homeless families pilot project evaluation

Homeless FamiliesPilot ProjectEvaluation

Submitted by the California Institute for Mental Healthto the Department of Mental Health

Adult Systems of Care—Special ProgramsCounty of Los Angeles

550 S. Vermont Ave. 11th FloorLos Angeles, CA 90020

August 2005

The authors of this report are Daniel Chandler, Joan Meisel and Pat Jordan. This report would nothave been possible without the gracious consent of participants in the Homeless Families PilotProject who allowed us to interview them and use data collected by project staff. We believe these

parents participated in the hope that their stories would influence funders and policy makers to increaseservices for homeless families; we share their hope.

We are also grateful to the following project staff members for their valuable assistance:

Department of Public Social Services staff members:Margaret QuinnFrances Godoy

Elvie MatiasMichael BonoDebora Gotts

Department of Mental Health staff members:Dennis MurataDolorese Daniel

Cynthia AlmazenJulia Carreon

Elizabeth Gross

Los Angeles Homeless Services Authority staff members:Jeannette RoweDavid Garcia

Downtown Mental Health Clinic staff members:Larry HurstLisa Wong

Marcie GibbsMadeline Rosario

PROTOTYPES program staff members:Anna Aithal

Amy Maddigan

Finally, we thank the project staff who met with us and shared their successes and challenges.

ACKNOWLEDGMENTS

HOMELESS FAMILIES PILOT PROJECT EVALUATION

TABLE OF CONTENTSEXECUTIVE SUMMARY ............................................................................................................................. i

Overview ............................................................................................................................................... i

Methodology.......................................................................................................................................... i

Major findings ....................................................................................................................................... i

RECOMMENDATION ................................................................................................................................ii

PART 1: THE HOMELESS FAMILIES PILOT PROJECT IN CONTEXT .................................................... 1

Background of the project .................................................................................................................... 1

The definition of “CalWORKs mental health clients” .......................................................................... 1

Dimensions of the problem ..................................................................................................................1

The broader context .............................................................................................................................3

PART 2: DESCRIPTION OF THE HOMELESS PILOT PROJECT SERVICE MODEL ................................ 4

Open issues ...........................................................................................................................................6

PART 3: PROJECT PARTICIPANTS AND THE OUTCOMES OF THEIR SERVICES ................................ 7

A. CLIENT CHARACTERISTICS ............................................................................................................... 7

B. PRIMARY OUTCOMES .......................................................................................................................... 7

Program engagement and retention ..................................................................................................... 7

Housing ................................................................................................................................................8

Mental health status ........................................................................................................................... 10

Child well-being and parenting .......................................................................................................... 11

Employment, training and school ...................................................................................................... 13

C. SECONDARY OUTCOMES ................................................................................................................. 14

Material well-being ............................................................................................................................. 14

Parent health and health care ............................................................................................................. 14

Domestic abuse and traumatic violence ............................................................................................. 14

Problems with alcohol or other drugs ................................................................................................ 15

Criminal justice history ...................................................................................................................... 16

Learning disabilities ........................................................................................................................... 16

Child care and transportation............................................................................................................. 16

Social support .....................................................................................................................................17

D. CONCLUSION .....................................................................................................................................17

NOTES ...................................................................................................................................................... 18


EXECUTIVE SUMMARYOverview

The Los Angeles County Board of Supervisorshas established an interagency HomelessFamilies Pilot Project in order to respond to

the needs of CalWORKs-eligible homeless familiesin which a parent has mental health problems.

The Board has funded the project in three phases.

❑ Phase One (initial pilot): In 2002 the boardestablished an initial pilot project located in the“skid row” area to serve 26 homeless familiesthat included a parent who had mental healthproblems. Interagency partners included theDepartment of Public Social Services (DPSS),the lead agency, the Department of MentalHealth (DMH) and the Los Angeles HomelessServices Authority (LAHSA).

❑ Phase Two (two sites): In September 2003, theboard approved a resolution calling for twopilot projects—a continuation of the one inskid row and one in the San Gabriel Valley—toserve 20 clients each. The skid row projectoperates out of the Downtown Mental HealthClinic, and the San Gabriel Valley project islocated at the PROTOTYPES clinic.

❑ Phase Three (expansion): The two existing siteswere expanded and an additional four siteswere added in the spring of 2005. The skid rowsite has 100 participants and each other site has50 participants for a total of 350 families. Eachsupervisorial district contains one site.

At the request of the Board of Supervisors, theDepartment of Mental Health contracted with theCalifornia Institute for Mental Health (CIMH) toconduct an evaluation of the outcomes of the 40participants served in 2003-2004. The contract hassince been modified so that CIMH is also evaluatingthe expanded 350 participants program.

This report presents service outcomes fromPhase Two (two sites) and describes the servicemodel as it has evolved in all three phases, throughApril 2005.

Methodology

Data are drawn from clinician ratings through-out the project, DMH and DPSS electronic records,and a research interview with participants con-ducted approximately one year after admission tothe program. Information on the service modelstems from extensive interviews with staff andadministrators.

Major findings

The service model

❑ Key elements of the service model appearhighly effective, particularly the interagencycollaboration, the use of both clinical and casemanagement staff, and the improved access toSection 8 vouchers.

❑ While Section 8 vouchers seem critical tosuccess, they take time to obtain, not everyonequalifies, and there remain difficulties infinding and obtaining qualified housing withinthe 90 days allowed by Section 8.

Primary outcomes

❑ The program had significant but limited suc-cess (35%) in obtaining rental housing forclients; however, over half of the participantswere still receiving services in May 2005 andmay yet be successful in achieving permanenthousing.

❑ Stability of housing is associated with improvedparticipant mental health status.

❑ Mental health status and functioning in dailyliving improved somewhat, but improvementwas not found on all measures—in part becauseparticipation in therapy was not consistent foralmost half of the participants.

❑ Clients are generally rated by therapists as goodparents, and their capacity to parent improvedduring the program.

❑ Despite project services, parents at follow-upinterviews reported not always being able to getmedical and dental care for their children andthat they sometimes did not have enough food.

❑ The number of school changes was reduced forchildren in the project compared to the previ-ous year.

i


❑ About 45% of participating parents workedduring the study period, and staff consideredthe capacity to work of most participants tohave improved during the program.

Secondary outcomes

❑ About a third of the parents had serious func-tional limitations due to health conditions.

❑ Domestic violence was a problem for 12 moth-ers.

❑ A learning disability is likely for almost onethird of the participants, and these personsreport more disability days and fewer job skills.

❑ Despite receiving assistance with transporta-tion, it remains a significant barrier to obtain-ing services, maintaining children in school,and finding employment.

The broader context

❑ Most participants became homeless for eco-nomic reasons.

❑ A substantial number of additional familiesmeet the criteria for the pilot project. Weestimate between 1,000 and 3,400 familiesreceiving CalWORKs are homeless in a year. Ifeligibility requirements were less stringent,many other homeless families would qualify.

Efforts to achieve long-term housing andemployment are now coming up against systemissues which are beyond the scope of the pilotproject service model. These system issues includelimited short-term housing for families, limitedhousing vouchers, and a growing discrepancybetween welfare income or income from low-wagejobs and the costs of rental housing suitable forfamilies.

RECOMMENDATIONProgram funding should be continued throughJune 2006.

A ll of the interviewed pilot project staffbelieve that it takes at least one year ofintervention for clients to have a reasonable

chance of achieving permanent housing andemployment goals. The results of this evaluationconfirm that a program length of only six to twelvemonths reduces the likelihood of positive out-comes. The current Phase Three program is sched-uled to be funded only until December 30, 2005,with subsequent funding dependent on stateallocations. Since most participants did not becomeenrolled until March 2005, or later, this wouldmean they will receive services for less than a year.Because this program is resource-intensive, it isprudent to continue operating for a sufficientlength of time to allow for a meaningful evaluation.We therefore recommend that program funding becontinued through June 2006.

ii

HOMELESS FAMILIES PILOT PROJECT EVALUATION 1

Background of the project

The Los Angeles County Board of Supervisorshas established an interagency HomelessFamilies Pilot Project in order to respond to

the needs of CalWORKs-eligible homeless familieswhose parent(s) have mental health problems.

The Board has funded the project in three phases.

❑ Phase One (initial pilot): In 2002 the boardestablished an initial pilot project located in the“skid row” area to serve 26 homeless familiesthat included a parent who had mental healthproblems. Interagency partners included theDepartment of Public Social Services (DPSS),the lead agency, the Department of MentalHealth (DMH) and the Los Angeles HomelessServices Authority (LAHSA).

❑ Phase Two (two sites): In September 2003, theboard approved a resolution calling for twopilot projects—a continuation of the one inskid row and one in the San Gabriel Valley—toserve 20 clients each. The skid row projectoperates out of the Downtown Mental HealthClinic and the San Gabriel Valley project islocated at the PROTOTYPES clinic.

❑ Phase Three (expansion): The two existing siteswere expanded and an additional four siteswere added in January of 2005. The skid rowsite has 100 participants and each other site has50 participants for a total of 350 families. Eachsupervisorial district contains one site.

At the request of the Board of Supervisors, theDepartment of Mental Health contracted with theCalifornia Institute for Mental Health (CIMH) toconduct an evaluation of the outcomes of the 40participants served in 2003-2004. The contract hasbeen modified so that CIMH is also evaluating theexpanded 350 participants’ program.

This report contains:

❑ A description of the service model as it hasevolved over three years, based on extensiveinterviews with project staff and related indi-viduals, current through early May 2005.

❑ Outcomes for the 40 clients served in 2003-2004. Primary outcomes include engagement

PART 1: THE HOMELESS FAMILIES PILOT PROJECT IN CONTEXT

and retention in the program, housing, mentalhealth status, status of the children, and em-ployment-related activities. Secondary out-comes comprise material well-being, domesticviolence, health, social support, learningdisabilities, criminal justice involvement, andchild care and transportation. Data are drawnfrom clinician ratings throughout the project,DMH and DPSS electronic records, and aresearch interview with participants conductedapproximately one year after admission to theprogram.

The definition of “CalWORKsmental health clients”

The Homeless Families Pilot Project is intendedto serve homeless CalWORKs-eligible parents whohave mental health problems severe enough toconstitute a barrier to finding housing and employ-ment. However, these problems should not be sosevere as to qualify for federal disability paymentsdue to severe and persistent mental illness. Personswith a diagnosis of schizophrenia, for example, arelikely to receive federal disability payments (SSI)and therefore not be on CalWORKs. Project eligi-bility is determined by the DPSS definition ofhomelessness, but other definitions may be appliedby related agencies, such as administrators of theSection 8 housing vouchers.

Dimensions of the problem

CalWORKs participants with mental healthproblems are three times as likely as CalWORKsparticipants overall to become homeless.

CalWORKs participants who have self-declaredmental health, substance abuse or domestic vio-lence problems are more likely to be homeless thanthose without such problems. The EconomicRoundtable has found that 8.1% of all CalWORKsparticipants with a welfare-to-work plan werehomeless during a one-year period. However,24.7% of those who self-declared they had a mentalhealth need were homeless in a year.1

As many as 3,450 adult family members withmental health problems are CalWORKs partici-pants and have been homeless during a year.

2 HOMELESS FAMILIES PILOT PROJECT EVALUATION

In March 2004, the California Institute forMental Health conducted a random-sample surveyof 310 of the CalWORKs mental health clientsreceiving services at that time. We asked, “Duringthe past 12 months were you homeless on the streetor in a shelter?” A total of 12.7% of the sampledparticipants reported being homeless. If this per-centage is applied to the 8,242 CalWORKs-fundedmental health clients seen in 2003-2004, we arriveat an estimate of 1,047 adults (parents) who werehomeless on the street or in shelters during the yearwho also had a mental health problem that poses abarrier to economic self-reliance.2

However, CalWORKs clients can seek mentalhealth treatment on their own, without it being intheir welfare-to-work plan. In fact, in 2001 (the lastdate for which we have information) a higherpercentage of clients are served in this “back door”mode than those who have the service in theirwelfare-to-work plan and are thus paid for throughCalWORKs funds. A total of 13,970 CalWORKsparticipants in 2000-2001 received mental healthservices—either directly funded by CalWORKs orfunded by other sources. Applying the 12.7% tothat group, we arrived at a figure of 1,774.

Staff profile of a client intemporary shelter:“The client is a 26-year-old single femaleand domestic violence survivor living withher two children ages 7 and 2. Herfiancé was incarcerated due to thedomestic violence and committed suicidein jail in 2003. She became homeless inMarch 2005 after being evicted due tonon-payment of rent. Recently, she and herchildren were placed in a four-monthresidential facility. The client reports thatdespite the difficulty in transferring one ofher children to a school in a differentarea, she is satisfied with the current livingarrangements.”

We have been using the survey definition ofhomelessness—homeless on the street or in shel-ters—but the homeless indicator used by DPSS isless strict, permitting doubling up to count ashomelessness, for example. As noted above, theEconomic Roundtable found that 24.7% of thosewith a mental health indicator in their GAIN recordalso had a homeless indicator. Applying this figure toboth the 8,242 and 13,970 figures discussed above asCalWORKs mental health clients (funded by DPSSor not), we arrived at a total potential number ofparents eligible for homeless family services ofbetween 2,036 and 3,450.

Thus, the number of parents who would beeligible within a year’s time for a homeless familiesprogram that consisted only of those who hadreceived DMH services varies from around 1,000 to3,450 persons per year, depending on both thedefinition of homelessness and how the populationof mental health service recipients is defined.

Among a random sample of CalWORKs mentalhealth clients, those who have been homeless areyounger, more functionally impaired, and lesslikely to achieve employment goals.

In the CIMH March 2003 staff survey of dis-charged CalWORKs mental health clients, a num-ber of differences appeared between persons staffreported had been homeless and persons staffreported had not been homeless (those whom staffwere uncertain about have been omitted).

❑ Demographics. Being homeless during themental health treatment episode was signifi-cantly associated in the staff survey of 2003with age (those homeless were younger) andwith race (Asian Americans and Hispanics wereless likely to be homeless)3, but not withnumber of children.

❑ Impairment. Staff ratings of functioning in dailylife upon admit4 showed significantly moreimpairment for those persons whom staffreported had been homeless during the treat-ment (44.6 vs. 49.6 on a 100-point scale).

❑ Service pattern. Parents that staff reported hadbeen homeless received mental health servicesfor a shorter period of time at discharge thanthose who had not been homeless (an averageof 7.7 vs. 10.5 months).

❑ Outcomes. Having been homeless was related towhether persons were employed at discharge.


According to staff, only 12.9% of clients whohad been homeless were employed when theyterminated services, vs. 27.5% of those who hadnot been homeless.

Homelessness among families is often temporary,especially when defined to include “doubling up.”

The 2002 winter shelter survey of homelessfamily members found 86% were homeless for lessthan six months. In a fall 2004 survey of 373 adultsapplying for DPSS Homeless Assistance, 51% hadbeen homeless for less than a month, 75% werehomeless for less than two months, and 84% werehomeless for less than six months—nearly identicalto the winter shelter figure.5

Eligibility requirements for the Homeless PilotProject exclude many who have similar needs.

The 1,000 to 3,400 persons potentially eligiblefor similar services discussed above referenceCalWORKs participants—and eligibility forCalWORKs is a pilot project requirement. However,many homeless parents and children are not receiv-ing or eligible for CalWORKs supportive services.LAHSA outreach workers in the Homeless FamiliesPilot Project reported that in both years they en-countered substantial numbers of homeless familieswith mental health problems who are not eligible forthe CalWORKs welfare-to-work program. Theseinclude undocumented persons and those with drugfelony records, as well as those who are no longereligible for welfare-to-work due to timing out. Also,if mental health problems were construed morebroadly to include substance use disorders, far morefamilies would be eligible—the Homeless in LA FinalReport6 revealed problems with alcohol and drugs tobe the most commonly reported cause ofhomelessness among both individuals and families.Finally, since the rate of homelessness amongCalWORKs recipients self-declaring domesticviolence is higher than among persons self-declaringeither mental health or substance abuse problems, anapparent need exists for collaborative services fordomestic violence victims as well.

Adults and children in homeless Los Angelesfamilies experience a wide range of behavioral andsocial difficulties.

Over the past 20 years, Los Angeles studiesregarding this population and their children havebeen limited but document the following:

❑ Homeless women having children with themhave high rates of stress, psychiatric disorders,domestic violence, and substance abuse; theyare at high risk for sexual assault.7

❑ Children in homeless families are at high riskfor missing school, delayed academic achieve-ment, the experience of divorce, separationfrom one or both parents, and witnessingviolence. Up to 25% have clinically significantmental health symptoms.8

The broader contextIt is important not to “pathologize” the families

in the Homeless Families Pilot Project, since theproblems they face finding housing are for the mostpart systemic and not due to mental health difficul-ties. While mental health problems may make itmore difficult to find or maintain stable housing,evidence shows that mental health symptoms inhomeless mothers can be the result of the conditionsof being homeless, particularly of living in shelters.9

The Homeless Families Pilot Project can beseen as the initial stages of building a safety net fora specific population numbering only a few thou-sand in Los Angeles. But the larger forces thatproduce families living in poverty on the edge ofhomelessness can not be ignored.10 In the expan-sion to 350 participants, project staff report that thedifficulties families encounter are for the most partdue to a) lack of low-income housing stock, b)limited and restrictive emergency and transitionalhousing, c) difficulty finding funds for first and lastmonth’s rent and moving expenses, and d) thedifficulty of pursuing employment or training whileliving in temporary situations.

In addition, a welfare reform experiment hasdemonstrated that wage supplements that increaseincome of welfare participants who find work to theequivalent of $17.50 per hour can do as much toreduce mental health problems as does treatment.11

Similarly, a recent study demonstrated that a signifi-cant increase in family income in itself is associatedwith a large reduction in child behavioral symptoms.12

So while planning specific services for homelessfamilies in which a parent has mental health problemsis important and will continue to be necessary, largerscale interventions to reduce poverty and promoteaffordable housing (such as expanded voucherprograms for renters and tax incentives for housingdevelopers) could obviate much of the current needfor these specialized services.


The service model has changed over the threephases of the project, based on learningfrom experience but also due to the exigen-

cies of funding. A summary of the basic principlesof the model appears below.

❑ Provide services for one to two years, based onindividual need;

❑ Provide services through a collaboration of theDepartment of Public Social Services, theDepartment of Mental Health, and the LosAngeles Homeless Services Authority;

❑ Co-locate staff to the extent possible;

❑ Provide therapeutic and case managementservices, as well as skills training groups (suchas money management);

❑ Provide active assistance in locating bothtemporary and permanent housing;

❑ Make use of Section 8 housing vouchers andother rent subsidies to the extent possible; and

❑ Provide transportation.

Experience in Phase One and Phase Two hasshown that programs need to be funded for at leasta year, preferably two, in order to have a reason-able expectation of achieving project goals.

The model was originally designed to be a six-month intervention. But experience in Phase Oneand Phase Two suggested that this is an insufficienttime frame within which to establish permanenthousing (with or without Section 8 certificates).Additionally, most participants cannot becomeactively involved in training or seeking employ-ment until they have some stability in their livingsituation. Staff have universally recommended aservice duration of 12 months to two years, de-pending on individual circumstances.

The interagency collaboration involved in thisproject is a crucial element necessary for success.

Families who receive public assistance, who arehomeless, and who also have mental health prob-lems require a comprehensive approach if they areto succeed in finding and retaining permanenthousing and employment. The initial board resolu-

PART 2: DESCRIPTION OF THE HOMELESS PILOT PROJECTSERVICE MODEL

tion required the interagency cooperation of theLos Angeles Homeless Services Authority (LAHSA),the Department of Public Social Services and theDepartment of Mental Health. LAHSA is respon-sible for identifying families in need, providingtransportation, and finding temporary housing.Mental health services (including psychotherapy,case management, crisis assistance, and findingpermanent housing) are provided by a county-operated or county-contracted agency. The leadagency is the Department of Public Social Services(DPSS), which provides CalWORKs eligibility andGAIN services and overall guidance for the pro-gram. Interagency meetings are held weekly totrack participant progress and identify needs. Asteering committee, composed of all agenciesrepresented, meets monthly to recap progress andrespond to agency issues.

CIMH interviews and meetings with agencystaff from all three components of the project overthe last two years have confirmed that this collabo-ration in the pilot project has been effective inmaking necessary decisions and in minimizing thedisagreements and misperceptions that often resultwith interagency efforts.

Co-location of LAHSA and DPSS staff at themental health sites has been very helpful, butpractical problems continue to pose challenges.

Having all of the major project services andresources in one location near the clients to beserved enables clients to access services more easilyand increases the chance that they will followthrough with the things they need to do to achievesuccess. In addition, the collaboration that occurswhen staff are all in the same location results in thedevelopment of more effective policies and proce-dures, as well as more comprehensive and coordi-nated service plans for participants.

Eligibility workers have been assigned to eachmental health site from the beginning of theproject; GAIN workers were assigned beginning inPhase Two. LAHSA staff were readily available inPhase One, since both they and the mental healthagency—the Downtown Mental Health Clinic—were located in the skid row area. The challengewas greater in having LAHSA staff co-located


during Phase Two given the distance from LAHSA’sdowntown location to the project site in the SanGabriel Valley.

Current challenges in Phase Three include a)inability of co-located DPSS workers to access thecomputerized data bases they need while co-located, and b) the time it takes each day forLAHSA staff to pick-up and return the vans theyuse from a downtown garage.

The model envisions mental health agenciesproviding clinical therapy, direct case manage-ment, and homeless assistance by staff dedicated tothis program.

Therapeutic staff, case management staff andhousing specialists are all part of the model. InPhase Two, the mental health provider in the SanGabriel Valley was able to provide both clinicaltherapy staff and case managers. In Phase Three, allsites have both capacities, although the designatedfunding for the project through DPSS does notpermit billing therapy services, so they must beprovided out of the general CalWORKs mentalhealth supportive services funding. This arrange-ment works well for some providers but not forothers.

A first objective of the programs is to establish safetemporary housing.

The Phase One pilot in skid row area utilizedemergency shelters located in the downtown area.Access to this resource was facilitated by thelocation of LAHSA in the area and the location wasconvenient for clients who were homeless in thedowntown area.

Difficulties arose regarding reliance on emer-gency shelters for immediate stabilization ofhousing in Phase Two in San Gabriel, which did nothave any emergency shelters located in the area.Families were reluctant to be housed in a shelteroutside the area, since it could mean changing theirchildren’s school and might jeopardize theirchances for obtaining permanent housing in theSan Gabriel Valley, where they had other estab-lished supports.

These complications with the use of emergencyshelters became even more apparent during thePhase Three expansion. In some instances sheltershad waiting lists to get in. In other cases, thefamilies were not eligible for the shelters because ofa lack of family friendly policies. (Examples in-

clude: not accepting families with sons over 12years old, not accepting infants, not accepting largefamilies, or requiring parents to be married.)

The challenges around the need for immediatestabilization of housing remains a problem and maywell depend on the circumstances within eachgeographic region. Having LAHSA as a housingpartner enhances the chances of making the sheltersystem more responsive to the needs of thesefamilies, but over time other alternatives to the useof shelters (e.g. hotel vouchers, or placing familiesdirectly in transitional housing) may be found to bemore beneficial.

Mental health staff are responsible for findingpermanent housing.

While LAHSA finds emergency shelter andtransitional housing (for up to two years), findingpermanent housing is the responsibility of the casemanagers and housing specialists employed by theparticipating mental health clinics. This is a new rolefor many mental health agencies, but they havetaken it on with enthusiasm and skill.

Section 8 housing vouchers and other forms ofhousing assistance are critical to the model.

A major improvement in the model occurred inPhase Two in which all 40 families were guaranteeda Section 8 voucher if they completed the applica-tion process and were eligible. As will be noted inthe Outcomes Section (see below), having a Section8 voucher has been a necessary but not sufficientcondition for Phase Two clients to obtain perma-nent rental housing.

The shortage of vouchers overall13 has led tohaving only 50 vouchers available for the 350clients in the pilot program’s Phase Three expan-sion. The design of the Phase Three evaluationincludes a test of the importance of the Section 8vouchers. Clients assigned a Section 8 voucher willbe compared with comparable clients on theSection 8 waiting list. This will provide an ex-tremely valuable assessment of whether the modelis viable without Section 8 vouchers.

Assistance from staff is needed even with theSection 8 vouchers. The application process is time-consuming and detailed and may require clients topay for obtaining birth certificates or other docu-mentation. Program staff have spent considerabletime in assisting clients to obtain all the necessarydocumentation and in facilitating all the pertinent


agencies to act expeditiously on the paperwork.

With the lack of Section 8 vouchers for mostclients in Phase Three, the collaborating partnershave sought out additional resources for PhaseThree programs. DPSS has added 14 days of Tem-porary Homeless Assistance and Rental SubsidyAssistance of up to $250 for four consecutivemonths for eligible families who have foundpermanent non-subsidized housing and gonethrough the STEP vocational training program.DPSS also has identified specialized workers torespond to issues relating to move-in costs, perma-nent housing, relocation and moving assistance.Clinical staff at each site are providing moneymanagement and other types of group education toassist clients in gaining the skills necessary forlong-term tenancy. And DMH has identified aHome Options Made Easy (HOME) team thatdedicates its time to networking with landlords andfinding permanent housing.

Efforts have been made to respond to children’s needsand other needs of the parents, but this is throughreferral rather than direct provision of services.

Mental health staff recognize the added risks tothe children in these families and attempt toidentify particular problems experienced by thechildren in the families they serve. But no provisionis in place for systematically screening or assessingchildren. The programs have established referralarrangements with mental health agencies (orsometimes other therapists within the same agency)that specialize in children’s services.

A large percentage of the clients have hadhistories of witnessing or being subjected to violencein their lives, and many have experienced recent orcurrent domestic abuse. A smaller percentage of theclients have active substance abuse issues. Whilecriminal justice involvement is infrequent, manyclients have other legal issues related to their familysituations or other problems. Having expertise torespond to multiple other issues within the scope ofthe programs would be ideal, but currently these arebeing referred to other organizations.

The provision of transportation is an explicitcomponent of the model.

Funds for transportation, vouchers and/or thedirect provision of transportation is a necessary partof these projects. At times the project staff have haddifficulty getting other parts of the system to accept

their responsibilities, e.g. schools are responsiblefor transporting homeless children to school.

Open issuesWhile the project has been remarkably flexible

in adapting the model based on actual experience,open questions remain. Given the complexity of thepersonal and system issues that arise in trying toassist these families, the model continues to evolve.Here is a set of issues that policy makers shouldconsider, particularly if the program is to be furtherexpanded.

❑ How should clients be selected for the programs?Given the scarcity of Section 8 vouchers and thehigh cost of the intervention, how should thetarget population be defined? Should preferencebe given to the clients who face the most barriers(e.g. longer periods of homelessness, or moreserious mental health issues)?

❑ What are the best alternatives for the initialstabilization of housing? Can shelters be mademore “family friendly”? What other optionsmight be used? How can families be supportedin maintaining their ties to the region fromwhich they come?

❑ How can the needs of children best be ad-dressed? Is it feasible to incorporate childspecialists within the service team, or will thismake the model too expensive?

❑ Can clients be moved more quickly into moreintensive education, training, work-relatedactivities to enhance their ability to obtainemployment? Would the addition of specialvocational rehabilitation specialists facilitatethis process?

❑ With the longer time frames, can the programsvary the intensity of the service over time sothat the average cost of the program per clientis reduced?

❑ How essential to the project model is thepresence of the Section 8 vouchers?

The CIMH evaluation of the Phase Threeprograms will provide additional insight into someof these open issues.


A. CLIENT CHARACTERISTICS

The homeless families represent a wide range offamily constellations and life circumstances.

Three of the 40 parents enrolled as mentalhealth clients are male; 37 are female. Thirtyseven have English as a primary language,

three are monolingual Spanish. The largest racial/ethnic group is Hispanic (42%)15 followed byAfrican American (38%), and Caucasian (15%). Amajority are over 35 years of age. Half of thefamilies have only one child, 40% have two orthree, and 10% have four or more, including afamily with eight and a family with nine children.Seventeen percent of the interviewees were married(and lived with their spouse), while 53% had neverbeen married.16

Parents participating in the project are moreeducated than are CalWORKs participants overall:three-fourths of the parents had finished highschool, compared to 49% of Los AngelesCalWORKs participants who are not homeless.17

Most parents received welfare prior to entering thepilot program.

Of the 40 participants, DPSS records show 33had received CalWORKs at some point prior toenrolling in the project. Four had initially receivedCalWORKs between 1990 and 1995; six initiatedCalWORKs between 1996 and 2000, four startedreceiving CalWORKs in 2001, 10 in 2002, andthree in 2003 (but prior to enrollment in the pilot).

Two measures indicate Homeless Pilot Projectparents have psychiatric impairments with roughlythe same degree of severity as other CalWORKsmental health supportive services clients.

In order to judge the severity of the mentalhealth problems of pilot project participants, wecompared them to a random sample of CalWORKssupportive service mental health clients.18

❑ A staff rating of Global Assessment of Func-tioning (GAF) at admission showed the pilotproject participants to be slightly less impairedthan the larger sample of CalWORKs mentalhealth clients.19 Note, though, that a wide rangeof functioning levels is shown in both samples.

PART 3: PROJECT PARTICIPANTS AND THE OUTCOMESOF THEIR SERVICES14

❑ The diagnostic profile of the two groups wassimilar. For example, the two groups have nearlyidentical percentages of depressive disorders:58% for pilot project parents and 59% for thelarger population. One difference, however, isthe somewhat larger percentage (12% vs. 7%) ofpilot project parents with serious mental illness(bipolar or other psychotic disorder).

❑ Clinical staff at both the Downtown MentalHealth Clinic and PROTOTYPES clinic haveworked with CalWORKs mental health clientsfor several years. At both sites they affirmedthat the mental health problems of pilot projectparticipants seem consistent with those ofCalWORKs mental health clients in general.

Taken together, the available informationindicates that the mental health problems of pilotproject parents are not more severe than of otherCalWORKs mental health participants.

B. PRIMARY OUTCOMES

Program engagement and retention

Pilot project parents were slightly more engaged intherapy than a random sample of CalWORKsmental health clients, but less so than is desirable.

The first and critical outcome is retention andparticipation in the program. According to theclinicians’ final summary for 38 clients, six termi-nated early. (Three of these moved away from theservice site.)

In addition, staff report that only 51% of theparticipants had “good” or “very good” participa-tion in treatment. In some cases, obvious reasonsaccounted for the lack of participation (such asworking full time, receiving cancer treatments, orhaving moved a considerable distance away), butsome simply did not use the therapeutic resourcesavailable to them. Although disappointing, thispattern is actually somewhat better than for arandom sample of CalWORKs mental health clientsoverall.20 Figure 1 shows the pattern for bothgroups.


12%12%

2020%%

333%3%

3131%%

335%5%

334%4%

220%0%

1414%%

All MH Supportive Services Clients (N=362)

Homeless Pilot Project Clients (N=35)

Participation in virtually all sessions

Participation in most sessions

Participation sporadic

Participation rare

Figure 1: Participation in mental health therapy:homeless pilot project parents vs. random sample of allCalWORKs mental health clients

76%

7%7%

112%2%

667%7%

112%2%

SGV

Downtown

1 year or less

1 to 12 years

13 to 35 years

27%

Figure 2: Time participants had lived in downtown orSan Gabriel Valley (N=32)

Housing

Families varied widely in their previous housingstability; San Gabriel residents had been muchmore stable.

The number of moves occurring during thestudy period21 that families reported in follow-upinterviews varied from none (four families) to 20(one family). Downtown and San Gabriel Valleyfamilies were very different in their relationship tothe region in which they were served. At admission,downtown participants had been in that area amean of 3.2 years (median of six months) vs. the

14.4 years (median of 16 years) San Gabriel Valleyparticipants reported living in that area.22 Figure 2shows this graphically. Thus, it is very important tohelp most of the San Gabriel Valley residentsremain in the area.

Most families were homeless due to loss of job orincome.

At the interview conducted by clinicians twomonths into the program, clients were asked whythey were homeless when they entered the pro-gram. Many parents reported eviction or rentincreases (13), loss of job or income (4), conflictwith a partner or relative (6), or physical healthproblems (4). Only two persons mentioned amental health problem as contributing tohomelessness.

As of May 15, 2005, 35% of the 40 participantswere living in rented apartments or houses, 20%were in temporary shelters, and 20% were livingwith friends or family.

Homeless pilot project staff kept records of thehousing situation for all 40 clients. Staff distinguishtime-limited shelters from other housing that is nottime limited—referred to as “permanent” housingbut meaning in practice rental housing.

As seen in Figure 3, rental housing supportedby Section 8 vouchers was found by 14 of the 40participants (35%). The other major categorieswere shelters (eight persons or 20%) and friends or

Parents in two familiesdescribe how theybecame homeless:“Where I was residing the heirs sold theproperty and I had to move after theproperty was sold. I had been there for14 years.”

“My work hours got cut, I wasn’t gettingchild support for my son, so I couldn’tpay the rent any more.”


20%

13%

Rented Apartment or house

Emergency or transitional shelter

Family or friends

Other

Unknown

35%

20%

12%

Figure 3: Housing of 40 Year Two Homeless Pilot Projectparticipants as of May 1, 2005

family (also eight persons or 20%). The livingsituation of five persons (13%) was unknown. Inaddition, one person rented a room, two were innursing homes, and one was in a sober living homeafter completing drug treatment. Note that in May2005 none of the participants were living in arented apartment or house unless they were using aSection 8 voucher.

San Gabriel Valley participants were more likelyto find rental housing.

Only four persons of the 20 from the down-town area were in rented apartments or houses. Thedetailed site-by-site report of housing status as ofMay 1, 2005 is shown in Table 1.23

Table 1: Housing status as of May 1, 2005

Type of housing SGV SGV Downtown DowntownN Percent N Percent

Emergency shelter 1 5% 2 10%Transitional shelter 2 10% 3 15%Rented room 0 0% 1 5%Rented house or apart 0 0% 0 0%Rented using Section 8 10 50% 4 20%With parents 0 0% 2 10%With other relatives 2 10% 4 20%With friends 0 0% 1 5%Nursing home 1 5% 1 5%Sober living home 0 0% 1 5%Unknown 4 20% 1 5%TOTAL 20 100% 20 100%

Obtaining a Section 8 housing voucher is animportant and perhaps critical component ofsuccess in achieving permanent housing. Sixty-fivepercent of the pilot participants received a Section8 voucher.

Section 8 vouchers (which provide federal rentsupplements sufficient to keep the renters’ paymentat 30% of their income) were considered a criticalelement of the Year Two Homeless Families PilotProject. A total of 26 or 65% of the 40 participantsreceived the vouchers. Table 2 shows the December31 status of all 40 participants with respect toSection 8 vouchers.

Table 2: Section 8 voucher application status

Status N %

Voucher issued 26 65%Case closed in Homeless Project 4 10%Denied Section 8 4 10%Ineligible 2 5%Failed criminal background check 2 5%Section 8 applied for prior to project 2 5%TOTAL 40 100%

Once a Section 8 housing voucher is obtained,a rental must be found and approved within 90days. A number of problems may make this diffi-cult, including a lack of housing stock, landlords’reluctance to accept vouchers, and delays in inspec-tion of the rental by housing authorities. Although

26 families were granted avoucher, only 14 were invouchered housing as of May 1,2005.24

A majority of parents in follow-up interviews were satisfied withtheir housing situation.

In follow-up research inter-views with 28 parents, 13 werevery satisfied and eight weresomewhat satisfied with theirhousing situation. Not surpris-ingly those in rented housingwere very satisfied while thoseliving with relatives and inemergency shelter were morelikely to be dissatisfied.


Mental health status

Mental health status showed improvement, butonly some results were statistically significant.25

Client-reported mental health status on the SF-12 scale increased to a statistically significantdegree over time. We compared the scores of clientson the scale at two months after admission withscores at the follow-up research interview. Themean score increased (higher scores are better)from 35.6 to 41.1, a statistically significantchange.26

Staff-rated Global Assessment of Functioningscores improved little between ratings made at thesecond month and ratings made at the tenth month(54.5 to 56.6, not a statistically significant change).Ratings on the 17-item Multnomah CommunityAbilities Scale show improvements from twomonths to 10 months; however, this change is notstatistically significant.

Nonetheless, clinical staff were optimistic thatmany participants would be successful in overcom-ing their mental health issues. Staff reported 77% ofparents had made at least “some positive change”on their primary mental health problem. Staff alsowere asked, “How successful do you think theclient will be in overcoming mental health issues?”

Mother of nine children:“Finding a place has been hard becausepeople don’t want to rent to so manykids.”

For 13 clients, staff reported a positive prognosis.Staff indicated 12 clients would be successful onlyif they continued their mental health therapy (or inone case substance abuse treatment). Doubtful ornegative prognoses were offered for 10 clientsbecause of a pattern of non-participation in treat-ment.

Disability days due to mental health symptomsalso declined, but not to a statistically significantdegree.

We compared client-reported mental health-related days of disability27 at two months into theprogram with days of disability at the time of theresearch interview. The overall means were 8.7 attwo months vs. 6.5 at the time of the researchinterview. (For the 26 persons taking part in bothinterviews no statistically significant changeoccurred over time.)

Staff were asked how much positive changeclients made in the ability to carry out the tasks ofdaily life. At the final summary “some positivechange” in functioning was reported for 78%.

Mental health status and otheroutcomes

Among participants in the follow-up interviews,having fewer psychiatric symptoms was associatedwith greater housing stability.

Scores on an extensive self-report symptomscale (the BASIS-32) were statistically significantlybetter if participants lived in their own rentalhousing.28 See figure 4.

In addition, a low number of disability days

Staff on how the programhas helped a parent:“Stabilization has allowed this client tofocus on employment and educationalgoals. Prior to permanent housing shebegan researching training programs andrequirements. She has set a specific datefor when to begin working towardeducational/employment goals.”


0 5 10 15 20 25 30 35 40Average symtom scale score

Living in shelter (N=9)

Rental housing

Friends/family/ other

28.2

12.5

35.4

Figure 4: Symptom scale scores at follow-up, by type ofresidence (lower scores indicate fewer symptoms)

was correlated with having one’s own rentedhousing. These associations suggest two alterna-tive hypotheses: a) those who achieve permanenthousing had fewer symptoms to begin with; or b)achieving permanent housing reduces symptoms.

We used the SF-12 mental health scale to test thesetwo hypotheses.29

❑ At two months after admission, the SF-12mental health scores for persons who wouldeventually end up in permanent housing wereno different from scores of those who did notobtain permanent housing. That is, the lowernumber of symptoms at follow-up of those inpermanent housing are not just a reflection offewer symptoms to begin with.30

❑ Positive change on SF-12 mental health statuswas associated with having a place of one’s ownat follow up.31 Thus, the data are consistentwith the hypothesis that positive change inmental health status results from having perma-nent housing.

❑ However, the finding that positive change inmental health status is associated with havingpermanent housing is also consistent with thepossibility that the causal sequence starts withresponding to treatment which then leads tobetter mental health status. That, in turn, leadsto obtaining permanent housing.

Without a series of measurements over time weare not able to say definitively which factor (havinghousing vs. making rapid progress in treatment)better explains the association of permanenthousing with positive mental health status.

Child well-being and parentingAt intake, staff rated parents’ ability to meet theneeds of their children as “good” or “very good” ina majority of cases; these ratings improved sub-stantially for many clients during the project.

Case managers rated a parent’s ability to meetchildren’s needs as “good” or “very good” at intakefor 21 out of 36 families (58%). None of the parentswere rated as “deficient” or “unsafe.” Final assess-ment information from staff on the same measurewas available for 34 of the families. As seen inFigure 5, “good” or “very good” care increasedfrom 58% to 71% during the study period, a statisti-cally significant improvement.32

The most common reasons for improvementmentioned by staff were increased housing stabilityand decreased stress.

For many parents, child support was not paid; andfor a minority of parents, child custody became anissue.

At the follow-up research interview, 27 families

Inconsistent

Good

Very good

At Admit

At Discharge3%

22%

59%

56%

38%

22%

Figure 5: Change in staff-rated parenting capacities fromadmission to discharge

had a child living with them who had one parentliving elsewhere; only six of these received financialcontributions from the absent parent althougheight were covered by a child support order; onlytwo parents reported receiving full payments.

For various reasons, a number of the familysituations were unstable. Three parents had chil-dren restored to them by child welfare servicesduring the course of the project,33 and one family


was in the process of “reunification.”

Children had health, dental, and mental healthneeds that were not always met.

We asked follow-up interviewees, “How oftenin the past 12 months did your children get themedical care they need?” While 70% respondedthat needed care was always obtained, in severalfamilies obtaining care was quite problematic. Thepercentage saying “all of the time” did not improvebetween two months into the program and thefollow-up interview.

Dental care was even more problematic, as 37%of the families reported at follow-up that theirchildren received the care they needed “very littleof the time.” Figure 6 shows the percentages ineach category receiving medical and dental care.

Failure to get medical or dental care was notrelated to child care, but was related to transporta-tion (it was the reason most frequently given forboth medical and dental). Problems with insurance(including finding a dentist who would take Medi-Cal), lack of money, and getting off work all con-tributed to some degree.

A number of children had serious health andmental health problems, some of which are likelyto reduce the ability of their parents to find em-ployment.

While the average rating on two standardizedchild behavior scales was in the “higher function-

ing” range, some children had behavioral or physi-cal difficulties. In both the interview at two monthsand in the follow-up research interviews, sixparents said that a child in the household had anon-going physical, mental, or emotional problem ordisability that limits the parent’s activities (e.g., byrequiring special care at home or preventing theparent from undertaking a job).

In their final assessment of participants, clini-cians reported parents had told them about anumber of problems concerning their children. Onehad on-going health problems, two had learningproblems or school difficulties, five had symptomsof depression, and three had anger, defiance oroppositional behavior. Two teenagers were hospital-ized during the project due to suicide risk.

Project staff did not systematically assess andserve children in the families, but referrals formental health services were common.

Of the 36 families reported on in staff finalassessments, none of the children had been seen foran assessment or therapy by project therapists. Staffdid, however, refer children in 14 families forspecialized services. In 12 cases the referrals werefollowed-up by parents, at least initially. In researchinterviews with parents, 11 said a child had neededmental health care in the prior year, and eight saidit had been provided.

Children had relatively little absence from school,and changing schools decreased during the studyperiod.

Very little of the time

Some of time

Most of time

All the time70%

56%

13%

7%

3%

0%

13%

37%

Medical Dental

Figure 6: Children’s receipt of medical and dental careover past year, reported by parents (N=30)

Two parents describingwhy getting medicalservices is difficult:“The county assigned me to an HMO, it’sa county facility. [You have to go] at 7:00am whenever you need to be seen, but Ineed to take kids to school then.”

“Medi-Cal was a problem because I wason an HMO. The HMO didn’t cover if youare homeless—you have to live in thearea before they would treat you.”


The 30 participants in follow-up interviews hada total of 53 children over age 5, and 51 werereported by the parent to attend school regularly.34

The percentage of children who had changedschools two or more times in the previous yeardecreased from 80% at an initial interview to 32%in the follow-up research interview.

Employment, training and school

DPSS and client follow-up data shows that 18project participants worked during the studyperiod.

Of the 37 persons with a DPSS GAIN record, 11(30%) had a record of employment during theperiod from October 1, 2003 to September 30,2004. One person left one job to take a better one,otherwise those employed had only one job. Of the27 participants in follow-up interviews (about threemonths after the DPSS data ended), 13 (48%)

“Family counseling should be expandedand directed toward families.... Treatingthe family system should be the primaryfocus of any intervention aimed atchildren.”

From: Cousineau, M. R., Nocella, K.,Cross, T. A., & Jefferson, M.-E. (2003).Neglect on the Streets: The Health andMental Health Status & Access to Care forthe Homeless Adults & Children in CentralLos Angeles.

reported working in the prior 12 months, andseven were working at the time of the interview.However, because the DPSS and interview datapertained to different persons (and time periods),the total number having worked during the studyperiod (according to either one or the other source)was 18 persons, or 45% of all participants.

Follow-up interview participants who workedin the past year reported wages of no more than$10 per hour, and a majority earned less than $7.50

per hour. Only three persons received any kind ofbenefit (sick days, vacation, health insurance orretirement).

About 40% of participants took part in employ-ment-related activities.

According to DPSS records, 14 persons (38%)took part in community service, work experience,remedial education, Self Initiated [educational]Program (SIP), or vocational training during theperiod from October 1, 2003 to September 30,2004.

Out of 29 participants in follow-up interviews,13 (45%) said they had participated in some formof schooling or training within the past year.Neither the time frame nor the categories in theresearch interview coincide exactly with the GAINdata, but the conclusions are similar: 38% of GAINrecords and 45% of participants in follow-upinterviews show some vocational training orschooling activity.

Project staff felt that capacity to work hadimproved for 18 clients.

About two fifths of the 36 clients for whom wehad final assessment information were judged bycase managers to have good or very good capacityto work at the time of the last or most recent visit.Project staff noted that three clients had workedthroughout the project, and staff felt that capacityhad improved for 18 of the other 33 clients.

However, project staff also recorded somesetbacks experienced by clients. For five clients,mental health symptoms still reduced capacity towork. Three of these clients did not engage intreatment. One had an increase in psychotic symp-toms, and a staff member reported, “depression andanxiety have increased due to denial of Section 8voucher; client has been less stable and therefore Ibelieve client’s capacity to work has decreased.”One parent’s capacity was reduced by having toundergo treatment for cancer and another’s byhaving a new infant. In all, eight were noted ashaving negative change in capacity to work, andfour others showed no change.

Overall, in their final assessments project staffdetermined that eventual success in employmentwas likely for about three-fourths of the clients. For15 clients, the prognosis was unambiguous; for


another 12 the prognosis was positive but condi-tional on getting stable housing or continuingmental health treatment.

C. SECONDARY OUTCOMES

Material well-being

In follow-up interviews, 30% of families reportedhunger, and fewer than half the children were per-ceived to be safe in their neighborhoods all the time.

Participants in follow-up interviews were askedwhether there was a time when they or theirchildren were hungry because of inability to affordfood. Of 29 families, the parents in nine families(31%) indicated they or their children had beenhungry in the past 60 days due to inability to affordfood. Seven of these parents said the family hadbeen hungry for two to three days, and two familiesreported they were hungry for four to 10 days ofthe prior 60 days.

Staff describe a client who isgetting permanent housing:“The client is a 35-year-old survivor ofdomestic violence with two children,currently seeking services for symptoms ofpanic attack. She reported becominghomeless after she decided to move to LosAngeles to enroll her children in a schoolprogram for the gifted. She moved underthe impression that her friends in LosAngeles would assist her until sheobtained her own apartment, but thoseplans fell through. She was referred to ashelter program and has just finished herlast interview for their apartmentprogram. She will be placed in one oftheir apartments, which can be paid forby her Section 8 voucher, which shereceived prior to moving to Los Angeles.She has decided to go back to school andis currently enrolled as a ChildDevelopment and Honors English major.”

Participants in follow-up interviews were askedhow often they feel safe in their neighborhood,taking into account things like gangs, drug dealing,street fights, and crime. They also were asked howoften their children are safe from physical harm inthe neighborhood. Less than a third of parents(31%) and only a half of children (52%) wereperceived to be safe all the time—certainly not asituation that any parent would find satisfactory.

Parent health and health care

Almost a third of the participants in follow-upinterviews had health conditions that might affectthe capacity to work.

At follow-up interviews, we used the standard-ized SF-12 scale to measure limitations of function-ing due to health (either mental or physical): 31%of the 26 clients had a score so low on the healthsubscale that it indicated a substantial probabilityof being unable to work.35

Domestic abuse and traumaticviolence

A history of traumatic violence was reported forover half of the participants.

Eighteen of the 30 participants (60%) in follow-up interviews had experienced some sort of trau-matic violence in their lifetime. The highest per-centage (47% of the 30) had witnessed violence.However, 40% had experienced emotional abusefrom a partner, and 37% had experienced physicalabuse; a third had experienced physical abuse as achild, and a quarter had experienced sexual abuseas a child; 10 percent had experienced sexual abuseas an adult.

Five of the 18 participants who had experiencedthese traumatic events had told a physician orcounselor about post-traumatic symptoms. Five hadtaken medications for the post-traumatic symptomscaused by these events; six said the post-traumaticsymptoms had interfered with their activities in thepast year; and one said she had used alcohol ordrugs in response to these symptoms.

About a third of the participants in follow-upinterviews reported some adult domestic abuse inthe prior 12 months; serious physical abuse wasreported by six of 30.


A total of 12 of 30 participants in follow-upinterviews reported at least one abusive act in thepast year. When the 12 respondents were askedhow serious the abuse was, five said, “Not veryserious,” five said, “Moderately serious,” and twosaid, “Very serious.” Victims reported that theabuse either stayed the same (6 persons) or got lessserious (5) over the previous year. Two victimsreported they were currently afraid of a past orpresent partner. Three of the persons were still in arelationship with the abuser while eight were nolonger in the relationship.

Participants in follow-up interviews were alsoasked about whether they had sought help for thedomestic abuse; nine of the 12 had. Parents whoreported abuse were asked if they had been told bya welfare staff person about the domestic abuseoption: 76% reported they had. Five parents saidthat a reason they had received cash aid fromwelfare in the previous 12 months was because they“were trying to get away from an abusive situation.”Two parents reported that they had, in the past twoyears, chosen not to apply for child support (or notto complain if it was not received) due to fear ofpartner abuse. Five parents said they had appliedfor a domestic violence waiver due to partnerabuse. Three reported receiving a referral fordomestic violence services from a welfare worker.

In their final assessments of client status,clinical staff reported that two women had to takeout a restraining order against ex-partners; addi-tionally, police were called to intervene in a conflictbetween a client and her husband. Staff reportedthe degree of change in domestic violence for 12clients: three had strong positive change, five had“some” positive change in their situation; and fourhad no change. Three persons received a “domesticviolence waiver” regarding CalWORKs requirements.

Problems with alcohol or other drugs

Alcohol and drug problems affected a minority ofparticipants, and several made positive changes.

Participants in follow-up interviews were askeda set of alcohol- and drug-screening questions.None had ever lost a job or had job problems dueto alcohol, marijuana or other drug use; andalthough seven had to take drug tests in the pastyear, they all passed.

One participant characterized herself as cur-rently substance dependent, three other partici-pants have had problems with marijuana in the pastsix months, and six are recovering alcoholics and/or drug addicts. In addition, the partners of twowomen are dependent on alcohol and/or drugs, andfive other partners are recovering alcoholics and/ordrug addicts.

Project staff reported making three referrals forsubstance abuse treatment during the course of thestudy period. Two follow-up interviewees reportedgetting substance abuse treatment within the pastsix months—one through a counselor at a shelterand one at a residential program. In each case,methamphetamine was the primary drug. Eachentered treatment voluntarily (as opposed to being

Staff describe a client whofound full-time work whilelooking for housing:“The client is a 47-year-old separatedsurvivor of domestic violence with fouradult children and one minor child wholives with her. Her 17-year-old son wasshot when he was 14; he is angry and hasbeen physically abusive towards her. Shebecame homeless in August 2004 whenshe left her rented property due to abusivemale neighbors and because theneighborhood was dangerous. The clientwas employed at the time that she left herhome and began living in various hotelsand with family members. She eventuallylost her job due to missing multiple days atwork to deal with her son’s school, courtand mental health issues. She now hasobtained full-time employment, isparticipating in an academic program,and is volunteering at her church. Sheand her son are temporarily living withone of her adult daughters.”


mandated to treatment by a court or other agency).Both said they had successfully completed treatment.

In their final assessment of change for the sevenclients they regarded as having substance abuseissues, clinicians rated two as having strong posi-tive change, two had some positive change, one didnot change, and two had negative change duringthe study period.

Criminal justice history

Criminal justice problems affected a small numberof participants.

A criminal justice history can be of consequencefor homeless families in several ways. It can, as it didfor two pilot project parents, disqualify applicantsfrom receiving a Section 8 voucher. If it is a felonydrug conviction, it can disqualify parents fromreceiving Medi-Cal. It can contribute to havingchildren removed from the home by Child ProtectiveServices. Finally, many job application forms askabout criminal justice history.

Of 28 respondents, seven reported having beenarrested and charged with a crime at some time intheir lives. None had been incarcerated in theprevious 12 months, however, and in the two yearsprior to entering the pilot project, only two personshad been in jail—for periods of 42 days or less.

During the study period, one person wasarrested and sentenced to residential treatment ondrug charges. A second person was discovered tohave an outstanding bench warrant when applica-tion was made for a Section 8 voucher. She wassentenced to probation and ordered to take aparenting class.

Learning disabilities

About one third of the participants in follow-upinterviews are likely to have learning disabilities.

DPSS does not automatically screen for learningdisabilities, so the follow-up research interviewcontained the State of Washington learning disabil-ity screening tool—which has an overall 74%accuracy rate for identifying learning disabilities.Of the 29 persons answering these questions, nine(31%) screened positive, including six who re-ported having been in a special program or givenextra help in school.

Respondents who reported being disabled forfive or more days of the previous 30 due to mentalhealth symptoms were significantly more likely tohave high learning disability scores as were thosewho had less than four of the nine work skills whichwe inquired about in the research interview.36

Child care and transportation

For a minority of clients, child care access, costs orquality cause difficulties in working or going totraining, or school.

Only 21 of the 30 clients who participated inthe follow-up research interview said that theyneeded child care for work, training or school. Fora minority of respondents, obtaining child care isan issue. While only four of the 21 said that childcare was very difficult to obtain, another five said itis “somewhat difficult” to obtain. Three parentswere paying for child care ($30, $50 and $100 aweek, respectively). And five of the 21 parents hadto travel a half hour or more in order to take a childto child care.

Several parents said child care problems hadcaused interference with work, school or training.The most common type of interference (30% of allresponses and 100% of the seven cases) was caus-ing the parent to be late or absent. However, fourparents said they had not participated in work,training or school due to child care problems, threesaid they had turned down a job, and two said theyhad to quit or had been fired. All in all, four ofseven who reported child care problems had turneddown, or lost, an opportunity to work, train or goto school due to problems with child care.

Lack of transportation caused problems for asubstantial minority.

Pilot project participants had access to trans-portation for project activities through LAHSA. Ofthe 30 participants in follow-up interviews, 21 saidthey received some assistance with transportationto work, school or training (such as bus passes).However, we were told by staff that transportationto school was a problem for some of the families(especially if they were in temporary housing awayfrom the child’s usual school). Twelve of the 30participants in follow-up interviews reported thatduring the prior 12 months, lack of transportationinterfered with work, school or training. In addi-


tion, only 13 of the parents had a California driver’slicense and only 10 had access to a car if needed forchild care, training, school or work. (Of thosereporting their mode of transportation when theywere working, five drove their own car, four took abus, and one walked to work.)

Social support

A majority of participants in follow-up interviewsreported having support for a variety of needs.

Approximately 80% of the participants infollow-up interviews felt they had someone to turnto for emotional support or help with a personalproblem. However, a majority said they would beunlikely to have anyone willing to provide themwith several hundred dollars if it was needed (forcar repair, say) and 37% said they would not havesomeone who could help take care of them if theywere sick in bed for several weeks. Family weremost likely to provide little support while staff inthe mental health agency were perceived by amajority as providing a lot of support.

Staff describe a case in whichthere were two transportationproblems:“One mother and her son who werehomeless in Tarzana were placed in ashelter in South Central. The boy waspicked up at a nearby school andtransported by the Tarzana school districtto his existing school. The mother and boywere moved to another shelter in SanPedro, where they had to get up at 4:00am each day to drive to the South Centralschool where the Tarzana bus picked theboy up. When the mother’s car brokedown, she used savings to fix it so her soncould continue in his school. However, theshelter required residents to save theirmoney, so using the savings for autorepair resulted in being evicted from theshelter.”

D. CONCLUSION

Pilot project direct service staff—whether theywork for mental health clinics, the Los AngelesHomeless Services Authority, or for the Departmentof Public Social Services—are dedicated and skill-ful. Over time they have gotten better and better atfiguring out ways to help homeless parents withmental health problems. They are actively sup-ported by capable administrative staff at all threeagencies.

We hope that this report contains enough detailto show the reader the many difficulties homelessparents in the pilot project face but also somethingof the resourcefulness they have brought to thetask. Staff can assist, but parents must take action.Parents in the pilot project have confronted, andare overcoming, substance abuse, domestic vio-lence, inadequate social support, limited incomesand other material resources, and homelessnessitself.

The pilot program’s coordinated approach toservices is resource-intensive but pays off in theexperience parents have of support and helpfulnessas well as in the outcomes presented here.


NOTES1 Personal communication from Dan Flaming,

Economic Roundtable. The definition ofhomelessness used is the “indicator” or “flag”entered into the DPSS GAIN data system whenhomelessness is reported.

2 The 95% confidence interval—a range that takesaccount of possible sampling error—for the1,047 figure is 732 to 1,356.

3 In the expanded Phase Three project a majority ofthe participants are Hispanic, thus differing inthis regard from the overall population ofmental health supportive services clients whowere homeless during the year.

4 Global Assessment of Functioning Scale, or GAF.

5 Flaming, D., Burns, P., & Haydamack, B. (Septem-ber 2004). Homeless in LA: Final ResearchReport. Los Angeles: Economic Roundtable 315West Ninth Street, Suite 1209, Los Angeles,California 90015.Bono and colleagues., op cit. These figures areconfirmed by the CIMH study of a randomsample of CalWORKs participants in Kern andStanislaus counties: 60% of those who werehomeless in the prior year were homeless for amonth or less and this time did not differ bywhether the parent had mental health needs(unpublished data).

6 Flaming, D., Burns, P., & Haydamack, B., op cit.

7 Dennison, B., Mendizabal, A., & White, P. (2004).Many Struggles, Few Options: Findings andRecommendations From the 2004 DowntownWomen’s Needs Assessment. Los Angeles: Down-town Women’s Action Coalition.Cousineau, M. R., Nocella, K., Cross, T. A., &Jefferson, M.-E. (2003). Neglect on the Streets:The Health and Mental Health Status & Access toCare for the Homeless Adults & Children inCentral Los Angeles. Los Angeles: Division ofCommunity Health, USC Keck School ofMedicine, Dept. of Family Medicine 3375South Hoover St., Suite H201b, Los Angeles,CA 90007.Zima, B. T., Wells, K. B., Benjamin, B., & Duan,N. (1996). Mental health problems amonghomeless mothers: relationship to service use

and child mental health problems. Arch GenPsychiatry, 53(4), 332-338; Gamst, G. (2002).City of Pomona, La Verne, and Claremont 2002Homeless Count: Final Report. La Verne, Tri-CityMental Health Center; Wenzel, S., Leake, B., &Gelberg, L. (2002). Risk factors for majorviolence among homeless women. Journal ofInterpersonal Violence, 16(8), 739.Nyamathi, A., Wenzel, S. L., Lesser, J.,Flaskerud, J., & Leake, B. (2001). Comparisonof psychosocial and behavioral profiles ofvictimized and nonvictimized homeless womenand their intimate partners. Research in NursingAnd Health, 24(4), 324-335.

8 Zima, B. T., Bussing, R., Forness, S. R., & Ben-jamin, B. (1997). Sheltered homeless children:their eligibility and unmet need for specialeducation evaluations. Am J Public Health,87(2), 236-240.Zima, B. T., Bussing, R., Forness, S. R., &Benjamin, B. (1997). Sheltered homelesschildren: their eligibility and unmet need forspecial education evaluations. Am J PublicHealth, 87(2), 236-240.Zima, B. T., Bussing, R., Bystritsky, M.,Widawski, M. H., Belin, T. R., & Benjamin, B.(1999). Psychosocial stressors among shelteredhomeless children: relationship to behaviorproblems and depressive symptoms. Am JOrthopsychiatry, 69(1), 127-133.

9 Bogard, C. J., McConnell, J. J., Gerstel, N., &Schwartz, M. (1999). Homeless mothers anddepression: misdirected policy. J Health SocBehav, 40(1), 46-62.

10 In California, the percentage of children inpoverty is increasing, with rates of 34% forHispanic children and 24% for African Ameri-can children. Palmer, J., Song, Y., & Lu, H.-H.(2002). The Changing Face of Child Poverty inCalifornia. New York City: National Center forChildren in Poverty. http://www.nccp.org/pub_cpc02.html. Median income for rentershas been decreasing and rental prices have beenincreasing leaving persons on welfare or work-ing in low-wage jobs in progressively moredifficult circumstances.. City of Los AngelesHousing Department, Building Healthy Com-munities 101, available at: http://www.lacity.org/lahd/curriculum/index.html


“For the rest of 2005, rents [in Los AngelesCounty] should increase almost 4% to anaverage of approximately $1,200 for a one-bedroom, $1,500 for a two-bedroom and$1,700 for a three-bedroom unit.” (April 2005).Source: http://www.usc.edu/schools/sppd/lusk/press/item.php?id=548

11 Cremieux, P., Greenberg, P., Kessler, R., Merrigan,P., & Audenrode, M. V. (2004). Employment,Earnings Supplements, and Mental Health: AControlled Experiment. Ottawa: Social Researchand Demonstration Corporation.

12 Costello, E. J., Compton, S. N., Keeler, G., &Angold, A. (2003). Relationships betweenpoverty and psychopathology: a natural experi-ment. JAMA, 290, 223-2029.

13 There have been long waiting lists for Section 8vouchers for some time. Recent changes in thefunding formula have exacerbated shortages,and future Section 8 reductions have beenproposed. Center on Budget and Policy Priori-ties, February 18, 2005, retrieved June 15,2005: http://www.cbpp.org/2-18-05hous-app.htm

14 The number of cases for analysis varies by type ofinformation. Some of the basic demographic,eligibility and housing data is available (in a de-identified format) for all 40 participants. Fourof the 40 participants chose not to be inter-viewed or make their data available to CIMH.Final summaries were completed by clinicalstaff for 38 participants. Data collected byclinicians at two-month intervals during thestudy period is available for 32 participants.Thirty parents participated in the follow-upresearch interview. Participation in the follow-up interviews was reduced because, in additionto the four pilot participants who refused toparticipate in the evaluation at all, another fourwere lost to contact prior to the beginning ofthe interviews. Interviews could not be ar-ranged with two other participants despitemany attempts. Twenty-six persons participatedboth in the initial interview conducted byclinicians at two months into the program andin the follow-up research interview.

15 Compare to 41% among homeless or at risk ofbeing homeless among all CalWORKs partici-pants. Bono, M., et al, op cit.

16 The 53% never married compares to 51% state-wide. Characteristics Survey. CalWORKs.Federal Fiscal Year 2002. http://www.dss.cahwnet.gov/research/CalWORKsCh_1450.htm

17 Bono, op cit.

18 Chandler, D., Meisel, J., & Jordan, P. (2005).Outcomes of CalWORKs Supportive Services inLos Angeles County, Phase Two: Mental Health.Sacramento: California Institute for MentalHealth, 2030 J Street, Sacramento, CA 95814.

19 Global Assessment of Functioning (GAF) scoresat admission were compared for 32 Pilotparticipants with 323 randomly sampledCalWORKs mental health participants. Thescale goes from 1 to 100, with 50 or lowerindicating considerable impairment. The meanswere 50.4 for the Pilot and 48.3 for the randomsample participants. The difference betweenthem is marginally statistically significant:t-score=1.73, p=0.09. (Statistical significancetells us how likely it is that our findings mighthave occurred by chance, particularly due torandom sampling differences. By convention wecall “significant” a between-group differencethat would occur by chance only one out of 20times, i.e. , p≤0.05.)

20 Chandler, D., Meisel, J., & Jordan, P., op cit.

21 The questions referred to the 12 months prior tothe interview date. For some this includedsome period of time prior to entering the Pilot.

22 From the initial interview at two months into theprogram.

23 The minimum time between the time participantsentered the program and May 15, 2005 was 13months; the maximum was 19 months.

24 Reasons are diverse and include: lost to contact(1), serious medical condition (1), still lookingfor housing or in the process of finalizing arental (3), did not find affordable housing solost voucher (2), and moved into a five-yeartransitional program instead (1).


25 Statistical significance is affected by sample size.With groups this small, differences have to bevery large to be statistically significant.

26 The means above are for all 31 clients for whomwe have a two-month score and the 26 forwhom we have a research interview score. Atotal of 22 clients took part in both interviews.Their means were 35.3 at two months and 42.3eight to 10 months later (t-score –2.72,p<0.012.

27 We used a standard two-item scale asking a) howmany days the respondent had been totallyunable to carry out activities of daily life,including work, and b) how many days therespondent had had to “cut down” on theseactivities. Kessler, R. C., Greenberg, P. E.,Mickelson, K. D., Meneades, L. M., & Wang, P.S. (2001). The effects of chronic medicalconditions on work loss and work cutback.Journal of Occupational and EnvironmentalMedicine, 43(3), 218-225. The article describesthe evidence for the way we used of relating the“unable” and “cut down” elements (i.e., tocount each “cut down” day as .40 of an “un-able” day and add that to the unable days).

28 T-test comparing living in rental housing vs.shelter or other temporary housing:t-score=2.21, p=0.04. (Results from a non-parametric ranksum test were nearly identical.)

29 We reported the BASIS-32 results in addition tothese SF-12 findings because the BASIS-32 has32 items rather than five, and is thus likely tobe more reliable; the drawback is that there isno change score, hence the switch to the SF-12and to change scores.

30 The relationship between initial SF-12 scores andlater housing status at follow-up was notsignificant; paralleling the BASIS-32 relation-ship, scores of the SF-12 at follow-up weresignificantly different by housing status atfollow-up (One way ANOVA: F=4.06, df=25, p=0.03.)

31 This relationship is statistically significant inmultiple regression analysis. Higher socialsupport was also associated with positivechange while the number of children in thefamily was associated with negative change.

32 Overall, the rating for 1 person was worse atfollow-up (3%), for 19 (59%) it did not change,and for 12 (38%) it improved. This a statisti-cally significant change: t=-3.57, p=0.001. (Anon-parametric ranksum test was nearlyidentical.)

33 Not the same parents whose children wereremoved.

34 Even early in the program high attendance wasthe rule.

35 The four persons with a score between 35 and 44have, based on national norms, a 27% chance ofbeing unable to work; while the four personswith scores below 35 have a 57% chance ofbeing unable to work. Ware, J. E., Jr., Kosinski,M., & Keller, S. D. (1995). SF-12: How to Scorethe SF-12 Physical and Mental Health SummaryScales. The Health Institute, New EnglandMedical Center, Boston, MA.

36 For disability days t-score=-3.03, p<0.01; forwork skills t-score=-2.2873 p<0.03; for arrests,t-score =1.78, p<0.09.

The California Institute for Mental Health is a non-profit public interestcorporation established for the purpose of promoting excellence inmental health. CIMH is dedicated to a vision of “a community and mentalhealth service system which provides recovery and full social integrationfor persons with psychiatric disabilities; sustains and supports families andchildren; and promotes mental health wellness.”

Based in Sacramento, CIMH has launched numerous public policyprojects to inform and provide policy research and options to both policymakers and providers. CIMH also provides technical assistance, trainingservices, and the Cathie Wright Technical Assistance Center undercontract to the California State Department of Mental Health.

California Institute for Mental Health • (916) 556-3480 • www.cimh.org

California Institute for Mental Health

2125 19th Street, 2nd Floor

Sacramento, CA 95818

homeless families pilot project evaluation

Documents