Skip to main content


  • Research article
  • Open Access
  • Open Peer Review

History of foster care among homeless adults with mental illness in Vancouver, British Columbia: a precursor to trajectories of risk

  • 1Email author,
  • 1 and
  • 1
BMC Psychiatry201515:32

  • Received: 29 May 2014
  • Accepted: 13 February 2015
  • Published:
Open Peer Review reports



It is well documented that a disproportionate number of homeless adults have childhood histories of foster care placement(s). This study examines the relationship between foster care placement as a predictor of adult substance use disorders (including frequency, severity and type), mental illness, vocational functioning, service use and duration of homelessness among a sample of homeless adults with mental illness. We hypothesize that a history of foster care predicts earlier, more severe and more frequent substance use, multiple mental disorder diagnoses, discontinuous work history, and longer durations of homelessness.


This study was conducted using baseline data from two randomized controlled trials in Vancouver, British Columbia for participants who responded to a series of questions pertaining to out-of-home care at 12 months follow-up (n = 442). Primary outcomes included current mental disorders; substance use including type, frequency and severity; physical health; duration of homelessness; vocational functioning; and service use.


In multivariable regression models, a history of foster care placement independently predicted incomplete high school, duration of homelessness, discontinuous work history, less severe types of mental illness, multiple mental disorders, early initiation of drug and/or alcohol use, and daily drug use.


This is the first Canadian study to investigate the relationship between a history of foster care and current substance use among homeless adults with mental illness, controlling for several other potential confounding factors. It is important to screen homeless youth who exit foster care for substance use, and to provide integrated treatment for concurrent disorders to homeless youth and adults who have both psychiatric and substance use problems.

Trials registration numbers

Both trials are registered with the International Standard Randomized Control Trial Number Register and were assigned ISRCTN57595077 (Vancouver At Home Study: Housing First plus assertive community 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) on September 9, 2012.


  • Adjust Odds Ratio
  • Foster Care
  • Adverse Childhood
  • Homeless Youth
  • Adverse Childhood Experience


The childhoods of homeless adults are disproportionately characterized by persistent poverty, residential mobility, school problems, and other stressful and/or traumatic events [1,2], particularly among homeless individuals with mental illness [3]. One of the earliest identifiable precursors to homelessness may be placement in out-of-home or foster care. Studies of youth “aging out” of the foster care system indicate that between 11% and 36% [4,5] experience street homeless and approximately one-third live with family, friends or acquaintances because they cannot afford permanent housing [6]. In fact, many youth are discharged from foster care without a place to live [6].

In addition to challenges around finding stable housing, many youth exiting foster care do not complete high school and have difficulty finding employment [7,8]. Underlying difficulties related to the transition to adulthood among youth leaving foster care is a cumulative history of trauma, abuse and neglect that results in a number of neurobiological and behavioural deficits including poor social support and coping skills [9,10] as well as risk for further abuse and victimization [8]. Once homeless or “drifting”, street culture promotes antisocial behavior, including association with deviant peers and means of subsistence (e.g., sex trade, theft, drug dealing) [11]. The cumulative impact of these factors places youth at high risk for a wide range of health and psychosocial problems including substance misuse and chronic homelessness [12].

A large body of evidence suggests that adverse childhood experiences, which typically include physical, sexual, and emotional abuse, neglect, dysfunctional family environments, and unstable family structure, are linked to later biological, psychological and social functioning and may affect multiple domains of health and well-being [12-14]. Moreover, it appears that adverse childhood experiences tend to cluster together [15,16] and the number of adverse experiences may be more predictive of negative adult outcomes than particular categories of events. Tsai, Edens and Rosenheck [17] examined the childhood profiles of 738 homeless adults and found three clusters: numerous childhood problems (21%); disrupted family structure (44%); and few childhood problems (35%). Participants with histories of foster care were significantly more likely to fall into the ‘numerous childhood problems’ category, which in turn predicted younger age when first homeless and more severe drug use than other participants. Tam, Zlotnick and Robertson [18] examined the long-term effects of adverse childhood experiences on adult substance use, social service use, and employment among homeless adults in Oakland. Adverse childhood experiences were positively correlated with consistent substance use over 15 months of follow-up as well as social service use. Further, consistent substance use was negatively associated with employment and social service use. However, this latter study did not separate foster care from other adverse childhood events, therefore, the contribution of foster care to various outcomes is not clear.

More specifically, research has documented the association between childhood foster care placement and subsequent homelessness [7,11]. Further, research with homeless populations has found that childhood histories of foster care are significantly correlated with psychiatric disorders in adulthood [3,19]. Roos et al. [19] examined the association between foster care history and a range of health and social outcomes among homeless adults with mental disorders, 75% of whom were Aboriginal, in Winnipeg, Canada. Compared to participants without histories of foster care, those placed in care were significantly more likely to be young, female, less educated, married/cohabitating, Aboriginal, and reported a longer lifetime duration of homelessness. A history of foster care independently predicted a range of adult outcomes including depression, psychosis, substance dependence and interpersonal violence. While various forms of interpersonal violence were examined, different types and frequencies of substance use were not included. A more in depth examination of the impact of foster care on various aspects of substance use is required.

In this study, we focus on the relationship between foster care placement as a predictor of adult homelessness and associated health and service outcomes among a sample of homeless adults with mental disorders. Specifically, we examine the relationship between foster care history and adult substance use (including age of initiation, frequency, severity and type of substance), mental illness, vocational functioning, service use and duration of homelessness. We hypothesize that a history of foster care predicts earlier, more severe and more frequent substance use, multiple mental disorder diagnoses, discontinuous work history and longer durations of homelessness.


The Vancouver At Home Study includes two randomized controlled trials involving homeless adults with mental illness in Vancouver, British Columbia. Vancouver At Home is one of five experimental projects investigating Housing First in Canada, funded by the Federal Government [20]. Details related to the trials including eligibility criteria, recruitment, randomization, retention, measures, interventions, and hypotheses have been reported elsewhere [21]. The current study focuses on baseline data from both trials in Vancouver and does not incorporate any longitudinal findings. The two trials are distinguished based on participants’ level of need. Based on an algorithm that included information about diagnosis, community functioning, and service use, participants were assigned to a “high” or “moderate ” need group [20,21].

Eligibility criteria included legal adult status (19 years and older), current mental disorder on the MINI International Neuropsychiatric Interview (MINI) [22], and being absolutely homeless or precariously housed. The MINI is a structured diagnostic interview that can be administered by non-clinicians and inquires about symptoms related to the following categories: Psychotic Disorder, Mood Disorder with Psychotic Features, Hypo(manic) Episode, Major Depressive Episode, Suicide Risk, Panic Disorder, Post-traumatic Stress Disorder, Alcohol Dependence, and Substance Dependence. Absolute homelessness was defined as living on the streets or in an emergency shelter for at least the past seven nights with little likelihood of obtaining secure accommodation in the upcoming month. Precariously housed was defined as living in a rooming house, hotel or other transitional housing; in addition, individuals must have experienced at least two episodes of absolute homelessness in the past year, or one episode lasting for at least four weeks in the past year.

Participants were recruited through referral from over 40 agencies available to homeless adults in Vancouver; the majority was recruited from homeless shelters, drop-in centers, homeless outreach teams, hospitals, community mental health teams, and criminal justice programs. Referral was typically initiated by service providers and a preliminary screening for eligibility (e.g., duration of homelessness, mental health and substance use problems), was conducted via telephone with the referral agent. All participants met face-to-face with a trained research interviewer who explained procedures, obtained informed consent, and confirmed study eligibility. A cash honorarium of $5 was provided to the participant for the screening process. Institutional ethics board approval was obtained through Simon Fraser University and the University of British Columbia. Data derived through the shared cross-site protocol are owned by the study sponsor (Mental Health Commission of Canada). Approval for use of baseline data from both trials used in this study was granted by the study sponsor. Data that are specific to the Vancouver site (site-specific instruments, administrative data) are retained at the host institution (Simon Fraser University).

If the individual met all study criteria, they were enrolled as a participant and the baseline interview commenced, consisting of a series of interviewer-administered questionnaires including socio-demographic characteristics, psychiatric symptoms, substance use, physical health, service use, and quality of life [21]. Participants received a further cash honorarium of $30 upon completion of the baseline interview, which typically took 90 minutes to complete. The following analyses are based upon data from the baseline questionnaires of 497 participants recruited from October 2009 to June 2011 and data from a series of questions related to out-of-home care, which were administered 12 months after baseline.

Variables of interest

Details associated with out-of-home care were elicited by a series of questions related to personal experiences of out-of-home care and having one’s biological children placed in out-of-home care. First, participants were asked if they ever lived away from their parents prior to age 18 and, if so, the reason for living away from home. Second, participants were asked if they were ever placed in foster care, the reason for the placement, and their age and the duration of the most recent placement. Finally, participants were asked if they had biological children currently or previously placed in foster care or other forms of out-of-home care.

Mental disorders were assessed using the MINI and were grouped, based on convention [23], into a severe cluster, comprised of (at least one) current Psychosis, Mood Disorder with Psychotic Features, and Hypo(manic) Episode, and a less severe cluster comprised of Major Depressive Episode, Panic Disorder, and Post-traumatic Stress Disorder (PTSD). Frequency and type of substance use over the past month were recorded using the Maudsley Addiction Profile (MAP) [24], a structured interview which inquires about a wide range of substances and the number of days used in the past month, the amount used on a typical day, and the route of administration. Physical illness was assessed by self-report using a checklist of 30 chronic health conditions (lasting longer than six months). Blood-borne infectious disease consisted of positive self-report diagnosis of HIV, Hepatitis B or Hepatitis C. Vocational functioning was assessed via two questions: (1) have you ever had a job that lasted for at least one year? (yes/no) and (2) are you currently employed in paid work? (yes/no) Psychometric properties for all measures are provided in previous manuscripts [20,21].

Statistical analyses

Descriptive statistics (median and inter-quartile range for continuous variables; frequency and percentage for categorical variables) were used to characterize the sample. All continuous variables included in statistical analyses (i.e., age at randomization, duration of homelessness, age first homeless, age of first drug/alcohol use) were transformed into categorical variables. For dichotomous variables, median values were used as the cut-off point. Comparisons of categorical data between participants who completed or did not complete the foster care items, and those who did or did not report a history of foster care, were conducted using Pearson’s chi-square test.

Univariate and multivariable logistic regression analyses were used to model the independent associations between placement in foster care and a series of outcome variables that were established a priori based on previous literature [6,8,25,26]. Each outcome variable was modeled in both univariate and multivariate settings using placement in foster care as an independent risk factor. Outcome variables that were significant at the p ≤ 0.05 level were considered for the multivariable logistic regression analyses using the same set of controlling variables (i.e., age at randomization, gender, ethnicity, marital status, and level of need), chosen based on previous literature [11]. Both unadjusted and adjusted odds ratios and 95% confidence intervals (CI) are reported as effect sizes and all p-values are two-sided. IBM SPSS Statistics (version 22) was used to conduct the analyses. Missing values that ranged from zero to 1% for all outcome and controlling variables in the regression analysis were excluded, with the exception of age of first alcohol and drug use. A small minority of participants who reported never using alcohol and/or drugs did not have valid information for these variables (5% and 9% respectively).


In total, 497 participants completed the baseline questionnaires. Of the total sample, 449 participants (90%) were located for the 12 month interview and 442 of those participants provided a valid response when asked if they had ever been placed in foster care. Among 442 participants, only 5 (1%) responded “Don’t know” which was categorized as “No”. An additional 1% of participants declined to respond to the foster care questions and was excluded from the analysis. Table 1 presents the baseline characteristics for the full baseline sample (n = 497) and for participants who completed all of the foster care items (n = 442). At baseline, the majority of participants who responded to the foster care items were male (72%) and White (57%); the mean age at randomization was 41.0 (SD = 10.9) years; and the mean age when first homeless was 30.1 (SD = 13.4) years. The median duration of lifetime homelessness was 36 months (IQR: 12–84 months). Compared to participants who did not complete the foster care items (n = 55), those who did were more likely to engage in early alcohol use (before age 14) and less likely to report weekly alcohol use in the month prior to enrollment (p ≤ 0.05). Otherwise, there were no significant differences at baseline between participants who completed vs. did not complete the foster care items.
Table 1

Socio-demographic, mental disorder, and drug-use related characteristics for Vancouver At Home participants (n = 497)


All participants (n =  497) N (%)

Participants with valid response ( n =  442) N (%)

Participants with missing response (n  = 55 1 ) N (%)

P value

Need level (high)

297 (60)

267 (60)

30 (55)


Housing first interventions

297 (60)

274 (62)

23 (42)


Male gender

359 (73)

317 (72)

42 (76)


Age at randomization


19-24 years

36 (7)

29 (7)

7 (13)


25-44 years

281 (57)

254 (57)

27 (49)


>44 years

180 (36)

159 (36)

21 (38)





77 (15)

68 (15)

9 (16)



280 (56)

250 (57)

30 (55)



140 (28)

124 (28)

16 (29)


Incomplete high school

280 (57)

249 (57)

31 (56)


Single marital status

343 (70)

305 (69)

38 (73)


Have children (under age 18)

122 (25)

109 (25)

13 (24)


Hospitalized for mental illness (>6 months) in past 5 years

57 (12)

51 (12)

6 (11)


Hospitalized for mental illness (>2 times) in past 5 years

253 (53)

222 (52)

31 (60)


Worked continuously at least one year in the past

323 (65)

285 (65)

38 (69)



457 (92)

406 (92)

51 (93)


Jail in last 6 months

68 (14)

57 (13)

11 (20)


Precariously housed

109 (22)

96 (22)

13 (24)


Duration of homelessness - lifetime (>36 months)2

234 (48)

211 (48)

61 (42)


Duration of homelessness - longest single period (>12 months)

245 (50)

222 (51)

23 (42)


Age of first homelessness (<25 years)

214 (44)

192 (44)

22 (40)


Overall health (poor)

67 (13)

60 (14)

7 (13)


Less severe mental illness

264 (53)

235 (53)

29 (53)


Severe mental illness

363 (73)

323 (73)

40 (73)


Multiple (≥2) mental disorders

240 (48)

213 (48)

27 (49)


Alcohol dependence

121 (24)

104 (23)

17 (31)


Substance dependence

288 (58)

258 (58)

30 (55)


High suicidality

87 (17)

79 (18)

8 (14)


Blood-borne infectious diseases

157 (32)

142 (32)

15 (28)


Multiple (≥3) physical illness

344 (69)

309 (70)

35 (64)


Early initiation of alcohol (≤13 years)

215 (46)

198 (47)

17 (32)


Early initiation of drugs (≤13 years)

183 (40)

167 (41)

16 (30)


Daily substance use

143 (29)

127 (29)

16 (29)


Daily drug use

126 (25)

113 (26)

13 (24)


Daily illicit drug use3

74 (15)

69 (16)

5 (9)


Injection drug use

88 (18)

83 (19)

5 (9)


Use of multiple (≥2) substances

257 (52)

231 (53)

26 (47)


Use of multiple (≥2) drugs

188 (38)

173 (39)

15 (27)


Daily use of alcohol

26 (5)

23 (5)

3 (5)


Daily use of cannabis

70 (14)

61 (14)

9 (16)


*p ≤ 0.05.

**p ≤ 0.01.

1Among 55 participants with missing values, 7 participants declined to respond and 48 participants could not be located for the 12 months interview.

2Dichotomized based on median value (3 years).

3Does not include alcohol and marijuana.

As shown in Table 2, 59% of participants reported that they lived away from their parents prior to age 18. The primary reasons for living away from home were placement in foster care (37%) and running away from home (25%). Thirty percent of participants reported being placed in foster care as a child, primarily due to parental abuse and/or neglect (38%). The median number of foster care placements was 1.0, the median age of the most recent placement was 10 years, and the median duration of the most recent placement was 18 months. Further, 9% of participants reported having children currently in foster care and 23% reported having children in foster care in the past (see Table 2).
Table 2

Foster care experiences among Vancouver At Home participants


N (%)

Lived away from parents before age 18 (n = 445)

264 (59)

Reasons for living away (n = 264)4


Lived with relatives or friends

36 (14)

Foster home

98 (37)

Group home

23 (9)

Jail or detention

18 (7)

Ran away or kicked out

75 (28)

Moved out

28 (11)


36 (14)

Placed in foster care (n = 442)

135 (30)

Reasons (n = 135)5


Parental illness or accident

19 (14)

Parental incarceration

5 (4)

Parental abuse or neglect

54 (40)

Behavioural problems

34 (25)

Don’t know

19 (14)


14 (10)

Number of foster care placements (n = 125)


Median (IQR)

1 (1–4)

Age at time of most recent placement (n = 131)


Median (IQR)

10 (3–13)

Duration (months) of most recent placement (n = 122)


Median (IQR)

18 (8–48)

Biological children (currently or in the past) in foster care (n = 447)

41 (9)

Biological children (currently or in the past) adopted or living with someone else (n = 446)

105 (23)

4Proportions were derived using 264 as the denominator. Participants could endorse more than one reason.

5Proportions were derived using 135 as the denominator. Participants could endorse more than one reason.

Bivariate comparisons by foster care history (yes/no) are summarized in Table 3. Participants with histories of foster care were significantly more likely to share a particular socio-demographic profile (i.e., female, Aboriginal ethnicity, incomplete high school, younger age at enrollment, have children under age 18, no continuous work history); they also first experienced homelessness at a younger age (before 25 years) and, due to their younger age, reported shorter durations of homelessness (both lifetime and longest single episode). Participants with histories of foster care were significantly more likely to meet criteria for: the less severe cluster of mental disorders (i.e., major depressive episode, panic disorder, PTSD); two or more mental disorders; substance dependence; high suicide risk; early initiation of alcohol and/or drug use (before age 14); current daily drug use; injection drug use; and poly-drug use.
Table 3

Comparisons of socio-demographic, mental disorders and drug-use related characteristics by history of foster care among Vancouver At Home participants (n = 442)


Foster care - no

Foster care - yes

P value

Need level



186 (70)

81 (30)



121 (69)

54 (31)

Study arm


Housing first (HF)

190 (69)

84 (31)


Treatment as usual (TAU)

117 (70)

51 (30)




231 (73)

86 (27)



73 (60)

48 (40)

Age at randomization


19-24 years

18 (62)

11 (38)


25-44 years

164 (65)

90 (35)

>44 years

125 (79)

34 (21)





26 (38)

42 (62)


188 (75)

62 (25)


93 (75)

31 (25)

Educational attainment



Completed high school

152 (80)

38 (20)

Incomplete high school

153 (61)

96 (39)

Marital status



Single (never married)

213 (70)

92 (30)


94 (69)

42 (31)

Have children (under age 18)




237 (73)

87 (27)


62 (57)

47 (43)

Worked continuously at least one year in the past




89 (58)

65 (42)


215 (75)

70 (25)





26 (77)

8 (23)


279 (69)

127 (31)

Jail (past 6 months)




273 (71)

112 (29)


34 (60)

23 (40)

Housing status



Precariously housed

63 (66)

33 (34)

Absolutely homeless

244 (71)

102 (29)

Duration of homelessness - lifetime



>36 months6

168 (75)

57 (25)

≤36 months

134 (64)

77 (36)

Duration homeless - longest single episode



>12 months7

164 (77)

50 (23)

≤12 months

138 (62)

84 (38)

Age of first homelessness



≥25 years

192 (79)

52 (21)

<25 years

111 (58)

81 (42)

Overall health




265 (70)

116 (30)


42 (70)

18 (30)

Less severe mental illness




160 (77)

47 (23)


147 (63)

88 (37)

Severe mental illness




82 (69)

37 (31)


225 (70)

98 (30)

Multiple (≥2) mental disorders




176 (77)

53 (23)


131 (62)

82 (38)

Alcohol dependence




235 (70)

103 (30)


72 (69)

32 (31)

Substance dependence




140 (76)

44 (24)


167 (65)

91 (35)

Suicide risk




260 (72)

103 (28)


47 (60)

32 (40)

Blood-borne infectious diseases




207 (70)

89 (30)


97 (68)

45 (32)

Multiple (≥3) physical illness




96 (72)

37 (28)


211 (68)

98 (32)

Initiation of alcohol



Before 14 years

173 (78)

48 (22)

14 years or after

115 (58)

83 (42)

Initiation of drugs



Before 14 years

179 (76)

57 (24)

14 years or after

93 (56)

74 (44)

Frequency of substance use



Less than daily/none

231 (73)

84 (27)


76 (60)

51 (40)

Frequency of drug use



Less than daily/none

241 (73)

88 (27)


66 (58)

47 (42)

Injection drug use




257 (73)

95 (27)


47 (57)

36 (43)

Poly-substance use



Two or less

157 (76)

51 (24)

Three or more

149 (65)

82 (35)

Poly-drug use



Two or less

198 (74)

68 (26)

Three or more

108 (62)

65 (38)

Frequency of alcohol use



Less than daily/none

291 (70)

124 (30)


14 (61)

9 (39)

Frequency of cannabis use



Less than daily/none

271 (71)

110 (29)


36 (59)

25 (41)

*p ≤ 0.05.

**p ≤ 0.01.

***p ≤ 0.001.

+p > 0.05 & p ≤ 0.10.

6Dichotomized based on median value (3 years).

7Dichotomized based on median value (1 year).

Unadjusted (UOR) and adjusted odds ratios (AOR) and 95% CI for variables included in the univariate and multivariable analyses are presented in Table 4. Results from the multivariable logistic regression analyses indicate that foster care history independently predicted incomplete high school (AOR: 2.08), duration of homelessness – lifetime (AOR: 1.60) and longest single episode (AOR: 1.95), age of first homelessness (AOR: 2.36), continuous work history (AOR: 0.62); meeting criteria for the less severe cluster of mental disorders (AOR: 1.62) and two or more mental disorders (AOR: 1.69); early initiation of alcohol (AOR: 2.54) and/or drugs (AOR: 2.49) and current daily drug use (AOR:1.64). A significant positive trend (p > 0.05 and p ≤ 0.10) was observed between a history of foster care and meeting criteria for panic disorder (AOR: 1.63), PTSD (AOR: 1.60), high risk of suicide (AOR: 1.59), daily substance use (AOR: 1.53), daily cannabis use (AOR: 1.74), as well as incarceration (AOR: 1.70) and court appearance (AOR: 1.48) in the past six months.
Table 4

Logistic regression analyses for socio-demographic, mental disorder and substance use related outcomes of foster care experiences (n = 442)

Outcome Variable 8

Unadjusted OR ( 95 % CI )

P value

Adjusted 9 OR ( 95 % CI )

P value

Incomplete high school

2.51 (1.62, 3.89)


2.08 (1.29, 3.35)




Lifetime duration (>3 years)10

1.69 (1.12, 2.55)


1.60 (1.02, 2.50)


Longest single episode (>1 year)11

2.00 (1.32, 3.03)


1.95 (1.25, 3.05)


Age first homeless (<25 years)

2.69 (1.77, 4.10)


2.36 (1.42, 3.91)


Vocational functioning


Worked continuously (≥1 year) in the past

0.45 (0.29, 0.68)


0.62 (0.39, 0.98)



1.50 (0.65, 3.36)




Mental disorders


Major depressive episode

1.44 (0.96, 2.17)



Panic disorder

1.56 (0.97, 2.51)



Post-Traumatic Stress Disorder

2.22 (1.43, 3.47)


1.60 (0.98, 2.62)


Less severe cluster

2.04 (1.34, 3.10)


1.62 (1.03, 2.57)


(Hypo) manic episode

1.25 (0.76, 2.06)



Mood disorder with psychotic features

1.21 (0.72, 2.04)



Psychotic disorder

0.69 (0.46, 1.04)



Severe cluster

0.97 (0.61, 1.52)




Alcohol dependence

1.01 (0.63, 1.63)




Substance dependence

1.73 (1.13, 2.65)


1.18 (0.74, 1.88)


High suicide risk

1.72 (1.04, 2.84)


1.59 (0.92, 2.72)


Multiple mental disorders (≥2)

2.08 (1.38, 3.14)


1.69 (1.08, 2.66)


Physical health


Blood-borne infectious diseases

1.08 (0.70, 1.66)



Multiple physical illness (≥3)

1.21 (0.77, 1.89)



Overall health (poor)

0.98 (0.54, 1.77)



Substance use


Early initiation of alcohol (<14 years)12

2.60 (1.70, 3.99)


2.54 (1.60, 4.05)


Early initiation of drugs (<14 years)13

2.50 (1.63, 3.83)


2.49 (1.56, 3.96)


IV drug use

2.07 (1.27, 3.40)


1.50 (0.87, 2.59)


Daily substance use

1.85 (1.20, 2.85)


1.53 (0.96, 2.44)


Daily drug use (no alcohol)

1.95 (1.25, 3.05)


1.64 (1.01, 2.65)


Daily illicit drug use14

1.70 (1.00, 2.89)


1.23 (0.69, 2.19)


Daily alcohol use

1.51 (0.64, 3.58)



Daily marijuana use

1.71 (0.98, 2.98)



Poly-substance (≥2) use

1.69 (1.12, 2.57)


1.22 (0.77, 1.94)


Poly-drug (≥2) use (no alcohol)

1.75 (1.16, 2.65)


1.32 (0.84, 2.09)


Justice System Involvement


Incarceration (past 6 months)

1.65 (0.93, 2.93)



Been arrested (past 6 months)

1.41 (0.92, 2.15)



Court appearance (past 6 months)

1.50 (0.98, 2.28)



*p ≤ 0.05.

**p ≤ 0.01.

***p ≤ 0.001.

+p > 0.05 & p ≤ 0.10.

8Separate binary logistic regression analyses (univariate and multivariable) were conducted for each outcome using Foster Care (No vs. Yes) as an independent variable.

9Controlled for age (continuous), gender (Male, Female), ethnicity (Aboriginal, White, Other), need level (High, Moderate), and marital status (Single, Other).

10Dichotomized based on median value (3 years).

11Dichotomized based on median value (1 year).

12Dichotomized based on median value (14 years).

13Dichotomized based on median value (14 years).

14Does not include alcohol and marijuana.


Among our sample of homeless adults with mental illness, we found a significant relationship between a history of childhood foster care placement and a number of indicators of socio-demographic risk, poor mental and physical health, and problematic substance use in adulthood. Foster care history independently predicted incomplete high school, duration of homelessness, discontinuous work history, meeting criteria for a current mental disorder in the less severe cluster, multiple mental disorders, early initiation of alcohol and/or drug use, and daily drug use. Our study is the first to examine a range of different adult substance use outcomes (i.e., type, frequency, poly-substance use, age of initiation) among homeless adults with mental illness and histories of foster care placement.

Foster care placement independently predicted substance use problems in our adult sample, specifically early initiation (before age 14) of drug and/or alcohol use and daily drug use. Abuse of alcohol and other drugs places adolescents at greater risk of homelessness, although it may not be a direct causal factor [27]. Compared to their non-homeless counterparts, homeless youth use substances earlier and with greater frequency [28,29]. Further, Thompson and Hasin [30] reported that homeless youth with histories of foster care were almost nine times more likely to have been in drug treatment than homeless youth without such histories, even after adjusting for demographics, other adverse childhood events, prior arrest, unemployment and educational attainment.

Our findings suggest that daily drug use may be a common mediator for a range of early risk factors [31]. Previous research using our sample of homeless adults with mental disorders found that daily drug use was independently predicted by self-report of childhood learning disabilities [32] and adverse childhood experiences [33]. In the latter study, we found a strong relationship between the breadth of exposure to abuse or household dysfunction during childhood (i.e., the number of adverse childhood events) and a number of indicators of poor mental and physical health as well as problematic substance use in adulthood (e.g., daily drug use). In turn, previous work with the same sample has shown that daily drug use significantly predicts the duration of adult homelessness [34] and the severity of mental health symptoms [35].

The current study did not examine other potential mediators such as self-esteem or shame; however, one could assume that strong manifestations of these psychological states might manifest in the mental disorders examined in our study. Daily drug use and other potential mediators (e.g., emotion regulation skills, access to social support and early intervention, antisocial behaviors) need to be examined in future studies using path-type analysis. Cross-sectional, retrospective data cannot disentangle the unique predictors of homelessness and mental illness, but it is likely that adverse childhood experiences, particularly foster care placement, have both direct and indirect (perhaps mediated by substance use) effects on a range of adult health and social outcomes. Thus, among homeless adults with mental disorders, a history of early and persistent substance abuse may be a key indicator of risk. Our findings support the need for more accessible and integrated care for homeless adults with concurrent disorders.

In addition to substance use problems, a history of foster care independently predicted meeting criteria for a less severe type of mental disorder (i.e., major depressive episode, panic disorder, PTSD) and more than one mental disorder. It makes sense that foster care placement would contribute to mood and anxiety disorders more so than psychosis. However, Roos et al. [19] found that a history of foster care predicted adult psychotic disorders as well as depression and multiple mental disorders. While the disorders comprising our less severe cluster may appear less serious than those in the severe cluster, the disorders in the former cluster can be very debilitating. The fact that foster care predicts meeting criteria for multiple mental disorders suggests that poor psychiatric outcomes in adulthood are related to foster care, and further strengthens the case for integrated treatment for concurrent disorders for this population.

Given that over 1,000 youth exit foster care without a family each year in Canada, services and supports need to be in place to help youth address concurrent psychiatric and substance use problems prior to and after discharge. Organizations that provide services to homeless youth should target those with histories of adverse childhood events, particularly foster care, and tailor services to their collective and individual needs. Approaches such as brief motivational interviewing have been shown to reduce substance use among homeless youth in a variety of settings [36,37] and should be made more widely available as part of integrated treatment for concurrent disorders.

The significant association between foster care history and the duration of homelessness in adulthood is also of concern. Both researchers and advocates have called for more evidence-based interventions to reduce homelessness among youth exiting foster care [8,11]. Preliminary research on transitional living programs for youth exiting foster care is associated with reduced rates of street homelessness and increased employment, particularly when paired with vocational skills training [38].

In previous work, we have emphasized the importance of viewing homelessness in the context of cumulative adversity [39] and Problem Behavior Theory [40,41]. These models conceptualize homelessness as the result of successive environmental disruptions, each of which places individuals at greater risk for homelessness and associated risk factors. In addition to the effects of progressive risk, risk may be multiplied at each stage of development because the environmental risks associated with later phases present increasingly greater challenges to development and adaptation. Problem Behavior Theory hypothesizes that various early risk factors may comprise a cluster of risky behaviors that mediate the link from childhood adversity to a variety of adverse adult outcomes, rather than distinct independent pathways. Our findings, alongside others’, suggest that the relationship between foster care placement and adult homelessness (and the associated health and social problems we identified) may be mediated by problem behaviours such as heavy substance use. Further, it is likely that the number of adverse events (e.g., adverse childhood events in addition to foster care, incomplete high school, early problems with substance use and/or the law, lack of institutional or family support) may be more important in predicting negative adult health and social outcomes than any one event. Daily or heavy substance use must be tested as a mediating factor alongside other potential mediators.

Placement in out-of-home care is a significant and salient event that should indicate appropriate interventions for both children and parents. Early interventions in childhood might change or moderate the cycle of homelessness across generations because early risk factors are often longstanding and drive a trajectory of cumulative risk, potentially leading to severe psychopathology and social exclusion. Among our sample of homeless adults, 31% had children currently or previously in foster care, suggesting that foster care placement is a form of trauma that has inter-generational effects.


Despite the strengths of our study design (i.e., large sample size, structured diagnostic interviews), several limitations must be considered. First, all variables were based on participant self-report. Given that participants were selected based on presence of a current mental disorder, accuracy of recall may have been compromised. Further, the baseline interview occurred before randomization to a housing intervention and therefore may have led some participants to modify their responses. It is unlikely that participants purposefully modified their responses to the foster care items given that they were administered after the baseline interview. Furthermore, foster care is a highly salient event that is likely to be remembered; however, recall error could have occurred in questions pertaining to the reason for being placed in care.

Additionally, data were not available to examine whether participants’ biological mothers had abused substances during pregnancy, the age participants first entered foster care, the total length of time they spent in foster care, and number and types of residential and educational placements. As each of these factors has been shown to increase the risk for subsequent psychiatric disorders, the effect of these factors on the risk for psychiatric and substance use disorders should be examined in future research. In addition to the above noted variables, we did not have access to several other important variables necessary for conducting a robust path analysis in order to more fully investigate the links between foster care placement and adverse outcomes among homeless adults (e.g., early problem behaviors, social support, affect regulation skills, self-esteem/self-concept, and the availability of treatment/institutional supports). Future longitudinal studies should examine foster care in the context of other adverse childhood events, early problem behaviors that have been identified in the literature (e.g., adolescent substance use, involvement with the law, incomplete high school), as well as key adult outcomes, in order to investigate the key pathways that link foster care to adverse outcomes among homeless adults with mental illness.


This is the first Canadian study to investigate the relationship of a history of foster care with current substance use among homeless adults with mental illness, controlling for several potential confounding factors. The high rates of foster care history observed among our sample of homeless adults with mental illness highlight the dire consequences that can result from this trajectory of risk. Given the salience of removing children from the family home and placing them in foster care, this should be a sentinel event that triggers a range of interventions aimed at preventing negative adult psychiatric, health and social outcomes.



This research was funded by Health Canada and the Mental Health Commission of Canada. The authors thank the At Home/Chez Soi Project collaborative at both national and local levels; National project team: J. Barker, PhD (2008–2011) and C. Keller, National Project Leads; P. Goering, RN, PhD, Research Lead; approximately 40 investigators from across Canada and the US; 5 site coordinators; numerous service and housing providers; and persons with lived experience.

Authors’ Affiliations

Faculty of Health Sciences, Simon Fraser University, 8888 University Drive, Burnaby, British, Columbia, Canada


  1. Herman D, Susser ES, Struening EL, Link BL. Adverse childhood experiences: are they risk factors for adult homelessness? Am J Pub Health. 1997;87(2):249–55.View ArticleGoogle Scholar
  2. Koegel P, Malamid E, Burnam MA. Childhood risk factors for homelessness among homeless adults. Am J Pub Health. 1995;85(12):1642–9.View ArticleGoogle Scholar
  3. Sullivan G, Burnam A, Koegel P. Pathways to homelessness among the mentally ill. Soc Psychiatry Epidemiol. 2000;35:444–50.View ArticleGoogle Scholar
  4. Brandford C, English D: Foster youth transition to independence study. Seattle: Office of Children’s Administration Research, Washington Department of Social and Health Services; 2004Google Scholar
  5. Reilly T. Transition from Care: The Status and Outcome of Youth Have ‘Aged out’ of the Foster Care System in Clark County, Nevada. Las Vegas: Nevada KIDS COUNT; 2001.Google Scholar
  6. Fowler PJ, Toro PA, Miles BW. Pathways to and from homelessness and associated psychosocial outcomes among adolescents leaving the foster care system. Am J Pub Health. 2009;99:1453–8.View ArticleGoogle Scholar
  7. Courtney ME, Dworksy A. Early outcomes for youth people transitioning from out-of-home care in the USA. Child Fam Soc Work. 2006;11:209–19.View ArticleGoogle Scholar
  8. Dworsky A, Courtney ME. Homelessness and the transition from foster care to adulthood. Child Welfare. 2009;88:23–56.PubMedGoogle Scholar
  9. DeBellis M, Zisk A. The biological effects of childhood trauma. Child Adol Psychiatr Clinics N Amer. 2014;23(2):185–222.View ArticleGoogle Scholar
  10. Galletly C, VanHoof M, McFarlane A. Psychotic symptoms in young adults exposed to childhood trauma: a 20-year follow-up study. Schizophrenia Res. 2011;127(1–3):76–82.View ArticleGoogle Scholar
  11. Penzerro RM. Drift as adaptation: foster care and homeless careers. Child Youth Care Forum. 2003;32(4):229–44.View ArticleGoogle Scholar
  12. Zlotnick C. What research tells us about the intersecting streams of homelessness and foster care. Am J Orthopsychiatr. 2009;79:319–25.View ArticleGoogle Scholar
  13. Corso PS, Edwards VJ, Fang X, Mercy J. Health-related quality of life among adults who experienced maltreatment during childhood. Am J Pub Health. 2008;98:1094–100.View ArticleGoogle Scholar
  14. Dube S, Felitti VJ, Dong M, Chapman D, Giles W, Anda R. Childhood abuse, neglect and household dysfunction and the risk of illicit drug use: the Adverse Childhood Experience Study. Pediatrics. 2003;111:564–72.View ArticlePubMedGoogle Scholar
  15. McLaughlin K, Green J, Gruber MJ, Sampson N, Zaslavsky A, Kessler R. Childhood adversities and adult psychiatric disorders in the national comorbidity survey replication II: associations with persistence of DSM-IV disorders. Arch Gen Psychiatry. 2010;67:124–32.View ArticlePubMedPubMed CentralGoogle Scholar
  16. Green J, McLaughlin K, Berglund P, Gruber MJ, Sampson N, Kessler R. Childhood adversities and adult psychiatric disorders in the national comorbidity survey replication I: associations with the onset of DSM-IV disorders. Arch Gen Psychiatry. 2010;67:113–23.View ArticlePubMedPubMed CentralGoogle Scholar
  17. Tsai J, Edens E, Rosenheck R. A typology of childhood problems among chronically homeless adults and its association with housing and clinical outcomes. J Health Care Poor Underserved. 2011;22:853–70.View ArticlePubMedGoogle Scholar
  18. Tam T, Zlotnick C, Robertson M. Longitudinal perspective: adverse childhood events, substance use, and labor force participation among homeless adults. Am J Drug Alcohol Abuse. 2003;29:829–46.View ArticlePubMedGoogle Scholar
  19. Roos L, Distasio J, Bolton S, Katz L, Afifi T, Isaak C, et al. A history in-care predicts unique characteristics in a homeless population with mental illness. Child Abuse Negl. 2014;38:1618–27.View ArticlePubMedGoogle Scholar
  20. Goering PN, Streiner DL, Adair C, Aubry T, Barker J, Distasio J, et al. The At Home/Chez Soi trial protocol: a pragmatic, multi-site, randomised controlled trial of a housing first intervention for homeless individuals with mental illness in five Canadian cities. BMJ Open. 2011;1(2):e000323.View ArticlePubMedPubMed CentralGoogle Scholar
  21. Somers J, Patterson M, Moniruzzaman A, Currie L, Rezansoff S, Palepu A, et al. Vancouver At Home: pragmatic randomized trials investigating housing first for homeless and mentally ill adults. Trials. 2013;14:365.View ArticlePubMedPubMed CentralGoogle Scholar
  22. Sheehan DV, Lecrubier Y, Sheehan KH, Amorim P, Janavs J, Weiller E, et al. The Mini-International Neuropsychiatric Interview (M.I.N.I.): the development and validation of a structured diagnostic psychiatric interview for DSM-IV and ICD-10. J Clin Psychiatry. 1998;59 Suppl 20:22–33.PubMedGoogle Scholar
  23. Ruggeri M, Lesse M, Thornicroft G, Bisoffi G, Tansella M. Definition and prevalence of severe and persistent mental illness. Brit J Psychiatr. 2000;177:149–55.View ArticleGoogle Scholar
  24. Marsden J, Gossop M, Stewart D, Best D, Farrell M, Strang J: The Maudsley Addiction Profile (MAP): a brief instrument for treatment outcome research. Development and User Manual. London UK: National Addiction Centre/ Institute of Psychiatry; 1998. Available at:
  25. Zlotnick C, Tam T, Soman L. Life course outcomes on mental and physical health: the impact of foster care on adulthood. Am J Pub Health. 2012;102:534–40.View ArticleGoogle Scholar
  26. Montgomery P, Donkoh C, Underhill K. Independent living programs for young people leaving the care system: the state of the evidence. Child Youth Serv Rev. 2006;28:1435–48.View ArticleGoogle Scholar
  27. Sosin MR, Bruni M. Homelessness and vulnerability among adults with and without alcohol problems. Subst Use Misuse. 1997;32(7–8):939–68.View ArticlePubMedGoogle Scholar
  28. McLean MG, Paradise MJ, Cauce AM. Substance use and psychological adjustment in homeless adolescents: a test of three models. Am J Community Psychol. 1999;27:405–27.View ArticleGoogle Scholar
  29. Barczyk A, Thompson S. Alcohol/drug dependency in homeless youth. Alcoholism. Clin Experimental Res. 2008;32(s1):367A–367A.View ArticleGoogle Scholar
  30. Thompson RG, Hasin D. Cigarette, marijuana, alcohol use and prior drug treatment among newly homeless young adults in New York City: relationship to a history of foster care. Drug Alc Dependence. 2011;117:66–9.View ArticleGoogle Scholar
  31. Brook DW, Brook JS, Zhang C, Cohen P, Whiteman M. Drug use and the risk of major depressive disorder, alcohol dependence, and substance use disorder. Arch Gen Psychiatry. 2002;59(11):1039–44.View ArticlePubMedGoogle Scholar
  32. Patterson ML, Moniruzzaman A, Frankish CJ, Somers JM. Missed opportunities: childhood learning disabilities as early indicators of risk among homeless adults with mental illness in Vancouver, British Columbia. BMJ Open. 2012;2:e001586.View ArticlePubMedPubMed CentralGoogle Scholar
  33. Patterson ML, Moniruzzaman A, Somers J. Setting the stage for chronic health problems: cumulative childhood adversity among homeless adults with mental illness in Vancouver, British Columbia. BMC Pub Health. 2014;4:350.View ArticleGoogle Scholar
  34. Patterson M, Somers JM, Moniruzzaman A. Prolonged and persistent homelessness: multivariable analyses in a cohort experiencing current homelessness and mental illness in Vancouver, British Columbia. Ment Health Subst Use. 2012;5:85–101.View ArticleGoogle Scholar
  35. Palepu A, Patterson M, Strehlau V, Moniruzzaman A, Tan De Bibiana J, Frankish C, et al. Daily substance use and mental health symptoms among a cohort of homeless adults in Vancouver, British Columbia. J Urban Health. 2013;90:740–6.View ArticlePubMedGoogle Scholar
  36. Baer J, Garrett S, Beadnell B, Wells E, Peterson P. Brief motivational intervention with homeless adolescents: evaluating effects on substance use and service utilization. Psychol Addict Behav. 2007;21:582–6.View ArticlePubMedGoogle Scholar
  37. Peterson P, Baer J, Wells E, Ginzler J, Garrett S. Short-term effects of a brief motivational intervention to reduce alcohol and drug risk among homeless adolescents. Psychol Addict Behav. 2006;20:254–64.View ArticlePubMedGoogle Scholar
  38. Lenz-Rashid S. Employment experiences of homeless young adults: are they different for youth with a history of foster care? Child Youth Serv Rev. 2006;28:235–59.View ArticleGoogle Scholar
  39. Whitbeck LB, Simons RL. Life on the streets: the victimization of runaway and homeless adolescents. Youth Soc. 1990;221:108–25.View ArticleGoogle Scholar
  40. Jessor R. Risk behavior in adolescence: a psychosocial framework for understanding and action. J Adol Health. 1991;12:597–605.View ArticleGoogle Scholar
  41. Jessor R, Jessor SL. Problem Behavior and Psychosocial Development: A Longitudinal Study of Youth. NY: Academic Press; 1977.Google Scholar


© Patterson et al.; licensee BioMed Central. 2015

This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver ( applies to the data made available in this article, unless otherwise stated.