Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2020 Oct 1.
Published in final edited form as: J Res Pers. 2019 Aug 29;82:103863. doi: 10.1016/j.jrp.2019.103863

Unemployment and the Relationship between Borderline Personality Pathology and Health

Patrick J Cruitt 1, Thomas F Oltmanns 1
PMCID: PMC7448725  NIHMSID: NIHMS1539008  PMID: 32863466

Abstract

Evidence suggests that employment may buffer against the negative health outcomes associated with borderline personality disorder (BPD). The purpose of the current analyses was to examine unemployment and the BPD-health relationship prospectively. Participants were 1,536 older adults in a longitudinal study of health and aging, with repeated measures of physical health, depressive symptoms, and life satisfaction. We measured BPD features using multiple sources at baseline, and used principal components analysis to obtain latent scores. Multilevel models indicated that unemployment experiences did not moderate the prospective relationship between BPD features and physical health or life satisfaction, but did strengthen the positive relationship between BPD features and depressive symptoms. These findings provide insight into mechanisms of recovery for individuals with BPD.

Keywords: borderline personality disorder, unemployment, physical health, depressive symptoms, life satisfaction

Personality does not exist in a vacuum. Work, or its absence, forms an important context that has the potential to interact with an individual’s level of personality functioning, for better or worse. Although work is a broad category of activity of which employment is only one salient example, research that compares employment to unemployment offers insight into the psychosocial benefits of working (Jahoda, 1982). According to both the latent deprivation model (Jahoda, 1982) and the vitamin model (Warr, 1987), environmental features of employment fulfill important psychosocial functions beyond financial gain. Both models point to employment (vis-à-vis unemployment) as providing social contact outside the family and a sense of valued status/identity. These models theorize that the loss of these latent functions or ‘vitamins’ partially explains the established relationships between unemployment and both psychological and physical health, including mortality (McKee-Ryan, Song, Wanberg, & Kinicki, 2005; Paul & Moser, 2009; Roelfs, Shor, Davidson, & Schwartz, 2011). Recent research in general lends partial support to these models (Paul & Batnic, 2010; Selenko, Batinic, & Paul, 2011; Sousa-Ribeiro, Sverke, & Coimbra, 2014), and aspects of the latent deprivation model have been validated cross-culturally (Gnambs, Stiglbauer, & Selenko, 2015).

Both models point to presumably universal human needs, but few studies have examined the role of personality in relation to unemployment and health outcomes in this context (Creed & Evans, 2002; Creed, Muller, & Machin, 2001). Nevertheless, loss of the latent functions served by employment may be particularly harmful when an individual suffers from personality pathology that interferes with interpersonal and identity functioning, such as borderline personality disorder (BPD). As defined in the Diagnostic and Statistical Manual for Mental Disorders, Fifth Edition (DSM-5), BPD is characterized by features such as impulsivity, emotional lability, unstable interpersonal relationships, identity disturbances, and self-harm behaviors (American Psychiatric Association, 2013). The population prevalence of BPD appears to be between 0.7 and 2.7 percent (Coid, Yang, Tyrer, Roberts, & Ullrich, 2006; Lenzenweger, Lane, Loranger, & Kessler, 2007; Trull, Jahng, Tomko, Wood, & Sher, 2010), and this rate is associated with substantial societal costs due to healthcare utilization and lost productivity (Soeteman, Roijen, Verheul, & Busschbach, 2008). However, these rates likely underestimate the total impact of BPD features, as even subthreshold levels are associated with impaired functioning (Ellison, Rosenstein, Chelminski, Dalrymple, & Zimmerman, 2016). Given the public health significance of BPD features, it is important to explore the way in which they interact with the employment context. The psychosocial functions of employment appear to address central problems associated with BPD, suggesting that the context of employment and its absence may have particular effects for individuals high on BPD features beyond the universal.

There are a number of reasons to suspect that individuals with high levels of BPD features may be particularly affected by the context of employment and its loss. Given the identity disturbance often seen in individuals with BPD features, they may stand to benefit the most from employment’s ability to facilitate identity formation through collective purpose and opportunity for skilled activity. In addition, the interpersonal difficulties they experience may be lessened by regular social contact that is less emotionally charged relative to their frequently volatile relationships with family and close others. Clinical research supports this possibility, suggesting that employment is one predictor of recovery in individuals with psychopathology more generally (Strauss & Davidson, 1997; Strauss, Harding, Silverman, Eichler, & Lieberman, 1988) and with BPD in particular (Paris, 2003). Although remission of BPD symptoms is an important outcome in its own right, extensive research indicates that BPD features are associated with poor physical health (Dixon-Gordon, Whalen, Layden, & Chapman, 2015) as well as other indicators of poor mental health, such as depressive symptoms (Galione & Oltmanns, 2013; Lenzenweger et al., 2007; Skodol et al., 1999; Zanarini et al., 1998). If the latent functions of employment are particularly beneficial for individuals with BPD features, one might expect employment status to moderate the relationship between BPD features and associated health outcomes.

Much of the research examining employment in individuals with high levels of BPD pathology focuses on the extent to which occupational or vocational functioning is impaired in these individuals. Follow-up studies of BPD patients in general show poor employment outcomes (Sansone & Sansone, 2012). As mentioned above, lost productivity in the workplace represents a meaningful part of the overall societal cost due to BPD (Soeteman et al., 2008), as does the fact that BPD features are associated with risk for receiving disability benefits (Amundsen Østby et al., 2014; Knudsen et al., 2012; Zanarini, Jacoby, Frankenburg, Reich, & Fitzmaurice, 2009). Trait levels of BPD pathology are associated with lower educational attainment and increased rates of conflict at work, dismissal or demotion, and unemployment (Hengartner, Müller, Rodgers, Rössler, & Ajdacic-gross, 2014). Furthermore, the employment histories of individuals with BPD features show greater number of jobs in adulthood, lower time employed, higher rates of “under the table” jobs, and higher rates of being fired from a job (Sansone, Leung, & Wiederman, 2012). Although BPD is highly comorbid with other psychiatric disorders such as major depression (see above), BPD appears to have effects on occupational functioning above and beyond those of mood disorders. Patients with comorbid depression and BPD are more likely than patients with depression alone to be unemployed, exhibiting similar rates of unemployment as individuals with bipolar disorder (Zimmerman, Martinez, Young, Chelminski, & Dalrymple, 2012). In the Collaborative Longitudinal Personality Disorders Study (CLPS), patients with BPD showed decreased rates of full-time employment over ten years of follow-up compared to individuals with another PD diagnosis (either avoidant PD or obsessive-compulsive PD) or major depressive disorder (Gunderson et al., 2011). At the same time, patients’ scores on a measure of social adjustment in the employment domain showed modest improvement over time, and compared favorably to the course of employment functioning in mood disorders. These data indicate that finding and maintaining stable employment is a difficult, but not hopeless, prospect for individuals with BPD features. As such, it is a potential target for interventions aiming to promote health and well-being in these individuals.

Additional evidence suggests that the relationship between BPD symptoms and occupational functioning may be bidirectional. In the McLean Study of Adult Development (MSAD), patients with a BPD diagnosis who showed a good vocational record at baseline were prospectively more likely to experience symptomatic remission, which was defined as no longer exhibiting sufficient symptoms to meet the arbitrary DSM-5 threshold (Zanarini et al., 2006). Although rates of functional recovery were high, 40% of patients failed to achieve a recovery in both interpersonal and occupational functioning over the course of ten years, even though the vast majority of these individuals experienced a symptomatic remission of at least two years (Keuroghlian & Zanarini, 2015; Zanarini, Frankenburg, Reich, & Fitzmaurice, 2012). Poor occupational functioning (i.e., not reliably working or going to school full-time) was predominately to blame for failure to recover (Keuroghlian & Zanarini, 2015; Zanarini, Frankenburg, Reich, & Fitzmaurice, 2010). The latter finding is particularly interesting in light of the fact that individuals who failed to recover experienced worse health outcomes over the course of follow-up (Keuroghlian, Frankenburg, & Zanarini, 2013). Taken together, the results of MSAD suggest that work may act as a buffer against the worst health effects of BPD features. By the same token, experiencing unemployment may result in particularly negative health outcomes for individuals with high levels of BPD features compared to others. However, few studies have directly examined employment as a potential moderator of the relationship between BPD and physical/mental health outcomes.

We previously explored this question using data from the St. Louis Personality and Aging Network (SPAN) study (Cruitt, Boudreaux, Jackson, & Oltmanns, 2018). A representative community sample of older adults in the St. Louis metropolitan area, the SPAN study is uniquely situated to examine the relationships between personality pathology, employment, and health outcomes. We used data obtained during the baseline assessment to examine whether current employment status moderated the cross-sectional association between a multisource assessment of BPD features and self-and informant reported physical health. Employment status did moderate this association, such that individuals who were unemployed showed a stronger negative association between BPD features and physical health. These findings potentially support the hypothesis that employment status plays a buffering role for individuals with BPD features. However, the cross-sectional nature of these analyses prevented causal conclusions. It may be that individuals with worse health and higher levels of BPD features were more likely to be unemployed. To address this issue, the current analyses build on our previous research by examining the longitudinal relationships between BPD features, new instances of unemployment, and subsequent physical health. In addition, whereas the previous study was focused solely on the link between BPD features and physical health, the current study considers the possibility that unemployment moderates the relationship between BPD features and psychological well-being as well.

The SPAN study has a number of methodological strengths that lend themselves to the study of BPD, employment, and health outcomes. First, given that impairments in functioning and mental health comorbidities may be associated with treatment seeking for individuals who exhibit features of personality pathology, researchers have increasingly recognized the utility of examining BPD features in community samples to obtain a more representative understanding of the relationship between BPD features and various functional outcomes (Gleason, Weinstein, Balsis, & Oltmanns, 2014; Javaras, Zanarini, Hudson, Greenfield, & Gunderson, 2017). Furthermore, because maladaptive and normal-range personality exist along the same underlying dimensions, it is important for studies of the general population to incorporate explicit measures of maladaptive personality features in order to obtain more information at the extremes of personality (Samuel, Simms, Clark, Livesley, & Widiger, 2010; Suzuki, Samuel, Pahlen, & Krueger, 2015). By examining factors that buffer against, or enhance, the negative effects of maladaptive personality, implications may also be drawn regarding normal-range personality traits that may not have as pronounced of effects. Finally, individuals with high levels of personality pathology may have poor insight into the problems that their personality causes themselves and others (Carlson & Oltmanns, 2015; Oltmanns & Powers, 2012). Informants may endorse BPD features at lower levels of the overall latent trait than the self, thereby providing more information with regard to subthreshold manifestations of BPD pathology (Balsis, Loehle-Conger, Busch, Ungredda, & Oltmanns, 2018). Both our previous and current analyses address these methodological issues by examining the interaction between BPD features and employment with regard to health outcomes in a community sample of older adults while measuring BPD with both maladaptive and adaptive measures of personality obtained from multiple sources of information (self, informant and interviewer report).

Based on the previous literature and our previous analyses, we hypothesized that both BPD features and unemployment experiences would predict poorer outcomes with regard to physical health and depressive symptoms. We also hypothesized that those participants who reported a new experience of unemployment in the six months preceding follow-up would show a stronger negative relationship between BPD features and physical health, and stronger positive relationship between BPD features and depressive symptoms, compared to those individuals who did not experience unemployment in the past six months. This pattern of results would suggest that certain features of the employment context (versus the unemployment context) buffer against the negative effects of BPD features on well-being. To further examine this possibility, we also conducted an exploratory analysis using an additional outcome with fewer waves of data: life satisfaction. The current analyses have implications for treatment targets. If the employment context is protective, it may be particularly important to explore ways to enhance occupational opportunities and functioning in individuals with BPD. In addition, interventions may seek to provide alternative opportunities for accessing the latent functions of employment (e.g., volunteering).1

Method

Participants and Procedure

The current analyses used data from the SPAN study, a longitudinal study of personality, health and aging (see Oltmanns, Rodrigues, Weinstein, & Gleason, 2014 for characteristics of the full baseline sample). Of 1,630 participants at baseline, 1,536 had relevant follow-up data for the current analyses. Individuals with follow-up data were representative of the full baseline sample: 845 (55%) were women and 1,021 (66.5%) were White. At baseline, they ranged in age from 54 to 65 (M = 59.55, SD = 2.74). On average, participants with follow-up data had 15.39 years of education (SD = 2.53), and a median household income of $40,000 to $59,999. We asked participants to nominate an informant who knew them well to report on their personality and health throughout the study (again, see Oltmanns, Rodrigues, Weinstein, & Gleason, 2014 for informant characteristics). For the purpose of the current analyses, baseline informant reports were used to help produce the BPD component scores using principal components analysis. We choose to not examine informant reported outcomes to maximize our sample size.

Participants had between one and nine waves of follow-up data with the relevant measures for the current analyses. The median number of waves completed was seven, with 1,219 participants (79%) completing five or more waves. Baseline and three of the follow-up assessments were composed of interviews and questionnaires and were conducted in-lab. Participants completed the remaining follow-up assessments online or through the mail. The first seven follow-ups occurred at approximately six-month intervals. A gap in funding interrupted data collection, resulting in an average gap between the seventh and eighth follow-ups of 3.31 years. After the funding was renewed, we increased the interval between follow-ups. As a result, the interval between the eighth and ninth follow-ups used in the current analyses was approximately 2.27 years, on average.

Measures

Structured Interview for DSM-IV Personality (Pfohl, Blum, & Zimmerman, 1997).

The Structured Interview for DSM-IV Personality (SIDP-IV) is an 80-item semi-structured interview that assesses the PD criteria found in both DSM-IV and DSM-5. Trained interviewers rate each criterion on a scale from 0 (not present) to 3 (strongly present). We then summed across the nine items corresponding to BPD criteria to compute a continuous score of BPD symptoms. Absolute range for this score was 0 to 27; observed range at baseline using the full sample was 0 to 16 (M = 1.14, SD = 1.86). Interrater reliability was obtained using 265 interviews randomly selected from the full sample; one-way random, average measures intraclass correlation coefficient was .67 for the full interview and .77 for the BPD subscale (Shrout & Fleiss, 1979).

Multisource Assessment of Personality Pathology (Oltmanns & Turkheimer, 2006).

Developed to obtain self- and informant ratings of personality pathology, the 80-item Multisource Assessment of Personality Pathology (MAPP) consists of lay translations of the DSM-IV and DSM-5 PD criteria. The five-point response scale ranges from 0 (I am/he or she is never like this) to 4 (I am/he or she is always like this). As with the SIDP-IV, the nine BPD items were summed to create a continuous score (absolute range = 0 – 36). The observed range at baseline using the full sample was 0 to 29 for the self-report MAPP (M = 3.85, SD = 3.55), and for the informant report MAPP it was 0 to 28 (M = 4.97, SD = 4.82). Coefficient alpha for the BPD subscale was .70 for the self-report version and .79 for the informant report.

NEO Personality Inventory-Revised (Costa & McCrae, 1992).

Previous research has identified a method of generating PD scores from the NEO Personality Inventory-Revised (NEO PI-R), a 240-item measure of the five-factor model of normal-range personality (Miller, Bagby, Pilkonis, Reynolds, & Lynam, 2005). We administered both self- (Form S) and informant (Form R) report versions of the NEO PI-R. The following facets combine to produce the BPD count score: anxiety, angry hostility, depression, impulsiveness, vulnerability, openness to feelings, openness to actions, compliance (reverse scored), and deliberation (reverse scored). The mean of the self-report NEO BPD count score was 71.11 at baseline (SD = 2.73, range = 63.38 to 84.13), and the mean of the informant report version was 126.13 (SD = 29.79, range = 40 to 239). Baseline coefficient alphas for the self-report facets included in the BPD count score ranged from .60 (compliance) to .82 (depression), and from .59 (openness to actions) to .86 (depression) for the informant report facets.

List of Threatening Events Questionnaire (Brugha, Bebbington, Tennant, & Hurry, 1985).

We used the LTE-Q to assess whether participants experienced a new instance of unemployment between follow-ups. The LTE-Q consists of 12 items representing stressful life events that typically produce long-term negative effects. Each item required a response of either yes or no regarding whether the participant had experienced that event within the last six months. The item we used in the current analyses asked if the participant had experienced “unemployment or looking for a job for more than one month.” Trained interviewers administered the LTE-Q during in-lab assessments. During at-home follow-ups, participants completed a self-report version of the LTE-Q either online or using a paper-and-pencil questionnaire that was mailed to the person’s home. If the participant endorsed at least one event on the questionnaire, interviewers then arranged to conduct a structured interview over the phone in order to determine if the event was a major, distinct event that actually occurred to the participant in the specified time period. We then adjusted responses on the LTE-Q based on that interview. This hybrid assessment procedure for the assessment of stressful life events was employed in order to correct for bias associated with BPD features; individuals with higher levels of BPD features showed greater rates of adjustment (Gleason, Powers, & Oltmanns, 2012; Harkness & Monroe, 2016).

In addition to unemployment, the LTE-Q asks participants whether or not they have been fired from a job in the last six months. We considered examining this event as well, with the idea that individuals with high levels of BPD features may respond more negatively to a termination due to emotion regulation and interpersonal difficulties. However, only 1-13 instances of being fired from a job were reported in any given wave of data collection, representing between 0.00-0.01% of that wave’s number of participants. Given the extremely low rates of endorsement, we determined that we would be unable to detect meaningful effects. Therefore, we did not include being fired from a job in any analyses.

RAND-36 Health Status Inventory (Hays & Morales, 2001).

A 36-item self-report measure of subjective physical and mental health, the RAND-36 Health Status Inventory (HSI) consists of eight subscales. Four of these subscales, physical functioning, role limitations due to physical health problems, pain, and general health perceptions, are weighted and combined to produce a physical health composite score. Across all time points, the coefficient alphas ranged from .90 to .92 for physical functioning, from .87 to .90 for role limitations due to physical health problems, from .70 to .78 for pain, and from .81 to .83 for general health perceptions.

Beck Depression Inventory II (Beck, Steer, & Brown, 1996).

The Beck Depression Inventory II (BDI-II) is a 21-item self-report measure of depressive symptoms experienced during the two weeks prior to assessment. Participants pick the statement for each item that best matches their experience, with each statement corresponding to a score from 0 to 3. Coefficient alphas for the BDI-II ranged from .89 to .93 across all waves of data collection.

Satisfaction with Life Scale (Diener, Emmons, Larsen, & Griffin, 1985).

To assess subjective well-being, we began administering the Satisfaction with Life Scale (SWLS) at the fifth wave of data collection. It contains five items that use a Likert-style response scale from one (Strongly disagree) to seven (Strongly agree). Coefficient alphas across all administrations of the SWLS ranged from .89 to .92.

Demographic covariates.

All covariates were assessed at baseline. Gender was assessed dichotomously, with women coded as 0 and men coded as 1. Race was assessed categorically, but the low number of individuals identifying as other than White/Caucasian or Black/African American provided insufficient power to conduct analyses using all categories. Therefore, we coded race as 0 for White/Caucasian and 1 for Black/African American/Other for the purposes of the current analyses. Age was coded as years of age. Education was assessed categorically, with nine response options. These response options were recoded in terms of years of education as follows: Less than high school (6.5), high school or GED (12), some college (14), vocational school (14), 2-year college degree (14), 4-year college degree (16), master’s degree (18), doctoral degree (20), and professional degree (20). Household annual income was measured on a scale from 1 (under $20,000) to 8 (over $140,000), with categories increasing in $20,000 increments (e.g., 2 = 20,000 – 39,999).

Data Analytic Plan

We used R to conduct all analyses (R Core Team, 2018). First, we used principal components analysis to obtain a latent BPD score from the SIDP-IV, self- and informant reported MAPP, and self- and informant reported NEO BPD count scores from baseline, using the full SPAN sample of 1,630 participants (Le, Josse, & Husson, 2008). Prior to conducting the principal components analysis, we imputed missing data using the imputePCA function from the missMDA package (Josse & Husson, 2016). Then, we examined whether an experience of unemployment in the prior six months moderated the relationship between latent BPD scores and physical and mental health outcomes at each follow-up by running a series of multilevel models using the lme4 package (Bates, Maechler, Bolker, & Walker, 2015). We first entered the following second-level covariates: gender, race, years of education, household annual income, age, and grand-mean centered physical health or depressive symptoms at baseline. Next, we entered the grand-mean centered latent BPD score and person-mean centered unemployment.2 Finally, we ran the model including the interaction term between latent BPD score and unemployment.3

Results

Table 1 presents means, standard deviations, and ranges for the HSI physical health composite, the BDI-II and the SWLS, as well as the frequencies of unemployment experiences as reported on the LTE-Q (after adjustment following the interviews). A total of 177 participants (11.52%) reported at least one period of unemployment during the nine waves of follow-up, with 52 (3.39%) reporting more than one (M = 0.16, SD = 0.51, Range: 0 – 5). These participants were demographically representative of the full sample: 97 (55%) were women and 80 (45%) were men, 112 (63%) were White/Caucasian and 65 (37%) were Black/African American/Other.

Table 1.

Descriptive Statistics for Depressive Symptoms, Physical Health, and Unemployment

Variables N HSI-Physical Health Composite BDI-II SWLS Unemployment
Mean SD Range Mean SD Range Mean SD Range Frequency (% of sample)
Baseline 1,536a 59.30 9.90 29-72 5.12 5.92 0-43 - - - -
Wave 1 1,430 59.00 10.03 29-74 5.25 6.42 0-54 - - - 73(5.10)
Wave 2 1,366 59.37 9.88 29-74 5.43 6.87 0-52 - - - 56(4.10)
Wave 3 1,326 59.52 9.85 29-74 5.39 6.77 0-55 - - - 20(1.51)
Wave 4 1,242 59.48 10.10 28-74 5.16 6.54 0-46 - - - 40(3.22)
Wave 5 1,256 59.41 10.02 29-74 4.76 6.06 0-45 24.59 6.83 5-35 20(1.59)
Wave 6b 342a 60.42 9.88 31-74 5.32 7.18 0-49 24.47 7.32 5-35 5(1.46)
Wave 7b 359a 60.71 9.75 30-74 4.72 6.10 0-36 24.64 7.29 5-35 4(1.11)
Wave 8 1,065 59.54 9.83 31-74 5.36 6.63 0-45 24.54 6.72 5-35 16(1.50)
Wave 9 936 54.27 5.88 34-67 5.78 6.75 0-54 24.92 6.78 5-35 14(1.50)
Open in a new tab

Note. HSI = RAND-36 Health Status Inventory. BDI-II = Beck Depression Inventory II. SWLS = Satisfaction with Life Scale.

a

Number of participants at baseline with later follow-up data.

b

Waves 6 and 7 were ongoing when funding was interrupted, accounting for their smaller sample sizes.

We imputed missing data and submitted the intercorrelations among the five BPD measures to a principal components analysis. The eigenvalues of the first five components were 6.99, 1.68, 0.59, 0.47, and 0.28, and the first component accounted for 69.94% of the variance in the measures. Component loadings for the first component ranged from .68 (self-reported MAPP) to .77 (self-reported NEO PI-R BPD count score). These results suggest that the first component adequately captures the shared variance among the five BPD measures. Therefore, we computed component scores for each participant to use in subsequent analyses. BPD component scores for the current subsample ranged between −4.82 and 15.60 (M = −0.06, SD = 2.59).

Next, we ran a series of multilevel models with physical health as the outcome. Table 2 presents the results of the final multilevel model. We found a negative main effect of BPD features on physical health at follow-up, controlling for covariates and physical health at baseline. However, we did not find a main effect of unemployment experiences on changes in physical health, or a meaningful interaction between BPD features and unemployment. The final model with the interaction did not produce a significantly better fit of the data than the model without the interaction (χ2(1) = 0.31, p = .577).

Table 2.

Summary of Multilevel Model Predicting Physical Health

Parameter Estimate SE t value 95% CI
Random effects
Intercept 14.62 3.82
Residual 26.69 5.17
Fixed effects
Intercept 48.28 2.75 17.53 [42.89, 53.67]
Level 1
 Unemployment 0.10 0.42 0.24 [−0.72, 0.92]
Level 2
 Gender 0.20 0.24 0.82 [−0.27, 0.67]
 Race −1.24 0.28 −4.45 [−1.78, −0.69]
 Years of education 0.16 0.06 2.96 [0.06, 0.27]
 Household annual income 0.19 0.07 2.88 [0.06, 0.32]
 Age 0.12 0.04 2.79 [0.04, 0.21]
 Initial physical health 0.68 0.01 49.50 [0.66, 0.71]
 BPD features −0.36 0.05 −7.15 [−0.46, −0.26]
Cross-level Interaction
 BPD features x Unemployment −0.08 0.14 −0.56 [−0.35, 0.20]
Open in a new tab

Note. Based on 8,576 observations of 1,432 participants. All level 2 predictors were measured at baseline. BPD = Borderline personality disorder.

We then reran the multilevel models with depressive symptoms as the outcome (Table 3). BPD features were positively associated with increases in depressive symptoms from baseline, controlling for all covariates. As with physical health, unemployment experiences did not exhibit a main effect on depressive symptoms. However, we did find an interaction between BPD features and unemployment experiences. Adding the interaction to the model produced a significantly better model fit over the model with only the main effects of BPD features, unemployment experiences, and the covariates (χ2(1) = 4.17, p = .041).

Table 3.

Summary of Multilevel Model Predicting Depressive Symptoms

Parameter Estimate SE t value 95% CI
Random effects
Intercept 10.95 3.31
Residual 13.26 3.64
Fixed effects
Intercept 5.19 2.26 2.30 [0.77, 9.62]
Level 1
 Unemployment −0.10 0.29 −0.36 [−0.68, 0.47]
Level 2
 Gender 0.11 0.20 0.55 [−0.28, 0.50]
 Race 0.81 0.23 3.60 [0.37, 1.26]
 Years of education −0.09 0.04 −2.01 [−0.18, −0.00]
 Household annual income −0.06 0.05 −1.07 [−0.16, 0.05]
 Age 0.03 0.04 0.78 [−0.04, 0.10]
 Initial depressive symptoms 0.59 0.02 29.76 [0.55, 0.63]
 BPD features 0.56 0.04 12.56 [0.47, 0.65]
Cross-level Interaction
 BPD features x Unemployment 0.20 0.10 2.04 [0.01, 0.39]
Open in a new tab

Note. Based on 8,833 observations of 1,454 participants. All level 2 predictors were measured at baseline. BPD = Borderline personality disorder.

Exploratory Analysis

The findings above may suggest that short-term unemployment is more impactful for individuals’ psychological, rather than physical, well-being. In order to examine this possibility further, we conducted an additional exploratory analysis with life satisfaction as the outcome.4 We did not originally plan to examine life satisfaction due to the fact that it was added at the fifth follow-up, and therefore had fewer waves of data. Therefore, these results should be interpreted with caution. The results of the final multilevel model predicting life satisfaction from BPD features, unemployment experiences, and their interaction are presented in Table 4. As with physical health, BPD features predicted declines in life satisfaction relative to baseline and adding the interaction term did not produce a better model fit (χ2(1) = 0.37, p = .544). However, there appeared to be a weak main effect of unemployment on life satisfaction, although this did not reach statistical significance at the p = .05 level (χ2(1) = 3.63, p = .057).

Table 4.

Summary of Multilevel Model Predicting Satisfaction with Life

Parameter Estimate SE t value 95% CI
Random effects
Intercept 8.14 2.85
Residual 12.09 3.48
Fixed effects
Intercept 23.34 2.72 8.58 [18.03, 28.65]
Level 1
 Unemployment −1.48 0.85 −1.74 [−3.15, 0.19]
Level 2
 Gender −0.38 0.24 −1.55 [−0.85, 0.10]
 Race −0.90 0.28 −3.21 [−1.45, −0.35]
 Years of education 0.09 0.05 1.62 [−0.02, 0.20]
 Household annual income 0.22 0.07 3.26 [0.09, 0.35]
 Age −0.01 0.04 −0.25 [−0.09, 0.07]
 Initial satisfaction with life 0.60 0.02 30.89 [0.57, 0.64]
 BPD features −0.55 0.05 −10.90 [−0.65, −0.45]
Cross-level Interaction
 BPD features x Unemployment −0.19 0.32 −0.60 [−0.82, 0.43]
Open in a new tab

Note. Based on 2,362 observations of 1,038 participants. All level 2 predictors except initial satisfaction with life were measured at baseline. Initial satisfaction with life was measured at the fifth wave of data collection. BPD = Borderline personality disorder.

Discussion

The results of the current study offer mixed support for our hypotheses. Although latent BPD features showed a main effect with regard to repeated measures of physical health, depressive symptoms, and life satisfaction, experiences of unemployment in the preceding six months did not exhibit a main effect on well-being, with the possible exception of a weak effect on life satisfaction. The hypothesis that unemployment would moderate the longitudinal relationship between BPD features and self-reported physical health was unsupported - contrary to the previously reported cross-sectional analyses (Cruitt et al., 2018). We did uncover a moderating effect of unemployment experiences on the prospective relationship between BPD features and depressive symptoms. Specifically, experiencing an instance of unemployment in the six months prior to follow-up was associated with a stronger, positive association between latent BPD features and depressive symptoms at follow-up.

The moderating effect of unemployment on the relationship between BPD features and depressive symptoms is modest, but it does offer some insight into potential mechanisms linking BPD and other mental health outcomes. This finding appears to suggest that the loss of the functions of employment, whether manifest (e.g., financial) or latent (e.g., structure, self-esteem, social contact), leads to worse mental health outcomes for those individuals who exhibit trait vulnerability in the form of BPD features. Longitudinal studies of BPD indicate that the features of the disorder follow a remitting/relapsing course (Keuroghlian & Zanarini, 2015). The current findings suggest that, just as employment may be one factor in helping individuals experience a recovery in functioning (Paris, 2003), unemployment may play a role in exacerbating the negative effects of BPD symptoms. Interventions that seek not simply to achieve, but also maintain, recovery may need to take into account the role of employment. Fortunately, at least one treatment study has demonstrated the feasibility of adapting standard treatment for BPD to specifically target improved occupational outcomes (Comtois, Kerbrat, Atkins, Harned, & Elwood, 2010). Alongside efforts to help individuals with BPD features find and maintain employment, it may also be important to identify alternative opportunities to access some of the latent functions of employment (e.g., hobbies, volunteering, parenthood).

Although our data seem to have some implications for the role of employment in the relationship between BPD features and negative health outcomes, the mixed nature of our findings must be interpreted with caution. It is somewhat surprising that we did not find a main effect of unemployment experiences on physical health and depressive symptoms. These results may indicate that it is particularly important to determine for whom the loss of employment is most harmful, such as individuals who exhibit features of BPD. The possible small main effect of unemployment on life satisfaction that we found supports this idea. It may be that short-term unemployment is associated with declines in positive psychological well-being for people in general, whereas increases in psychological distress following unemployment may be more specific to high-risk individuals (e.g., those who have high levels of BPD features). Future research would benefit from considering personality, particularly in its maladaptive extremes, when examining the impact of environmental contexts on both positive and negative health and well-being outcomes. Another important factor to consider is the length of unemployment (McKee-Ryan et al., 2005). Given that long-term unemployment is conventionally defined as unemployment lasting more than six months, it may be that our duration of follow-up was too short to detect a main effect of chronic unemployment on mental or physical health. Our sample was also older, and unemployment experiences may be differentially related to health outcomes in this age group. To clarify the interaction between unemployment and personality, additional research will need to examine the length of unemployment and follow the chronically unemployed over extended periods of time, as well as in samples of young to middle aged adults.

In addition to the null result with regard to a main effect of unemployment on health, the lack of a moderating effect of unemployment experiences on the relationship between BPD features and physical health came as a surprise. The cross-sectional moderation effects previously reported still require replication (Cruitt et al., 2018), so one must consider the possibility that the current null results with regard to physical health are evidence that the previous findings were due to Type I error. However, a number of alternative explanations exist that deserve serious consideration in future research. As with the main effects of unemployment, it may be that the follow-up period was too short to detect the moderating effect of long-term unemployment on the BPD-health relationship. Another possibility is that the causal direction of the effect is in the opposite direction. The individuals who identified as unemployed at baseline included those on long-term disability. It may be that individuals with high levels of BPD features were more likely to become unemployed due to poor physical health. A final possible explanation that deserves exploration is that our sample consisted of late-middle age to older adults entering retirement age over the course of the study. Although the interview procedure we used alongside the LTE-Q in general excluded voluntary retirement from being included as an experience of unemployment, there may be relatively normative outlets for individuals of retirement age to fulfill certain latent functions of employment even when experiencing unemployment. Additionally, the effects of personality traits on physical health may be relatively stable in this age group, and less likely to be affected by stressful life events. For example, older adults show a lower mortality risk associated with unemployment than middle-aged adults (Roelfs et al., 2011). Future research should examine these potential explanations of the current null findings to determine if a moderating effect of unemployment on physical health can be identified over longer follow-up periods and at earlier stages of life.

Limitations

One limitation of the current analyses is that we were unable to directly measure access to the manifest and latent functions of employment. Many individuals may be able to satisfy these needs through alternative activities to paid employment (e.g., volunteering), especially in older adulthood when retirement is normative. Highly conscientious individuals who are retired or unemployed are much more likely to volunteer, apparently fulfilling their need for achievement (Mike, Jackson, & Oltmanns, 2014). Unfortunately, this avenue may be more difficult for individuals with BPD features characterized by disinhibition. It will be important to determine the degree to which (un)employment is predictive of changes in social contact, status, self-esteem, and other relevant environmental features for individuals with BPD, and whether these changes explain the moderation effect of unemployment on the relationship between BPD features and depressive symptoms. Another limitation is that we were unable to examine whether individuals with higher levels of BPD features reacted more strongly to unemployment events due to more frequently being fired (rather than unemployed for other reasons). Although the LTE-Q included an item about “being fired from a job,” the low rates of endorsement on this item prevented us from analyzing it further. One possible reason for the low rates of endorsement is the older age of the sample, and future research may benefit from examining this question in early middle adulthood.

The current analyses benefit from enhanced generalizability due to the community-based sample, and researchers increasingly recognize the importance of examining BPD features in the general population (Javaras et al., 2017). However, drawing implications for treatment will require following up these findings with research on clinical populations. Given that our sample was demographically representative of the diverse St. Louis metropolitan area, we anticipate that our findings will generalize to similar older adult populations across the United States. A potential constraint on the generality of these findings is the fact that baseline data collection began in 2007, and initial waves of data collection were run during the Great Recession. Future studies will need to determine whether our findings are limited to periods of economic downturn, or whether they replicate in times of prosperity as well. Finally, the field of personality pathology research is going through a period of definitional change, as dimensional diagnostic models are developed and implemented (American Psychiatric Association, 2013; World Health Organization, 2018). The current analyses benefit from a multi-source and dimensional approach to the study of BPD features, but it will be important for future research to see if these findings generalize to measures of maladaptive personality traits developed for these new classification systems.

Conclusions

The current study draws on the vast empirical and theoretical literature on the relationship between (un)employment and health outcomes to examine the role unemployment plays in the relationship between BPD pathology and health. Although our current findings are mixed, they point to the importance of considering the maladaptive extremes of personality in studying the role of environmental contexts in determining health outcomes. Just as personality does not exist in a vacuum, work does not affect people in the same ways uniformly. It is important to examine the role individual differences play in how people respond to the activity of work.

Highlights:

  • Borderline personality disorder (BPD) predicts poor health outcomes.

  • Unemployment does not predict physical health/depressive symptoms.

  • Unemployment may exhibit a small, negative effect on life satisfaction.

  • Unemployment does not moderate effect of BPD on physical health/life satisfaction.

  • Unemployment does strengthen effect of BPD on depressive symptoms.

Acknowledgments

Grants from the National Institute of Mental Health (RO1-MH077840-01), National Institute on Aging (RO1-AG045231), and National Institutes of Health (5 T32 AG000030-39) supported this work.

We would like to thank Michael Boudreaux for preparing the data for analysis, Patrick Hill and Joshua Jackson for their comments on the analyses, Merlyn Rodrigues, Christina Noel White, and Shawn Fraine for coordinating data collection, all of our research assistants for assisting in data collection, and our participants for volunteering their time and energy to research.

The first author (PC) assisted in data collection, ran analyses, and wrote the manuscript. The second author (TO) designed the study, oversaw data collection, and offered comments on the manuscript.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

All authors report no financial relationships with commercial interests.

1

We did not formally preregister our hypotheses and data analytic plan in an independent, institutional database. We submitted an abstract for consideration in this special issue that contained our data analytic plan prior to conducting analyses. Abstract is available on request from the first author.

2

Results were similar if the variables were not centered, with the exception of the main effect of unemployment on life satisfaction (see below).

3

We will be making the data used for the current analyses available publicly after the conclusion of the SPAN study, in accordance with NIH guidelines. R code is available upon request from the first author.

4

We would like to thank an anonymous reviewer, who suggested this analysis.

References

  1. American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). Washington, DC: Author. [Google Scholar]
  2. Amundsen Østby K, Czajkowski N, Knudsen GP, Ystrom E, Gjerde LC, Kendler KS, … Reichborn-Kjennerud T (2014). Personality disorders are important risk factors for disability pensioning. Social Psychiatry and Psychiatric Epidemiology, 49(12), 2003–2011. 10.1007/s00127-014-0878-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Balsis S, Loehle-Conger E, Busch AJ, Ungredda T, & Oltmanns TF (2018). Self and informant report across the borderline personality disorder spectrum. Personality Disorders: Theory, Research, and Treatment, 9(5), 429–436. 10.1037/per0000259 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Bates D, Maechler M, Bolker B, & Walker S (2015). Fitting linear mixed-effects models using lme4. Journal of Statistical Software, 67(1), 1–48. 10.18637/jss.v067.i01 [DOI] [Google Scholar]
  5. Beck AT, Steer RA, & Brown GK (1996). Manual for the Beck Depression Inventory-II. San Antonio, TX: Psychological Corporation. [Google Scholar]
  6. Brugha T, Bebbington P, Tennant C, & Hurry J (1985). The List of Threatening Experiences: A subset of 12 life event categories with considerable long-term contextual threat. Psychological Medicine, 15(1), 189–194. 10.1017/S003329170002105X [DOI] [PubMed] [Google Scholar]
  7. Carlson EN, & Oltmanns TF (2015). The role of metaperception in personality disorders: Do people with personality problems know how others experience their personality? Journal of Personality Disorders, 29(4), 449–467. 10.1521/pedi.2015.29.4.449 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Coid J, Yang M, Tyrer P, Roberts A, & Ullrich S (2006). Prevalence and correlates of personality disorders in Great Britain. British Journal of Psychiatry, 188, 423–431. 10.1192/bjp.188.5.423 [DOI] [PubMed] [Google Scholar]
  9. Comtois KA, Kerbrat AH, Atkins DC, Harned MS, & Elwood L (2010). Recovery from disability for individuals with borderline personality disorder: A feasibility trial of DBT-ACES. Psychiatric Services, 61(11), 1106–1111. 10.1176/ps.2010.61.11.1106 [DOI] [PubMed] [Google Scholar]
  10. Costa PT, & McCrae RR (1992). Revised NEO Personality Inventory (NEO PI-R) and the NEO Five-Factor Inventory (NEO-FFI) Professional Manual. Odessa, FL: Psychological Assessment Resources. [Google Scholar]
  11. Creed PA, & Evans BM (2002). Personality, well-being and deprivation theory. Personality and Individual Differences, 33(7), 1045–1054. 10.1016/S0191-8869(01)00210-0 [DOI] [Google Scholar]
  12. Creed PA, Muller J, & Machin MA (2001). The role of satisfaction with occupational status, neuroticism, financial strain and categories of experience in predicting mental health in the unemployed. Personality and Individual Differences, 30(3), 435–447. 10.1016/S0191-8869(00)00035-0 [DOI] [Google Scholar]
  13. Cruitt PJ, Boudreaux MJ, Jackson JJ, & Oltmanns TF (2018). Borderline personality pathology and physical health: The role of employment. Personality Disorders: Theory, Research, and Treatment, 9(1), 73–80. 10.1037/per0000211 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Diener E, Emmons RA, Larsen RJ, & Griffin S (1985). The satisfaction with life scale. Journal of Personality Assessment, 49(1), 71–75. 10.1207/s15327752jpa4901_13 [DOI] [PubMed] [Google Scholar]
  15. Dixon-Gordon KL, Whalen DJ, Layden BK, & Chapman AL (2015). A systematic review of personality disorders and health outcomes. Canadian Psychology/Psychologie Canadienne. 10.1037/cap0000024 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Ellison WD, Rosenstein L, Chelminski I, Dalrymple K, & Zimmerman M (2016). The clinical significance of single features of borderline personality disorder: Anger, affective instability, impulsivity, and chronic emptiness in psychiatric outpatients. Journal of Personality Disorders, 30(2), 261–270. 10.1521/pedi_2015_29_193 [DOI] [PubMed] [Google Scholar]
  17. Galione JN, & Oltmanns TF (2013). The relationship between borderline personality disorder and major depression in later life: Acute versus temperamental symptoms. The American Journal of Geriatric Psychiatry, 21(8), 747–756. https://doi.Org/10.1016/j.jagp.2013.01.026 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Gleason MEJ, Powers AD, & Oltmanns TF (2012). The enduring impact of borderline personality pathology: Risk for threatening life events in later middle-age. Journal of Abnormal Psychology, 121(2), 447–457. 10.1037/a0025564 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Gleason MEJ, Weinstein Y, Balsis S, & Oltmanns TF (2014). The enduring impact of maladaptive personality traits on relationship quality and health in later life. Journal of Personality, 82(6), 493–501. 10.1111/jopy.12068 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Gnambs T, Stiglbauer B, & Selenko E (2015). Psychological effects of (non)employment: A cross-national comparison of the United States and Japan. Scandinavian Journal of Psychology, 56(6), 659–669. 10.1111/sjop.12240 [DOI] [PubMed] [Google Scholar]
  21. Gunderson JG, Stout RL, McGlashan TH, Shea MT, Morey LC, Grilo CM, … Skodol AE (2011). Ten-year course of borderline personality disorder: Psychopathology and function from the collaborative longitudinal personality disorders study. Archives of General Psychiatry, 68(8), 827–837. 10.1001/archgenpsychiatry.2011.37 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Harkness KL, & Monroe SM (2016). The assessment and measurement of adult life stress: Basic premises, operational principles, and design requirements. Journal of Abnormal Psychology, 125(5), 727–745. 10.1037/abn0000178 [DOI] [PubMed] [Google Scholar]
  23. Hays RD, & Morales LS (2001). The RAND-36 measure of health-related quality of life. Annals of Medicine, 33(5), 350–357. [DOI] [PubMed] [Google Scholar]
  24. Hengartner MP, Müller M, Rodgers S, Rössler W, & Ajdacic-gross V (2014). Occupational functioning and work impairment in association with personality disorder trait-scores. Social Psychiatry and Psychiatric Epidemiology, 49(2), 327–335. 10.1007/s00127-013-0739-2 [DOI] [PubMed] [Google Scholar]
  25. Jahoda M (1982). Employment and unemployment: A social-psychological analysis. Cambridge, UK: Cambridge University Press. [Google Scholar]
  26. Javaras KN, Zanarini MC, Hudson JI, Greenfield SF, & Gunderson JG (2017). Functional outcomes in community-based adults with borderline personality disorder. Journal of Psychiatric Research, 89, 105–114. 10.1016/j.jpsychires.2017.01.010 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Josse J, & Husson F (2016). missMDA: A package for handling missing values in multivariate data analysis. Journal of Statistical Software, 70(1), 1–31. 10.18637/jss.v070.i01 [DOI] [Google Scholar]
  28. Keuroghlian AS, Frankenburg FR, & Zanarini MC (2013). The relationship of chronic medical illnesses , poor health-related lifestyle choices , and health care utilization to recovery status in borderline patients over a decade of prospective follow-up. Journal of Psychiatric Research, 47(10), 1499–1506. 10.1016/j.jpsychires.2013.06.012 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Keuroghlian AS, & Zanarini MC (2015). Lessons learned from longitudinal studies of personality disorders In Huprich SK (Ed.), Personality Disorders: Toward Theoretical and Empirical Integration in Diagnosis and Assessment (pp. 145–161). Washington DC: American Psychological Association; 10.1037/14549-007 [DOI] [Google Scholar]
  30. Knudsen AK, Skogen JC, Harvey SB, Stewart R, Hotopf M, & Moran P (2012). Personality disorders, common mental disorders and receipt of disability benefits: Evidence from the British national survey of psychiatric morbidity. Psychological Medicine, 42(12), 2631–2640. 10.1017/S0033291712000906 [DOI] [PubMed] [Google Scholar]
  31. Le S, Josse J, & Husson F (2008). FactoMineR: An R package for multivariate analysis. Journal of Statistical Software, 25(1), 1–18. 10.18637/jss.v025.i01 [DOI] [Google Scholar]
  32. Lenzenweger MF, Lane MC, Loranger AW, & Kessler RC (2007). DSM-IV personality disorders in the National Comorbidity Survey Replication. Biological Psychiatry, 62(6), 553–564. 10.1016/j.biopsych.2006.09.019 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. McKee-Ryan FM, Song Z, Wanberg CR, & Kinicki AJ (2005). Psychological and physical well-being during unemployment: A meta-analytic study. Journal of Applied Psychology, 90(1), 53–76. 10.1037/0021-9010.90.1.53 [DOI] [PubMed] [Google Scholar]
  34. Mike A, Jackson JJ, & Oltmanns TF (2014). The conscientious retiree: The relationship between conscientiousness, retirement, and volunteering. Journal of Research in Personality, 52, 68–77. 10.1016/j.jrp.2014.07.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Miller JD, Bagby RM, Pilkonis PA, Reynolds SK, & Lynam DR (2005). A simplified technique for scoring DSM-IV personality disorders with the five-factor model. Assessment, 12(4), 404–415. 10.1177/1073191105280987 [DOI] [PubMed] [Google Scholar]
  36. Oltmanns TF, & Powers AD (2012). Knowing our pathology In Vazire S & Wilson TD (Eds.), Handbook of self-knowledge (pp. 258–273). New York, NY: Guilford Press. [Google Scholar]
  37. Oltmanns TF, Rodrigues MM, Weinstein Y, & Gleason MEJ (2014). Prevalence of personality disorders at midlife in a community sample: Disorders and symptoms reflected in interview, self, and informant reports. Journal of Psychopathology and Behavioral Assessment, 36(2), 177–188. 10.1007/s10862-013-9389-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Oltmanns TF, & Turkheimer E (2006). Perceptions of self and others regarding pathological personality traits In Krueger RF & Tackett JL (Eds.), Personality and psychopathology (pp. 71–111). New York: Guilford Press. [Google Scholar]
  39. Paris J (2003). Personality disorders over time: Precursors, course and outcome. Quebec, Canda: American Psychiatric Publishing, Inc. [DOI] [PubMed] [Google Scholar]
  40. Paul KI, & Batnic B (2010). The need for work: Jahoda’s latent functions of employment in a representative sample of the German population. Journal of Organizational Behavior, 31, 45–64. 10.1002/job.622 [DOI] [Google Scholar]
  41. Paul KI, & Moser K (2009). Unemployment impairs mental health: Meta-analyses. Journal of Vocational Behavior, 74(3), 264–282. 10.1016/j.jvb.2009.01.001 [DOI] [Google Scholar]
  42. Pfohl B, Blum N, & Zimmerman M (1997). Structured interview for DSM-IV personality. Washington, DC: American Psychiatric Press. [Google Scholar]
  43. R Core Team. (2018). R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing; Retrieved from https://www.r-project.org/ [Google Scholar]
  44. Roelfs DJ, Shor E, Davidson KW, & Schwartz JE (2011). Losing life and livelihood: A systematic review and meta-analysis of unemployment and all-cause mortality. Social Science and Medicine, 72(6), 840–854. 10.1016/j.socscimed.2011.01.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Samuel DB, Simms LJ, Clark LA, Livesley WJ, & Widiger TA (2010). An item response theory integration of normal and abnormal personality scales. Personality Disorders: Theory, Research, and Treatment, 1(1), 5–21. 10.1037/a0018136 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Sansone RA, Leung JS, & Wiederman MW (2012). Employment histories among patients with borderline personality disorder symptomatology. Journal of Vocational Rehabilitation, 37(2), 131–137. 10.3233/JVR-2012-0606 [DOI] [Google Scholar]
  47. Sansone RA, & Sansone LA (2012). Employment in borderline personality disorder. Innovations in Clinical Neuroscience, 9(9), 25–29. Retrieved from http://oxfordmedicine.com/view/10.1093/med/9780199608386.001.0001/med-9780199608386-chapter-8 [PMC free article] [PubMed] [Google Scholar]
  48. Selenko E, Batinic B, & Paul K (2011). Does latent deprivation lead to psychological distress? Investigating Jahoda’s model in a four-wave study. Journal of Occupational and Organizational Psychology, 84(4), 723–740. 10.1348/096317910X519360 [DOI] [Google Scholar]
  49. Shrout PE, & Fleiss JL (1979). Intraclass correlations: Uses in assessing rater reliability. Psychological Bulletin, 86(2), 420–428. 10.1037/0033-2909.86.2.420 [DOI] [PubMed] [Google Scholar]
  50. Skodol AE, Stout RL, McGlashan TH, Grilo CM, Gunderson JG, Trade Shea M, … Oldham JM (1999). Co-occurrence of mood and personality disorders: A report from the Collaborative Longitudinal Personality Disorders Study (CLPS). Depression and Anxiety, 10(4), 175–182. [DOI] [PubMed] [Google Scholar]
  51. Soeteman DI, Roijen LH, Verheul R, & Busschbach JJV (2008). The economic burden of personality disorders in mental health care. Journal of Clinical Psychiatry, 69(2), 259–265. Retrieved from http://cataleg.ub.edu/record=b2217197~S1*cat [DOI] [PubMed] [Google Scholar]
  52. Sousa-Ribeiro M, Sverke M, & Coimbra JL (2014). Perceived quality of the psychosocial environment and well-being in employed and unemployed older adults: The importance of latent benefits and environmental vitamins. Economic and Industrial Democracy, 35(4), 629–652. 10.1177/0143831X13491840 [DOI] [Google Scholar]
  53. Strauss JS, & Davidson L (1997). Mental disorders, work, and choice In Bonnie RJ & Monahan J (Eds.), Mental Disorder, Work Disability, and the Law (pp. 105–130). Chicago, IL: University of Chicago Press. [Google Scholar]
  54. Strauss JS, Harding CM, Silverman M, Eichler A, & Lieberman M (1988). Work as treatment for psychiatric disorder: A puzzle in pieces In Vocational Rehabilitation of Persons with Prolonged Psychiatric Disorders (pp. 47–55). [Google Scholar]
  55. Suzuki T, Samuel DB, Pahlen S, & Krueger RF (2015). DSM-5 alternative personality disorder model traits as maladaptive extreme variants of the five-factor model: An item-response theory analysis. Journal of Abnormal Psychology, 124(2), 343–354. 10.1037/abn0000035 [DOI] [PubMed] [Google Scholar]
  56. Trull TJ, Jahng S, Tomko RL, Wood PK, & Sher KJ (2010). Revised NESARC personality disorder diagnoses: Gender, prevalence, and comorbidity with substance dependence disorders. Journal of Personality Disorders, 24(4), 412–426. 10.1521/pedi.2010.24.4.412 [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Warr P (1987). Work, unemployment, and mental health. New York: Oxford University Press. [Google Scholar]
  58. World Health Organization. (2018). International statistical classification of diseases and related health problems. Retrieved from https://icd.who.int/browse11/l-m/en
  59. Zanarini MC, Ed D, Frankenburg FR, Hennen J, Ph D, Reich DB, & Silk KR (2006). Prediction of the 10-year course of borderline personality disorder. American Journal of Psychiatry, 163(5), 827–832. 10.1176/ajp.2006.163.5.827 [DOI] [PubMed] [Google Scholar]
  60. Zanarini MC, Frankenburg FR, Dubo ED, Sickel AE, Trikha A, Levin A, & Reynolds V (1998). Axis I comorbidity of borderline personality disorder. American Journal of Psychiatry, 155(12), 1733–1739. 10.1176/ajp.155.12.1733 [DOI] [PubMed] [Google Scholar]
  61. Zanarini MC, Frankenburg FR, Reich DB, & Fitzmaurice G (2010). Time to attainment of recovery from borderline personality disorder and stability of recovery: A 10-year propsective follow-up study. The American Journal of Psychiatry, 167(6), 663–667. 10.1176/appi.ajp.2009.09081130 [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Zanarini MC, Frankenburg FR, Reich DB, & Fitzmaurice G (2012). Attainment and stability of sustained symptomatic remission and recovery among patients with borderline personality disorder and axis II comparison subjects: A 16-year prospective follow-up study. The American Journal of Psychiatry, 169, 476–483. 10.1176/appi.ajp.2011.11101550 [DOI] [PMC free article] [PubMed] [Google Scholar]
  63. Zanarini MC, Jacoby RJ, Frankenburg FR, Reich DB, & Fitzmaurice G (2009). The 10-year course of social security disability income reported by patients with borderline personality disorder and axis II comparison subjects. Journal of Personality Disorders, 25(4), 346–356. 10.1521/pedi.2009.23.4.346 [DOI] [PMC free article] [PubMed] [Google Scholar]
  64. Zimmerman M, Martinez JH, Young D, Chelminski I, & Dalrymple K (2012). Sustained unemployment in psychiatric outpatients with bipolar depression compared to major depressive disorder with comorbid borderline personality disorder. Bipolar Disorders, 14(8), 856–862. 10.1111/bdi.12014 [DOI] [PubMed] [Google Scholar]

RESOURCES