year +

2Year +

3Year +

4

CohortCohort

Joint Outreach Sites Comparison Sites

SSA-VA Joint Outreach: Rates of Award Among Applicants (N=3,952)

7175

69 67 68 6865

80

efits

5861

5560 57 5855

6167 65

5960

ved

Ben

20

40

nt R

ecei

v

Intervention

0

20

Perc

en

Year -4

Year -3

Year -2

Year -1

Year +1

Year +2

Year +3

Year +4

Year Before/After Program InitiationYear Before/After Program Initiation

Joint Outreach Sites Comparison Sites

SSA-VA Joint Outreach: Rates of Award Among All Outreach Veterans (N=34,431)

13

15

nefit

s

Intervention

1011 11

8

10

ived

Ben Intervention

4 45 5

4

6 6 67

87 7

5

ent R

ece

0Perc

e

Year -4

Year -3

Year -2

Year -1

Year +1

Year +2

Year +3

Year +4

Year Before/After Program Initiation

Joint Outreach Sites Comparison Sites

SSA-VA Outreach:Client Outcomes (Subsample) (1)

Conducted baseline assessment just before benefits decision (N=173).Of those interviewed and followed up 29 % (n=50) Of those interviewed and followed-up 29 % (n=50) received benefits, 71% (n=123) did not.

Compared 3-month outcomes among recipients and i i tnon-recipients.

Total monthly income $735 vs. $458 (p<.001) No difference in psychiatric, substance abuse, or No difference in psychiatric, substance abuse, or

medical status at baseline OR follow-up.

SSA-VA Outreach:Client Outcomes (2)

f b fBenefits No benefits Days homeless 9.4 17.0 ns Quality of Life 3 0 2 7 ** Quality of Life 3.0 2.7 Wants to work 2.9 3.2 ** Days worked 2.3 5.8 *

* p< 05* p<.05

** p<.01

***p<.001

d lSupported Employment Employment specialists work directly with

clients to identify, obtain and retain employmentemployment.

Pre-post intervention cohort design at 9 VA sites

Phase 1: before intervention (N=308) Phase 2: after intervention (N=322)

Employment Specialist Cost $2 063/ client Employment Specialist Cost=$2,063/ client Less productivity of $1,299 =

$764/client/year$ / /yRosenheck RA and Mares AS. Implementation of Supported Employment for Homeless Veterans with Psychiatric or Addiction Disorders: Two-Year Outcomes. Psychiatric Services 2007; 58: 315-324.

Days Competitive Employment inDays Competitive Employment in Supported Employment (Least Square Means)

10

12

6

8 SupportedEmploytmentControl

2

4Control

0bl 3-

mo6-mo

9-mo

12-mo

15-mo

18-mo

21-mo

24-momo mo mo mo mo mo mo mo

Days Housed in past 90 (Independent)(Independent)

(Least Square Means)

354045

38% independently housed

20253035

SupportedEmploymentControl

33% independently housed

5101520 Control

05

bl 3-mo

6-mo

9-mo

12-mo

15-mo

18-mo

21-mo

24-mo

Conclusions from h ( )Research (1)

Diverse intervention models are effective in reducing homelessness gamong people with mental illness.

Effect sizes are typically modest in yp ymagnitude and primarily center in one outcome domain (i.e. housing,

mptom in ome osymptoms, income, or employment). Li it d S i l Limited Social reintegration/recovery

Conclusions from h ( )Research (2)

Program costs can be substantial ($2,000-$9,000/client/year) and are entirely offset by savings only whenentirely offset by savings only when high-cost, high-risk populations are targeted, or when the duration oftargeted, or when the duration of treatment is limited. i.e. there is a cost-effectiveness generalizability tradeoff

l ff Progress is incremental: Cost-effective, generalizable ways to end homelessness among people withhomelessness among people with mental health problems have yet to emerge.