Improving Access to Disability Benefits Among the Chronically Homeless Population: SSA-HOPE evaluation outcomes

1
Improving Access to Disability Benefits Among the
Chronically Homeless Population SSA-HOPE
evaluation outcomes

• Marion McCoy, Ph.D.
• Mustafa Karakus, Ph.D.
• William Frey, PhD

2
Presidents Initiative End Chronic Homelessness
in 10 years (2013)

• To support this goal, the Social Security
  Administration (SSA) funds HOPE projects that
  provide outreach services and disability
  application assistance to individuals who meet
  the HUD definition of chronically homeless.

3
Previous relevant research

• Income support is a significant factor that
  differentiates people with mental illness who are
  domiciled from those who are homeless (Rossi,
  1989).
• Successful housing outcomes for veterans who were
  homeless and had mental illness are associated
  with receipt of public support payments
  (Rosenheck, et al., 1995).
• Inter-agency efforts to integrate services
  enhance access to disability benefits among
  people who are homeless and have mental illness
  (Rosenheck, et al., 1999).

4
SSA- Homeless Outreach Projects and Evaluation
(HOPE) Awards

• In 2004, 41 HOPE cooperative agreements to public
  or private organizations in every region of the
  country were awarded.
• Programs received 4 years of funding.
• Current HOPE programs end in September, 2008.

5
Project HOPE Goals

• Outreach and engage people who are chronically
  homeless and have a disability
• Assist enrollees with disability benefit
  applications ?
• Improve quality and timeliness of medical
  evidence
• Enable quick determinations and benefit payments
• Reduce processing time for SSA and Disability
  Determination Services (DDS)
• Provide enrollees with housing and support
  services ?
• Improve housing status and quality of life

6
Components of the SSA-HOPE intervention for
HOPE grantees

• A training manual about the SSA disability
  processing system
• Assigned liaisons in Disability Determination
  Services (DDS) SSA offices
• Technical assistance from liaisons staff at SSA
  HQ
• Annual grantee training conferences

7
Study objective

• Assess differences in client outcomes for HOPE
  grantees compared to comparison groups
• Focal outcomes Time to determination decisions
  (i.e., to allow or deny disability entitlements)

8
HUD definition of a person who is chronically
homeless

• Unaccompanied individual
• Has a disabling condition
• Has been continuously homeless for a year or more
  OR 4 episodes of homelessness in last 3 years
• Has been sleeping in a place not meant for human
  habitation or living in a homeless shelter

9
A disabling condition in HUD definition for
chronic homelessness

• Limits an individuals ability to work or perform
  1 or more activities of daily living
• Is diagnosable (e.g., substance abuse disorder,
  serious mental illness, developmental
  disability)
• Includes the co-occurrence of 2 or more of these
  conditions

10
Primary Evaluation Data Sources

• HOPE program quarterly reports
• HOPE Web site records
• Focus groups (with program staff, SSA FO, DDS)
• HOPE site visits and interviews with staff
  enrollees
• SSA 831 data (info that DDS used to make a
  decision about disability benefits)
• For HOPE enrollees
• For Comparison Agency participants

11
Comparison agency groups

• Comparison agencies do similar work with the same
  population but do not receive HOPE funding
• Comparison agencies
• Randomly assigned, 17 agencies received HOPE
  training manual (C1) 14 agencies did not (C2)
• 31 comparison agencies were recruited in the HOPE
  grantees locales 19 had 831 data available for
  the evaluation

12
HOPE Analysis Procedures

• Mixed methods both quantitative and qualitative
  analysis procedures were used
• Content analysis was used to analyze qualitative
  data from focus groups, quarterly reports, and
  site visits
• Descriptive and statistical procedures were used
  to analyze HOPE web site and 831 data

13
HOPE Final sample composition

• 831 data for HOPE enrollees on timing and final
  decisions for benefit allowances.
• N3,055
• 831 data for comparison agency clients on timing
  and final decisions for benefit allowances.
• N198
• Final sample size N 3,253

14
HOPE Final Sample Demographics(Total N 3,253
HOPE programs n3,055 Comp. Agencies n198)
Age HOPE Comp.
Under 30 14.1 10.1
30-39 18.2 18.7
40-49 34.9 33.3
50 32.6 37.9
Average 42.8 years 45.0 years
Median 45 years 47 years
Gender
Male 54.7 45.5
Female 23.0 24.2
No 831 data 22.3 30.3
Note All numbers in percentages except average
and median age. Gender distribution is
significantly different (Chi2 p0.016).
15
HOPE Final Sample Demographics (cont. 2)
Race HOPE Comp.
White 22.8 20.2
Black 14.0 10.6
Other 6.8 8.1
No 831 data 56.4 61.1
Education HOPE Comp.
Less than high school 25.7 36.9
High school 40.9 36.4
More than high school 26.0 20.7
No 831 data 7.4 6.0
Note All numbers in percentages. Education is
significantly different (Chi2 p0.007)
16
HOPE Final Sample Demographics (cont. 3)
Disability HOPE Comp.
Psychiatric/emotional 52.6 56.1
Musculoskeletal 14.3 11.6
Other disabilities 33.1 32.3
Note All numbers in percentages.
17
Cox Proportional Hazard Model
Variables Hazard Ratio Robust std. err. Prob.
HOPE Program 1.24 0.11 0.013
Age LT30 1.01 0.08 0.929
Age 30-39 1.13 0.04 0.001
Age 40-49 1.00 0.03 0.909
Male 1.07 0.06 0.214
Gender - N/A 1.00 0.07 0.962
Psychiatric/emotional 1.06 0.06 0.345
Musculoskeletal 0.93 0.03 0.029
White 1.09 0.10 0.355
Race - N/A 0.75 0.08 0.006
Less than high school 1.06 0.04 0.111
High school 1.03 0.04 0.522
Note Proportionality assumption cannot be
rejected at 0.01 percent. (chi2204.57)
18
Time to determination calculations
Propensity Sore Matching Propensity Sore Matching Propensity Sore Matching Propensity Sore Matching Propensity Sore Matching
N Treatment N Control Avg. Treatment Effect Std. Error T-stat
2,252 196 -0.8 month 0.30 -2.65

• TtoD the time between the date a disability
  application is filed and the date that DDS
  reaches a decision (to allow or deny benefits to
  the claimant).
• HOPE enrollees received determinations almost a
  month earlier than claimants in either of the
  comparison agency groups who received the HOPE
  manual (C1) or agencies that received no support
  (C2).

Note We apply Average effect of Treatment on the
Treated (ATT) estimation with the Radius Matching
method (bootstrapped standard errors) in our
propensity score estimation.
19
Summary

• Outcome analyses show that the SSA-HOPE
  intervention effectively reduced processing
  time for initial disability claims compared to
  local comparison agencies that did not receive
  SSA help.

20
Other findings

• Approximately 41 of initial or reconsidered
  claims were allowed benefits in the final sample
• No significant differences found for
• allowance rates for HOPE program vs. comparison
  agency (C1, C2) clients
• allowance rates for comparison agencies that
  received the HOPE training manual (C1) vs. those
  that didnt (C2)
• HOPE enrollees living status at 1 year was
  significantly improved over status at intake
  (HOPE website data)
• HOPE programs with successful outcomes report
  increased inter-agency coordination and
  cooperation

21
Thank you!