Characterizing Factors of Employment Status in Persons With Major Depressive Disorder

Abstract

Employment is fundamental to mental health recovery. The aim of this study is to construct a parsimonious profile indicating employment potential of people with major depressive disorder (MDD) to facilitate clinical assessment on employment participation. Data were drawn from Waves 1 (2001–2002) and 2 (2004–2005) of the National Epidemiologic Survey on Alcohol and Related Conditions. We included participants who had MDD at Wave 1 and were interviewed in both waves (N = 2,864). We conducted Classification and Regression Tree (CART) analysis to identify key characterizing factors of Wave 2 employment among 32 Wave 1 risk and protective factors. The results show that 82.1% of those employed at Wave 1 were likely to be employed at Wave 2. Among those unemployed at Wave 1, 51% of those motivated to work, measured by work-seeking behavior in the prior year, were likely to be employed at Wave 2. Among those unemployed and motivated to work, better functional mental health was associated with employment (>25.3 vs. ≤25.3). Results highlight the importance of motivation to work, shown in active work seeking, in facilitating employment despite clinical conditions.

Keywords

employment motivation to work functional mental health major depressive disorder CART

Employment is fundamental to mental health recovery (Barber, 2012; Davidson & Roe, 2007; Jacobson & Greenley, 2001). Russinova, Wewiorski, Lyass, Rogers, and Massaro (2002) introduced the term “vocational recovery” to specifically address how persons with severe mental illness may have the capacity to achieve stable and meaningful participation in the competitive workforce. More specifically, studies of supported employment programs have found that employment is associated with improved psychiatric symptoms (Bond et al., 2001) as well as increases in psychosocial functioning (Eklund, Hansson, & Ahlqvist, 2004). Other studies have found that employment contributes to self-reported increased recovery, empowerment, self-esteem, personal growth, and a sense of purpose or usefulness (Auerbach & Richardson, 2005; Fossey & Harvey, 2010; Provencher, Gregg, Mead, & Mueser, 2002; Repper & Carter, 2011; Salzer & Shear, 2002; van Dongen, 1996).

Research has identified risk and facilitating factors that may influence the likelihood for persons with mental illness to obtain and retain employment. Studies suggest that clinical factors, such as psychiatric diagnosis, self-reported psychosocial functioning, and co-occurring medical conditions or disabilities, are negatively associated with obtaining competitive employment (Elinson, Houck, Marcus, & Pincus, 2004; Razzano et al., 2005), while recent work history, educational level, and younger age have significant positive association with obtaining competitive employment (Cook et al., 2008). Macias, DeCarlo, Wang, Frey, and Barreira (2001) found that both self-reported interest in work and receipt of vocational services significantly predicted obtaining competitive employment and the length of time needed to obtain employment.

In addition, research indicates that self-confidence and motivation to work, applicable work skills, person-job fit, and social and vocational support contribute to the likelihood of obtaining and retaining employment (Dunn, Wewiorski, & Rogers, 2008; Fossey & Harvey, 2010). Specific to depressive disorders, a recent review identified disorder-related indicators (severity and length), personal factors (age and previous history of breaks from employment), and types of work to be associated with work participation and work functioning in workers with depression (Lagerveld et al., 2010).

Despite research attesting to the impact vocational training and support programs have on work outcomes (Cook et al., 2008; Fossey & Harvey, 2010; Macias, DeCarlo, Wang, Frey, & Barreira, 2001), clinical judgment and mental health consumers’ self-assessment regarding the need for employment services do not always match (Casper & Carloni, 2006; Cook, 2006; Crane-Ross, Roth, & Lauber, 2000). Studies show that most consumers express the desire to work but receive less employment guidance from providers compared to guidance in other life domains (McQuilken et al., 2003). Clinical decisions on referrals to supported employment programs also tend to be influenced by consumers’ co-occurring substance abuse (Biegel, Beimers, Stevenson, Ronis, & Boyle, 2009; Frounfelker, Wilkniss, Bond, Devitt, & Drake, 2011), despite the explicit eligibility criteria set forth by Substance Abuse and Mental Health Services Administration for supported employment programs which include substance abusers (Substance Abuse and Mental Health Services Administration, 2009). However, the relationship between substance abuse and employment of persons with mental illness is inconclusive. While some studies suggest that substance abuse has a negative impact on employment outcomes (Lehman et al., 2002; Meisler, Blankertz, Santos, & McKay, 1997; Pickett-Schenk, Cook, Jonikas, & Banghart, 2004; Razzano et al., 2005), others have found limited contribution of substance abuse to functional impairment beyond impairments due to severe mental illness (Drebing et al., 2002) nor direct association between lifetime or current substance use and work participation (Bell, Greig, Gill, Whelanhan, & Bryson, 2002; Macias et al., 2001).

Although there are numerous benefits of work, the array of risk and facilitating factors related to employment and inconclusive findings on co-occurring substance abuse issues make it challenging for service providers to assess consumers’ potential for work and to facilitate employment participation. To inform clinical judgment on providing vocational intervention for persons with major depressive disorder (MDD), we purposefully take an exploratory approach and examine a variety of risk and facilitating factors associated with employment in a general population sample. In this study, we used the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC), a nationally representative survey of all noninstitutionalized U.S. adult civilians (National Institute on Alcohol Abuse and Alcoholism [NIAAA], 2010). NESARC included a diagnostic interview that allowed for identification of MDD and co-occurring substance abuse issues as well as various factors associated with employment. We conducted Classification and Regression Tree (CART) analysis (Breiman, Friedman, Olshen, & Stone, 1984), a robust tree building statistical procedure that identifies key characterizing factors of a target outcome through data mining rather than testing predetermined hypotheses. CART has been applied to a range of health issues to aid clinical decision making, such as generating rules for nurse–patient assignments (Sundaramoorthi, Chen, Rosenberger, Kim, & Buckley-Behan, 2009), identifying characterizing factors of recurrent falling among community-dwelling seniors (Leclerc et al., 2009), and diagnosing chest pain (Cricheon, Hinde, & Marchini, 1997). With the application of CART, our study aims to identify the strongest baseline characterizing factors of the employment status 3 years later among persons with MDD.

Method

Sample

Data were drawn from Wave 1 (2001–2002) and Wave 2 (2004–2005) of NESARC. The NESARC overall response rate was 81%. Young adults aged 18–24 and Black and Hispanic individuals were oversampled. Data were weighted to reflect the demographic characteristics of the population based on the 2000 U.S. Census. The sample is comprised of NESARC participants who had MDD in the year prior to Wave 1 and who were interviewed 3 years later (N = 2,864). MDD was measured by using the NIAAA Alcohol Use Disorder and Associated Disabilities Interview Schedule–Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV; AUDADIS-IV, Grant & Dawson, 2000). This study was exempted from human subjects review by a university’s Institutional Review Board.

Outcome Variable

The outcome measures were employed or unemployed at the time of Wave 2 surveys. Employment was defined as: (a) working full time or part time; (b) employed but not at work due to temporary illness, injury, or paid vacation; or (c) employed but absent from work without pay.

Risk and Protective Factors

Thirty-two factors of employment at Wave 1 were identified in the following domains. Demographic characteristics included sex, age (aggregated into four age-groups: 24 and younger, 25–44, 45–64, and 65 and older), and race/ethnicity (collapsed into White and non-White groups). Employment status at Wave 1 was measured in the same way as the outcome variable. Education was measured by the highest grade or year of school completed. Health insurance was a dichotomized measure of having none or any of the following types of insurance: Medicare, Medicaid, military healthcare, and private health insurance. Social support was measured by the number of related adults in the household. Marital status was collapsed into a dichotomized measure of married and living with someone as if married or not. Motivation to work was operationalized as work-seeking behavior currently and/or at any time in the prior year.

Past year medical and mental health conditions were measured by the following. Past year medical condition was a dichotomized measure of having none or any of the six types of diseases: arteriosclerosis, hypertension, liver diseases, heart diseases, gastrointestinal diseases, and arthritis. Past year substance abuse and dependence were measured with the AUDADIS-IV (Grant & Dawson, 2000). The AUDADIS-IV includes detailed assessment of DSM-IV abuse and dependence criteria for alcohol and 10 drug classes. Items were collapsed into alcohol, drug, and nicotine subcategories. Past year manic/hypomanic disorders and anxiety disorders were also measured by the AUDADIS-IV. The AUDADIS-IV includes four anxiety disorders: social phobia, specific phobia, generalized anxiety, and panic disorder, all of which were collapsed into a dichotomized measure of having none or any type of anxiety disorders. Psychosis was measured by a dichotomized response to whether there was schizophrenia or psychotic illness or episode in the past year. The number of major depression episodes and attempted suicide over life time was also identified.

Lifetime use of alcohol or drug intervention was a dichotomized measure of having none or any of 13 (for alcohol issues) or 14 (for drug issues) types of treatment and services, such as 12-step self-help, professional interventions, and employee assistance program. Similarly, lifetime use of treatment for MDD, manic/hypomanic disorders, and anxiety disorders were dichotomized measures of seeking none or any help from professionals, hospitals, emergency room, and prescribed medication for their disease. Finally, a dichotomized response to whether most recent episode of major depression was in remission in past year was also included in the analysis.

Functional health status was measured with the SF-12v2 health survey, which provides norm-based summary scores of physical health and mental health in the prior 4 weeks. There were 8 subscales, including physical functioning, bodily pain, general health perceptions, energy/vitality, social functioning, mental health, role functioning due to emotional problems, and role functioning due to physical problems (Ware, Kosinski, Turner-Bowker, & Gandek, 2002). In norm-based scoring, each scale had an average score of 50 and the standard deviation of 10. An individual respondent’s scale score below 45 indicated that his or her health status was below the average range ( User’s Manual, 2007).

Analysis

We used IBM^® SPSS^® Statistics 20 software for analysis. All estimates were weighted by the sampling weights to represent the U.S. population. Univariate analysis was conducted on the outcome variable and risk and protective factors. The 32 factors of employment were compared between Wave 2 employed and unemployed groups using chi-square tests or t-tests.

We then conducted CART analysis (Breiman et al., 1984) to identify key characterizing factors of employment. CART is a procedure to generate a binary tree-structured solution containing a subset of characterizing factors in accordance to their importance in indicating a given outcome. The procedure starts with the entire sample. All of the factors are investigated to identify a factor and its value that splits the sample into binary subgroups with best external heterogeneity and internal homogeneity. For each resulting subgroup, this procedure repeats until no further splitting is possible according to prespecified tree growing rules. We preset tree depth to 5 levels, subgroups to contain no fewer than 50 cases, and the Gini index to .0001 for determining factors’ importance in reducing impurity within subgroups. The tree was then pruned according to the risk <10 (standard error) to identify a simpler tree, balancing the additional tree accuracy and tree complexity that a split would generate.

Finally, we performed additional logistic regression models to verify a suppression effect that we identified in the results from the CART analysis. We conducted this analysis by following the procedures set forth by MacKinnon, Krull, and Lockwood (2000, p. 176).

Results

Table 1 shows the characteristics of the weighted sample. To highlight, at Wave 1, 59.9% were employed, and 19.6% were looking for work. The average scores of the functional health scales are between 41.2 and 48.1 with the average score of 42.5 (SD = 13.6) for role functioning due to emotional problems. At Wave 2, 59.6% were employed.

Table 1.

Sample Characteristics.

Characteristics	Percentage or Mean (SD)
Age (years)	40.1 (15.3)
Sex (female)	66.6
Race
White	74.0
Black	9.5
Hispanic	10.0
Wave 1 characteristics
Had employment	59.9
Education (high school and above)	82.6
Had health insurance	78.4
Had social support	65.3
Married/cohabitated	48.8
Seeking work	19.6
Physical functioning	48.1 (12.0)
Bodily pain	45.1 (14.1)
General health perceptions	45.0 (14.2)
Energy/vitality	47.0 (11.3)
Social functioning	43.4 (13.7)
Mental health	41.2 (12.0)
Role functioning due to physical problems	46.8 (12.5)
Role functioning due to emotional problems	42.5 (13.6)
Health/mental health condition in the year prior to Wave 1
Medical condition	51.9
Alcohol abuse/dependence	15.8
Drug abuse/dependence	6.4
Nicotine dependence	28.6
Manic/hypomanic disorders	17.6
Anxiety disorders	40.5
Psychosis	2.2
Major depressive disorder in remission	63.9
Life time mental health condition
Had two or more major depression episodes	66.1
Had five or more major depression episodes	28.7
Had attempted suicide	13.6
Received alcohol intervention	10.0
Received drug intervention	6.0
Received major depressive disorder intervention	65.3
Received intervention for manic/hypomanic disorders	11.6
Received anxiety intervention	6.4
Employed at Wave 2	59.6

Note. Weighted sample N = 16,194.

Differences in the 32 factors between the employed and unemployed groups at Wave 2, using chi-square tests or t-tests, were all significant at p < .01 level (two sided/tailed), except the following: health insurance (p = .186), marital status (p = .442), work seeking (p = .170), past year drug abuse/dependence (p = .622), and past year manic/hypomanic disorders (p = .514).

The CART analysis showed that employment at Wave 1, work seeking as the measure of motivation to work, and the role emotional subscale of functional health status as measured by the SF-12v2, respectively, were the most important factors associated with future employment (Figure 1). A large majority of cases who were employed at Wave 1 (82.1%) were likely to be employed at Wave 2. Among those unemployed at Wave 1, more than half of those who were actively seeking work (51.1%) were likely to be employed at Wave 2. Among those who were unemployed and not seeking work at Wave 1, only 17.1% were likely to be employed at Wave 2. Among those who were unemployed and actively seeking work, better functional health status in the role–emotional subscale (i.e., disability in role functioning resulting from emotional problems such as depression or anxiety) was associated with employment (>25.3 vs. ≤ 25.3). These three characterizing factors accurately classified 80.4% of all cases according to their Wave 2 employment status, including 90.7% of those who were employed at Wave 2.

Figure 1.

Classification and Regression Tree Model of Wave 2 employment in persons with major depressive disorder. Note. n (%) represents the weighted number of subjects employed at Wave 2 and its percentage in its respective category.

Although not a significantly associated factor in bivariate analysis, work seeking stood out as one of the strongest characterizing factors in CART analysis, known as the suppression effect (MacKinnon, Krull, & Lockwood, 2000). We examined the suppression effect by following the approaches proposed by MacKinnon and colleagues (2000, p. 176). As shown in Table 2, we found that in a simple logistic regression model where Wave 2 employment was regressed on Wave 1 work seeking, the coefficient of Wave 1 work seeking was −.055 (Model 1). In a multiple logistic regression model where Wave 2 employment was regressed on both Wave 1 work seeking and Wave 1 employment, the coefficient of Wave 1 work seeking became .467, while the coefficient of Wave 1 employment was 2.646 (Model 2). This result shows that after controlling for Wave 1 employment, Wave 1 work-seeking coefficient not only increased its magnitude in its absolute value but also showed its effect in the opposite direction (from negative to positive). That is, the effect of Wave 1 work seeking on Wave 2 employment was suppressed by Wave 1 employment.

Table 2.

Suppression of W1 Employment on the Effect of W1 Work Seeking on W2 Employment.

Suppression of Wave 1 Work Seeking’s Effect on Wave 2 Employment by Wave 1 Employment	β₁	β₂
Model 1
W2 employment = β₀ + β₁W1 work seeking + ε	−.055	—
Model 2
W2 employment = β₀ + β₁W1 work seeking + β₂W1 employment + ε	.467	2.646
Model 3
W1 employment = β₀ + β₁W1 work seeking + ε	−.649	—

This result was verified by an alternative method also proposed by MacKinnon and colleagues (2000, p. 176). In a simple logistic regression model where Wave 1 employment was regressed on Wave 1 work seeking to identify the relation between the two, the coefficient of Wave 1 work seeking was −.649 (Model 3). The product of the coefficient of Wave 1 work seeking in Model 3 and the coefficient of Wave 1 employment in Model 2 (−1.717) has the opposite sign from the coefficient of Wave 1 work seeking in Model 2 (.467), showing the suppressor effect of Wave 1 employment. Additionally, the binary correlation between Wave 1 employment and Wave 1 work seeking was −.129 (p < .001), indicating a significant negative correlation.

Discussion

Wave 1 work history was found to be a significant characterizing factor of employment among persons with MDD. This finding is consistent with results of previous studies on employment among persons with a variety of mental health conditions (Anthony & Jansen, 1984; Burke-Miller et al., 2006; Campbell, Bond, Drake, McHugo, & Xie, 2010; Fossey & Harvey, 2010). Stability and duration of prior work are associated with future employment outcomes (Goldberg et al., 2001). In addition, Fossey and Harvey (2010) conclude from a review of qualitative studies that psychological effects of work history such as self-esteem and self-efficacy may influence future employment outcomes.

Another important finding is the association between work-seeking behavior and Wave 2 employment in persons with MDD. Our analysis on the suppression effect indicated that Wave 1 employment mediated/suppressed the effect of Wave 1 work seeking on Wave 2 employment in bivariate analysis, likely, because once people had work, they were not seeking work. This is supported by the binary correlation of –.129 (p < .001) between Wave 1 employment and Wave 1 work seeking. Because having work is most likely the result of seeking work, this finding shows that both seeking work and having work need to be considered to account for the potential of people who are motivated and in the process of seeking employment.

This finding has implications for future research on motivation to work. Prior research indicates that motivation to work is critical to obtaining and retaining employment (Auerbach & Richardson, 2005; Macias et al., 2001). Motivation to work might be expressed by desire to work and active work seeking. Both desire to work and active work seeking are consistent with a recovery-based approach to mental health services, as these factors underscore notions of consumer preference and empowerment, that is, gaining/regaining a sense of control and volition over life domains (Davidson & Roe, 2007; Jacobson & Greenly, 2001). However, emerging evidence suggests that active work seeking in persons with serious mental illness is associated with better employment outcomes, compared to a self-reported desire for employment (Alverson, Carpenter, & Drake, 2006). In their study, Alverson and colleagues found that all participants expressed a desire to find employment but only 15 of the 25 (60%) were “active work-seekers” (Alverson et al., 2006). They suggest that work-seeking behavior might be a better measure of motivation to work than expressed desire for work. Findings of a study by Macias and colleagues (2001) also suggest behavioral exposure to work, such as participation in supported employment services and work, is a stronger indicator than initial stated interest in work, in retaining employment. Our study adopted work seeking as behavioral manifestation of work motivation and again showed its significance in Wave 2 employment status.

The finding on the importance of functional role-emotion health status on employment outcomes echoes previous research. Perceived better health and functional (social, cognitive, and physical) characteristics are strongly associated with working status (Elinson et al., 2004). MDD is associated with pervasive and chronic psychosocial disability in day-to-day functioning. However, the severity of depressive symptoms varies over a long-term course, and when it decreases, psychosocial disability decreases as well (Judd et al., 2000). Our finding of the SF-12v2 role-emotion health associated with employment similarly suggests that functional mental health, rather than the clinical diagnosis of MDD, may more accurately reflect the potential of employment participation during the course of MDD, even when the score was much lower than the average of the scale (25.3 and 50, respectively).

The application of the CART procedure allows associated factors to emerge according to their relative importance to the target outcome (i.e., employment status). The top three factors—Wave 1 work history, work seeking, and functional role-emotion health—which emerged from the epidemiological data set are all strength-based factors. Clinical conditions, including substance use disorders, were also included in the analysis but none emerged as the most important factors in characterizing people with employment at Wave 2 in our analysis. This is consistent with an earlier review in which little or no correlation was found between symptomatology and functional skills (Anthony & Jansen, 1984) and with mental health consumers’ personal accounts of factors influencing their entry/reentry into the workforce (Dunn et al., 2008; Fossey & Harvey, 2010). Additionally, the supported employment literature provides little empirical support for excluding consumers from vocational services based on psychiatric history, diagnosis, or clinical work readiness (Bond, 2004). Our findings are consistent with research showing that substance abuse has limited (Drebing et al., 2002) to no direct impact on work outcomes among persons with mental health conditions (Bell et al., 2002; Macias et al., 2001).

The results of this study have multilevel implications. At the practice level, our findings suggest the absence of clinical conditions not be the sole determinant of consumers’ qualification for vocational services. Instead, service providers need to assess consumers’ strengths for vocational development (Anthony & Jansen, 1984). This aligns with defining rehabilitation readiness as one’s willingness to engage in the process of rehabilitation toward the accomplishment of a goal (Farkas, Sullivan Soydan, & Gagne, 2000). Service providers thus focus on assessing consumers’ perceived need for change; commitment to change; self-awareness of skills, interests, values, and preferences related to a desire role; awareness of options and demands; and willingness to establish therapeutic alliance for providing vocational services (Roberts & Pratt, 2007).

Moreover, for programming and service provision, we recommend vocational services made available to any consumer with a stated interest in employment or engagement in active job seeking, regardless of their type or severity of clinical condition. As previously discussed, work history has a significant influence on future employment. We thus argue for the importance of appropriate vocational services that provide exposure to work opportunities to consumers with limited work histories. Particularly, supportive, consumer-centered approaches can address issues that undermine both work motivation and actual work performance (Kirsh, Cockburn, & Gewurtz, 2005; Krupa, 2011). An example of such an approach is the standardized individual placement and support (IPS) model which considers consumers’ desire to work, includes rapid job placement, ongoing vocational assessment and supports, integrated vocational and mental health services, and targets competitive employment as the desired outcome (Becker & Drake, 1994; Bond, Drake, & Becker, 2008). The extensive evidence supporting IPS demonstrates that the holistic, consumer-centered programming leads to higher rates of competitive employment and longer job retention than other vocational rehabilitation models. These outcomes have been demonstrated across clinical diagnoses, including depressive disorders and schizophrenia (Bond et al., 2001, 2008).

On a policy level, our findings suggest that the scope of vocational services should be broadened to consumers who, based on their clinical conditions, may not have been previously assessed as strong candidates for vocational recovery. Many mental health consumers express interest in work, yet national surveys suggest only about one half of those with a serious mental illness (including MDD), are working at any given time (Elinson et al., 2004; Mechanic, Bilder, & McAlpine, 2002). Cook (2006) notes that the small scope of effort is a fundamental policy problem in vocational rehabilitation; mental health programming only involves limited vocational training and support. Models such as IPS could be integrated into mental health services, giving more consumers the option to work. Additionally, policy-based barriers to employment—such as stringent income eligibility requirements for Social Security Disability Insurance and Medicaid —should be overcome in both vocational rehabilitation and integrated models of care (O’Day & Killeen, 2002). Cook (2006) suggests that offering benefits planning and financial education to address consumers’ concerns about the effects of paid employment on cash and medical benefits does not affect their motivation to work.

For future studies, we support greater focus on the strengths-based research approach, which seeks to identify and enhance the inherent strengths of mental health consumers as well as to mobilize assets in their larger social environment (Maton et al., 2004). For persons with MDD, this involves extending the scope of investigations from a focus on pathology to a focus on personal and community characteristics that may foster greater vocational participation. Additionally, exploratory studies that allow understudied factors of vocational participation to emerge hold promise in expanding research, practices, and programming for improving vocational recovery of persons with MDD and other mental health conditions.

Our study has several limitations. First, employment in the NESARC was measured at only two time points. Future longitudinal studies that measure employment at multiple time points are needed to better capture employment trajectories of persons with MDD. Second, work-seeking behavior was measured by a single dichotomized item. Future research using more sophisticated measurement is needed to confirm our findings. Third, the timeframes for MDD and functional health status measurements did not overlap (past year and past month, respectively), thus their association was uncertain. Fourth, the survey relied on self-report. However, studies have shown good consistency between respondents’ and collateral reports of substance use (Connors & Maisto, 2003; Hjorthøj, Hjorthøj, & Nordendoft, 2012). Finally, unlike a conventional regression model, CART does not provide confidence intervals and thus precludes assessment of the reliability of an estimate.

Despite these limitations, the study has the strength of a prospective design and a large and nationally representative sample which supports the generalizability of the results. The use of CART analysis resulted in a parsimonious profile of factors indicating the employment potential of persons with MDD. The relative importance of the characterizing factors highlighted the significance of motivation to work manifested as active work-seeking behavior in facilitating employment. Remission of clinical conditions need not be a pre-condition to support vocational development of consumers with MDD.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflict of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Alverson

Carpenter

Drake

R. E.

(2006). An ethnographic study of job seeking among people with severe mental illness. Psychiatric Rehabilitation Journal, 30, 15–22. doi:https://dx-doi-org.web.bisu.edu.cn/10.2975/30.2006.15.22

Anthony

W. A.

Jansen

M. A.

(1984). Predicting the vocational capacity of the chronically mentally ill: Research and implications. American Psychologist, 39, 537–544. doi:https://dx-doi-org.web.bisu.edu.cn/10.1037//0003-066x.39.5.537

Auerbach

E. S.

Richardson

(2005). The long-term work experiences of persons with severe and persistent mental illness. Psychiatric Rehabilitation Journal, 28, 267–273. doi:https://dx-doi-org.web.bisu.edu.cn/10.2975/28.2005.267.273

Barber

M. E.

(2012). Recovery as the new medical model for psychiatry. Psychiatric Services, 63, 277–279. doi:https://dx-doi-org.web.bisu.edu.cn/10.1176/appi.ps.201100248

Becker

D. R.

Drake

R. E.

(1994). Individual placement and support: A community mental health center approach to vocational rehabilitation. Community Mental Health Journal, 30, 193–206. doi:https://dx-doi-org.web.bisu.edu.cn/10.1007/bf02188630

Bell

Greig

Gill

Whelanhan

Bryson

(2002). Work rehabilitation and patterns of substance use among persons with schizophrenia. Psychiatric Services, 53, 63–69. doi:https://dx-doi-org.web.bisu.edu.cn/10.1176/appi.ps.53.1.63

Biegel

D. E.

Beimers

Stevenson

L. D.

Ronis

R. J.

Boyle

(2009). Predictors of referral to supported employment among consumers with co-occurring mental and substance use disorders. Community Mental Health Journal, 45, 427–438. doi:https://dx-doi-org.web.bisu.edu.cn/10.1007/s10597-009-9242-3

Bond

G. R.

(2004). Supported employment: Evidence for an evidence-based practice. Psychiatric Rehabilitation Journal, 27, 345–359. doi:https://dx-doi-org.web.bisu.edu.cn/10.2975/27.2004.345.359

Bond

G. R.

Drake

R. E.

Becker

D. R.

(2008). An update of randomized control trials of evidence-based supported employment. Psychiatric Rehabilitation Journal, 31, 280–290. doi:https://dx-doi-org.web.bisu.edu.cn/10.2975/31.4.2008.280.290

10.

Bond

G. R.

Resnick

S. G.

Drake

R. E.

Xie

McHugo

G. J.

Bebout

R. R.

(2001). Does competitive employment improve nonvocational outcomes for people with severe mental illness. Journal of Consulting and Clinical Psychology, 69, 489–501. doi:https://dx-doi-org.web.bisu.edu.cn/10.1037//0022-006x.69.3.489

11.

Breiman

Friedman

J. H.

Olshen

R. A.

Stone

C. J.

(1984). Classification and regression trees. New York, NY: Chapman & Hall/CRC.

12.

Burke-Miller

J. K.

Cook

J. A.

Grey

D. G.

Razzano

L. A.

Blyler

C. R.

Leff

H. S.

… Carey

M. A.

(2006). Demographic characteristics and employment among people with severe mental illness. Community Mental Health Journal, 42, 143–159. doi:https://dx-doi-org.web.bisu.edu.cn/10.1007/s10597-005-9017-4

13.

Campbell

Bond

G. R.

Drake

R. E.

McHugo

G. J.

Xie

(2010). Client predictors of employment outcomes in high-fidelity supported employment: A regression analysis. The Journal of Nervous and Mental Disease, 198, 556–563. doi:https://dx-doi-org.web.bisu.edu.cn/10.1097/nmd.0b013e3181ea1e53

14.

Casper

E. S.

Carloni

(2006). Increasing utilization of supported employment services with the need for change scale. Psychiatric Services, 57, 1430–1434. doi:https://dx-doi-org.web.bisu.edu.cn/10.1176/appi.ps.57.10.1430

15.

Cricheon

N. J.

Hinde

Marchini

(1997). Models for diagnosing chest pain: Is CART helpful? Statistics in Medicine, 16, 717–727. doi:https://dx-doi-org.web.bisu.edu.cn/10.1002/(sici)1097-0258(19970415)16:7<717:: aid-sim504>3.0.co;2-e

16.

Connors

G. J.

Maisto

S. A.

(2003). Drinking reports from collateral individuals. Addiction, 98, 21–29. doi:https://dx-doi-org.web.bisu.edu.cn/10.1046/j.1359-6357.2003.00585.x

17.

Cook

J. A.

(2006). Employment barriers for persons with psychiatric disabilities: Update of a report for the President’s Commission. Psychiatric Services, 57, 1391–1405. doi:https://dx-doi-org.web.bisu.edu.cn/10.1176/appi.ps.57.10.1391

18.

Cook

J. A.

Blyer

C. R.

Leff

H. S.

McFarlane

W. R.

Goldberg

R. W.

Gold

P. B.

… Razzano

L. A.

(2008). The employment intervention demonstration program: Major findings and policy implications. Psychiatric Rehabilitation Journal, 31, 291–295. doi:https://dx-doi-org.web.bisu.edu.cn/10.2975/31.4.2008.291.295

19.

Crane-Ross

Roth

Lauber

B. G.

(2000). Consumers’ and case managers’ perceptions of mental health and community support service needs. Community Mental Health Journal, 36, 161–178. doi:https://dx-doi-org.web.bisu.edu.cn/10.1023/a:1001842327056

20.

Davidson

Roe

(2007). Recovery from versus recovery in serious mental illness: One strategy for lessening confusion plaguing recovery. Journal of Mental Health, 16, 459–470. doi:https://dx-doi-org.web.bisu.edu.cn/10.1080/09638230701482394

21.

Drebing

C. E.

Fleitas

Moore

Krebs

Ormer

A. V.

Penk

… Rosenheck

(2002). Patterns in work functioning and vocational rehabilitation associated with coexisting psychiatric and substance use disorders. Rehabilitation Counseling Bulletin, 46, 5–13. doi:https://dx-doi-org.web.bisu.edu.cn/10.1177/00343552020460010501

22.

Dunn

E. C.

Wewiorski

N. J.

Rogers

E. S.

(2008). The meaning and importance of employment to people in recovery from serious mental illness: Results of a qualitative study. Psychiatric Rehabilitation Journal, 32, 59–62. doi:https://dx-doi-org.web.bisu.edu.cn/10.2975/32.1.2008.59.62

23.

Eklund

Hansson

Ahlqvist

(2004). The importance of work as compared to other forms of daily occupations for wellbeing and functioning among persons with long-term mental illness. Community Mental Health Journal, 40, 465–477. doi:https://dx-doi-org.web.bisu.edu.cn/10.1023/b:comh.0000040659.19844.c2

24.

Elinson

Houck

Marcus

S. C.

Pincus

H. A.

(2004). Depression and the ability to work. Psychiatric Services, 55, 29–34. doi:https://dx-doi-org.web.bisu.edu.cn/10.1176/appi.ps.55.1.29

25.

Farkas

Sullivan Soydan

Gagne

(2000). Introduction to rehabilitation readiness. Boston, MA: Boston University, Center for Psychiatric Rehabilitation.

26.

Fossey

E. M.

Harvey

C. A.

(2010). Finding and sustaining employment: A qualitative meta-synthesis of mental health consumer views. Canadian Journal of Occupational Therapy, 77, 303–314. doi:https://dx-doi-org.web.bisu.edu.cn/10.2182/cjot.2010.77.5.6

27.

Frounfelker

R. L.

Wilkniss

S. M.

Bond

G. R.

Devitt

T. S.

Drake

R. E.

(2011). Enrollment in supported employment services for clients with a co-occurring disorder. Psychiatric Services, 62, 545–547. doi:https://dx-doi-org.web.bisu.edu.cn/10.1176/ps.62.5.pss6205_0545

28.

Goldberg

R. W.

Lucksted

McNary

Gold

J. M.

Dixon

Lehman

(2001). Correlates of long-term unemployment among inner-city adults with serious and persistent mental illness. Psychiatric Services, 52, 101–103. doi:https://dx-doi-org.web.bisu.edu.cn/10.1176/appi.ps.52.1.101

29.

Grant

B. F.

Dawson

D. A.

(2000). The alcohol use disorder and associated disabilities interview schedule–DSM-IV Version. Rockville, MD: National Institute on Alcohol Abuse and Alcoholism.

30.

Hjorthøj

C. R.

Hjorthøj

A. R.

Nordendoft

(2012). Validity of timeline follow-back for self-reported use of cannabis and other illicit substances - Systematic review and meta-analysis. Addictive Behaviors, 37, 225–233. doi:https://dx-doi-org.web.bisu.edu.cn/10.1016/j.addbeh.2011.11.025

31.

Jacobson

Greenley

(2001). What is recovery? A conceptual model and explication. Psychiatric Services, 52, 482–485. doi:https://dx-doi-org.web.bisu.edu.cn/10.1176/appi.ps.52.4.482

32.

Judd

L. L.

Akiskal

H. S.

Zeller

P. J.

Paulus

Leon

A. C.

Maser

J. D.

… Keller

M. B.

(2000). Psychosocial disability during the long-term course of unipolar major depressive disorder. Achieve of General Psychiatry, 57, 375–380. doi:https://dx-doi-org.web.bisu.edu.cn/10.1001/archpsyc.57.4.375

33.

Kirsh

Cockburn

Gewurtz

(2005). Best practice in occupational therapy: Program characteristics that influence vocational outcomes for people with serious mental illness. Canadian Journal of Occupational Therapy, 72, 265–280. doi:https://dx-doi-org.web.bisu.edu.cn/10.1177/000841740507200503

34.

Krupa

(2011). Approaches to improving employment outcomes for people with serious mental illness. In Schultz

I. J.

Rogers

E. S.

(Eds.), The handbook of job accommodation and mental health (pp. 219–231). New York, NY: Springer.

35.

Lagerveld

S. E.

Bultmann

Franche

R. L.

van Dijk

F. J.

Vlasveld

M. C.

van der Feltz-Cornelis

C. M.

… Nieuwenhuijsen

(2010). Factors associated with work participation and work functioning in depressed workers: A systematic review. Journal of Occupational Rehabilitation, 20, 275–292. doi:https://dx-doi-org.web.bisu.edu.cn/10.1007/s10926-009-9224-x

36.

Leclerc

B. S.

Begin

Cadieux

Goulet

Allaire

J. F.

Meloche

… Kergoat

. (2009). A classification and regression tree for predicting recurrent falling among community-dwelling seniors using home-care services. Canadian Journal of Public Health, 100, 263–267. doi:https://dx-doi-org.web.bisu.edu.cn/10.1016/j.respe.2009.10.008

37.

Lehman

A. F.

Goldberg

Dixon

L. B.

McNary

Postrado

Hackman

McDonnell

(2002). Improving employment outcomes for persons with severe mental illness. Archives of General Psychiatry, 59, 165–172. doi:https://dx-doi-org.web.bisu.edu.cn/10.1001/archpsyc.59.2.165

38.

Macias

DeCarlo

L. T.

Wang

Frey

Barreira

(2001). Work interest as a predictor as competitive employment: Policy implications for psychiatric rehabilitation. Administration and Policy in Mental Health, 28, 279–297. doi:https://dx-doi-org.web.bisu.edu.cn/10.1023/a:1011185513720

39.

MacKinnon

D. P.

Krull

J. L.

Lockwood

C. M.

(2000). Equivalence of the mediation, confounding and suppression effect. Prevention Science, 1, 173–181. doi:https://dx-doi-org.web.bisu.edu.cn/10.1023/a:1026595011371

40.

McQuilken

Zahniser

J. H.

Novak

Starks

R. D.

Olmos

Bond

G. R.

(2003). The work project survey: Consumer perspectives on work. Journal of Vocational Rehabilitation, 18, 59–68.

41.

Maton

K. I.

Dodgen

D. W.

Leadbeater

B. J.

Sandler

I. N.

Schellenbach

C. J.

Solarz

A. L.

(2004). Strengths-based research and policy: An introduction. In Manton

K. L.

Schellenbach

C. J.

Leadbeater

B. J.

Solarz

A. L.

(Eds.), Investing in children, youth, families, and communities: Strengths-based research and policy (pp. 3–12). Washington, DC: American Psychological Association.

42.

Mechanic

Bilder

McAlpine

D. D.

(2002). Employing persons with serious mental illness. Health Affairs, 21, 242–253. doi:https://dx-doi-org.web.bisu.edu.cn/10.1377/hlthaff.21.5.242

43.

Meisler

Blankertz

Santos

A. B.

McKay

(1997). Impact of assertive community treatment on homeless persons with co-occurring severe psychiatric and substance use disorders. Community Mental Health Journal, 33, 113–122. doi:https://dx-doi-org.web.bisu.edu.cn/10.1023/a:1022419316396

44.

National Institute on Alcohol Abuse and Alcoholism. (2010). Alcohol use and alcohol use disorders in the United States, a 3-year follow-up: Main findings from the 2004-2005 Wave 2 National Epidemiologic Survey on Alcohol and Related Conditions (NESARC). In U.S. alcohol epidemiologic data reference manual (Vol. 8), (NIH Publications No. 10-7677). Bethesda, MA: NIAAA.

45.

O’Day

Killeen

(2002). Does U.S. federal policy support employment and recovery for people with psychiatric disabilities? Behavioral Sciences and the Law, 20, 559–583. doi:https://dx-doi-org.web.bisu.edu.cn/10.1002/bsl.514

46.

Pickett-Schenk

S. A.

Cook

J. A.

Jonikas

J. A.

Banghart

(2004). Business as usual: Work experiences of homeless persons with mental illness. In Fisher

W. H.

(Ed.), Research in community mental health (Vol. 13, Employment for Persons With Severe Mental Illness, pp. 63–69). Oxford, England: Elsevier.

47.

Provencher

H. L.

Gregg

Mead

Mueser

K. T.

(2002). The role of work in the recovery of persons with psychiatric disabilities. Psychiatric Rehabilitation Journal, 26, 132–144. doi:https://dx-doi-org.web.bisu.edu.cn/10.2975/26.2002.132.144

48.

Razzano

L. A.

Cook

J. A.

Burke-Miller

J. K.

Mueser

K. T.

Pickett-Schenk

S. A.

Grey

D. D.

… Ann Carey

(2005). Clinical factors associated with employment among people with severe mental illness: Findings from the employment intervention demonstration program. Journal of Nervous and Mental Disease, 193, 705–713. doi:https://dx-doi-org.web.bisu.edu.cn/10.1097/01.nmd.0000185939.11282.3e

49.

Repper

Carter

(2011). A review of the literature on peer support in mental health services. Journal of Mental Health, 20, 392–411. doi:https://dx-doi-org.web.bisu.edu.cn/10.3109/09638237.2011.583947

50.

Roberts

M. M.

Pratt

C. W.

(2007). Putative evidence of employment readiness. Psychiatric Rehabilitation Journal, 30, 175–181. doi:https://dx-doi-org.web.bisu.edu.cn/10.2975/30.3.2007.175.181

51.

Russinova

Wewiorski

Lyass

Rogers

E. S.

Massaro

J. M.

(2002). Correlates of vocational recovery for persons with schizophrenia. International Review of Psychiatry, 14, 303–331. doi:https://dx-doi-org.web.bisu.edu.cn/10.1080/0954026021000016941

52.

Salzer

M. S.

Shear

S. L.

(2002). Identifying consumer-provider benefits in evaluations of consumer-derived services. Psychiatric Rehabilitation Journal, 25, 281–288. doi:https://dx-doi-org.web.bisu.edu.cn/10.1037/h0095014

53.

Substance Abuse and Mental Health Services Administration. (2009). Supported employment: Building your program (DHHS Pub. No. SMA-08-4364). Rockville, MD: Center for Mental Health Services, Substance Abuse and Mental Health Services Administration, U.S. Department of Health and Human Services.

54.

Sundaramoorthi

Chen

V. C. P.

Rosenberger

J. M.

Kim

S. B.

Buckley-Behan

D. F.

(2009). A data-integrated simulation model to evaluate nurse–patient assignments. Health Care Management Science, 12, 252–268. doi:https://dx-doi-org.web.bisu.edu.cn/10.1007/s10729-008-9090-7

55.

User’s Manual for the SF-36v2 Health Survey (2nd ed.). (2007). Lincoln, RI: QualityMetric.

56.

van Dongen

C. J.

(1996). Quality of life and self-esteem in working and non-working persons with mental illness. Community Mental Health Journal, 32, 535–548. doi:https://dx-doi-org.web.bisu.edu.cn/10.1007/bf02251064

57.

Ware

J. E.

Kosinski

Turner-Bowker

D. M.

Gandek

(2002). How to score Version 2 of the SF-12 Health Survey. Lincoln, RI: QualityMetrics.