Homeless Veterans: VA Expands Partnerships, but Effectiveness of Homeless
Programs Is Unclear (Testimony, 06/24/99, GAO/T-HEHS-99-150).

Pursuant to a congressional request, GAO discussed its recent report on
the Department of Veterans Affairs (VA) homeless programs, focusing on:
(1) VA's programs to address homelessness, including efforts made in
partnership with community-based organizations; and (2) what VA knows
about the effectiveness of its homeless programs.

GAO noted that: (1) in addition to the need for housing, homeless
veterans typically have multiple problems, which may include medical and
mental health problems, limited work skills, and long-standing social
isolation; (2) research suggests that effective interventions for the
homeless involve comprehensive, integrated services to address their
multiple needs; (3) VA provides medical, mental health, and substance
abuse treatment to homeless veterans through its health care facilities;
(4) in addition, VA's targeted homeless programs address a variety of
nonmedical needs by providing services such as case management,
employment assistance, and transitional housing; (5) to leverage its
efforts, VA has developed partnerships with other federal departments,
state and local government agencies, and community-based organizations;
(6) while much activity has occurred and many millions have been spent,
VA has little information about the long-term effectiveness of its
homeless programs; (7) VA has conducted some research over the years to
identify program outcomes, but methodological weaknesses in those
studies have limited the extent to which they can be used to assess
program effectiveness; (8) as a result, little is known about whether
veterans served by VA's homeless programs remain housed or employed, or
whether they instead relapse into homelessness; (9) for this reason, GAO
recommended that VA initiate a series of program evaluation studies
designed to clarify the effectiveness of its homeless programs; (10) VA
concurred with this recommendation; and (11) it has one study of
outcomes for veterans judged ready for permanent housing under way and
plans several more on its new homeless initiatives.

--------------------------- Indexing Terms -----------------------------

 REPORTNUM:  T-HEHS-99-150
     TITLE:  Homeless Veterans: VA Expands Partnerships, but
	     Effectiveness of Homeless Programs Is Unclear
      DATE:  06/24/99
   SUBJECT:  Homelessness
	     Veterans
	     Performance measures
	     Program evaluation
	     Federal aid programs
	     Interagency relations
	     Housing programs
	     Disadvantaged persons
IDENTIFIER:  VA Homeless Chronically Mentally Ill Program
	     VA Homeless Providers Grant and Per Diem Program
	     VA Domiciliary Care for Homeless Veterans Program

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO report.  Delineations within the text indicating chapter **
** titles, headings, and bullets are preserved.  Major          **
** divisions and subdivisions of the text, such as Chapters,    **
** Sections, and Appendixes, are identified by double and       **
** single lines.  The numbers on the right end of these lines   **
** indicate the position of each of the subsections in the      **
** document outline.  These numbers do NOT correspond with the  **
** page numbers of the printed product.                         **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
** A printed copy of this report may be obtained from the GAO   **
** Document Distribution Center.  For further details, please   **
** send an e-mail message to:                                   **
**                                                              **
**                                            **
**                                                              **
** with the message 'info' in the body.                         **
******************************************************************

Cover
================================================================ COVER

Before the Subcommittee on Oversight and Investigations, Committee on
Veterans' Affairs, House of Representatives

For Release on Delivery
Expected at 10:00 a.m.
Thursday, June 24, 1999

HOMELESS VETERANS - VA EXPANDS
PARTNERSHIPS, BUT EFFECTIVENESS OF
HOMELESS PROGRAMS IS UNCLEAR

Statement of Cynthia A.  Bascetta, Associate Director
Veterans' Affairs and Military Health Care Issues
Health, Education, and Human Services Division

GAO/T-HEHS-99-150

GAO/HEHS-99-150T

(406172)

Abbreviations
=============================================================== ABBREV

  DCHV - Domiciliary Care for Homeless Veterans
  GPD - Homeless Providers Grant and per Diem
  HCMI - Homeless Chronically Mentally Ill
  HHS - Department of Health and Human Services
  HUD - Department of Housing and Urban Development
  NEPEC - Northeast Program Evaluation Center
  VA - Department of Veterans Affairs

HOMELESS VETERANS:  VA EXPANDS
PARTNERSHIPS, BUT EFFECTIVENESS OF
HOMELESS PROGRAMS IS UNCLEAR
============================================================ Chapter 0

Mr.  Chairman and Members of the Subcommittee: 

We are pleased to be here today to discuss our recent report on VA's
homeless programs.\1 Homelessness is a complex and difficult problem. 
The exact number of homeless is unknown, but on any given night an
estimated 500,000 to 600,000 homeless people live on the streets or
in shelters.\2 The Department of Veterans Affairs (VA) reports that
approximately one-third of the adult homeless population are
veterans, and these homeless veterans suffer about the same
relatively high rates of psychiatric and substance abuse disorders as
the general homeless population.  Over the past decade or so, VA has
established several programs to address the special needs of homeless
veterans; these targeted programs supplement the health care services
provided through VA's medical facilities.  In fiscal year 1997, VA
obligated approximately $84 million to these programs targeted to
homeless veterans.  Other federal departments and agencies have also
developed programs to assist the homeless.  In fiscal year 1997, the
federal government, including the Departments of Education, Health
and Human Services (HHS), Housing and Urban Development (HUD), Labor,
and VA, and the Federal Emergency Management Agency, spent
approximately $1.2 billion on targeted homeless assistance.\3

Federal agencies serving the homeless, including VA, have begun to
coordinate their activities with each other and with community-based
service providers.  These collaborative efforts are intended to
minimize barriers to service, avoid unnecessary duplication of
services, and enhance service provision.  The development of these
programs and the investment in them have generated questions about
their effectiveness.  As you requested, my remarks today will focus
on (1) VA's programs to address homelessness, including efforts made
in partnership with community-based organizations, and (2) what VA
knows about the effectiveness of its homeless programs.  To develop
this information, we conducted work at VA headquarters and VA's
Northeast Program Evaluation Center (NEPEC) in West Haven, Conn., and
reviewed reports from federally funded research programs.  We visited
VA and community-based homeless programs in Little Rock, Ark.;
Denver, Colo.; Washington, D.C.; Los Angeles and San Diego, Calif.;
and New York, N.Y. 

In brief, we found that in addition to the need for housing, homeless
veterans typically have multiple problems, which may include medical
and mental health problems, limited work skills, and long-standing
social isolation.  Research suggests that effective interventions for
the homeless involve comprehensive, integrated services to address
their multiple needs.  VA provides medical, mental health, and
substance abuse treatment to homeless veterans through its health
care facilities.  In addition, VA's targeted homeless programs
address a variety of nonmedical needs by providing services such as
case management, employment assistance, and transitional housing.  To
leverage its efforts, VA has developed partnerships with other
federal departments, state and local government agencies, and
community-based organizations.  While much activity has occurred and
many millions have been spent, VA has little information about the
long-term effectiveness of its homeless programs.  VA has conducted
some research over the years to identify program outcomes, but
methodological weaknesses in those studies have limited the extent to
which they can be used to assess program effectiveness.  As a result,
little is known about whether veterans served by VA's homeless
programs remain housed or employed, or whether they instead relapse
into homelessness.  For this reason, we recommended that VA initiate
a series of program evaluation studies designed to clarify the
effectiveness of its homeless programs.  VA concurred with this
recommendation.  It has one study of outcomes for veterans judged
ready for permanent housing under way and plans several more on its
new homeless initiatives. 

--------------------
\1 Homeless Veterans:  VA Expands Partnerships, but Homeless Program
Effectiveness Is Unclear (GAO/HEHS-99-53, Apr.  1, 1999). 

\2 Martha R.  Burt, Demographics and Geography:  Estimating Needs,
paper presented at the National Symposium on Homelessness Research: 
What Works, cosponsored by the Department of Housing and Urban
Development and the Department of Health and Human Services, Oct. 
1998. 

\3 Homelessness:  Coordination and Evaluation of Programs Are
Essential (GAO/RCED-99-49, Feb.  26, 1999) provides an inventory of
targeted and nontargeted federal programs that assist the homeless. 

   BACKGROUND
---------------------------------------------------------- Chapter 0:1

Veterans constitute about one-third of the adult homeless population
in the United States on any given day.  They form a heterogeneous
group and are likely to have multiple needs.  Many homeless veterans
need treatment for medical or psychiatric conditions in addition to
housing and other supportive services.  Although many questions
remain about how to treat homelessness, a series of research
initiatives launched in 1982 and funded primarily by HHS suggests
that effective interventions for the homeless involve comprehensive,
integrated services.  These initiatives also suggest that a range of
housing, treatment, and supportive-service options needs to be
available to the homeless, and that flexibility is needed to
appropriately match services to the individual needs of homeless
people. 

Although meeting the most basic needs of a homeless person for food,
clothing, and shelter is a first step, it is rarely sufficient to
enable a person to exit homelessness.  Instead, progress in achieving
housing stability requires comprehensive attention to the full range
of a homeless person's needs.  VA estimates that approximately
one-half of homeless veterans have a substance abuse problem,
approximately one-third have a serious mental illness (and of those,
about half also have a substance abuse problem), and many have other
medical problems.  Some homeless veterans need assistance in
obtaining benefits, managing their finances, resolving legal matters,
developing work skills, or obtaining employment.  Supportive services
such as transportation or child care may also be needed.  Problems in
any of these areas can interfere with progress.  As examples,
untreated mental illness may interfere with a person's ability to
retain housing, and lack of transportation may limit access to
medical appointments or job interviews. 

Research suggests that positive outcomes are promoted by integration
of services, as well as by comprehensive services.  Attempts to
address the needs of a homeless person sequentially, or
simultaneously but without coordination, seem less effective than
strategies that involve integrated efforts to address multiple needs. 
For example, homeless people who have both a mental illness and a
substance abuse problem have been found to benefit more from
integrated treatment programs than from programs that approach these
problems separately.  Similarly, the effectiveness of employment and
training programs for the homeless is enhanced by linkage to housing
assistance and supportive services.  Integration is needed in part
because of fragmentation of the homeless service-delivery system,
which involves different organizations that address different needs. 
Case managers can facilitate integration by helping the homeless
obtain services in ways that complement rather than conflict with one
another.  In addition, organizations that serve the homeless can
collaborate to promote integrated, comprehensive service provision. 

Experts suggest that in terms of housing, the goal of homeless
assistance programs should be stable residence in a setting that
allows the highest level of independence each person can achieve. 
For some homeless veterans, independent housing and economic
self-support are reasonable goals.  But for others, including many
seriously mentally ill homeless people, neither full-time work nor
independent housing may be feasible.  Instead, for these individuals,
residence in a supportive environment, such as a group home, may be
the most reasonable outcome.  In addition, transitional housing may
be necessary before a more permanent housing arrangement can be
achieved.  Thus, efforts to assist the homeless require a range of
housing options (including emergency shelter as well as transitional
and permanent housing); treatment for medical, mental health, and
substance abuse problems; and supportive services such as
transportation and case management.  This spectrum of options is
referred to as the continuum of care.  Because the homeless have
diverse needs and local resources vary, flexibility in arranging
partnerships among organizations optimizes the development of a
continuum of care at the local level. 

      VA PROVIDES KEY SERVICES,
      BUILDS CAPACITY THROUGH
      PARTNERSHIPS
-------------------------------------------------------- Chapter 0:1.1

VA provides key services to homeless veterans through its mainstream
health care programs.  In addition, VA has established several
programs specifically targeted to homeless veterans, providing
veterans at some VA facilities services such as case management, work
rehabilitation, or residential treatment for mental illness or
substance abuse.  Because it does not have sufficient resources to
address all the needs of homeless veterans, VA has expanded its
partnerships with community-based providers.  Thus, VA is working
with other agencies to identify and prioritize gaps in service
availability and to develop strategies for meeting those needs--that
is, to develop a continuum of care for homeless veterans. 

Many homeless veterans receive medical, mental health, and substance
abuse services through VA's mainstream health care programs. 
Although VA does not know the extent to which its annual health care
appropriations are spent on medical care and other treatment services
for homeless veterans, recent estimates suggest the amount is
substantially greater than the level of funding for VA's targeted
homeless programs.  VA's targeted homeless efforts include additional
services, such as outreach to identify homeless veterans, case
management to assess the needs of homeless veterans and link them
with appropriate VA or community-based service providers, job
counseling and placement assistance, and referral to residential
treatment programs to address clinical disorders. 

Since establishing its targeted homeless programs, VA has worked with
other service providers and expanded its relationships with
community-based organizations.  This commitment to partnering is
reflected in annual meetings among VA homeless program staff and
other homeless service providers and organizations.  These meetings
are intended to promote a collaborative effort to assess, plan for,
and address the needs of homeless veterans.  VA has acknowledged that
it alone cannot meet all the needs of homeless veterans.  Not only
are its resources insufficient, but VA's homeless programs are not
available in all locations.  By partnering with other providers, VA
increases its potential for stretching its resources to provide
needed services to homeless veterans and ensure better coordination
of services. 

The specific services available to homeless veterans vary across VA
facilities and may be offered through VA or through arrangements made
by VA with community-based service providers.  Through VA's Homeless
Chronically Mentally Ill (HCMI) program, 62 VA medical facilities
contract with existing community-based providers to provide
time-limited residential treatment to mentally ill or substance
abusing homeless people.  For example, some homeless veterans seen at
San Diego's VA Medical Center are referred to the Veterans
Rehabilitation Center operated by Vietnam Veterans of San Diego. 
This facility specializes in treating substance abusing homeless
veterans with post-traumatic stress disorder or serious depression. 
As another example, some homeless veterans with substance abuse
problems or mental illness receive convalescent medical care at
Christ House through a contract with the VA medical center in
Washington, D.C.  Veterans served through these contracts receive
case management from VA staff and may receive some of their medical
or mental health treatment through VA. 

As part of VA's effort to expand its partnerships with
community-based providers and increase the availability of
transitional housing, VA developed the Homeless Providers Grant and
per Diem (GPD) program.  In contrast with the HCMI program, which
involves contracting with existing community-based residential
treatment facilities, the GPD program awards grants and per diem
payments to public and nonprofit organizations that establish and
operate new supportive housing and services for homeless veterans. 
When grants awarded during this program's first 5 years (1994 through
1998) become fully operational, VA estimates that over 2,700 new
community-based transitional housing beds will be available for
homeless veterans.  Moreover, VA has indicated its intention to
continue expanding this program.  To date, a heterogeneous group of
programs has been funded.  In some cases, veterans who have completed
a residential treatment program through VA's HCMI contract program
move on to a GPD facility, which offers transitional housing in
conjunction with supportive services.  As an example, at the West Los
Angeles VA Medical Center, homeless veterans may first be referred
for residential substance abuse treatment and then, once they have
completed such a program, be referred to L.A.  Vets' welfare-to-work
program, where they receive housing and assistance in obtaining and
maintaining employment through a GPD program.  In a few instances, VA
has awarded GPD funds to programs with more unique missions.  For
example, the Veterans Hospice Homestead in Leominster, Massachusetts,
provides housing and support for terminally ill homeless veterans. 

In addition, the Veterans Programs Enhancement Act of 1998 (P.L. 
105-368) authorized VA to guarantee up to $100 million in loans to
construct, rehabilitate, or acquire land for multifamily transitional
housing projects for homeless veterans. 

      EFFECTIVENESS OF VA HOMELESS
      PROGRAMS IS UNCLEAR
-------------------------------------------------------- Chapter 0:1.2

VA's NEPEC monitors and evaluates VA's homeless programs.  Although
NEPEC collects extensive descriptive data, it has only limited
information about the effectiveness of VA's homeless programs. 
Homeless program sites routinely submit data to NEPEC, but this
information is generally used for monitoring program activities
rather than for evaluating program effectiveness.  That is, the data
routinely collected by NEPEC are used primarily to provide program
managers with information about aspects of specific homeless program
sites, such as characteristics of the veterans served and length of
stay in treatment.  This information is used for comparison with
other program sites or with standards established by legislation or
VA policy.  Research designed to evaluate program effectiveness
requires more rigorous and costly data collection methods than those
NEPEC routinely uses for monitoring purposes.  For example, NEPEC
collects some data about program participants upon discharge from a
homeless program, including information about housing and employment
status and changes in substance abuse and mental health problems. 
These data are of limited use, however, in assessing program
effectiveness, because the measures are relatively imprecise and do
not indicate what happens after a veteran is discharged from
treatment.  As a result, VA cannot use this information to determine
whether veterans served by its homeless programs remain employed or
stably housed over the long term.  NEPEC has conducted several
studies in which additional data, sometimes collected on follow-up,
were obtained from program participants.  Results of these studies
led NEPEC to conclude that veterans served by VA's major homeless
programs, the HCMI and Domiciliary Care for Homeless Veterans (DCHV),
derived substantial benefit from their participation.  We found,
however, that methodological shortcomings in that research prevent
firm conclusions about program effectiveness. 

         PROGRAM DATA FOCUS ON
         DESCRIPTIVE
         CHARACTERISTICS, STATUS
         AT DISCHARGE
------------------------------------------------------ Chapter 0:1.2.1

NEPEC collects and analyzes a wide range of descriptive information
regarding program structure, veteran characteristics, program
processes, and veteran status at discharge for specific sites. 
Program managers use this information to monitor and compare program
sites.  These data would indicate if programs failed to conform to
intended guidelines.  For example, by monitoring diagnostic
information, NEPEC can determine whether programs designed for
homeless mentally ill veterans are serving that population. 

When discharged from a VA homeless program involving transitional
housing or residential treatment, a veteran's reported housing and
employment status are recorded.  In addition, participants are rated
for changes in alcohol, drug, and mental health problems, but the
rating system that VA has been using has allowed case managers, at
most, to indicate that the problem has worsened, remains unchanged,
or has improved.\4 These assessments are made at the time that the
veteran is discharged from a DCHV program or at the time that VA
stops paying a per diem fee to a contract residential treatment
facility or a GPD facility.  If VA pays for only part of a veteran's
course of treatment, and the veteran remains in treatment with a
community-based provider after discharge from VA's homeless program,
then the veteran's status upon completion of treatment (which may
occur some time later) is not captured in NEPEC's data. 

In fiscal year 1997, about 8,500 veterans were discharged from VA's
two largest and oldest residential treatment programs, the DCHV
program (in which homeless veterans receive rehabilitative services
while occupying dedicated beds at VA medical centers) and the
contract-based HCMI program.  NEPEC reported that of the homeless
veterans served through the DCHV program, 62 percent successfully
completed the program (that is, the veteran and clinician agreed that
program goals had been met).  It also said that 57 percent of DCHV
veterans were housed at discharge, and 52 percent reported full- or
part-time employment at discharge.  NEPEC reported that of those
served through the HCMI program, 52 percent successfully completed
the program.  It further said that 39 percent of HCMI veterans
reported having their own apartment, room, or house at discharge, and
43 percent reported full- or part-time employment at discharge. 
About three-fourths of participants in each program were rated by VA
as improved in drug, alcohol, and mental health problems.  How to
interpret these ratings, however, is not entirely clear.  Almost all
participants who were deemed to have completed a program successfully
were rated as improved in these domains.  It is difficult to
interpret a rating of improved with regard to drug or alcohol use
when that assessment is made at the end of a program that requires
participants to avoid alcohol and drugs (as VA residential treatment
programs do), especially when the only alternative ratings are
unchanged or worse.

During fiscal year 1997, over 1,000 veterans were discharged from
VA's GPD program.  Reported outcomes were less favorable for these
veterans; in particular, the proportion of unsuccessful discharges
from GPD programs was high.  As VA noted, however, the GPD program is
relatively new, and early data may not provide a clear basis for
evaluation.  For example, veterans who were benefiting from their
placements might not have been discharged from the GPD program yet,
so no information about them would have been included in the data. 

--------------------
\4 NEPEC has indicated its intention to begin using a 5-point rating
scale to assess changes in alcohol, drug, and mental health problems. 

         LIMITED INFORMATION
         AVAILABLE ABOUT PROGRAM
         EFFECTIVENESS
------------------------------------------------------ Chapter 0:1.2.2

Although outcome research can be difficult and costly, VA has
acknowledged the need for program evaluation and includes such
efforts in its strategic plan under the Government Performance and
Results Act of 1993.  In addition to routine monitoring of homeless
programs, NEPEC has conducted studies that suggest that veterans
served through VA's homeless programs are better off after receiving
program services than before admission.  Methodological shortcomings
in that research, however, prevent strong conclusions regarding
program effectiveness.  NEPEC does not typically collect or examine
data in a way that clarifies the long-term effectiveness of its
programs, the effect of specific interventions in comparison with
alternative treatments, or which interventions work for specific
populations.  We noted in our April report that program effectiveness
could be clarified by additional evaluation research. 

To identify the benefits associated with program participation, NEPEC
conducted pilot follow-up projects at a sample of its homeless
program sites between 1987 and 1990, using more detailed outcome
measures than VA typically collects from program participants. 
Follow-up is needed to determine whether veterans are still employed,
housed, or successfully dealing with substance abuse or mental health
problems after program completion.  NEPEC concluded that, compared
with their status at admission, veterans showed improvements in
housing, employment, mental health, and substance abuse problems 6
months after discharge from DCHV treatment and that, with the
exception of alcohol use, these improvements remained evident 1 year
after discharge.  Similarly, veterans who participated in the HCMI
program were assessed from 1 month to 2 years after their initial
contact with VA homeless staff.  On average, these veterans were last
interviewed 8.3 months after their first contact.  About two-thirds
were admitted to residential treatment; of these, some were still in
residential treatment when last interviewed.  NEPEC concluded that
veterans who participated in VA's HCMI program (including both those
who were and those who were not provided with contract residential
treatment) showed improvements in terms of housing, employment,
psychiatric problems, and substance abuse upon follow-up relative to
initial contact. 

These follow-up studies represented a major undertaking in terms of
resources and effort, and they suggest that the DCHV and HCMI
programs are worthy of further investigation.  However, these studies
had two major shortcomings that NEPEC acknowledged in its reports and
that limit the extent to which firm conclusions can be drawn about
program effectiveness.  First, post-program outcome data were not
obtained from a substantial number of veterans.  As a result,
interview data were not collected from a fully representative sample. 
Follow-up interviews were conducted with only 72 percent of the
veterans who agreed to participate in these studies.  Although the
status of those veterans who were not reinterviewed is not known, it
is possible that the veterans who were doing the poorest were also
less likely to be reinterviewed.  As a result, the data from those
who were reinterviewed could suggest more positive outcomes than
would be true for the program as a whole.  Second, no data were
obtained from veterans who did not participate in the DCHV or HCMI
programs.  Data from such groups would have allowed an estimate of
the degree of improvement attributable to the DCHV or HCMI programs. 
It is possible that some of the improvements noted among those
veterans who were reinterviewed would have occurred in the absence of
DCHV or HCMI treatment.  Other research suggesting that some
improvement over time may occur among the homeless, even in the
absence of intensive treatment, highlights the importance of
comparison data.  Without data from an appropriate comparison group
of veterans who were not served through VA's homeless programs, VA
cannot determine how much veterans benefited from those programs. 

In addition, NEPEC analyzed data from small subsamples of
participants in the HCMI follow-up study to examine relationships
between measures of program participation and improvement.  These
analyses suggested that certain aspects of participation in the
program, such as longer stays in residential treatment, were
associated with greater improvement.  Again, these findings are
promising, but NEPEC acknowledged that strong conclusions could not
be reached because of methodological limitations.  Research designed
to clarify the processes that make interventions effective, or what
aspects of treatment are associated with positive results for
different clinical groups (for example, those with serious mental
illnesses or those with a substance abuse disorder), can yield
information relevant to efforts to improve programs or to optimize
program outcomes.  NEPEC officials have occasionally conducted such
analyses, which require them to supplement their data files with
additional information (for example, about treatment approaches). 
Clear conclusions about what treatment strategies are most strongly
associated with achieving housing stability, and about which
strategies work best for which veterans, require more rigorous and
costly research methods than NEPEC has typically employed. 

NEPEC officials stated that they have not conducted additional
evaluation research on VA's core HCMI and DCHV programs because
obtaining follow-up information on this hard-to-serve population is
difficult and expensive.  A NEPEC official estimated that if it were
to conduct another follow-up study, the cost would be about $60,000
per site per year, and noted that multiple sites would be needed to
ensure generalizability.  Total cost would thus depend on the number
of sites sampled and the length of the follow-up interval. 

NEPEC is not currently conducting evaluation research on its largest
residential treatment and transitional housing programs (the DCHV,
HCMI, and GPD programs).  It is, however, studying some of VA's other
programs.  Follow-up data are being collected from participants in
one of VA's smaller programs called the Housing and Urban
Development-VA Supported Housing program.  In this program, intensive
case management and vouchers for permanent, subsidized housing are
made available to homeless veterans through a cooperative arrangement
between VA and HUD.  To evaluate this program, NEPEC has collected
follow-up information from a sample of program participants, as well
as from a comparison group of veterans who were considered
appropriate candidates for permanent housing but who were randomly
assigned to receive either intensive case management without a
housing voucher or more traditional case management through VA's HCMI
program, again without a housing voucher.  In addition, veterans who
have participated in the Compensated Work Therapy/Transitional
Residence and VA Supported Housing programs are also reinterviewed
periodically.\5 VA has recently indicated its intention to initiate
three new homeless programs and to evaluate each of those programs
using follow-up procedures similar to those it has used in the past. 
These new initiatives involve using a promising case management
strategy called Critical Time Intervention, developing programs for
homeless women veterans, and implementing a vocational service called
Therapeutic Employment, Placement and Support.  A NEPEC official
acknowledged that to minimize the cost of these evaluative efforts,
the methods used to evaluate Critical Time Intervention and the
homeless women programs are likely to be less rigorous than would be
ideal. 

In our April report, we recommended that a series of program
evaluation studies be conducted to clarify the effectiveness of VA's
core homeless programs and provide information about how to improve
those programs.  We concluded that this series of studies should
address long-term effects, processes associated with positive
outcomes, and program impact.  Thus, VA could design follow-up
studies to examine, for example, the stability of housing and
employment in the year or two after discharge from transitional
housing or residential treatment.  VA could also undertake outcome
evaluations designed to assess program processes to better understand
the factors that produce desirable outcomes and how they could be
replicated.  Such studies could also identify aspects of treatment
that are associated with positive outcomes for veterans with
different conditions.  Finally, VA could estimate how program
outcomes differ from outcomes that would be likely in the absence of
the program.  For example, results observed for a sample of homeless
veterans who received a particular kind of treatment could be
compared with results for a control group who did not receive that
treatment.  We also recommended that, where appropriate, VA should
make decisions about these studies (including the type of data needed
and the methods to be used) in coordination with other federal
agencies with homeless programs, including HHS, HUD, and Labor. 

Even though evaluation research can be difficult and expensive to
conduct, we concluded that such studies are necessary to ensure that
VA directs its resources to those efforts with the greatest potential
for beneficial effects.  VA concurred with our recommendation and
described plans to initiate evaluations of several new homeless
projects and to supplement NEPEC's budget with $600,000 from the
additional $50 million VA requested for its homeless programs in its
fiscal year 2000 budget. 

In summary, VA provides medical, mental health, and substance abuse
treatment to homeless veterans through its mainstream health care
programs, and it offers additional specialized services for homeless
veterans at many of its medical centers and through partnerships with
community-based service providers.  As VA facilities attempt to
develop a continuum of care for homeless veterans, variations in
local needs and resources will result in different patterns of
involvement for VA and its partners.  Because homeless veterans
differ from one another in their needs, no single treatment program
can serve all veterans with equal effectiveness.  Local programs
designed to serve groups with different needs are likely to be
important components of any continuum of care for the homeless.  VA
has obtained some information about outcomes for veterans who have
participated in its programs, but methodological shortcomings of that
research prevent clear conclusions about program effectiveness. 
Further research on program effectiveness could provide the
information needed to make decisions about how to direct VA's limited
resources and improve its homeless programs. 

--------------------
\5 In addition, a small sample of veterans who participate in the GPD
program are being surveyed within a few months of their discharge
from that program, but the questions focus on verification of the
services received, rather than on outcomes. 

-------------------------------------------------------- Chapter 0:1.3

Mr.  Chairman, this concludes my prepared statement.  I will be happy
to answer any questions you or other Members of the Subcommittee may
have. 

   GAO CONTACT AND ACKNOWLEDGMENTS
---------------------------------------------------------- Chapter 0:2

For future contacts regarding this testimony, please call Cynthia A. 
Bascetta at (202) 512-7207.  Individuals making key contributions to
this testimony included George Poindexter, Kristen Anderson, Timothy
Hall, Jean Harker, and Deborah Edwards. 

*** End of document. ***