Open Collections

UBC Faculty Research and Publications

Vancouver At Home: pragmatic randomized trials investigating Housing First for homeless and mentally… Somers, Julian M; Patterson, Michelle L; Moniruzzaman, Akm; Currie, Lauren; Rezansoff, Stefanie N; Palepu, Anita; Fryer, Karen Nov 1, 2013

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-13063_2013_Article_1553.pdf [ 414.79kB ]
JSON: 52383-1.0074664.json
JSON-LD: 52383-1.0074664-ld.json
RDF/XML (Pretty): 52383-1.0074664-rdf.xml
RDF/JSON: 52383-1.0074664-rdf.json
Turtle: 52383-1.0074664-turtle.txt
N-Triples: 52383-1.0074664-rdf-ntriples.txt
Original Record: 52383-1.0074664-source.json
Full Text

Full Text

RESEARCH Open AccessVancouver At Home: pragmatic randomized trialsinvestigating Housing First for homeless andmentally ill adultsJulian M Somers1*, Michelle L Patterson1, Akm Moniruzzaman1, Lauren Currie1, Stefanie N Rezansoff1,Anita Palepu2 and Karen Fryer1AbstractBackground: Individuals with mental illnesses are overrepresented among the homeless. Housing First (HF) hasbeen shown to promote positive outcomes in this population. However, key questions remain unresolved,including: how to match support services to client needs, the benefits of housing in scattered sites versus singlecongregate building, and the effectiveness of HF with individuals actively using substances. The present studyaimed to recruit two samples of homeless mentally ill participants who differed in the complexity of their needs.Study details, including recruitment, randomization, and follow-up, are presented.Methods: Eligibility was based on homeless status and current mental disorder. Participants were classified as eithermoderate needs (MN) or high needs (HN). Those with MN were randomized to HF with Intensive Case Management(HF-ICM) or usual care. Those with HN were randomized to HF with Assertive Community Treatment (HF-ACT),congregate housing with support, or usual care. Participants were interviewed every 3 months for 2 years. Separateconsent was sought to access administrative data.Results: Participants met eligibility for either MN (n = 200) or HN (n = 297) and were randomized accordingly. Bothsamples were primarily male and white. Compared to participants designated MN, HN participants had higher rates ofhospitalization for psychiatric reasons prior to randomization, were younger at the time of recruitment, younger whenfirst homeless, more likely to meet criteria for substance dependence, and less likely to have completed high school.Across all study arms, between 92% and 100% of participants were followed over 24 months post-randomization.Minimal significant differences were found between study arms following randomization. 438 participants (88%) providedconsent to access administrative data.Conclusion: The study successfully recruited participants meeting criteria for homelessness and current mental disorder.Both MN and HN groups had high rates of substance dependence, suicidality, and physical illness. Randomizationresulted in no meaningful detectable differences between study arms.Trial registration: Current Controlled Trials: ISRCTN57595077 (Vancouver at Home study: Housing First plus AssertiveCommunity Treatment versus congregate housing plus supports versus treatment as usual) and ISRCTN66721740(Vancouver At Home study: Housing First plus Intensive Case Management versus treatment as usual).Keywords: Housing First, Homelessness, Mental illness, Concurrent disorders* Correspondence: jsomers@sfu.ca1Somers Research Group, Faculty of Health Sciences Simon Fraser University,8888 University Drive, Burnaby V5A 1S6, CanadaFull list of author information is available at the end of the articleTRIALS© 2013 Somers et al.; licensee BioMed Central Ltd. This is an open access article distributed under the terms of the CreativeCommons Attribution License (, which permits unrestricted use, distribution, andreproduction in any medium, provided the original work is properly cited.Somers et al. Trials 2013, 14:365 who are homeless and mentally ill are het-erogeneous in their health and social challenges. Effect-ive models of service must be responsive to individualneeds, which may vary across time and space, and areconstrained by pragmatic factors, including local standardsof care, housing availability, and funding. The VancouverAt Home (VAH) project has implemented two random-ized controlled trials (RCTs) involving homeless mentallyill adults in Vancouver, BC, Canada. VAH is collaboratingwith similar projects in four other Canadian cities [1].Each collaborating center has incorporated a commonmethodology, with pragmatic adaptations in each site.Site-specific adaptations were influenced by the character-istics of each local population (for example, ethno-racialservices in Toronto, ON, Aboriginal focus in Winnipeg,MB), as well as the structural features of each locale (forexample, rural service models in Moncton, NB, congregatehousing in Vancouver, BC). The purpose of the presentarticle is to describe the unique features of VAH, includingmeasures, interventions, and sample characteristics in ac-cordance with the Consolidated Standards of ReportingTrials (CONSORT) statement for the reporting of prag-matic trials.We briefly describe the physical setting of Vancouver,the local population who are both homeless and men-tally ill, and the Housing First (HF) program. These fac-tors influenced both the design and implementation ofVAH with the goal of maximizing the effectiveness andrelevance of the project.For decades, the city of Vancouver has struggled tomeet the needs of a visibly homeless and inadequatelysheltered population in a central downtown neighbor-hood. The same neighborhood has been afflicted by highcrime rates, an open market for illicit drugs, infectiousdiseases, and premature mortality [2]. For many years,the most affordable housing in the neighborhood hasconsisted of single room occupancy (SRO) hotels, manyof which earned a reputation for hazards ranging frombed bugs to criminal predation [3,4]. Individuals withmental illness are prominent among the homeless,particularly following ‘deinstitutionalization’, wherebyregional psychiatric facilities were downsized as prom-ises to implement community-based support, such asAssertive Community Treatment (ACT), were unful-filled. A diverse array of services emerged over timeto support individuals in the neighborhood, includingoutreach services, meal programs, shelters, drug-relatedservices (for example, needle exchange, supervisedinjection site [5]), and varied forms of supportedhousing.The City of Vancouver has implemented a plan withthe goal to ‘end street homelessness’ by 2015. A keyelement of the plan involves the construction ofapartment buildings to provide housing and support forthe homeless. However, key questions remain unresolvedregarding the appropriate mix of occupants in thesebuildings and the type of support that would be requiredto promote stable occupancy and recovery among indi-viduals who are leaving homelessness and who have dif-fering needs.Prior to VAH, no study had systematically examinedthe health and housing status of individuals who werehomeless and mentally ill in Vancouver. A considerableamount of anecdotal and descriptive information wasavailable from sources, such as shelter operators, streetoutreach clinicians, police, and from research involvingsamples that included homeless individuals (for example,patients with HIV/AIDS, survival sex workers [4,6,7]).The available evidence suggests that the local populationof homeless mentally ill individuals struggles with com-plex social and medical problems, including infectiousdiseases, frequent polysubstance use, cognitive impair-ment, trauma, victimization, and poor food security[4,8]. It had also been reported that homeless individualswere using emergency and hospital services due to inad-equate community care, and were frequently involvedwith the justice system [7,9]. Based on these consider-ations, it was anticipated that VAH would extend the HFmodel to clients with more complex needs than thosedescribed in previous trials, including participants withconcurrent substance use disorders.A growing literature supports the effectiveness of theHF model for individuals who are both homeless andmentally ill. HF emphasizes the value of client choiceand has been shown to promote residential stability[10,11], community integration [12], and high levels ofclient satisfaction [13]. Originating from the Pathwaysmodel in New York City, NY, USA, HF involves buildinga portfolio of rental accommodations (typically apart-ments) scattered throughout different neighborhoods,thereby providing clients with meaningful choices con-cerning the location and setting of their residence [14].Clients are then supported in their homes by either anHF with Assertive Community Treatment (HF-ACT)team or HF with Intensive Case Management (HF-ICM),depending on their level of needs. ACT was originallycreated to constitute a ‘hospital without walls’ enablingindividuals who might otherwise have been admitted topsychiatric facilities to instead pursue recovery in com-munity settings [15]. ACT teams are available 24/7, andinclude varied expertise across multiple disciplines. Theeffectiveness of ACT has been well established amongindividuals who reflect the original target population(that is, individuals with psychotic disorders or bipolardisorder), including specific outcomes, such as reduc-tions in hospital admissions and criminal justice involve-ment [16,17]. However, the effectiveness of HF-ACT isSomers et al. Trials 2013, 14:365 Page 2 of 20 well known among sub-populations who also havecognitive impairments, complex addictions, or multiplephysical and mental illnesses.The majority of studies examining HF have followedparticipants for up to 24 months (for example, Gulcur[11], Tsemberis and Eisenberg [14]). Longer-term re-search involving diverse residential interventions forhomeless individuals in New York City found that hous-ing stability decreased after 1, 2, and 5 years (75%, 64%,and 50%, respectively) [18]. Several studies have reportedthat substance use disorders are predictive of lowerhousing stability regardless of residence type [19,20].Preliminary evidence suggests that congregate HF (thatis, a single supported building) may achieve housing sta-bility and cost savings among homeless men who are al-cohol dependent [21]. However, it is not known whetherthese results would be replicated among mentally illusers of illicit (or multiple) substances and who receivescattered site HF.Compared to ACT, the effectiveness of ICM has re-ceived less empirical attention in the context of sup-ported housing for individuals with mental illness, andits definition varies widely [22]. Unlike HF-ACT, whichprovides a broad range of specialized services directly toclients, HF-ICM operates as a liaison connecting clientswith community services based on their expressedneeds. The success of HF-ICM is therefore a function ofthe complexity of client needs, as well as the availabilityand appropriateness of relevant community resources.HF-ICM may be appropriate for clients with less severemental illness or as a step-down from HF-ACT followingsuccessful stabilization. For example, as part of a multi-center trial, the Boston McKinney study [22] random-ized homeless adults with mental illness to independentapartments or small group homes, both of which re-ceived ‘comprehensive case management’. Findings indi-cated that housing availability, regardless of type, wasthe primary predictor of subsequent ability to avoidhomelessness, while enhanced services reduced therisk of homelessness if housing was also available.Substance abuse was the strongest single predictor ofdays homeless [22,23].The primary objectives of this article are:1. To report study details including measures andinterventions that are unique to VAH.2. To present details of recruitment and follow-uprates for participants in VAH including primary datacollection and administrative data.3. To present baseline characteristics and examinepotential non-equivalence between randomizationarms in each trial.4. To examine differences in the complexity of needsbetween participants in the two trials.MethodsCommunity engagementDevelopment of the research protocol was preceded byseveral community meetings and by six focus groupswith individuals who had experienced homelessness andmental illness in the Vancouver area. In total, 58 individ-uals were convened with the assistance of communityagencies, and met privately with an experienced aca-demic facilitator who took notes and prepared reports ofproceedings. Focus group participants were asked to ad-vise on procedures to enhance the relevance of the re-search, to minimize risks and maximize benefits toparticipants, and ways to incorporate the expertise of in-dividuals with direct experiences of homelessness in theresearch project. Narrative feedback from respondents(for example, amounts of honoraria, the need to includeindividuals who had experienced homelessness as mem-bers of service teams and on the research team) were in-cluded in the grant application and later implemented aspart of the project. Service provider representatives wereconsulted extensively during the design of the researchand were invited to respond to a request for proposalsto implement the major services that comprised the in-terventions tested in the study: HF-ACT, HF-ICM, scat-tered site housing portfolio management, and congregatehousing with support.RecruitmentParticipants were recruited through service providersand agencies serving individuals who are homeless andmentally ill in Vancouver, including shelters, drop-incenters, street outreach workers, hospitals, police, andcourts. An effort was made to locate individualsthroughout Vancouver, while recognizing that the major-ity of visible homelessness and related services wereconcentrated in one area.Eligibility and level of needsEligible participants were Canadian citizens at least19 years of age who met criteria for homelessness orprecarious housing and current mental disorder status.Informed consent required that individuals were madeaware that randomization would involve assignment toeither a pre-specified intervention that included housingor to usual care consisting of existing services and sup-port. Participants and interviewers were therefore notblinded to the results of randomization.Operational definitionsHomelessness was defined as having no fixed place tosleep or live for more than 7 nights and little likelihood ofobtaining accommodation in the coming month. Precar-ious housing was defined as currently residing in marginalaccommodation, such as a SRO hotel, and having two orSomers et al. Trials 2013, 14:365 Page 3 of 20 episodes of homelessness (as defined above) duringthe past 12 months. These were minimal criteria, and par-ticipants with more long-standing homelessness were eli-gible for inclusion. Current mental illness was assessedusing the Mini-International Neuropsychiatric Interview(MINI) [24] for the following: major depressive episode,manic or hypomanic episode, post-traumatic stress dis-order, mood disorder with psychotic features, and psych-otic disorder. Where possible, mental disorder status wascorroborated by physician diagnosis. Participants werecategorized as moderate needs (MN) or high needs (HN).Inclusion in the HN study was based on MultnomahCommunity Ability Scale (MCAS) [25] score of 62 orlower and current bipolar or psychotic disorder, as well asone of the following: legal involvement in the past year,substance dependence in the past month, and two ormore hospitalizations for mental illness in any one of thepast 5 years. All other eligible participants were includedin the MN study.Retention strategiesA team of full-time and part-time field interviewers wasrecruited to follow participants at 3-month intervals. In-terviewers received in-depth training and supervision inthe administration of measures, and scales and itemswere pre-tested with a sample of participants. Interviewswere considered ‘on time’ if they occurred within 2 weeksof the designated anniversary date. Participants werepaid C$35 for the baseline interview and approximatelyC$30 for each subsequent interview. Scales were admin-istered verbally and responses entered immediately onlaptop computers. Major interviews conducted at 6-month intervals required between 90 to 180 minutes tocomplete in most cases. A field research office was opendaily throughout the study period, and participants wereencouraged to drop-in regardless of their interviewschedule. Interviewers obtained periodic updates regard-ing participants’ routines and typical whereabouts, andcollateral contact information was sought in order to aidwith relocation. Interviews were conducted in various lo-cations based on randomization arm and participantpreference, including participants’ homes, field researchoffice, and public settings, such as restaurants, parks,and drop-in centers.RandomizationAfter establishing eligibility for either the MN or HNstudy, a computerized adaptive randomization procedurewas followed to assign participants to study arms. Inter-viewers used laptop computers with secure live connec-tions to upload data and receive randomization resultsprior to notifying participants of the outcome. Samplesizes of 100 participants in each study arm were derivedbased on effect size estimates of 0.5 for the majoroutcome variables, power of 0.80 (β = 0.20), an attritionrate of 40%, and significance levels of 0.05 (two-tailed).Additional details of sample size estimates are reportedseparately [1]. Based on their first 30 clients, the HF-ACT team determined that they would be able to sup-port no more than 90 clients, and the upper limit of thisarm was revised accordingly.MeasuresVAH researchers collaborated with investigators in fourother study centers and the study funder to develop acommon battery of repeated measures. In addition, anumber of site-specific measures were implemented atdifferent intervals during the study (Table 1). Bothshared and site-specific measures were selected based onthe review of existing literature and toward addressingmajor gaps in knowledge. Measures were administeredat single time points if their results were historical orhighly stable (for example, adverse childhood events).Repeated measures (for example, quality of life, commu-nity integration) were hypothesized to be subject to vari-ation over time based on the randomization arm, withsuperior outcomes in experimental conditions comparedto treatment as usual (TAU). We similarly hypothesizethat the models of service introduced through the studywould cause superior outcomes when compared to TAUon measures of hospitalization, emergency departmentvisits, and justice system involvement.The domains addressed by the cross-site scales are:housing and vocational status, psychiatric and physicalhealth, level of independent community integration andfunctioning, quality of life, and use of community ser-vices. The following questionnaires were administered atbaseline and at 6-month intervals: health service accessitems (ACC) [26,27]; community integration scale (CIS)[33]; Colorado Symptom Index (modified) (CSI) [36];EuroQol 5D (EQ-5D) [45]; Global Appraisal of Individ-ual Needs, Substance Problem Scale (GAIN-SPS) [50];SF-12 Health Survey (SF-12) [72]; Quality of Life Index,20-item (QoLI-20) [65]; social support items and foodsecurity (FS); health, social, and justice service use inven-tory (HSJSU); and MCAS [25]. Two instruments were ad-ministered at 3-month intervals, Residential Time-LineFollow-Back (RTLFB) [71] and Vocational Time-LineFollow-Back (VTLFB) [75], to produce a continuous time-line of housing status and vocational status, respectively.The Recovery Assessment Scale, 22-item (RAS-22) [68]was administered at baseline and 24 months, and the Ad-verse Childhood Experiences (ACE) questionnaire [28]was administered once at 18 months after baseline. Anumber of other measures were implemented at a singletime-point: comorbid conditions list (CMC) [41], landlordrelations (LR), Observer-rated Housing Quality Scale(OHQS), mobility history (MH), and core serviceSomers et al. Trials 2013, 14:365 Page 4 of 20 1 Vancouver At Home (VAH) questionnaire detailsAcronym Full name Timeline Key domain/topics ReferencesACC Health service access items BL, 6, 12, 18, 24 Use of community services: regular family physician or healthclinic use, and perceived unmet healthcare needs.[26,27]ACE Adverse Childhood Experiences 18 Psychiatric and physical health, traumatic early life events. [28]CI Cognitive impairment 6, 24 Level of independent community functioning, psychiatricand physical health: Hopkins Verbal Learning Test,Trail Making Test, and Digit Symbol Test.[29-32]CIS Community integration scale BL, 6, 12, 18, 24 Level of independent community functioning, qualityof life: community participation and sense of belonging.[33-35]CSI Colorado Symptom Index(modified)BL, 6, 12, 18, 24 Psychiatric and physical health: frequency of past month’spsychiatric symptoms.[36-39]CTS1 Conflict tactics scale 24 Level of independent community functioning: frequencyand severity of interpersonal conflict.[40]CMC Comorbid conditions list BL Psychiatric and physical health: presence of chronicand infectious diseases.[41]C-SSS Core service satisfaction scale 24 Quality of life, use of community services: participantsatisfaction with services provided by Vancouver AtHome (VAH) intervention teams.[42,43]DSHH Demographics, housing, vocational,and service use historyBL Housing status, use of community services: sociodemographicdetails, lifetime duration of homelessness, long-term health,social and justice service use, and vocational history.[44]EQ-5D EuroQol 5D BL, 6, 12, 18, 24 Psychiatric and physical health, health-related quality of life. [45-47]FS Social support items andfood securityBL, 6, 12, 18, 24 Use of community services, psychiatric and physical health:type, quality, availability and source of food, and recenthistory of food insecurity.[48,49]GAIN-SPS Global Appraisal of IndividualNeeds, Substance Problem ScaleBL, 6, 12, 18, 24 Substance-related problems. [50,51]HSJSU Health, social, and justiceservice use inventoryBL, 6, 12, 18, 24 Use of community services, psychiatric and physical health:nature and frequency of health, social, and justicesystem services.[42,52-57]III Interviewer impression items BL, 3, 6, 9, 12,15, 18, 21, 24Level of independent community functioning: interviewerassessment of validity and reliability of self-report data.LR Landlord relations 18 Housing status: specific to landlord relationship.PHQL Perceived housing quality 6, 12, 18, 24 Housing status: subjective housing quality assessedby participants.[58,59]MCAS Multnomah CommunityAbility ScaleBL, 6, 12, 18, 24 Level of independent community functioning: interviewerassessed level of functioning across range of domains.[25,60]MINI Mini-International NeuropsychiatricInterviewBL Psychiatric and physical health: current major Axis Idisorders and suicidality.[24,61-63]MH Mobility history 21 Housing status: geographic mobility.MoCA1 Montreal Cognitive Assessment 21 Level of independent community functioning: assessmentof cognitive domains indicated for the screening ofneurological deficits.[64]OHQS Observer-rated Housing Quality Scale 24 Housing status: objective ratings of physical characteristicsof participant dwellings.PAIN1 Chronic pain screener 21 Quality of life, psychiatric and physical health.QoLI-20 Quality of Life Index, 20-item BL, 6, 12, 18, 24 Quality of life, psychiatric and physical health: subjectivequality of life across range of domains.[65-67]RAS-22 Recovery Assessment Scale,22-itemBL, 24 Quality of life, psychiatric and physical health. [68-70]RTLFB Residential Time-Line Follow-Back 3, 6, 9, 12, 15,18, 21, 24Housing status: detailed chronology of housing status,including frequency of moves, type of accommodation,and household composition.[39,71]Somers et al. Trials 2013, 14:365 Page 5 of 20 scale (C-SSS) [42,43]; or at other intervals:cognitive impairment (CI) (Hopkins Verbal Learning Test,Trail Making Test, Digit Symbol Test) [29-32], perceivedhousing quality (PHQL) [58,59], and Working Alliance In-ventory (WAI) [76-80]. Semi-structured narrative inter-views were scheduled at baseline and 18-month timepoints with approximately fifty participants (10 from eachstudy arm in the two VAH trials).Site-specific measures were selected based on studyhypotheses and the anticipated characteristics of theVancouver homeless population. Major areas of hypoth-esis testing were: addictions, cognitive impairment, andpsychiatric severity would negatively influence housingstability; HF would result in superior outcomes whencompared to TAU, including reduced use of crisis ser-vices and justice system encounters, superior housingstability, and quality of life; and HF would produce su-perior health outcomes compared to TAU.The Maudsley Addiction Profile (MAP) [74] is a multi-dimensional instrument assessing alcohol and drug useand related harms, administered at 6-month intervals.The Montreal Cognitive Assessment (MoCA) [64] as-sesses several cognitive domains and is indicated for thescreening of neurological deficits in younger populations(for example, traumatic brain injury, brain tumors, vas-cular cognitive impairment). The foster care history(VFC) was administered once at 12 months after base-line. The MoCA, conflict tactics scale [40], and painscales (assessing acute and chronic pain; Schutz, unpub-lished) were administered at 21 months only.Ten participants in each study arm (n = 50) were in-vited to participate in open-ended, qualitative interviewsplanned for baseline and 18 months after recruitment.Participants were selected purposively in order to repre-sent differences across gender, ethnicity, duration ofhomelessness, and degree of functional impairment.Interview questions were organized around the followingthemes: pathways into and out of homelessness; high,low, and turning points in life; and challenges and enab-ling factors related to recovery.In addition, fifty participants were asked to provideconsent to undergo physical health examinations involv-ing basic physician assessment and blood analysis (forexample, hepatitis B/C, HIV/AIDS). These assessmentswere included to examine the possibility of undetectedillness among members of the study cohorts. Finally, allparticipants were asked to provide consent for the re-searchers to send their identifying details to public agen-cies in order to then receive administrative data regardingtheir use of health, justice, and social welfare services (sep-arate consent was sought for each category of agency). Aninter-agency data sharing protocol was created by a priorproject and was used as the basis for the current data ex-tract. The fields of data specified for inclusion were: phys-ician services; hospital services; pharmaceutical services;community mental health and substance use services; vitalstatistics; justice events, including convictions and sen-tences; and financial assistance.InterventionsParticipants in both the MN and HN studies were ran-domized to either an intervention based on the princi-ples of HF or to TAU. In both studies, TAU participantsdid not receive any housing or support services throughTable 1 Vancouver At Home (VAH) questionnaire details (Continued)SCNR Eligibility screening instrument BL Housing status, psychiatric and physical health: determinesparticipation eligibility based on legal adult status (>19 yearsin British Columbia (BC)), absolute homelessness or precarioushousing status, and current mental illness.[24,61-63]SF-12 SF-12 Health Survey BL, 6, 12, 18, 24 Psychiatric and physical health, level of independentcommunity functioning: assessment of the extent ofimpairment caused by both physical and mental illness.[45,72,73]VFC1 Foster care history 12 Quality of life, psychiatric and physical health: details earlylife involvement in the child welfare system.VMAP1 Maudsley Addiction Profile BL, 6, 12, 18, 24 Psychiatric and physical health: past month’s substance use,including drug type, mode of administration, frequencyof use, and drug-related harms.[74]VTLFB Vocational Time-LineFollow-Back3, 6, 9, 12, 15,18, 21, 24Level of independent community functioning: detailedchronological recent history of paid work and educationalor skills training. Quantifies income and income sources.[75]WAI-PAR Working Alliance Inventory,participant6, 12, 18, 24 Use of community services and quality of life: participant(WAI-PAR) and service provider (WAI-PRO) assessment ofworking relationship with key service provider. Assessmentof perception of client-provider relationship, support,confidence, and trust.[76-80]WAI-PROWorking Alliance Inventory,provider121Vancouver site-specific scales. Timeline: BL, baseline; 3, 3-month visit; 6, 6-month visit; 9, 9-month visit; 12, 12-month visit; 15, 15-month visit; 18, 18-month visit;21, 21-month visit; 24, 24-month visit.Somers et al. Trials 2013, 14:365 Page 6 of 20 study, but were able to access existing services andsupport for individuals who are homeless and mentallyill in Vancouver. The resources comprising TAU includeshelters, SRO hotels, and community services describedearlier. No research has previously examined the respon-siveness and effectiveness of these services by followinga cohort of homeless individuals prospectively. Coincidentwith this study, the City of Vancouver and the provincialhousing authority were in the process of expanding ser-vices for the homeless [81]. The quality and type of hous-ing received by participants in TAU will be carefullydocumented alongside the receipt of services and supportattendant to housing. In addition to TAU, HN participantswere randomized to either HF-ACT or Congregate Hous-ing with Support (CONG). MN participants receivedeither TAU or HF-ICM.Service providers for each intervention were selectedthrough a competitive request for proposals. Applica-tions were reviewed by a panel of senior individualsdrawn from homelessness research, management of ser-vices, and community granting agencies. The criteria forassessment included the delineation of organizationalexperience, plans for implementation, and budget. Eachselected service provider received specific training in thedelivery of HF, and underwent fidelity assessments byexternal review teams at two points during the study(see below). Services were based on the model definedby Pathways to Housing [82], including expertise thatanticipated the needs of local clients (for example, addic-tion severity), and configured to support participants inboth scattered and congregate housing configurations.Participants randomized to HF were transitioned to acase manager within 2 days of study recruitment.An inventory of apartments was developed in a varietyof neighborhoods throughout the city. These apartmentswere drawn from private market rentals with numerouslandlords. In order to promote community integration, amaximum of 20% of the units in any building could beallocated to program participants. Consistent with theprinciples of HF, participants were provided with achoice of housing units [82]. A housing portfolio managerwas responsible for building and maintaining relationshipswith landlords, including relocating participants to moresuitable locations when needed. Participants in the scat-tered site conditions (HF-ACT and HF-ICM) receivedsupport in their homes and were expected to meet withprogram staff on a weekly basis. The CONG conditionwas mounted in a single vacant hotel with the capacity tohouse approximately 100 occupants in independent suitesbut without full kitchens. The building was located in amixed residential and commercial neighborhood, adjacentto numerous amenities. The building was equipped with anumber of facilities to support residents and to promotethe development of a positive community culture,including: central kitchen and meal area, medical examin-ation room and formulary, and recreational areas (yoga,basketball, road hockey, lounge). Tenants were providedwith opportunities to engage in part-time work both withinthe building (for example, meal preparation, laundry) andin the community (for example, providing a graffiti removalservice). A reception area and front desk were staffed24 hours a day. Tenancy in any of the experimental housingconditions was not contingent on compliance with specifictherapeutic objectives (for example, addiction treatment).Program staff in each intervention condition participated ina series of training events in person in order to enhanceconsistency in practices. Subsidies were provided throughthe study to ensure that participants paid no more than30% of the total income on rent. Fidelity assessments wereconducted by an external team, with representatives fromPathways to Housing, the study funder, and individualswho had experienced homelessness. Assessments wereconducted at two time points (12 and 24 months afterimplementation) using a HF fidelity scale [82], and involvedmeetings with staff as well as participants in each of the HFinterventions. The assessment team provided verbal andwritten feedback to the staff at each intervention.VAH outcomesThe primary outcome domains for both trials are: hous-ing stability, health status, quality of life, and service use.Secondary outcome domains are: cost avoidance andcost effectiveness. Primary outcomes will be comparedbetween HF and TAU, including examination of similar-ities and differences between congregate and scatteredsite configurations of HF in the HN sample. Previous re-search has reported greater reductions in homelessnesswith group housing than with placement in independentapartments [22]. Particular attention will be paid to therole of substance use in relation to primary outcomes.Service use outcomes and economic analyses will beconducted using administrative data sources as specified.Data collection and analysis planRepeated measures (3- and 6-month scales) were collectedover 24 months. Based on previous studies and the resultsreviewed above, approximately 2 years of follow-up wasregarded as sufficient to detect changes in the major out-come domains. Fifteen months after commencing recruit-ment, the cross-site protocol was shortened from 24 to21 months. Following this change, VAH revised the com-position of the 21-month interview in order to providemore complete data for cross-site end point analyses;however, VAH continued to collect data through to24 months and preserved the original protocol as specifiedin the trials’ registration. Differences between sites in theprotocol change were primarily due to differences in fi-nancial resources between study centers. Preserving theSomers et al. Trials 2013, 14:365 Page 7 of 20 slightly longer follow-up in Vancouver was deemedimportant owing to the high levels of comorbidity, sub-stance dependence, and chronicity of homelessness withinthe sample.Descriptive statistics (such as mean, median, standarddeviation, and proportions) were calculated for all quan-titative instruments administered at baseline. Compari-sons of variables between groups were conducted usingparametric tests (t-tests or one-way analysis of variance(ANOVA) for continuous variables) and non-parametrictests (Mann–Whitney or Kruskal–Wallis test for con-tinuous variables; Pearson’s chi-squared or Fisher’s exacttests for categorical variables), as appropriate. All re-ported P values were two-sided.Longitudinal analyses of VAH data were planned for12 and 24 months using several analytic methods,such as hierarchical linear modeling (HLM), general-ized estimating equations (GEE), and time-to-eventanalysis (such as Cox regression and negative binomialregression), as appropriate. All longitudinal analysesare based on intention-to-treat. The major domains oflongitudinal analysis examine the overall robustness ofinterventions to promote health and recovery amonggroups of participants, and investigate individual char-acteristics that may predict different responses to in-terventions. Planned analyses also include examinationof service use and cost outcomes using administrativedata in combination with interview results. Sensitivityanalysis will be conducted to evaluate the effect ofmissing data using several methods, including meansubstitution, multiple imputations, and last observationcarried forward.The results of narrative interviews will be analyzed ac-cording to the organizing themes of pathways into andout of homelessness, and high, low, and turning pointsin life. These thematic analyses are expected to comple-ment results from questionnaires. Post hoc analyses willbe informed by qualitative findings, and will also exam-ine the characteristics of individuals who appear to exhibitbetter (and worse) outcomes regardless of randomizationFigure 1 Participant flow through eligibility, screening, needs level assessment, and allocation to study arm. αIncludes approximately 100participants deemed ineligible via an informal telephone screen, and 94 participants who were ineligible after formal in-person screening. βIncludes 11participants who were unable to be located after assignment or left within 1 month of entering.Somers et al. Trials 2013, 14:365 Page 8 of 20 In a related vein, ideographic analyses will be per-formed to examine whether CONG, HF-ACT, and TAUmay be associated with better outcomes for differingsubpopulations.Administrative data will be analyzed to provide long-term (that is, up to 15 years prior to randomization) his-torical perspectives on trajectories of service use prior toand following homelessness. Administrative data willalso be used as key outcome measures (for example,changes in hospitalization) and to validate specific itemsalso collected via self-report (for example, hospitalizedin past 6 months).Data derived through the shared cross-site protocolare owned by the study sponsor. Data that are specific toVAH (unique instruments, administrative data) will beretained at the host institution (Simon Fraser University,Burnaby, BC, Canada). The use and storage of provincialadministrative data are governed by Information SharingAgreements between the Government of British Columbiaand Simon Fraser University. The research protocolunderwent institutional ethics review and was approvedby Simon Fraser University.ResultsRecruitment was carried out between October 2009 andApril 2011. Follow-up interviews were completed on aschedule following each individual’s anniversary andwere completed by May 2013. Approximately 800 individ-uals were screened by telephone. Referral sources (n = 40)represented about thirteen different types of servicesavailable to homeless adults with mental illness. The ma-jority of potential participants were referred from home-less shelters, drop-in centers, homeless outreach teams,hospitals, community mental health teams, and criminaljustice programs. Approximately 100 individuals were ex-cluded via telephone due to clear ineligibility. A further200 were excluded through the baseline interview proced-ure due to ineligibility (n = 94), unable to contact for base-line interview (n = 100), declining to participate (n = 3), orincomplete interview (n = 3) (Figure 1). Of the total num-ber of participants randomized (n = 497), 200 met criteriafor MN and 297 met criteria for HN. Retention rates bystudy arm after 24 months are illustrated in Table 2. Dif-ferent rates are indicated in relation to scales administeredevery 3 months and every 6 months.The primary reasons for loss to follow-up over24 months were death (n = 5 for 3-month scales and n = 9for 6-month scales) or inability to locate the participant(n = 9 for 3-month scales and n = 12 for 6-month scales).Some deaths occurred after participants had completed atleast one follow-up interview and these data are eligiblefor analysis. The overall retention rate through 24 monthswas 97% (Table 2). No significant differences betweenstudy arms were observed in terms of follow-up (6-monthscales) completion rates (Fisher’s exact P values for HNand MN samples were 0.074 and 0.082, respectively). Forthe 3-month scales, follow-up completion rates betweenHN study arms were significantly different (CONG, 100%;HF-ACT, 100%; TAU, 94%; Fisher’s exact P value = 0.002),but no significant difference was observed for the MNTable 2 Follow-up status for ‘At Home’ participants after 24 months by need level6 months questionnaire 3 months questionnaireStudy arm No follow-up1 At least onefollow up2No follow-up3 At least onefollow up4Reason Total(n = 23)Total(n = 474)Reason Total(n = 16)Total(n = 481)No contact Death5 WithdrewconsentNo contact Death6 WithdrewconsentHigh Needs 5 5 - 10 287 3 3 - 6 291CONG (n = 107) 1 1 - 2 105 - - - - 107ACT (n = 90) - 1 - 1 89 - - - - 90HNTAU (n = 100) 4 3 - 7 93 3 3 - 6 94Moderate Needs 7 4 2 13 187 6 2 2 10 190ICM (n = 100) 2 1 - 3 97 2 0 - 2 98MNTAU (n = 100) 5 3 2 10 90 4 2 2 8 921 - No follow-up data collected at any of 6, 12, 18 & 24 months.2 - No significant differences between study arms was observed in terms of follow up completion rate (Fishers’ exact p values for HN & MN sample were 0.074and 0.082 respectively).3 - No follow-up data collected at any of 3, 6, 9, 12, 15, 18, 21 & 24 months visit.4 - In terms of follow up completion rate, significant differences between study arms was observed for the HN sample, but no significant difference was observedfor the MN sample (Fishers’ exact p values for HN & MN sample were 0.002 and 0.101 respectively).5 - Total number of deaths (n = 29) was as follows: CONG-4; ACT-7; HNTAU-5; ICM-6; & MNTAU-7. However, 20 participants completed at least one follow up visitbefore death and the remaining 9 participants whose follow up data was not available died with seven months of randomization.6 - Out of 29 deaths, 5 participants whose follow up data was not available died within five months of randomization and the rest 24 participants completed atleast one follow up visit before death.Somers et al. Trials 2013, 14:365 Page 9 of 20 3 Socio-demographic and mental disorder related characteristics for the ‘At Home’ participants by need statusVariables Overall N (%) HN N (%) MN N (%) P valueSocio-demographicsAge at enrolment visit (years)Mean (SD) 40.8 (11.0) 39.7 (11.2) 42.6 (10.5) 0.004Median (IQR) 41 (32–48) 39 (31–47) 44 (36–49) 0.002Male Gender 359 (73) 218 (74) 141 (71) 0.420Place of birth (Canada) 431 (87) 256 (87) 175 (88) 0.743EthnicityAboriginals 77 (16) 44 (15) 33 (16) 0.844White 280 (56) 170 (57) 110 (55)Other 140 (28) 83 (28) 57 (29)Incomplete High School 280 (57) 179 (61) 101 (51) 0.022Single (never married) 343 (70) 214 (73) 129 (65) 0.043Have children (under18) 122 (25) 69 (24) 53 (27) 0.483Native Language (English) 392 (80) 236 (80) 156 (78) 0.696HomelessnessPrecariously housed 109 (22) 65 (22) 44 (22) 0.976Lifetime duration of homelessness (months)Mean (SD) 60.2 (70.3) 62.0 (67.0) 57.5 (74.9) 0.489Median (IQR) 36 (12–84) 42 (12–84) 36 (12–84) 0.179Longest duration of homelessness (months)Mean (SD) 30.9 (40.1) 32.2 (40.8) 28.9 (39.1) 0.358Median (IQR) 12 (6–36) 18 (6–45) 12 (6–36) 0.236Age of first homelessness (years)Mean (SD) 30.3 (13.3) 28.7 (12.5) 32.6 (14.1) 0.002Median (IQR) 28 (19–41) 26 (19–36) 34 (20–44) 0.003EmploymentCurrently employed 18 (4) 10 (3) 8 (4) 0.722Worked continuously (>1 year) in past 323 (65) 185 (63) 138 (69) 0.164History of any wartime services 27 (5) 17 (6) 10 (5) 0.697Willingness to have paid job 384 (87) 217 (84) 167 (90) 0.102Hospitalized for mental illness (last 5 years)*Over 6 months 57 (12) 47 (16) 10 (5) <0.001More than two times 253 (53) 197 (69) 56 (29) <0.001MINI InternationalNeuropsychiatric Interview diagnosisPsychotic Disorder/Schizophrenia* 263 (53) 211 (71) 52 (26) <0.001Major Depressive Episode 199 (40) 95 (32) 104 (52) <0.001Post Traumatic Stress Disorder (PTSD) 129 (26) 63 (21) 66 (33) 0.003Manic or Hypomanic Episode* 97 (19) 68 (23) 29 (14) 0.021Panic Disorder 104 (21) 59 (20) 45 (22) 0.479Mood disorder with psychotic feature 84 (17) 56 (19) 28 (14) 0.152Substance Dependence 288 (58) 183 (62) 105 (52) 0.043Alcohol Dependence 121 (24) 72 (24) 49 (24) 0.948Somers et al. Trials 2013, 14:365 Page 10 of 20 (HF-ICM, 98%; TAU, 92%; Fisher’s exact P value =0.101). Of the 497 participants randomized, 438 (88%)gave consent to access administrative data from publicly-funded agencies.Adverse events of all kinds were reported to a moni-toring committee as well as to the Research Ethics Boardat Simon Fraser University. Apart from mortality, ad-verse events typically involved episodes of interpersonalconflict, such as abusive language or offensive behaviorinvolving participants.Sociodemographic characteristics of participants aredetailed in Table 3. Most were male (73%), white (56%),never married (70%), had a current medical illness (91%),were substance dependent (58%), and met criteria for‘absolute homelessness’ (78%).A number of significant differences between the MNand HN samples were observed (Table 3 and 4). Severaldifferences were expected based on inclusion criteria foreach study and are reflected in the results. HN partici-pants were more likely to have a psychotic disorder, havebeen hospitalized for psychiatric reasons, meet criteria forsubstance dependence, and have justice system involve-ment. HN participants also had lower MCAS scores.Beyond differences that were directly related to inclu-sion criteria (indicated with asterisk in Table 3 andTable 4), a number of additional significant differencesbetween MN and HN were observed. MN participantswere older at recruitment and when first homeless, weremore likely to have been married, and more likely tohave completed high school than those in the HN sam-ple. Participants in the MN sample were more likely toreport multiple physical illnesses, asthma, and HIV/AIDS than those in the HN sample.Results of standardized questionnaires indicate broadsimilarities between the MN and HN samples (Table 4).No significant differences between groups were observedon measures of: community integration (CIS total score),health-related quality of life (EQ-5D), food security (FS),overall health (SF-12, physical or mental health scores),overall quality of life (QoLI-20), and personal recovery(RAS-22). The MN sample reported significantly greaterphysical integration in the community (CIS physical) anda significantly lower level of externalizing or substance-related needs (GAIN-SPS) than the HN sample.A series of comparisons tested for potential non-equivalence between randomization arms at baseline.Sociodemographic and diagnostic characteristics for thethree HN and two MN study arms are shown in Table 5.Within the HN sample, there were no significant baselinedifferences of sociodemographics and mental disordersbetween groups. In the MN sample, those randomized toTAU had longer durations of homelessness (P = 0.037)and were more likely to be absolutely homeless (P = 0.041)at the time of recruitment. Further comparisons of ques-tionnaire results indicate no meaningful differences be-tween randomization arms in either the MN or HN study,except for several comorbid medical conditions (HIV,hepatitis B, cancer). HN participants randomized toCONG had a significantly higher prevalence of HIV andhepatitis B, but when all blood-borne diseases (HIV, hepa-titis B and C) were combined, no significant differenceswere observed between groups (Table 6).DiscussionAs expected, inclusion criteria led to a number of signifi-cant differences between the MN and HN samples.Table 3 Socio-demographic and mental disorder related characteristics for the ‘At Home’ participants by need status(Continued)Suicidality (high or moderate) 168 (34) 93 (31) 75 (37) 0.153Two or more mental disorders 240 (52) 148(53) 92 (51) 0.402Three or more mental disorders 114 (25) 78 (28) 36 (20) 0.032Referral sourcesShelter or transitional housing 143 (29) 82 (28) 61 (31) <0.001Housing Lists 19 (4) 9 (3) 10 (5)Outreach 86 (17) 44 (15) 42 (21)Hospitals 47 (9) 35 (12) 12 (6)Aboriginal groups 15 (3) 6 (2) 9 (4)Criminal justice 70 (14) 59 (20) 11 (6)Drop-in-centers 65 (13) 33 (11) 32 (16)Mental health teams 19 (4) 13 (4) 6 (3)Other 16 (3) 6 (2) 10 (5)Not specified 17 (3) 10 (3) 7 (3)* Variables used to determine eligibility for the HN sample.Somers et al. Trials 2013, 14:365 Page 11 of 20 4 Questionnaire related characteristics for ‘At Home’ participants by need status at enrolment visitQuestionnaire Overall mean (SD) HN mean (SD) MN mean (SD) P valueCommunity Integration Scale (CIS)Physical subscale score 2.1 (1.7) 1.9 (1.7) 2.4 (1.8) <0.001Psychological subscale score 10.9 (3.5) 11.0 (3.5) 10.7 (3.6) 0.368Colorado Symptom Index (CSI)Total score 37.2 (12.5) 38.0 (13.1) 36.0 (11.7) 0.098Comorbid Conditions List (CMC)1 N (%) N (%) N (%)Asthma 103 (21) 50 (17) 53 (26) 0.009Hepatitis C 139 (30) 78 (28) 61 (31) 0.302HIV/AIDS 43 (9) 18 (6) 25 (12) 0.012Hepatitis B 25 (5) 13 (5) 12 (6) 0.412Blood-borne infectious diseases2 157 (32) 87 (30) 70 (35) 0.224Epilepsy or seizure 67 (13) 49 (16) 18 (9) 0.016Stroke 27 (5) 19 (6) 8 (4) 0.248Cancer 18 (4) 14 (5) 4 (2) 0.117Head Injury 324 (65) 191 (64) 133 (67) 0.563Presence of any physical illness 453 (91) 268 (90) 185 (93) 0.384Multiple (≥ 2) physical illness 402 (81) 231 (78) 171 (86) 0.032Multiple (≥ 3) physical illness 344 (69) 189 (64) 155 (78) 0.001EuroQuol 5D (EQ5D)Overall health 61.0 (22.5) 61.8 (23.1) 60.0 (21.5) 0.382Food Security (FS)Total score 4.6 (2.6) 4.5 (2.5) 4.8 (2.7) 0.214Global Assessment of Individual need –Substance Problem Scale (GAIN-SPS)Total score (last month) 2.1 (2.0) 2.3 (2.0) 1.8 (2.0) 0.007Age of first alcohol use 14.1 (6.3) 14.2 (5.0) 14.1 (4.9) 0.751Age of first drug use 15.7 (5.0) 15.5 (5.6) 16.0 (7.2) 0.438Health Service Access Items (ACC)Have a regular medical doctor 320 (65) 177 (60) 143 (72) 0.008Place to go when you are sick 395 (81) 231 (79) 164 (83) 0.342Needed health care, but didn’t receive it 209 (43) 129 (45) 80 (40) 0.269Health, Social Justice Service Use Inventory (HSJSU)Seen by a health/social service provider 389 (79) 216 (74) 173 (89) <0.001Visited psychiatrist 134 (27) 89 (30) 45 (22) 0.066Talked with a health/social service provider 112 (29) 58 (20) 54 (27) 0.065Emergency room visit 281 (58) 163 (56) 118 (60) 0.483Ambulance 195 (40) 118 (40) 77 (39) 0.748Contacts with police (no arrest) 254 (52) 154 (53) 100 (51) 0.573Held in a police cell (≤24 hours) 112 (23) 80 (28) 32 (16) 0.002Arrested 173 (36) 128 (44) 45 (23) <0.001Court appearance 174 (36) 123 (43) 51 (26) <0.001Interviewer Impression Items (III) N (%) N (%) N (%)Signs of difficulty in reading response card (a lot) 20 (4) 17 (6) 3 (1) 0.019Signs of drug or alcohol intoxication (a lot) 10 (2) 7 (2) 3 (1) 0.505Somers et al. Trials 2013, 14:365 Page 12 of 20 to those in MN, members of the HN cohortwere significantly more likely to meet criteria for psych-osis or mania/hypomania, have multiple recent psychi-atric hospitalizations, and be severely compromised intheir community functioning. Other significant differ-ences between MN and HN were not directly related tothe inclusion/exclusion criteria. Members of the HN co-hort had lower educational achievement and were morelikely to have multiple mental disorders than thoseassigned to the MN study. In numerous other respects,participants assigned to MN and HN did not differ sig-nificantly. Overall, participants were white, male, ‘abso-lutely homeless’, physically ill, and met criteria forsubstance dependence and alcohol dependence.Participants assigned to MN were significantly morelikely to meet criteria for post-traumatic stress disorder,major depression, and report having HIV/AIDS thanthose assigned to HN. Thirty percent of the MN cohorthad been homeless for more than 60 months in theirlifetime and 26% met criteria for a psychotic disorder.These results suggest that the descriptor ‘moderate’ is amisnomer that understates the complexity of needswithin the MN cohort. As hypothesized, the profile ofboth study cohorts (MN and HN) included level ofneeds (for example, substance dependence, physicalillness) that have not been included in previous studies ofHF. VAH is therefore capable of generating new know-ledge regarding the effectiveness of HF-ICM and HF-ACTfor clients with a broad range of presenting characteristics.Randomization successfully minimized differences be-tween study arms. We tested for differences on sociode-mographic and mental health-related variables, as wellas all other measures (total score or subscale score or in-dividual item) administered at baseline. In the MNstudy, those assigned to the TAU arm were significantlymore likely to be ‘absolutely homeless’ and to have lowerlifetime duration of homelessness. No other significantdifferences were observed in either study, except forseveral comorbid medical conditions (HIV, hepatitis B,cancer) with very low prevalence. HN participants ran-domized to CONG had a significantly higher prevalenceof HIV and hepatitis B, but when all blood-borne dis-eases (HIV, hepatitis B and C) were combined, no sig-nificant differences were observed between groups.Despite differences between groups, the low prevalenceof these conditions in the sample is not expected to in-fluence results. However, tests will be designed to con-trol for relevant differences identified at baseline.Previous research has indicated that homeless mentallyill individuals may have a preference for independentTable 4 Questionnaire related characteristics for ‘At Home’ participants by need status at enrolment visit (Continued)Signs of psychiatric symptoms (a lot) 66 (13) 58 (19) 8 (4) <0.001Validity of information (no confidence) 14 (3) 13 (4) 1 (<1) 0.010Multnomah Community Ability Scale (MCAS)*Total score 56.1 (9.6) 50.7 (6.8) 64.1 (7.3) <0.001SF-12 Health Survey (SF-12)Physical health 45.9 (12.3) 46.5 (12.4) 45.1 (12.2) 0.233Mental health 35.4 (13.7) 35.8 (13.6) 34.8 (13.9) 0.445Quality of Life Index 20 Item (QOLI-20)Total score 73.6 (21.9) 74.4 (21.5) 72.5 (22.4) 0.337Recovery Assessment Scale 22 item (RAS-22)Total score 79.5 (12.0) 79.2 (11.5) 79.9 (12.8) 0.563Maudsley Addiction Profile (MAP) N (%) N (%) N (%)Use of alcohol 225 (46) 142 (48) 83 (42) 0.174Use of heroin 96 (19) 59 (20) 37 (19) 0.732Use of Cocaine 83 (17) 57 (19) 26 (13) 0.074Use of Cocaine-crack base 160 (32) 97 (33) 63 (32) 0.805Use of Amphetamine 61 (12) 44 (15) 17 (7) 0.035Use of Cannabis 205 (45) 133 (47) 72 (42) 0.256Injection drug use 88 (18) 54 (18) 34 (17) 0.727Daily drug use (excluding alcohol) 126 (25) 82 (28) 44 (22) 0.159Poly drug (≥ 3) use (excluding alcohol) 108 (22) 72 (24) 36 (18) 0.0651 - Response ‘Do not know’ was considered as no.2 - Included HIV, Hepatitis C & Hepatitis B.* Variables used to determine eligibility for the HN sample.Somers et al. Trials 2013, 14:365 Page 13 of 20 5 Comparisons of Socio-demographic and mental disorder related characteristics between study armsHigh need (n = 297) Moderate need (n = 200)Variable CONG(n = 107)N (%)ACT(n = 90)N (%)TAU(n = 100)N (%)P value1 ICM(n = 100)N (%)TAU(n = 100)N (%)P value2Socio-DemographicsAge at randomization (years)Mean (SD) 40.0 (11.6) 39.5 (10.8) 39.5 (11.2) 0.910 42.1 (10.4) 43.1(10.6) 0.475Median (IQR) 41 (30–48) 38 (31–47) 39 (32–48) 0.920 43 (34–50) 45 (36–49) 0.610Male gender 82 (77) 66 (74) 70 (71) 0.696 71 (71) 70 (71) 0.964EthnicityAboriginals 21 (20) 11 (12) 12 (12) 0.469 19 (19) 14 (14) 0.060Caucasian 60 (56) 53 (59) 57 (57) 60 (60) 50 (50)Mixed/Other 26 (24) 26 (29) 31 (31) 21 (21) 36 (36)Incomplete high school 70 (66) 47 (53) 62 (62) 0.192 56 (56) 45 (45) 0.120Single/Never married 76 (72) 63 (70) 75 (77) 0.591 66 (66) 63 (63) 0.658Birth country (Canada) 94 (88) 80 (89) 82 (83) 0.417 90 (90) 85 (85) 0.285Have children (under18) 24 (23) 21 (24) 24 (25) 0.938 28 (28) 25 (25) 0.692Native Language (English) 87 (81) 76 (84) 73 (73) 0.125 13 (13) 16 (16) 0.733HomelessnessAbsolutely homeless 88 (82) 72 (80) 72 (72) 0.179 72 (72) 84 (84) 0.041Lifetime duration of homelessness (months)Mean (SD) 52.2 (63.5) 61.5 (69.1) 67.6 (69.0) 0.541 68.5 (92.1) 46.4 (50.7) 0.037Median (IQR) 36 (12–72) 42 (12–84) 48 (13–109) 0.575 48 (14–93) 24 (12–72) 0.051Longest duration of homelessness (months)Mean (SD) 32.9 (39.3) 30.5 (42.7) 33.0 (41.0) 0.901 31.7 (43.7) 26.0 (33.9) 0.304Median (IQR) 20 (7–48) 12 (6–40) 12 (6–48) 0.623 48 (6–40) 12 (5–34) 0.321Age of first homelessness (years)Mean (SD) 29.9 (13.1) 28.0 (11.9) 28.0 (12.4) 0.474 30.8 (13.8) 34.3 (14.3) 0.083Median (1QR) 27 (20–39) 26 (19–35) 24 (18–36) 0.463 29 (18–42) 35 (21–45) 0.118EmploymentCurrently employed 5 (5) 1 (1) 4 (4) 0.368* 6 (6) 2 (2) 0.279*Worked continuously (>1 year) in past 66 (62) 58 (69) 61 (61) 0.773 66 (66) 72 (72) 0.987Wartime services in past 4 (4) 7 (8) 6 (6) 0.448 5 (5) 5 (5) 0.987Willingness to have paid job 79 (82) 68 (90) 70 (82) 0.353 80 (89) 87 (91) 0.696Hospitalized for mental illness (last 5 years)Over 6 months 14 (13) 18 (20) 15 (15) 0.406 4 (4) 6 (6) 0.506More than two times 73 (70) 57 (68) 67 (71) 0.927 25 (25) 31 (32) 0.299MINI International Neuropsychiatric Interview diagnosisMajor Depressive Episode 35 (33) 31 (34) 29 (29) 0.710 52 (52) 52 (52) 1.00Manic or Hypomanic Episode 25 (23) 23 (26) 20 (20) 0.654 11 (11) 18 (18) 0.160Post-Traumatic Stress Disorder 27 (25) 17 (19) 19 (19) 0.445 34 (34) 32 (32) 0.802Panic Disorder 20 (19) 15 (17) 24 (24) 0.418 19 (19) 26 (26) 0.236Mood Disorder with psychotic feature 20 (19) 17 (19) 19 (19) 0.997 16 (16) 12 (12) 0.415Psychotic Disorder 79 (74) 59 (66) 73 (73) 0.385 25 (25) 27 (27) 0.747Alcohol dependence 28 (26) 19 (21) 25 (25) 0.695 25 (25) 24 (24) 0.869Somers et al. Trials 2013, 14:365 Page 14 of 20 over group housing when offered a choice[83], and that these preferences may change based onexperiences after the initiation of housing [84]. In thepresent study, individuals randomized to HF-ACT had achoice of apartments, but those randomized to CONGwere limited to selecting from among the available unitsin one building. Narrative interviews will be examined inorder to assess whether participants experienced mean-ingful differences concerning their choice of housing ineither of these two settings, and whether the experienceof choice was related to outcomes of interest.A high percentage of participants in each study arm(92% to 100%) were successfully followed through24 months of interviews. Twenty-nine participants diedduring the 24 months following randomization. Unsur-prisingly, participants assigned to TAU were most likelyto be lost to follow-up during the 24 months post-randomization. However, the differences were not statisti-cally significant. Differences in follow-up for the 3-monthscales were significant in the HN study only. Despite sta-tistically significant differences, the high overall follow-uprate in each group (94% in one arm, 100% in two others)is expected to yield valid and generalizable results. Thishigh follow-up rate is attributable to diverse strategies,including extensive outreach, a welcoming field office,relationships with service providers in the field, and main-taining updated information regarding collateral contactsand daily routines. More generally, the recruitment andretention of a knowledgeable and committed team of in-terviewers is a critical factor.VAH shares several important methodological featureswith studies in four other Canadian sites, primarily:inclusion/exclusion criteria, randomization to HF orusual care, a common battery of cross-site measures,and semi-structured qualitative interviews with a subsetof participants [1]. At the same time, the current studyhas a number of important site-specific elements. VAHis the only RCT to compare different configurations ofHF (congregate and scattered sites) alongside TAU. Theresults of this comparison will offer guidance to many cit-ies, including Vancouver, that include congregate varia-tions of HF as part of their strategies to addresshomelessness [85]. In addition, this trial has recruitedsamples with a broader range and severity of symptomsthan those reflected in previous studies. For example, thehigh prevalence of substance use disorders in this study willhelp to fill a specific gap in knowledge regarding the robust-ness of HF for individuals with concurrent disorders [86].A number of unique measures were incorporated inthe protocol based on their expected relevance to thelocal population, including measures of addiction fre-quency and severity, neuropsychological functioning,and physical examinations. In addition, VAH incorpo-rates a large array of administrative data spanning di-verse publicly-funded services and interventions acrosstime. These data generate opportunities to study the tra-jectories of service involvement over 10 years prior torecruitment in VAH, and enable follow-up of partici-pants after the completion of the intervention. The in-clusion of longitudinal data from multiple relevantsectors (justice, health, financial assistance) provides aunique opportunity for cost-based analyses.This study combines models of housing with supportin each of the intervention conditions. This may make itTable 5 Comparisons of Socio-demographic and mental disorder related characteristics between study arms (Continued)Substance dependence 67 (63) 55 (61) 61 (61) 0.965 51 (51) 54 (54) 0.671Suicidality (moderate or high) 34 (32) 28 (31) 31 (31) 0.992 33 (33) 42 (42) 0.189Two or more mental disorders 53 (49) 41 (46) 54 (54) 0.507 44 (44) 48 (48) 0.547Three or more mental disorders 34 (32) 22 (24) 22 (22) 0.250 17 (17) 19 (19) 0.713Referral sourcesShelter or transitional housing 31 (29) 26 (29) 25 (25) 0.468 31 (31) 30 (30) 0.250Housing Lists 3 (3) 4 (4) 2 (2) 6 (6) 4 (4)Outreach 13 (12) 15 (17) 16 (16) 21 (21) 21 (21)Hospitals 11 (10) 11 (12) 13 (13) 4 (4) 8 (8)Aboriginal groups 2 (2) 2 (2) 2 (2) 6 (6) 3 (3)Criminal justice 27 (25) 14 (16) 18 (18) 5 (5) 6 (6)Drop-in-centers 15 (14) 9 (10) 9 (6) 20 (20) 12 (12)Mental health teams 4 (4) 3 (3) 6 (6) 0 (0) 6 (6)Other 0 (0) 1 (1) 5 (5) 4 (4) 6 (6)Not specified 1 (1) 5 (6) 4 (4) 3 (3) 4 (4)1 - Bold indicates p value ≤ 0.05 and Italic indicates p value between > 0.05 and ≤ 0.1.2 - Bold indicates p value ≤ 0.05 and Italic indicates p value between > 0.05 and ≤ 0.1.* - P value from Fisher’s Exact Test.Somers et al. Trials 2013, 14:365 Page 15 of 20 6 Comparisons of questionnaire related characteristics between study arms at enrolment visitHigh need (n = 297) Moderate need (n = 200)Questionnaire CONG(n = 107) N (%)or Mean (SD)ACT(n = 90) N (%)or Mean (SD)TAU(n = 100) N (%)or Mean (SD)P value1 ICM(n = 100) N (%)or Mean (SD)TAU(n = 100) N (%)or Mean (SD)P value2Community Integration Scale (CIS)Physical subscale score 2.1 (1.8) 1.6 (1.5) 1.8 (1.7) 0.148 2.5 (1.8) 2.4 (1.7) 0.884Psychological subscale score 10.6 (3.7) 11.3 (3.5) 11.1 (3.2) 0.384 10.4 (3.5) 11.0 (3.6) 0.241Colorado Symptom Index (CSI)Total score 37.1 (13.0) 36.4 (13.4) 40.2 (12.6) 0.093 36.1 (12.2) 36.0 (11.2) 0.954Comorbid Conditions List (CMC)3 N (%) N (%) N (%) N (%) N (%)Asthma 18 (17) 14 (16) 18 (18) 0.904 27 (27) 26 (26) 0.873Hepatitis C 26 (24) 23 (26) 29 (29) 0.732 32 (32) 29 (29) 0.645HIV/AIDS 12 (11) 2 (2) 4 (4) 0.025* 16 (16) 9 (9) 0.134Hepatitis B 9 (8) 1 (1) 3 (3) 0.039* 7 (7) 5 (5) 0.552Blood-borne infectious diseases4 33 (32) 23 (26) 31 (32) 0.572 37 (37) 33 (33) 0.553Epilepsy or seizure 20 (19) 10 (11) 19 (19) 0.256 5 (5) 13 (13) 0.081*Stroke 11 (10) 2 (2) 6 (6) 0.069 2 (2) 6 (6) 0.279*Cancer 4 (4) 1 (1) 9 (9) 0.036* 0 (0) 4 (4) 0.058*Head Injury 66 (62) 62 (69) 63 (63) 0.544 61 (61) 72 (73) 0.079Presence of any physical illness 98 (92) 81 (90) 89 (89) 0.818 90 (90) 95 (95) 0.179Multiple (≥ 2) physical illness 82 (77) 69 (77) 80 (80) 0.806 84 (84) 87 (87) 0.547Multiple (≥ 3) physical illness 69 (65) 52 (58) 68 (68) 0.334 78 (78) 77 (77) 0.866EuroQuol 5D (EQ5D)Overall health 59.5 (23.7) 64.2 (22.9) 62.0 (22.5) 0.361 58.4 (23.4) 61.5 (19.5) 0.325Food Security (FS)Total score 4.3 (2.6) 4.4 (2.6) 4.8 (2.4) 0.454 4.7 (2.8) 4.8 (2.7) 0.799Global Assessment of Individual need –Substance Problem Scale (GAIN-SPS)Total score (last month) 2.4 (2.0) 2.1 (1.9) 2.3 (2.0) 0.591 1.8 (2.0) 1.8 (2.0) 0.842Age of first alcohol use 14.3 (6.5) 14.2 (3.8) 14.2 (4.0) 0.979 13.8 (4.8) 14.3 (4.9) 0.523Age of first drug use 15.6 (6.5) 15.3 (5.0) 15.7 (5.0) 0.892 15.4 (6.3) 16.5 (7.9) 0.304Health Service Access Items (ACC)Have a regular medical doctor 64 (60) 53 (59) 60 (61) 0.971 71 (71) 72 (72) 0.876Place to go when you are sick 82 (80) 72 (81) 77 (79) 0.913 83 (84) 81 (82) 0.706Needed health care, but didn’t receive it 45 (43) 36 (41) 48 (51) 0.371 39 (39) 41 (41) 0.817Health, Social Justice Service UseInventory (HSJSU)N (%) N (%) N (%) N (%) N (%)Seen a health/social service provider 70 (66) 67 (75) 79 (80) 0.075 84 (84) 89 (90) 0.217Visited psychiatrist 28 (26) 27 (30) 34 (34) 0.470 20 (20) 25 (25) 0.397Talked a health/social service provider 12 (20) 18 (20) 19 (20) 0.992 31 (31) 23 (23) 0.218Emergency room visit 59 (57) 51 (57) 53 (54) 0.912 64 (64) 54 (55) 0.202Ambulance 47 (44) 33 (37) 38 (38) 0.535 35 (35) 42 (38) 0.282Contacts with police (no arrest) 53 (52) 42 (47) 59 (60) 0.226 45 (45) 55 (56) 0.118Held in a police cell (≤24 hours) 25 (24) 27 (31) 28 (30) 0.556 14 (14) 18 (19) 0.386Arrested 50 (48) 32 (36) 46 (48) 0.152 19 (19) 26 (27) 0.179Court appearance 51 (50) 34 (38) 38 (40) 0.189 22 (22) 29 (30) 0.222Somers et al. Trials 2013, 14:365 Page 16 of 20 to isolate the relative contribution of distinctcomponents (for example, housing alone) that best ac-count for any observed differences or improvements.Each of the data sources included in the VAH protocolis subject to sources of bias. Several questionnaires havenot been well validated with homeless mentally ill sam-ples, and a small number were adapted or developed forthe present study. Narrative interviews with selected par-ticipants may yield findings that are unrepresentative.Administrative data sources are subject to limitationsthat include the ability to match all subjects across alldatabases, and the accuracy and completeness of theresulting extracts. However, the combination of adminis-trative data, narrative interviews, and questionnaires en-ables the application of mixed-methods approaches thatenrich understanding beyond the scope of each individ-ual data type. Finally, while other sites abbreviated theirstudy durations to 21 months, VAH maintained theoriginal 24-month protocol, and therefore preserved agreater opportunity to detect changes that may require alonger period of observation.ConclusionsThe present results confirm that VAH has successfully im-plemented experimental protocols that promise to generatenew knowledge regarding interventions for individuals whoare both homeless and mentally ill. Participants were suc-cessfully recruited and retained through the follow-upperiod, and randomization effectively minimized differencesbetween study arms in each trial. Diverse data sources andrelatively long follow-up provide opportunities for multi-method approaches to longitudinal data analysis. VAH addsto previous research on HF by including a sample withcomplex comorbidities and concurrent substance use disor-ders, and is the first experiment to include congregatehousing alongside scattered site HF.Table 6 Comparisons of questionnaire related characteristics between study arms at enrolment visit (Continued)Interviewer Impression Items (III) N (%) N (%) N (%) N (%) N (%)Signs of difficulty in reading card (a lot) 6 (6) 6 (7) 5 (5) 0.883 2 (2) 1 (1) 1.00*Signs of drug or alcohol intoxication(a lot)1 (1) 2 (2) 4 (4) 0.349* 1 (1) 2 (2) 1.00*Signs of psychiatric symptoms (a lot) 25 (23) 11 (12) 22 (22) 0.108 4 (4) 4 (4) 1.00*Validity of information (no confidence) 7 (6) 4 (4) 2 (2) 0.287* 1 (1) 0 (0) 1.00*Multnomah Community AbilityScale (MCAS)Total score 49.9.0 (6.7) 51.6 (6.5) 50.6 (7.0) 0.195 64.1 (7.6) 64.1 (7.1) 0.962SF-12 Health Survey (SF-12)Physical health 47.4 (13.1) 46.7 (12.3) 45.3 (11.6) 0.466 43.9 (12.1) 46.4 (12.9) 0.140Mental health 34.8 (15.1) 36.9 (13.0) 35.8 (12.6) 0.551 35.7 (13.0) 33.9 (14.8) 0.371Quality of Life Index 20 Item (QOLI-20)Total score 72.6 (21.7) 76.2 (21.3) 74.7 (21.4) 0.497 72.2 (21.6) 72.8 (23.3) 0.851Recovery Assessment Scale 22 item(RAS-22)Total score 78.2 (12.1) 80.7 (11.5) 79.1 (10.7) 0.308 80.3 (11.3) 79.5 (14.1) 0.685Maudsley Addiction Profile (MAP) N (%) N (%) N (%) N (%) N (%)Use of alcohol 50 (47) 44 (49) 48 (49) 0.936 37 (38) 46 (46) 0.240Use of heroin 24 (22) 13 (14) 22 (22) 0.295 18 (18) 19 (19) 0.909Use of Cocaine 24 (22) 14 (16) 19 (19) 0.476 12 (12) 14 (14) 0.715Use of Cocaine-crack base 36 (34) 26 (29) 35 (36) 0.596 36 (37) 27 (27) 0.141Use of Amphetamine 12 (11) 17 (18) 15 (15) 0.346 10 (10) 7 (7) 0.421Use of Cannabis 46 (45) 40 (48) 47 (49) 0.848 35 (42) 37 (42) 0.960Injection drug use 19 (18) 16 (18) 19 (20) 0.944 18 (18) 16 (16) 0.682Daily drug use (excluding alcohol) 31 (29) 19 (21) 32 (32) 0.227 27 (27) 17 (17) 0.088Poly drug (≥ 3) use (excluding alcohol) 30 (28) 17 (29) 25 (25) 0.318 20 (20) 16 (16) 0.4211 -Bold indicates p value ≤ 0.05 and Italic indicates p value between > 0.05 and ≤ 0.1.2 -Bold indicates p value ≤ 0.05 and Italic indicates p value between > 0.05 and ≤ 0.1.2 -Response ‘Do not know’ was considered as no.4 -Included HIV, Hepatitis C & Hepatitis B.* -P value from Fisher’s Exact Test.Somers et al. Trials 2013, 14:365 Page 17 of 20 Health service access items; ACE: adverse childhood experiences;ACT: Assertive community treatment; ANOVA: Analysis of variance;CI: Cognitive impairment; CIS: Community integration scale; CMC: Comorbidconditions list; CONG: Congregate housing with support; CONSORT: Consolidatedstandards of reporting trials; CSI: Colorado symptom index (modified);C-SSS: Core service satisfaction scale; CTS: Conflict tactics scale; EQ-5D: EuroQol 5D;FS: Social support items and food security; GAIN-SPS: Global appraisal of individualneeds, substance problem scale; GEE: Generalized estimating equation;HF: Housing first; HLM: Hierarchical linear modeling; HN: High needs;HSJSU: health, social, and justice service use inventory; ICM: Intensive casemanagement; LR: Landlord relations; MAP: Maudsley addiction profile;MCAS: Multnomah community ability scale; MH: Mobility history; MINI:Mini-international neuropsychiatric interview; MN: Moderate needs;MoCA: Montreal cognitive assessment; OHQS: Observer-rated housing qualityscale; PHQL: Perceived housing quality; QoLI-20: Quality of life index, 20-item;RAS-22: Recovery assessment scale, 22-item; RCT: Randomized controlled trial;RTLFB: Residential time-line follow-back; SF-12: SF-12 health survey; SRO: Singleroom occupancy; TAU: Treatment as usual; VAH: Vancouver at home; VFC: Fostercare history; VTLFB: Vocational time-line follow-back; WAI: Working allianceinventory.Competing interestsThe authors declare that they have no competing interests.Authors’ contributionsJMS is lead investigator of the study, and drafted and finalized the manuscript.MLP oversaw field research and wrote sections of the manuscript. AMconducted statistical analyses and prepared results for the manuscript. LCadministered interviews and prepared sections of the manuscript. SNRcontributed to data analysis and prepared sections of the manuscript.AP contributed to the design of the study and edited the manuscript. KFcontributed to the design and implementation of the study. All authors readand approved the final manuscript.AcknowledgmentsSpecial thanks to the Vancouver At Home participants, service providers, andfield research team members. The authors also thank the At Home/Chez Soiproject collaborative. This research was funded by a grant to Simon FraserUniversity from Health Canada and the Mental Health Commission ofCanada. The views expressed herein solely represent the authors.Author details1Somers Research Group, Faculty of Health Sciences Simon Fraser University,8888 University Drive, Burnaby V5A 1S6, Canada. 2Department of Medicine,University of British Columbia, 2775 Laurel Street, Vancouver V5Z 1M9,Canada.Received: 16 May 2013 Accepted: 16 October 2013Published: 1 November 2013References1. Goering PN, Streiner DL, Adair C, Aubry T, Barker J, Distasio J, Hwang SW,Komaroff J, Latimer E, Somers J, Zabkiewicz DM: The At Home/Chez Soitrial protocol: a pragmatic, multi-site, randomised controlled trial of aHousing First intervention for homeless individuals with mental illness infive Canadian cities. BMJ Open 2011, 1:e000323–e000323.2. Campbell L, Boyd N, Culburt L: A Thousand Dreams: Vancouver’s DowntownEastside and the Fight for its Future. Vancouver, BC: Greystone Books; 2009.3. Foley D: Hellish conditions at single-room occupancy hotels. Body Posit1998, 11:18–23.4. Shannon K, Ishida T, Lai C, Tyndall MW: The impact of unregulated singleroom occupancy hotels on the health status of illicit drug users inVancouver. Int J Drug Policy 2006, 17:107–114.5. Kerr T, Tyndall M, Li K, Montaner J, Wood E: Safer injection facility use andsyringe sharing in injection drug users. Lancet 2005, 366:316–318.6. Harvey E, Strathdee SA, Patrick DM, Ofner M, Archibald CP, Eades G,O’Shaughnessy MV: A qualitative investigation into an HIV outbreakamong injection drug users in Vancouver, British Columbia. AIDS Care1998, 10:313–321.7. Palepu A, Tyndall MW, Leon H, Muller J, O’Shaughnessy MV, Schechter MT,Anis AH: Hospital utilization and costs in a cohort of injection drug users.CMAJ 2001, 165:415–420.8. Strathdee SA, Patrick DM, Archibald CP, Ofner M, Cornelisse PGA, Rekart M,Schechter MT, O’Shaughnessy MV: Social determinants predict needle-sharing behaviour among injection drug users in Vancouver, Canada.Addict 1997, 92:1339–1348.9. Kushel MB, Perry S, Bangsberg D, Clark R, Moss AR: Emergency departmentuse among the homeless and marginally housed: results from acommunity-based study. Am J Public Health 2002, 92:778–784.10. Gilmer TP, Stefancic A, Ettner SL, Manning WG, Tsemberis S: Effect of full-service partnerships on homelessness, use and costs of mental healthservices, and quality of life among adults with serious mental illness.Arch Gen Psychiatry 2010, 67:645–652.11. Gulcur L, Stefancic A, Shinn M, Tsemberis S, Fischer SN: Housing,hospitalization, and cost outcomes for homeless individuals withpsychiatric disabilities participating in continuum of care and housingfirst programmes. J Comm App Soc Psychol 2003, 13:171–186.12. Yanos PT, Felton BJ, Tsemberis S, Frye VA: Exploring the role of housingtype, neighborhood characteristics, and lifestyle factors in thecommunity integration of formerly homeless persons diagnosed withmental illness. J Ment Health 2007, 16:703–717.13. Siegel C, Samuels J, Tang D-I, Berg I, Jones K, Hopper K: Tenant outcomesin supported housing and community residences in New York City.Psychiatr Serv 2006, 57:982–991.14. Tsemberis S, Eisenberg RF: Pathways to housing: supported housing forstreet-dwelling homeless individuals with psychiatric disabilities.Psychiatr Serv 2000, 51:487–493.15. Stein LI, Test MA: Alternative to mental hospital treatment. I. Conceptualmodel, treatment program, and clinical evaluation. Arch Gen Psychiatry1980, 37:392–397.16. Bond G, Drake R, Mueser K, Latimer E: Assertive community treatment forpeople with severe mental illness: critical ingredients and impact onpatients. Dis Manag Health Outcomes 2001, 9:141–159.17. Marshall M, Lockwood A: Assertive Community Treatment for People withSevere Mental Disorders. Chichester: John Wiley & Sons; 1998.18. Lipton FR, Siegel C, Hannigan A, Samuels J, Baker S: Tenure in supportivehousing for homeless persons with severe mental illness. Psychiatr Serv2000, 51:479–486.19. Fichter MM, Quadflieg N: Intervention effects of supplying homelessindividuals with permanent housing: a 3-year prospective study.Acta Psychiatr Scand 2006, 113:36–40.20. Wong Y-LI, Hadley TR, Culhane DP, Poulin SR, Davis MR, Cirksey BA, BrownJL: Predicting Staying in or Leaving Permanent Supportive Housing that ServesHomeless People with Serious Mental Illness. US Department of Housing andUrban Development: Washington, DC; 2006.21. Larimer ME, Malone DK, Garner MD, Atkins DC, Burlingham B, Lonczak HS,Tanzer K, Ginzler J, Clifasefi SL, Hobson WG, Marlatt GA: Health care andpublic service use and costs before and after provision of housing forchronically homeless persons with severe alcohol problems. JAMA 2009,301:1349–1357.22. Schutt RK: Homelessness, Housing, and Mental Illness. Cambridge, MA:Harvard University Press; 2011.23. Schutt RK, Hough RL, Goldfinger SM, Lehman AF, Shern DL, Valencia E,Wood PA: Lessening homelessness among persons with mental illness: Acomparison of five randomized treatment trials. Asian J, Psychiatry 2009,2:100–105.24. Sheehan DV, Lecrubier Y, Sheehan KH, Amorim P, Janavs J, Weiller E,Hergueta T, Baker R, Dunbar GC: The Mini-International NeuropsychiatricInterview (MINI): the development and validation of a structured diag-nostic psychiatric interview for DSM-IV and ICD-10. J Clin Psychiatry 1998,59:20–33.25. Barker S, Barron ND, McFarland BH, Bigelow DD, Carnahan TM: Acommunity ability scale for chronically mentally Ill consumers: Part II.Applications. Community Ment Health J 1994, 30:459–472.26. Khandor EE, Mason KK, Chambers CC, Rossiter KK, Cowan LL, Hwang SWS:Access to primary health care among homeless adults in Toronto, Canada:results from the Street Health survey. Open Med 2011, 5:e94–e103.27. Hwang SW, Ueng JJM, Chiu S, Kiss A, Tolomiczenko G, Cowan L, LevinsonW, Redelmeier DA: Universal health insurance and health care access forhomeless persons. Am J Public Health 2010, 100:1454–1461.Somers et al. Trials 2013, 14:365 Page 18 of 20 Felitti VJ, Anda RF, Nordenberg D, Williamson DF, Spitz AM, Edwards V, KossMP, Marks JS: Relationship of childhood abuse and householddysfunction to many of the leading causes of death in adults. Am J PrevMed 1998, 14:245–258.29. Benedict RHB, Schretlen D, Groninger L, Brandt J: Hopkins Verbal LearningTest – Revised: Normative data and analysis of inter-form and test-retestreliability. Clin Neuropsychol 1998, 12:43–55.30. Brandt J: The hopkins verbal learning test: Development of a newmemory test with six equivalent forms. Clin Neuropsychol 1991, 5:125–142.31. Reitan RM: Validity of the trail making test as an indicator of organicbrain damage. Percept Mot Skills 1958, 8:271–276.32. Wechsler D: Wechsler Adult Intelligence Scale. 4th edition. London: Pearson;2008.33. Aubry T, Myner J: Community integration and quality of life: acomparison of persons with psychiatric disabilities in housing programsand community residents who are neighbors. Can J Commun Ment Health1996, 15:5–20.34. Segal SP, Aviram U: The Mentally Ill in Community-Based Sheltered Care.Chichester: John Wiley & Sons; 1978.35. Chavis DM, Hogge JH, McMillan DW, Wandersman A: Sense of communitythrough Brunswick’s lens: A first look. J Community Psychol 1986, 14:24–40.36. Boothroyd RA, Chen HJ: The psychometric properties of the ColoradoSymptom Index. Adm Policy Ment Health 2008, 35:370–378.37. Shern DL, Wilson NZ, Coen AS, Patrick DC, Foster M, Bartsch DA, Demmler J:Client outcomes II: Longitudinal client data from the Colorado treatmentoutcome study. Milbank Q 1994, 72:123–148.38. Conrad KJ, Yagelka JR, Matters MD, Rich AR, Williams V, Buchanan M:Reliability and validity of a modified Colorado Symptom Index in anational homeless sample. Ment Health Serv Res 2001, 3:141–153.39. Greenwood RM, Schaefer-McDaniel NJ, Winkel G, Tsemberis SJ: Decreasingpsychiatric symptoms by increasing choice in services for adults withhistories of homelessness. Am J Community Psychol 2005, 36:223–238.40. Straus MA, Hamby SL, Boney-McCoy S, Sugarman DB: The Revised ConflictTactics Scales (CTS2): Development and preliminary psychometric data.J Fam Issues 1996, 17:283–316.41. Mares AS, Greenberg GA, Rosenheck RA: HUD/HHS/VA Collaborative Initiativeto Help End Chronic Homelessness: National Performance OutcomesAssessment: Is System Integration Associated with Client Outcomes?. WestHaven, CT: Northeast Program Evaluation Center; 2007.42. Adair CE, McDougall GM, Mitton CR, Joyce AS, Wild TC, Gordon A, CostiganN, Kowalsky L, Pasmeny G, Beckie A: Continuity of care and healthoutcomes among persons with severe mental illness. Psychiatr Serv 2005,56:1061–1069.43. Maruish ME: The Use of Psychological Testing for Treatment Planning and OutcomesAssessment. London: Routledge: Instruments for Adults. 3rd edition; 2004.44. Goering P: Making a Difference: Ontario’s Community Mental HealthEvaluation Initiative. Community Mental Health Evaluation Initiative: Toronto,ON; 2004.45. Lamers LM, Bouwmans CAM, van Straten A, Donker MCH, Hakkaart L:Comparison of EQ-5D and SF-6D utilities in mental health patients.Health Econ 2006, 15:1229–1236.46. EuroQol Group: EuroQol A new facility for the measurement of health-related quality of life. Health Policy 1990, 16:199–208.47. Brooks R: EuroQol: the current state of play. Health Policy 1996, 37:53–72.48. Dickens CM, McGowan L, Percival C, Douglas J, Tomenson B, Cotter L,Heagerty A, Creed FH: Lack of a close confidant, but not depression,predicts further cardiac events after myocardial infarction. Heart 2004,90:518–522.49. United States Department of Agriculture (USDA): USDA Adult Food SecuritySurvey Module. Washington, DC: USDA; 2008.50. Dennis ML, Chan Y-F, Funk RR: Development and validation of the GAINShort Screener (GSS) for internalizing, externalizing and substance usedisorders and crime/violence problems among adolescents and adults.Am J Addict 2006, 15:80–91.51. Dennis ML, White MK, Titus JC, Unsicker JI: Global Appraisal of IndividualNeeds (GAIN). Bloomington, IL: Chestnut Health Systems: AdministrationGuide for the GAIN and Related Measures (Version 5); 2006.52. Guerriere DN, Ungar WJ, Corey M, Croxford R, Tranmer JE, Tullis E, Coyte PC:Evaluation of the ambulatory and home care record: Agreementbetween self-reports and administrative data. J of Inter Tech of HealthCare 2006, 22:203–210.53. Kashner TM, Stensland MD, Lind L, Wicker A, Rush AJ, Golden RM, HenleySS: Measuring use and cost of care for patients with mood disorders.Med Care 2009, 47:184–190.54. Sirey JA, Meyers BS, Teresi JA, Bruce ML, Ramirez M, Raue PJ, Perlick DA,Holmes D: The Cornell Service Index as a measure of health service use.Psychiatr Serv 2005, 56:1564–1569.55. Browne GB, Arpin K, Corey P, Fitch M, Gafni A: Individual correlates ofhealth service utilization and the cost of poor adjustment to chronicillness. Med Care 1990, 28:43–58.56. Forchuk C, Brown SA, Schofield R, Jensen E: Perceptions of health andhealth service utilization among homeless and housed psychiatricconsumer/survivors. J Psychiatr Ment Health Nurs 2008, 15:399–407.57. Chisholm D: Client Socio-Demographic and Service Receipt Inventory -European Version: development of an instrument for internationalresearch: EPSILON Study 5. Br J Psychiatry 2000, 177:28s–33.58. Tsemberis SJ, Moran L, Shinn M, Asmussen SM, Shern DL: Consumerpreference programs for individuals who are homeless and havepsychiatric disabilities: a drop-in center and a supported housingprogram. Am J Community Psychol 2003, 32:305–317.59. Toro PA, Passero Rabideau JM, Bellavia CW, Daeschler CV, Wall DD, ThomasDM, Smith SJ: Evaluating an intervention for homeless persons: results ofa field experiment. J Consult Clin Psychol 1997, 65:476–484.60. Dickerson FB, Origoni AE, Pater A, Friedman BK, Kordonski WM: Anexpanded version of the Multnomah Community Ability Scale: anchorsand interview probes for the assessment of adults with serious mentalillness. Community Ment Health J 2003, 39:131–137.61. Sheehan DV, Lecrubier Y, Harnett Sheehan K, Janavs J, Weiller E, Keskiner A,Schinka J, Knapp E, Sheehan MF, Dunbar GC: The validity of the MiniInternational Neuropsychiatric Interview (MINI) according to the SCID-Pand its reliability. Eur Psychiatry 1997, 12:232–241.62. Lecrubier Y, Sheehan DV, Weiller E, Amorim P, Bonora I, Harnett Sheehan K,Janavs J, Dunbar GC: The Mini International Neuropsychiatric Interview(MINI). A short diagnostic structured interview: reliability and validityaccording to the CIDI. Eur Psychiatry 1997, 12:224–231.63. Amorim P, Lecrubier Y, Weiller E, Hergueta T, Sheehan D: DSM-IH-RPsychotic Disorders: procedural validity of the Mini InternationalNeuropsychiatric Interview (MINI). Concordance and causes fordiscordance with the CIDI. Eur Psychiatry 1998, 13:26–34.64. Nasreddine ZSZ, Phillips NAN, Bédirian VV, Charbonneau SS, Whitehead VV,Collin II, Cummings JLJ, Chertkow HH: The Montreal CognitiveAssessment, MoCA: a brief screening tool for mild cognitive impairment.J Am Geriatr Soc 2005, 53:695–699.65. Lehman AF: Measures of quality of life among persons with severe andpersistent mental disorders. Soc Psychiatry Psychiatr Epidemiol 1996, 31:78–88.66. Uttaro T, Lehman A: Graded response modeling of the Quality of LifeInterview. Eval Program Plann 1999, 22:41–52.67. Lancon CC, Auquier PP, Launois RR, Toumi MM, Llorca PMP, Bebbington PP,Lehman AA: [Evaluation of the quality of life of schizophrenic patients:validation of the brief version of the Quality of Life Interview]. Encephale2000, 26:11–16.68. Giffort D, Schmook A, Woody C, Vollendorf C, Gervain M: Construction of a Scaleto Measure Consumer Recovery. Office of Mental Health: Springfield, IL; 1995.69. Corrigan PW, Giffort D, Rashid F, Leary M, Okeke I: Recovery as apsychological construct. Community Ment Health J 1999, 35:231–239.70. Corrigan PW, Salzer M, Ralph RO, Sangster Y, Keck L: Examining the factorstructure of the recovery assessment scale. Schizophr Bull 2004,30:1035–1041.71. Tsemberis S, McHugo G, Williams V, Hanrahan P, Stefancic A: Measuringhomelessness and residential stability: The residential time-line follow-back inventory. Am J Community Psychol 2007, 35:29–42.72. Ware JJ, Kosinski MM, Keller SDS: A 12-Item Short-Form Health Survey:construction of scales and preliminary tests of reliability and validity.Med Care 1996, 34:220–233.73. Larson CO: Use of the SF-12 instrument for measuring the health ofhomeless persons. Health Serv Res 2002, 37:733–750.74. Marsden J, Gossop M, Stewart D, Best D, Farrell M, Lehmann P, Edwards C,Strang J: The Maudsley Addiction Profile (MAP): a brief instrument forassessing treatment outcome. Addiction 1998, 93:1857–1867.75. Latimer J, Morton-Bourgon K, Chrétien JA: A Meta-Analytic Examination ofDrug Treatment Courts: Do they Reduce Recidivism?. Department of JusticeCanada: Ottawa, ON; 2006.Somers et al. Trials 2013, 14:365 Page 19 of 20 Horvath AO, Greenberg LS: Development and validation of the WorkingAlliance Inventory. J Couns Psychol 1989, 36:223–233.77. Horvath AO, Greenberg LS: (Eds): The Working Alliance: Theory, Research, andPractice. Chichester: John Wiley & Sons; 1994.78. Tracey TJ, Kokotovic AM: Factor structure of the Working AllianceInventory. Psychol Assess 1989, 1:207–210.79. Tichenor V, Hill CE: A comparison of six measures of working alliance.Psychother Theory Res Practice 1989, 26:195–199.80. Busseri MAM, Tyler JDJ: Interchangeability of the Working AllianceInventory and Working Alliance Inventory, Short Form. Psychol Assess2003, 15:193–197.81. City of Vancouver: Housing and Homelessness Strategy 2012–2021.Vancouver, BC: City of Vancouver; 2011. Tsemberis SJ: Housing First: The Pathways Model to End Homelessness forPeople with Mental Illness and Addiction. Hazelden: Center City, MN; 2010.83. Schutt RK, Goldfinger SM: Housing preferences and perceptions of healthand functioning among homeless mentally ill persons. Psychiatr Serv1996, 47:381–386.84. Schutt RK, Goldfinger SM: The contingent rationality of housingpreferences: Homeless mentally ill persons' housing choices before andafter housing experience. In Research in Community and Mental Health,Volume 11. Emerald Group Publishing Limited: Morrissey JP. Bingley;2000:131–156.85. Levitt AJ, Jost JJ, Mergl KA, Hannigan A, DeGenova J, Chung SY: Impact ofchronically street homeless tenants in congregate supportive housing.Am J Orthopsychiatry 2012, 82:413–420.86. Kertesz SG, Crouch K, Milby JB, Cusimano RE, Schumacher JE: Housing firstfor homeless persons with active addiction: are we overreaching?Milbank Q 2009, 87:495–534.doi:10.1186/1745-6215-14-365Cite this article as: Somers et al.: Vancouver At Home: pragmaticrandomized trials investigating Housing First for homeless and mentallyill adults. Trials 2013 14:365.Submit your next manuscript to BioMed Centraland take full advantage of: • Convenient online submission• Thorough peer review• No space constraints or color figure charges• Immediate publication on acceptance• Inclusion in PubMed, CAS, Scopus and Google Scholar• Research which is freely available for redistributionSubmit your manuscript at et al. Trials 2013, 14:365 Page 20 of 20


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items