Homelessness is a multifactorial problem and some homeless-related factors have been researched more than others. In particular, researchers have identified consistent links between homelessness and severe mental illness, including psychotic, bipolar, and depressive disorders.1–6 At the same time, however, there is need to explore pathways that might connect mental illness to homelessness in order to inform treatment and programming. In this regard, although psychiatric symptoms of severe mental illness are often studied, financial strain is sometimes overlooked, even though adults with severe mental illness have lower income, worse employment, and greater financial barriers compared with adults without severe mental illness.7–10 As a result, greater financial strain accompanying severe mental illness could contribute to risk of homelessness.
Financial strain itself relates to homelessness. A national survey found “economic difficulty in the past 12 months” was one of the strongest factors associated with lifetime experience of homelessness in young adults.11 Similarly, in another study, participants who reported recent overwhelming debt were three times more likely to report past history of homelessness than those without.3 Two studies have examined the link between financial strain and risk of homelessness prospectively: one found that poverty predicted first-time homelessness in the general population12 whereas the other found money mismanagement quadrupled rates of Veteran homelessness in the following year.13
To our knowledge, relatively little empirical research has examined the intricate links between financial strain, mental illness, and homelessness. In addition, few studies on homelessness examine various types of financial strain variables together, which is important given that history of homelessness, unemployment, financial debt, and low income do not occur in a vacuum but exist within the context of an individual’s socioeconomic circumstances. Finally, we are unaware of any studies that have examined the extent to which financial strain may mediate the relationship between severe mental illness and homelessness.
In this regard, there is a need for nationally representative, longitudinal datasets to examine whether severe mental illness as well as financial strain might predict future homelessness and whether financial strain affects the impact of mental illness on homelessness. If a mediating effect of financial strain were found, such knowledge would support efforts to prevent homelessness by assessing financial well-being and formalizing financial education in the context of homeless services. The current study aimed to examine financial strain as a predictor of homelessness and as a mediator of the link between severe mental illness and homelessness.
METHODS
Sample
This report analyzes data from waves 1 and 2 of the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC), a face-to-face interview conducted by the National Institute on Alcohol Abuse and Alcoholism.14,15 Wave 1, conducted between 2001 and 2002, included 43,093 US adults residing in the 50 states and District of Columbia. Of these individuals, 34,653 (86.7%) completed the wave 2 interview between 2004 and 2005. Demographics have been detailed in other publications of the NESARC.16 The current sample was 60% female, 42% non-White, and had a median age of 46 years, with 83% reporting having completed high school.
Measures
Severe mental illness specifically bipolar disorder, psychotic disorder, or major depressive disorder (0=no; 1=yes) and alcohol and/drug abuse and/or dependence denoting substance use disorder (0=no; 1=yes) in the past 12 months were assessed at wave 1 using the National Institute on Alcohol Abuse and Alcoholism Alcohol Use Disorder and Associated Disabilities Interview Schedule—DSM-IV Version, a structured diagnostic interview designed for lay interviewers, shown to have good to excellent reliability.17
Financial Debt/Crisis
Respondents were asked at wave 1, “In the past 12 months, have you experienced a major financial crisis, declared bankruptcy or more than once been unable to pay your bills on time?” and to respond no or yes (coded as 0 or 1, respectively).
Unemployment
Respondents were asked to respond yes or no to the following items: (1) “present situation includes unemployed or laid off and looking for work,” (2) “present situation includes unemployed or laid off and not looking for work,” (3) “present situation includes unemployed and permanently disabled,” and (4) “unemployed and looking for work for >1 month in last 12 months.” If respondents affirmed any of these items, unemployment was coded as “yes.”
Annual Income
Respondents were asked to indicate “total household income in last 12 months (including any income from food stamps)” and provided income ranges (0=below median of $35,000; 1=at or above median of $35,000).
History of Homelessness
Respondents were asked “Did you ever have a time lasting 1 or more months when you had no regular place to live?” and “Did you ever have a time lasting 1 or more months when you had to live with others because you had no place of your own?” Those answering “yes” to either question were coded as having history of homelessness.
Other covariates included participant age, sex (0=female and 1=male), race (0=non-White and 1=White), marital status (1=married; 0=other), educational level (1=high school or greater; 0=less than high school), and criminal justice involvement (0=none; 1=indication of arrests or incarceration in the past).
Homelessness Between Waves 1 and 2
Respondents were asked the following items: (1) “Since the last interview, did you ever have a time lasting 1 or more months when you had no regular place to live?” and (2) “Since the last interview, did you ever have a time lasting 1 or more months when you had to live with others because you had no place of your own?” Participants who endorsed either question were classified as experiencing homelessness between waves 1 and 2.
Statistical Analysis
χ2 and logistic regression analyses examining predictors of homelessness were conducted in SAS 9.4 with SUDAAN, which adjusts for population-level variables and methodological sampling bias (eg, survey design, nonresponse, and sample attrition; Research Triangle Institute, 2008) so that poststratification weights were applied to be representative of the US population. Multivariable logistic regression models using logit transformation and Taylor Series (linearization) variance estimation were used to identify associations between risk factors at wave 1 and homelessness between waves 1 and 2. For multivariate analyses, we controlled for the length of time between waves 1 and 2 (Minterval=3 y, 18 d).
Predicted probabilities of homelessness between waves 1 and 2 were conducted as a function of the number of financial strain variables at wave 1, stratified by presence or absence of severe mental illness. Mediation analysis using PROCESS18 was conducted in which severe mental illness at wave 1 served as the independent variable, number of financial strain variables at wave 1 served as the mediating variable, and homelessness (as a continuous variable, 0=none; 1=living with others; 2=literal homelessness) between waves 1 and 2 served as the dependent variable.
RESULTS
With respect to financial variables, 12% of the study sample reported a financial debt/crisis in the past 12 months, 12% had a history of homelessness, and 14% were unemployed. In regards to clinical variables in the past 12 months, 7.47% (n=2587) had major depressive disorder, 0.39% (n=135) psychotic disorder, 1.76% (n=609) bipolar disorder, and 8.59% (n=2975) met criteria for alcohol or drug abuse/dependence. Demographically, the median age of study participants was 46 years and 58% reported White race/ethnicity. In total, 83% graduated from high school, half were married.
Table 1 shows bivariate relationships between covariates at wave 1 and homelessness at wave 2. Increased odds of homelessness at wave 2 were predicted by multiple variables including: being young, unmarried, previously incarcerated, no high school education, and substance use.
TABLE 1 -
Association Between Covariates at Wave 1 and Homelessness at Wave 2
|
Homelessness Between Waves 1 and 2 |
|
|
| Variable at Wave 1 |
Unweighted Frequency of Variable at Wave 1 (n) |
Unweighted Frequency Homelessness Between Waves 1 and 2 (n) |
Weighted Percentage Homelessness Between Waves 1 and 2 (%) |
χ2
|
P
|
| Age |
| Younger than 35 y old |
16,248 |
1303 |
8.50 |
215.26 |
<0.001 |
| At least 35 y old |
18,138 |
420 |
2.12 |
|
|
| Sex |
| Male |
14,452 |
774 |
5.53 |
3.12 |
0.08 |
| Female |
19,934 |
949 |
4.92 |
|
|
| High school education |
| Completed high school |
28,711 |
1366 |
4.86 |
20.59 |
<0.001 |
| Some high school or less |
5675 |
357 |
7.30 |
|
|
| Race |
| Non-White |
14,357 |
779 |
6.21 |
14.07 |
<0.001 |
| White |
20,029 |
944 |
4.81 |
|
|
| Married |
| Yes |
17,285 |
496 |
2.87 |
188.63 |
<0.001 |
| No |
17,101 |
1227 |
8.70 |
|
|
| Severe mental illness |
| Yes |
2909 |
328 |
11.98 |
61.55 |
<0.001 |
| No |
31,477 |
1395 |
4.63 |
|
|
| Substance use |
| Yes |
2958 |
342 |
12.37 |
77.15 |
<0.001 |
| No |
31,428 |
1381 |
4.49 |
|
|
| History of arrests/incarceration |
| Yes |
4050 |
400 |
10.45 |
74.96 |
<0.001 |
| No |
30,336 |
1323 |
4.51 |
|
|
Table 2 shows bivariate relationships between financial variables at wave 1 and homelessness at wave 2. Increased odds of homelessness at wave 2 were predicted by several variables at wave 1, including: being unemployed, financial debt/crisis low income, and past homelessness.
TABLE 2 -
Association Between Financial Strain at Wave 1 and Homelessness at Wave 2
|
Homelessness Between Waves 1 and 2 |
|
|
| Variable at Wave 1 |
Unweighted Frequency of Variable at Wave 1 (n) |
Unweighted Frequency Homelessness Between Waves 1 and 2 (n) |
Weighted Percentage Homelessness Between Waves 1 and 2 (%) |
χ2
|
P
|
| Financial debt/crisis |
| Yes |
4064 |
478 |
12.84 |
105.54 |
<0.001 |
| No |
30,123 |
1233 |
4.30 |
|
|
| Unemployed |
| Yes |
4783 |
572 |
13.43 |
137.49 |
<0.001 |
| No |
29,603 |
1151 |
4.02 |
|
|
| Low income |
| Yes |
16,122 |
1062 |
7.14 |
96.82 |
<0.001 |
| No |
18,264 |
661 |
3.94 |
|
|
| History of homelessness |
| Yes |
4076 |
501 |
13.34 |
131.89 |
<0.001 |
| No |
29,655 |
1184 |
4.15 |
|
|
As expected, severe mental illness significantly predicted homelessness. When specific diagnoses are examined, 18.8% of respondents with psychotic disorder, 11.46% of respondents with major depressive disorder, and 18.15% of respondents with bipolar disorder reported homelessness between waves.
In addition, non-White respondents had higher levels of homelessness than White respondents. When examining specific race/ethnicity groups, the percentage who reported being homelessness between waves 1 and 2 were as follows: White, not Hispanic or Latino (4.81%); Black, not Hispanic or Latino (7.89%); American Indian/Alaska Native, not Hispanic or Latino (6.99%); Asian/Native Hawaiian/Pacific Islander, not Hispanic or Latino (3.31%) and Hispanic or Latino (5.51%).
Table 3 presents risk factors at wave 1 and homelessness at wave 2. The final model (χ2=1840.39, df=13, P<0.001) was significant. Risk factors reported at wave 1 indicating higher probability of homelessness at wave 2 included being young, unmarried, history of homelessness, unemployed, financial debt/crisis, low income, severe mental illness, and substance abuse.
TABLE 3 -
Multivariable Logistic Regression of Risk Factors at Wave 1 Predicting Homelessness Between Waves 1 and 2
|
Homelessness Between Waves 1 and 2 |
| Variable at Wave 1 |
Adjusted Odds Ratio |
95% CI |
P
|
| Severe mental illness |
1.30 |
1.07–1.58 |
0.008 |
| Financial debt/crisis |
1.55 |
1.32–1.82 |
<0.001 |
| Unemployed |
1.86 |
1.60–2.16 |
<0.001 |
| Income (<median) |
1.29 |
1.12–1.49 |
<0.001 |
| History of homelessness |
2.24 |
1.93–2.59 |
<0.001 |
| Age |
0.96 |
0.95–0.96 |
<0.001 |
| Sex (male) |
1.00 |
0.86–1.16 |
0.97 |
| No high school education |
1.33 |
1.11–1.59 |
0.002 |
| Race (White) |
1.17 |
1.01–1.35 |
0.04 |
| Unmarried |
1.72 |
1.48–2.00 |
<0.001 |
| Substance abuse/dependence |
1.29 |
1.09–1.54 |
0.004 |
| History of arrests/incarceration |
1.25 |
1.04–1.49 |
0.02 |
Full model: AUC=0.79, χ2=1840.39, df=13, P<0.001.
AUC indicates area under receiver operating characteristic; CI, confidence interval.
Figure 1 shows predicted probability of homelessness between waves 1 and 2 as a function of the number of financial strain variables participants endorsed at wave 1, stratified by the presence of mental illness at wave 1. Not surprisingly, participants with no financial strain variables had the lowest risk of homelessness, regardless of if they had a mental illness. Conversely, participants with all 4 financial strain variables had significantly higher risk of homelessness. Of note, the rate of increase with each additional financial strain variable did not depend on whether participants reported having mental illness.
;)
FIGURE 1: Predicted probability of homelessness between waves 1 and 2 as a function of financial strain at wave 1, stratified by severe mental illness. Financial strain variables include financial debt/crisis, past homelessness, unemployment, and below median income.
Figure 2 presents pathways of direct and indirect effects that depict unstandardized path regression coefficients for severe mental illness at wave 1 predicting homelessness between waves 1 and 2 as mediated by the number of financial strain variables at wave 1. The direct effect of severe mental illness on homelessness was significant: unstandardized coefficient=0.0517, SE=0.0052, t=9.9609, P<0.001. The indirect effect of financial strain was also significant: unstandardized coefficient=0.0338, SE=0.0014, Z=24.36, P<0.001. Indirect effects (percentages of total effect) represent the proportion of the association between severe mental illness and homelessness that can be explained by financial strain. As such, the current mediation analysis indicates that financial strain accounts for 0.0338/0.0856=39% of the total effect of severe mental illness on homelessness.
FIGURE 2: Pathways of direct and indirect effects depict unstandardized path regression coefficients for severe mental illness at wave 1 predicting homelessness between waves 1 and 2 as mediated by financial strain at wave 1, controlling for history of homelessness as a covariate. Total effects of the model were as follows: unstandardized coefficient=0.0856, SE=0.0051, t=16.6417, P<0.001. The direct effect of severe mental illness (0.0517) and indirect effect of severe mental illness via financial strain (0.0338) on subsequent homelessness were both statistically significant (P<0.001). Mediation analysis revealed that financial strain accounted for 39% of the total effect of severe mental illness on homelessness. ***P<0.001.
DISCUSSION
This study shows that financial debt and crises, past homelessness, unemployment, and lower income each prospectively increase subsequent risk of homelessness. These financial strain variables were significant even when controlling for well-studied risk factors for homelessness including substance abuse, criminal justice involvement, and severe mental illness. With respect to severe mental illness, the analysis demonstrated that its link to future homelessness was significantly mediated by financial strain. This finding is important and suggests that addressing not only severe mental illness but financial strain would be essential for permanently reducing an adult’s risk for homelessness.
The findings indicate that interventions could be proactively targeted at improving an individual’s financial literacy and well-being such that they could prevent situations that may contribute to future homelessness. Indeed, the analysis showed that 39% of the link between mental illness and future homelessness is accounted for by financial strain. Of note, the mediation analysis predicting future homelessness controlled for past homelessness, which was included in the formulation of financial strain. This finding suggests that addressing mental illness without consideration of financial strain may not lead to optimal reduction in homelessness risk. Conversely, the findings attest to the need to help individuals, with or without mental illness, achieve greater financial well-being so as to reduce chances of future homelessness. This conclusion is most strongly supported by Figure 1 which illustrates that participants with mental illness and few financial strain variables had a relatively similar homelessness risk compared with participants without mental illness and few financial strain variables.
The findings provide a fresh perspective from which to view homelessness prevention efforts. To illustrate, one of the most effective interventions to reduce homelessness is Housing First, which places homeless individuals immediately into residences without requiring mental health or substance abuse treatment.19 Studies demonstrate this approach is cost-effective and improves living stability.19–21 Of relevance to the current study, most Housing First programs require that participants use approximately one third of their income to pay rent for housing,19–22 and participants are also typically offered some type of money management assistance. These programs help participants to create a spending plan and ultimately, foster financial education. Although this may be considered so rudimentary that one might argue it is not a spending plan at all, but for many individuals, this may be the first budget they have ever made or been asked to follow. From this perspective, Housing First programs provide exactly the type of basic structure for money management that many participants may have previously lacked. At the very least, financial capability and well-being appears to be a crucial component of such programs.
Similarly, the Department of Veterans Affairs offers a wide array of homeless programs, ranging from helping Veterans at imminent risk of eviction to those who are literally homeless to those who have government vouchers and are in temporary or long-term housing. In each, case managers provide either formal or informal financial education with Veterans experiencing homelessness. In addition, case managers connect Veterans with employment services to increase income. Taken together, when considering both Housing First and Veterans Affairs homeless programs, the current empirical findings support efforts to offer homeless clients opportunities to grow financially through employment, increased financial knowledge, and judgment skills for money management. Case managers in these programs should examine how mental illness may impact a client’s spending habits and conversely, how the client’s money mismanagement might lead to stress and thereby increase symptoms of severe mental illness.
Although there is a plethora of empirical research that supports the link between socioeconomic status and poor mental health outcomes,7–10 it is important to consider the nuanced elements that constitute financial strain and the possibility that they may not be correlated with each other. Pearlin’s Stress Process Model23,24 offers a useful framework for understanding differential responses to financial strain. Pearlin asserts that life events or stressors are not isolated but rather exist within the greater context of an individual’s location in society and culture. Elements of the stress process model include social characteristics, stress exposure, as well as social and personal resources such as mastery and self-esteem. This model highlights the necessity of considering individual, family, and community level factors when predicting health outcomes. Of relevance for the current research, financial strain could act as a discrete life event, a secondary stressor, or a chronic stressor across the lifespan. Current findings on the mediating role of financial strain on the effect of mental illness on homelessness is thus consistent with this framework.
Correspondingly, though not explicitly measured in this study, it is also critical to consider the role stigma may play with respect to making it more difficult for people with severe mental illness to achieve financial wellness.25–27 Empirical studies have revealed both covert and overt discrimination against hiring people with mental illness and other disabilities.27–29 Moreover, studies show that people with mental illness who do work are more likely to be in low-wage positions25,30; thus, people with mental illness will be more likely to be living paycheck to paycheck, which increases chances of homelessness. At the same time, some public health approaches could be considered. For example, among individuals who are too mentally ill to work, representative payee (rep payee) programs have been shown to improve outcomes for individuals with severe mental illness and histories of homelessness.31 Thus, the current data indicate that focusing solely on individual-level interventions is only one piece—other important facets include thinking about opportunities at the structural/social levels for public health approaches to prevent or end homelessness for people with mental illness.
Additional consideration should be given to the impact of low wages, unstable employment, insufficient welfare support, medical bills, and other financial issues that may be out of people’s control and are dictated by the economy and governmental policies. Thus, teaching individuals financial skills to manage dwindling resource has limitations, and it is critical to recognize the broader economic status of society as the pathway to financial strain is more complex than only examining individual-level characteristics.
Nevertheless, Figure 1 shows that the more financial strain variables, the greater the risk of homelessness, regardless of mental illness status. Research has shown that interventions can help individuals with mental illness improve financial skills,32 so it is critical to recognize that mental illness does not inherently compromise an individual’s ability to manage their finances. In fact, the vast majority of people who receive disability from the government for psychiatric disorders are deemed competent to manage their money. For these reasons, the current study supports money management programs of people with mental illness as a means to teach them skills to reduce risk of future homelessness.
Study limitations should be considered. Although the NESARC was administered by highly trained interviewers, structured interview data were based on self-report; respondents may have underestimated the actual occurrence of domain variable behaviors or been hesitant to endorse homelessness due to social undesirability. Because the NESARC targeted noninstitutionalized subjects, individuals hospitalized for serious mental illness were excluded from the sampling frame and this may have contributed to an underestimate of the phenomena of interest. Within each of the 4 financial strain variables selected, there were many facets of each that would be important to examine in future research. For example, what behaviors led to bankruptcy? If someone is unemployed but not looking for work, does that necessarily mean that they need employment? For homelessness, we did not know how recently the respondent experienced housing instability; future work should assess recency of homelessness and its association with future homelessness as one could surmise that temporal proximity may be an issue.
In sum, the current study demonstrated that financial strain not only was associated with risk of homelessness, but also significantly mediated the relationship between severe mental illness and homelessness. This suggests that adding financial well-being as a focus of homelessness prevention efforts seems promising, both at the individual level and community level. The current study reveals that financial strain and mental illness should be considered together when strategizing how to make sure a given individual does not become homeless again in the future. The current study thus provides a new perspective on which to view the pathway toward homelessness by recognizing that financial well-being is prominent and, along with psychiatric symptoms and co-occurring substance abuse, should to be addressed to prevent homelessness among individuals with severe mental illness.
REFERENCES
1. Caton C, Shrout PE, Eagle PF, et al. Risk factors for homelessness among schizophrenic men: a case-control study. Am J Public Health. 1994;84:265–270.
2. Odell S, Commander M. Risk factors for homelessness among people with psychotic disorders. Social Psychiatry Psychiatr Epidemiol. 2000;35:396–401.
3. Tsai J, Hoff RA, Harpaz-Rotem I. One-year incidence and predictors of homelessness among 300,000 U.S. Veterans seen in specialty mental health care. Psychol Serv. 2017;14:203–207.
4. Rosenheck R, Fontana A. A model of homelessness among male veterans of the Vietnam War generation. Am J Psychiatry. 1994;151:421–427.
5. Copeland LA, Miller AL, Welsh DE, et al. Clinical and demographic factors associated with homelessness and incarceration among VA patients with bipolar disorder. Am J Public Health. 2009;99:871–877.
6. Metraux S, Cusack M, Byrne TH, et al. Pathways into homelessness among post-9/11-era veterans. Psychol Serv. 2017;14:229–237.
7. Nordt C, Müller B, Rössler W, et al. Predictors and course of vocational status, income, and quality of life in people with severe mental illness: a naturalistic study. Soc Sci Med. 2007;65:1420–1429.
8. Elbogen EB, Johnson SC. The intricate link between violence and mental disorder: results from the National Epidemiologic Survey on Alcohol and Related Conditions. Arch Gen Psychiatry. 2009;66:152–161.
9. Kessler RC, Bromet EJ. The epidemiology of depression across cultures. Annu Rev Public Health. 2013;34:119–138.
10. Fryers T, Melzer D, Jenkins R. Social inequalities and the common mental disorders. Social Psychiatry Psychiatr Epidemiol. 2003;38:229–237.
11. Shelton KH, Taylor PJ, Bonner A, et al. Risk factors for homelessness: evidence from a population-based study. Psychiatr Serv. 2009;60:465–472.
12. Thompson RG Jr, Wall MM, Greenstein E, et al. Substance-use disorders and poverty as prospective predictors of first-time homelessness in the United States. Am J Public Health. 2013;103(suppl 2):S282–S288.
13. Elbogen EB, Sullivan CP, Wolfe J, et al. Homelessness and money mismanagement in Iraq and Afghanistan veterans. Am J Public Health. 2013;103(suppl 2):S248–S254.
14. Grant B, Kaplan K, Stinson F. Source and accuracy statement for the wave 2 National Epidemiologic Survey on Alcohol and Related Conditions (NESARC). Rockville, MD: National Institute on Alcohol Abuse and Alcoholism. 2005.
15. Grant BF, Dawson DA. Introduction to the national epidemiologic survey on alcohol and related conditions. Alcohol Res Health. 2006;29:74–78.
16. Blanco C, Hasin DS, Wall MM, et al. Cannabis use and risk of psychiatric disorders: prospective evidence from a US national longitudinal study. JAMA Psychiatry. 2016;73:388–395.
17. Grant BF, Dawson DA, Stinson FS, et al. The Alcohol Use Disorder and Associated Disabilities Interview Schedule-IV (AUDADIS-IV): reliability of alcohol consumption, tobacco use, family history of depression and psychiatric diagnostic modules in a general population sample. Drug Alcohol Depend. 2003;71:7–16.
18. Hayes AF. Introduction to Mediation, Moderation, and Conditional Process Analysis: A Regression-based Approach. New York: Guilford Publications; 2017.
19. Tsemberis S, Gulcur L, Nakae M. Housing first, consumer choice, and harm reduction for homeless individuals with a dual diagnosis. Am J Public Health. 2004;94:651–656.
20. Sadowski LS, Kee RA, VanderWeele TJ, et al. Effect of a housing and case management program on emergency department visits and hospitalizations among chronically ill homeless adults: a randomized trial. J Am Med Assoc. 2009;301:1771–1778.
21. Larimer ME, Malone DK, Garner MD, et al. Health care and public service use and costs before and after provision of housing for chronically homeless persons with severe alcohol problems. JAMA. 2009;301:1349–1357.
22. Edens EL, Mares AS, Tsai J, et al. Does active substance use at housing entry impair outcomes in supported housing for chronically homeless persons? Psychiatr Serv. 2011;62:171–178.
23. Pearlin LI Aneshensel CS, Phelan JC. The stress process revisited. Handbook of the Sociology of Mental Health. Boston, MA: Springer; 1999:395–415.
24. Pearlin LI, Menaghan EG, Lieberman MA, et al. The stress process. J Health Soc Behav. 1981;22:337–356.
25. Krupa T, Kirsh B, Cockburn L, et al. Understanding the stigma of mental illness in employment. Work. 2009;33:413–425.
26. Hipes C, Lucas J, Phelan JC, et al. The stigma of mental illness in the labor market. Soc Sci Res. 2016;56:16–25.
27. Baldwin ML, Marcus SC. Perceived and measured stigma among workers with serious mental illness. Psychiatr Serv. 2006;57:388–392.
28. Drehmer DE, Bordieri JE. Hiring decisions for disabled workers: the hidden bias. Rehabil Psychol. 1985;30:157.
29. Gouvier WD, Sytsma-Jordan S, Mayville S. Patterns of discrimination in hiring job applicants with disabilities: the role of disability type, job complexity, and public contact. Rehabil Psychol. 2003;48:175.
30. Baron RC, Salzer MS. The career patterns of persons with serious mental illness: generating a new vision of lifetime careers for those in recovery. Psychiatr Rehabil Skills. 2000;4:136–156.
31. Rosenheck R, Lam J, Randolph F. Impact of representative payees on substance use by homeless persons with serious mental illness. Psychiatr Serv. 1997;48:800–806.
32. Elbogen EB, Hamer RM, Swanson JW, et al. A randomized clinical trial of a money management intervention for veterans with psychiatric disabilities. Psychiatr Serv. 2016;67:1142–1145.
