Emotion perception and executive functioning predict work status in euthymic bipolar disorder

Emotion perception and executive functioning predict work status in euthymic bipolar disorder

Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎ Contents lists available at SciVerse ScienceDirect Psychiatry Research journal homepage: www.elsevier.com/locat...

472KB Sizes 3 Downloads 82 Views

Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Contents lists available at SciVerse ScienceDirect

Psychiatry Research journal homepage: www.elsevier.com/locate/psychres

Emotion perception and executive functioning predict work status in euthymic bipolar disorder Kelly A. Ryan n, Aaron C. Vederman, Masoud Kamali, David Marshall, Anne L. Weldon, Melvin G. McInnis, Scott A. Langenecker Department of Psychiatry, University of Michigan, Ann Arbor, MI, USA

art ic l e i nf o

a b s t r a c t

Article history: Received 4 December 2012 Received in revised form 20 June 2013 Accepted 23 June 2013

Functional recovery, including return to work, in Bipolar Disorder (BD) lags behind clinical recovery and may be incomplete when acute mood symptoms have subsided. We examined impact of cognition on work status and underemployment in a sample of 156 Euthymic-BD and 143 controls (HC) who were divided into working/not working groups. Clinical, health, social support, and personality data were collected, and eight cognitive factors were derived from a battery of neuropsychological tests. The HC groups outperformed the BD groups on seven of eight cognitive factors. The working-BD group outperformed the not working-BD group on 4 cognitive factors composed of tasks of emotion processing and executive functioning including processing speed and set shifting. Emotion processing and executive tasks were predictive of BD unemployment, after accounting for number of mood episodes. Four cognitive factors accounted for a significant amount of the variance in work status among the BD participants. Results indicate that patients with BD who are unemployed/unable to work exhibit greater difficulties processing emotional information and on executive tasks that comprise a set shifting or interference resolution component as compared to those who are employed, independent of other factors. These cognitive and affective factors are suggested as targets for treatment and/or accommodations. & 2013 Elsevier Ireland Ltd. All rights reserved.

Keywords: Bipolar disorder Functioning Occupation Cognition

1. Introduction Functional recovery in Bipolar Disorder (BD) has been shown to lag behind recovery from clinical symptoms and may still be incomplete when mood symptoms have subsided (Coryell et al., 1993; Goldberg et al., 1995a; Keck, 2006; Tohen et al., 2000). Only recently have predictors of functional impairment in the specific areas of daily functioning been examined, above using global rating scales that include assessment of mood symptoms. One poorly captured area of functional recovery concerns employment, despite BD being the 6th leading cause of disability world-wide in ages 18–44 year olds (Murray and Lopez, 1997). Several clinical factors, including number of psychiatric hospitalizations and past episodes are weakly related to disability and occupational outcomes in BD, and they have not entirely accounted for disability (Altshuler et al., 2007; Dickerson et al., 2004; Goldberg et al., 1995b; Gutiérrez-Rojas et al., 2011; Staner et al., 1997). Although treatments to limit hospitalizations and reoccurrence of mood

n Corresponding author. 2101 Commonwealth Blvd, Suite C, Ann Arbor, MI 48105, USA. Tel.: +1 734 936 5524; fax: +1 734 936 9262. E-mail address: [email protected] (K.A. Ryan).

episodes likely impact occupational status, it is not clear if or how they could have great clinical utility in specifically improving work related performance and functional independence in BD. Beyond these clinical factors, social supports (Leonardi et al., 2007; Vieta et al., 2007), personality (Bieling et al., 2003; Medard et al., 2010; Pope et al., 2007) and cognitive functioning have been associated with low functioning and disability in BD, although these results also have been modest in nature (for review see Sanchez-Moreno et al., 2009). Despite persistent cognitive deficits in the remitted BD state (Clark et al., 2005; Dixon et al., 2004; El-Badri et al., 2001; Fleck et al., 2005; Mann-Wrobel et al., 2011; Martínez-Arán et al., 2004a, b), the impact of cognitive deficits on occupational outcomes is largely understudied. This is in contrast to other disorders like schizophrenia, where cognition has been shown as a predictor of employment status (McGurk and Mueser, 2004). In a recent study, working memory⁄attention and speed of processing, were associated with occupational recovery, with cognitive improvements being highly predictive of work recovery three months later (Bearden et al., 2011). Prior studies have demonstrated that better cognitive functioning and less severe clinical symptoms were associated with higher employment status (full time vs. part time or volunteer status) among BD patients, not accounting for mood state (Dickerson et al., 2004; O'Shea et al., 2010). In one small

0165-1781/$ - see front matter & 2013 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.psychres.2013.06.031

Please cite this article as: Ryan, K.A., et al., Emotion perception and executive functioning predict work status in euthymic bipolar disorder. Psychiatry Research (2013), http://dx.doi.org/10.1016/j.psychres.2013.06.031i

K.A. Ryan et al. / Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎

2

study, attention and fluency tasks predicted work/school recovery 12 months after initial stabilization from an acute mood episode (Jaeger et al., 2007). Self-report of executive functioning is weakly associated with unemployment status in veterans accounting for 1% of variance in employment (Altshuler et al., 2007). However, these previous studies are limited by small sample sizes, poor comparison groups, failure to account for current mood state, and/ or use of broad measures of functioning that do not capture occupational status. Additionally, these studies have demonstrated weak predictive value, and have not indicated treatment targets that might improve recovery and functional independence in BD. The present study examined the impact of clinical factors, health-related quality of life, social support, personality, and cognitive factors on occupational work status in a sample of euthymic BD patients compared to a healthy control group. Identifying significant predictors might be useful in devising targeted clinical treatment plans (and possibly better accommodations within the workplace) for those with greater risk for persistent un- or under-employment. 2. Method 2.1. Participants Participants were recruited into the Longitudinal Prechter Study of Bipolar Disorder at the University of Michigan. Using the longitudinal cohort consisting of 586 participants, the present study included 156 individuals with a DSM-IV diagnosis of BD in the euthymic state (133 BD-I, 23 BD-II) and 143 healthy controls (HC). Recruitment of all participants, including HC, occurred from 2005 to 2010 through advertisements on the web and in the newspaper, an outpatient specialty psychiatric clinic, and an inpatient psychiatric unit. Participants with BD were excluded from the study if they had a history of schizophrenia or schizoaffective disorder depressive type, any active or current substance dependence (within six months of baseline evaluation) or substance abuse (within two months of baseline evaluation), or a medical illness specifically associated with depressive symptoms (including but not limited to: terminal cancers, Cushing's disease, or stroke). HC participants were not eligible to participate if they had a history of any DSM-IV axis I disorder, active and current substance use disorder diagnosis, any medical illness specifically associated with depressive symptoms, or any first-degree family member who had been diagnosed or hospitalized for mental illness. All participants gave written informed consent prior to participation. This study was approved by the University of Michigan Institutional Review Board. All participants underwent an evaluation using the Diagnostic Interview for Genetic Studies (DIGS, Nurnberger et al., 1994), neuropsychological testing, psychiatric symptom and psychosocial questionnaires, Hamilton Depression Rating-17 item (HDRS, Hamilton, 1967) and Young Mania Rating Scale (YMRS, Young et al., 1978). A best estimate process by at least three of the authors was used to confirm diagnoses. We used conservative criteria at the time of the neuropsychological evaluation to validate euthymic state (HDRS (o 8) and YMRS ( o8)). Occupational data was obtained as part of the DIGS interview. The DIGS ascertains information on the participant's current job and classifies it according to occupational categories (e.g., professional specialty, administrative support, unemployed, disabled, student). All participants were divided into two groups based on current working status: working and not working. Participants in the working group consisted of individuals who were currently employed full-time (any occupational category) or enrolled as a full-time student. All participants in the not working group consisted of those who were currently unemployed or were disabled. Unless on disability, information about why a participant was not working was not collected. Participants who were retired (n¼12) or a “housewife” (n¼ 8) were excluded to clearly delineate the two working/not working groups. Information on “volunteering” was not collected. In addition, the DIGS collects information about the participant's most responsible occupational position (whether it was a past position or is their current position and based on the interviewee's report of “most responsible” position) and those positions were classified using the same occupational category as current occupation. A discrepancy variable was created based on current work classification and most responsible work classification, which provided a measure of who declined in job class. For example, if a participant's most responsible position was a “professional position” yet their current position is classified as “laborer,” they were considered to have declined in job class.

2.2. Clinical variables Number of psychiatric hospitalizations, age at onset of first depression/manic/ hypomanic mood episode, summation of number of mood episodes, years since

first episode and comorbid psychiatric diagnoses (substance abuse/dependence, anxiety disorders, eating disorders, and ADHD) taken from the DIGS were chosen based upon prior literature demonstrating that these factors have a potential impact upon functioning (Table 1). There were no significant differences between BD and HC groups for education, F(1, 263) ¼ 12.73, p ¼0.99 and gender χ2 (1, N¼ 289) ¼ 0.713, p¼ 0.47, but the BD group was significantly older than the HC group, F(1, 263) ¼13.34, p o 0.001, which is addressed specifically in later analyses, i.e., logistic regression. As expected, the HC had a significantly higher Global Assessment of Functioning (GAF) score than the BD group, F(1, 216)¼ 98.41, po 0.001. Among the BD individuals, the only significant difference between the working and not working BD groups for comorbid psychiatric diagnosis was for presence of social phobia, χ2 (1, N¼ 130) ¼12.04, p ¼0.001 with the not working BD group having a more individuals with a comorbid diagnosis of social phobia (n¼8) vs. the working BD group (n ¼5). There were no working/not working BD group differences for total number of psychiatric comorbidities, F(1, 289) ¼3.51, p ¼0.063. Individuals with BD were taking a number of medications across a range of classes. We examined the influence of medications by adopting a protocol often seen in the literature to assess total medication load. Antidepressant, anxiolytic, mood stabilizer, and antipsychotic medications were coded as absent ¼ 0, low¼1, or high ¼2 to convert each medication to a standardized dose (Almeida et al., 2009; Hassel et al., 2008; Sackeim, 2001). Antipsychotics were converted into chlorpromazine dose equivalents (Davis and Chen, 2004). We generated a composite measure of total medication load by summing all individual medication codes for each individual medication within categories for each BD participant (Hassel et al., 2008). There were no significant differences between the working and not working group on whether they were taking lithium, antiepileptic drugs, antipsychotics, sedatives, stimulants, or antidepressants (Supplemental Table S1). There were no differences between work status and total medication loading, F(1, 120)¼ 1.45, p¼ 0.230, and medication loading was not correlated with any of the cognitive factors (p 40.05; Supplemental Table S2).

2.3. Measures 2.3.1. Neuropsychological assessment Cognitive functioning was assessed via an extended neuropsychological test battery focused heavily upon areas known to be adversely affected in BD: Rey– Osterrieth Complex Figure Test (Meyers and Meyers, 1995), California Verbal Learning Test-II (Delis et al., 2000), Purdue Pegboard test (Lezak, 1995), Emotional Perception test (Green and Allen, 1997), Facial Emotion Perception test (Rapport et al., 2002), Wisconsin Card Sorting Test (Heaton, 1981), The Stroop Color and Word Test (Golden, 1978), the FAS verbal fluency task of the Controlled Oral Word Association Test and Animal Fluency (Benton and Hamsher, 1976), Digit Symbol from the Wechsler Adult Intelligence Scale-III (Wechsler, 1997), The Trail Making Test-Parts A and B (Armitage, 1946) and the Parametric Go/No-Go task (Langenecker et al., 2007a). Cognitive factor scores were created using standard data reduction techniques to reduce the tests using conceptually and theoretically categorized variables and based on existing knowledge regarding the cognitive constructs (Langenecker et al., 2010, for a full description). A confirmatory factor analysis was computed with the above variables, consistent with our prior work (Langenecker et al., 2010; Ryan et al., 2012). Although factor scores were used in main analyses, Supplemental Table S3 shows group differences for all neuropsychological tests and Supplemental Table S4 shows which cognitive subtests scores were included in each factor, including reliability of the factor scores. Reliabilities for seven cognitive factors with greater than three component variables were all strong. One factor is composed of only two variables; alpha reliability is not an appropriate metric for evaluation of that factor. The respective variables show moderate to strong internal validity using odd-even reliability estimates (Langenecker et al., 2007b). Eight factor scores were used; Auditory Memory, Visual Memory, Emotion Processing, Fine Motor Dexterity, and four executive functioning factors; Verbal Fluency and Processing Speed, Conceptual Reasoning and Set-Shifting, Processing Speed with Interference Resolution, and Inhibitory Control. The Vocabulary subtest from the Wechsler Abbreviated Scale of Intelligence (Wechsler, 1999), which was used as a methodological evaluation for premorbid ability, was not significantly different between the diagnostic groups or the work status groups (p 40.05; Table 1).

2.3.2. Psychosocial and personality measures The following measures were included as they have been associated with low functioning and disability in BD (Sanchez-Moreno et al., 2009). The Social Support and Undermining Scale (Abbey et al., 1985), a 17-item scale commonly used to examine relationship dynamics in stressful situations such as unemployment (Vinokur and Van Ryn, 1993). The Short Form-36 item Health Survey (SF-36) is a 36-item scale that measures a variety of health concepts. The two summary composite scores (physical and mental health composite scores) were used (Ware, 1996). The Revised NEO Personality Inventory (NEO-PI-R) captures five global dimensions of adult personality traits: neuroticism, extraversion, and openness to experience, agreeableness and conscientiousness (Costa and McCrae, 1992).

Please cite this article as: Ryan, K.A., et al., Emotion perception and executive functioning predict work status in euthymic bipolar disorder. Psychiatry Research (2013), http://dx.doi.org/10.1016/j.psychres.2013.06.031i

K.A. Ryan et al. / Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎

3

Table 1 Demographic characteristics and psychosocial measures for the bipolar and healthy control groups. Data are presented as Mean (SD). Work Status and Diagnostic Groups

Working BD n ¼123

Not working BD n ¼33

Working HC n¼125

Not working HC n¼ 18

p

Post-hoc

Age Education Gendera % Females WASI vocabulary HDRS-17 YMRS Years ill First age at BD onset Number of hospitalizations Number of all episodes % Psychosis history GAF

36.1(11.8) 15.6(2.08)

42.2(11.6) 15.2(2.2)

30.4(12.8) 15.8(2.2)

45.0(15.5) 14.2(2.0)

o 0.001 nwHC 4wBD,nwBD4wHC 0.02 wHC, wBD4 nwHC

0.050 o0.001

42.4 12.6 (2.8) 4.1(2.9) 1.5(2.0) 17.5(12.2) 18.5(7.07) 2.5(2.7)

39.3 11.52 (2.8) 4.85(3.0) 2.04(2.2) 21.4(10.3) 21.1(8.0) 3.3(2.9)

42.6 12.9 (2.9) 1.1(1.4) 0.2(0.5) – – –

61.1 12.1 (2.6) 0.9(1.2) 0.3(0.7) – – –

0.47 0.08 0.21 0.23 0.13 0.10 0.21

0.923 0.018 o0.001 0.049 o0.001 0.12

30.9(55.4)

81.9(118.5)





o 0.001 nwBD4wBD

75.0 76.2(10.9)

77.4 68.1(14.2)

– 86.6 (4.0)

– 84.1 (5.7)

PCS: SF-36 MCS: SF-36 Social support NEO neuroticism NEO extraversion NEO openness NEO conscientiousness NEO agreeableness Auditory memory Visual memory Fine motor

51.5(9.5) 43.6(11.9) 4.0(0.9) 56.6(12.6) 50.1(11.6) 59.5(11.8) 42.9(14.5)

45.4(11.2) 39.2(13.3) 3.6(1.1) 58.9(12.0) 52.3(11.7) 57.7(12.9) 37.4(11.8)

55.1(4.9) 54.3(5.9) 4.3(0.6) 45.8(9.2) 53.5(8.6) 56.8(11.2) 49.8(10.9)

52.7(5.1) 56.8(4.5) 4.2(0.8) 38.6(11.1) 52.6(10.7) 53.8(10.6) 47.5(10.8)

0.827 n.s. o 0.001 wHC,nwHC 4wBD, nw BD; wBD 4nwBD o 0.001 wHC 4wBD 4nwBD; nwHC 4nwBD o 0.001 wHC,nwHC 4 wBD,nwBD o 0.001 wHC 4wBD,nwBD o 0.001 wBD,nwBD4wHC, nwHC 0.11 n.s. 0.16 n.s. o 0.001 wHC 4wBD,nwBD; nwHC 4nwBD

0.065 0.016 0.033 0.006 0.007 o0.001 0.016

49.6(11.9)  0.2(0.7)  0.3(1.2)  0.6(0.7)

50.4(12.8)  0.5(0.6)  0.5(1.2)  1.1(1.1)

48.9(10.8)  0.2(0.5) 0.4(0.9)  0.1(0.9)

52.1(11.0)  0.2(0.6)  0.2(0.9)  0.3(0.8)

0.70 0.159 o 0.001 o 0.001

o0.001 0.006 0.077 0.104

Emotion processing CRSS VFPS

0.003(0.7)  0.1(0.9)  0.3(0.8)

 0.7(1.3)  0.3(0.8)  0.8(0.7)

0.2(0.7) 0.2(0.6) 0.04(0.6)

 0.2(0.7)  0.3(0.6)  0.1(0.7)

PSIR

 0.2(0.5)

 0.7(0.6)

0.2(0.5)

 0.06(0.6)

IC

 0.2(0.7)

 0.5(0.8)

0.05(0.6)

0.1(0.7)

n.s. n.s. n.s. n.s. n.s. n.s. n.s.

n.s. n.s. wHC 4w, nw BD wHC 4w, nwBD, nwHC 4 nwBD, wBD 4nwBD o 0.001 wHC 4nwBD 0.011 wHC 4w, nwBD o 0.001 wHC 4w, nwBD, nwHC 4 nwBD, wBD 4nwBD o 0.001 wHC 4w, nwBD, nwHC 4 wBD, wBD 4nwBD 0.008 wHC 4nwBD, nwHC 4nwBD

Partial eta squared

0.045

0.065

0.071 0.049 0.121 0.225 0.050

a

Chi-Square. WASI¼ Wecshler Adult Scale of Intelligence; HDRS-17¼ Hamilton Depression Rating Scale 17-item; YMRS ¼Young Mania Rating Scale; PCS: SF-36 ¼Physical Health Composite Score from SF-36; MCS: SF-36 ¼ Mental Health Composite Score from SF-36; NEO ¼ NEO Personality Inventory; VFPS ¼Verbal Fluency and Processing Speed; CRSS ¼ Conceptual Reasoning and Set-Shifting; PSIR ¼ Processing Speed with Interference Resolution; IC ¼ Inhibitory Control. 2.4. Data analyses All variables were screened for violations of assumptions associated with univariate and multivariate tests. The disproportionate influence of any given outlier was minimized by winsorizing (Tabachnick and Fidell, 2007). Analyses of variance (ANOVAs) or Chi-Square tests were used to test between-group differences in demographic variables. Multivariate analyses (MANOVA) were run to examine the effect of working group and bipolar status on clinical, personality, psychosocial, and cognitive variables. Statistical significance was set at 0.05 and Bonferroni corrections were made for multiple testing. Separate T-tests were run to evaluate the effect of job class stability and the key variables. Backward (Wald) logistic regressions were used to predict work status and estimate percent variance accounted for by variables of interest. Logistic regression models fit well according to the Hosmer–Lemeshow (2000) goodness-of-fit test. Each test was compared with the entire sample and then compared in with age-matched groups (by randomly removing younger participants from the working groups). There were no differences between age-matched groups and entire sub-group comparisons and so only the comparisons with entire subgroups are reported.

3. Results 3.1. Characteristics of working and nonworking BD participants Table 1 contains demographic characteristics of the working and not working BD and HC groups including sample size of each group. There was a significant difference between work status and diagnostic group (HC vs. BD), p ¼0.049, χ2(1, N ¼299) ¼3.87, with a greater proportion of BD participants in the not working group. Using all participants, work status groups did not differ based on gender, χ2(1, N ¼291) ¼0.41, p ¼ 0.68, or race,

χ2(1, N ¼273) ¼19.63, p ¼0.186. The majority of our sample (75.8%) identified as “White/Caucasian,” 8.4% as “Asian,” 9.5% as “Black/African American,” and 6.3% as “other race/more than one race.” Those who were in the working group were younger age, F (1, 295) ¼8.01, p ¼0.005, and more educated, F(1, 294) ¼24.1, p o 0.02. When examining just the BD group and just the HC groups alone, work status was not associated with significant differences on age or education among the BD patients. However, those HC participants who were in the working group were significantly younger and more educated than those in the not working HC group, p's o 0.015. As noted above, when a random subset of younger working HC and BD were removed to create a matched-groups analysis, the results were nearly identical for all analyses. As such, we elected to report the analyses with the entire sample. 3.2. Relationship between clinical, personality and psychosocial variables and work status. As shown in Table 1, BD participants in the not working group had greater number of mood episodes, lower GAF scores, and reported worse physical health on the SF-36 relative to the working BD group. There were no other significant differences between the working and not working BD participants on other clinical variables, including current mood symptoms on the HDRS and YMRS (p 40.05), proportion with history of psychosis or between groups for any of the five personality factors or the rating

Please cite this article as: Ryan, K.A., et al., Emotion perception and executive functioning predict work status in euthymic bipolar disorder. Psychiatry Research (2013), http://dx.doi.org/10.1016/j.psychres.2013.06.031i

4

K.A. Ryan et al. / Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎

of social support, F(8, 86) ¼1.76, p 40.09. There were no significant differences between the working HC and not working HC group on measures of subjective health, social support, or any of the five personality factors, F(8, 112) ¼ 1.19, p ¼0.314. Therefore, none of these variables were added as covariates in the comparison between groups for the cognitive factors. In a MANOVA using work status and both diagnostic groups, both HC groups reported better subjective health, more social support, and lower scores on the NEO neuroticism relative to the BD groups, as well as higher scores on the NEO conscientiousness, p's o0.001. A MANOVA with the eight cognitive factor scores as dependent variables and the four groups (working HC, not working HC, working BD, not working BD) as independent variables showed a significant group effect, F(24, 615) ¼ 3.06, p o0.001. All but Auditory Memory showed significant group differences (p ¼0.720). Post-hoc analyses showed that the working HC group performed similar to the not working HC group, except for lower scores on Visual Memory. The working HC outperformed both BD groups on the Fine Motor, Visual Memory, Verbal Fluency Processing Speed, Conceptual Reasoning Set-Shifting, and Processing Speed Interference Resolution factors and additionally the not working BD groups on the Inhibitory Control and Emotion Processing factors. The not working HC group was not significantly different from the working BD group on seven cognitive factors, performing better on Processing Speed Interference Resolution only. However, the not working BD performed worse than the not working HC on Fine Motor, Verbal Fluency Processing Speed, and Inhibitory Control factors (p's ¼0.009, 0.004, o0.001). The working BD group outperformed the not working BD group on the Fine Motor, Emotion Processing, Verbal Fluency Processing Speed and Processing Speed Interference Resolution factors (p's ¼ 0.010, 0.006, 0.001, and 0.002) (Fig. 1). A second analysis was completed in the BD group only. When using age and number of mood episodes as a covariate in comparing the working and not working BD groups, a follow-up MANCOVA showed the same effects. The working BD group outperformed the not working BD group on the Fine Motor, Emotion Processing, Verbal Fluency Processing Speed, and Processing Speed Interference Resolution factors (ps ¼ 0.041, 0.050, 0.006, and 0.020). The sample size for the not working BD group was modest, consisting of individuals who were both unemployed and disabled. In a targeted post-hoc analysis, the unemployed group (n ¼ 13) did not perform significantly different from the the disabled group (n ¼20), F(8, 16) ¼ 2.16, p¼ 0.09 for the cognitive factors. There also were no differences between the two not working BD subgroups on any of the clinical or personality variables or the measure of social support (p's 4 0.077), suggesting that these groups were appropriate to combine for analyses into one group. A logistic regression was used to evaluate the relative influences of significant variables (age, number of episodes, Fine Motor, Emotion Processing, Verbal Fluency Processing Speed, and Processing Speed Interference ) in predicting work status among the BD patients (working BD vs. not working BD). A test of the full model with the six predictors against the constant-only model was significant, χ2(4, N ¼ 103) ¼ 10.33, Nagelkerke R2 ¼0.14, p ¼0.025, indicating that the set of predictors reliably distinguished between working BD patients and not working BD patients. The model correctly classified 75.7% of cases, with 96.2% of the working BD patients correctly classified. Using the Wald criterion in follow-up, single regressor models, Emotion Processing, Verbal Fluency Processing Speed and Processing Speed Interference Resolution were predictive of work status (Table 2). A second logistic regression was used to evaluate the additional influence of age beyond the all cognitive factor scores in predicting diagnostic working group (working HC and working BD). A test of the full model with the nine predictors against the constant-only

model was significant, χ2(9, N ¼107)¼ 36.40, Nagelkerke R2 ¼0.255 p¼ 0.00003, indicating that the set of predictors reliably distinguished between working HC and working BD. The model correctly classified 69.2% of cases. Again, Processing Speed Interference Resolution made significant contributions above age and the other cognitive variables (Table 2) in predicting work status. 3.3. Job class stability/underemployment Among the BD participants who were in the working group, 25 (25%) declined in job class as compared to their most responsible position (underemployed), compared to 30.1% in the working HC group. The BD patients who declined in job class were not significantly different from BD patients in the not working group in terms of their most responsible job ever held, χ2(5, N ¼55) ¼ 6.44, p ¼0.27, or in terms of premorbid verbal intellect, F(1, 108) ¼ 3.845, p ¼0.10. However, those in the declined in job class group performed significantly better than those in the not working group on 3 of the 8 cognitive factors (Emotion Processing factor, t(48) ¼ 2.00, p ¼0.04, Verbal Fluency Processing Speed, t(48) ¼2.95, p¼ 0.005, and the Processing Speed Interference Resolution, t (45) ¼2.94, p ¼0.004). Those in the declined in job class group had significantly fewer lifetime mood episodes than the not working BD, t(39) ¼  2.10, p ¼0.04, but overall there was no difference between the two groups on other clinical and psychosocial factors.

4. Discussion Now, there is a limited understanding and few clinical tools that might be used to understand, predict, and improve life functioning, such as work status, in BD. The present study represents one of the few studies to show that specific aspects of cognitive dysfunction in euthymic BD are related to employment status using a well-powered sample, despite a more wellestablished literature base showing specific areas of cognitive dysfunction predicting employability in the schizophrenia population (e.g., attention, processing speed, and set-shifting) (McGurk and Mueser, 2004). Patients with BD who have better cognitive functioning are more likely to be employed. Those who are unemployed or unable to work exhibit greater difficulties in fine motor dexterity, processing emotional information, and on executive tasks comprising a set shifting or interference resolution component compared to BD patients who are gainfully employed. Cognitive factors, particularly those related to emotion processing, executive functioning, and auditory memory, account for a significant amount of variance in work status, independent of other clinical and demographic factors. BD patients have been shown to have greater dysfunction in the areas of executive functions, attention/concentration, and learning/memory compared to healthy controls (Martínez-Arán et al., 2004a,b), with persistent cognitive deficits in the remitted BD state (Clark et al., 2005; Dixon et al., 2004; El-Badri et al., 2001; Fleck et al., 2005; Martínez-Arán et al., 2004a,b); however, little is known how these persistent deficits impact specific areas of functional outcomes. The current findings add to the few studies indicating that cognitive dysfunction may negatively impact the specific area of employment, but the strength of our findings is linked to having a comparison, nonworking control group, such that effects of unemployment in the BD group can be dissociated from the effects related to greater impact of illness. Further, findings are strengthened by having a large sample of BD patients with an ability to examine the proportion those being unemployed in the BD group compared to a non-affected unemployed group.

Please cite this article as: Ryan, K.A., et al., Emotion perception and executive functioning predict work status in euthymic bipolar disorder. Psychiatry Research (2013), http://dx.doi.org/10.1016/j.psychres.2013.06.031i

K.A. Ryan et al. / Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎

5

Fig. 1. Cognitive factor difference scores between working and not working groups (working–nonworking) for each diagnosis.

Table 2 Odds ratios for binomial logistic regressions. Adjusted OR Work status (working BD/not working BD Age 0.975 # Mood episodes 0.993 Emotional processing 1.128 Fine motor 1.138 VFPS 2.004 PSIR 1.123 Diagnosis (working HC/working BD) Age 1.017 Fine motor 0.900 Visual memory 0.710 Auditory memory 1.091 Emotional processing 1.452 VFPS 1.506 CRSS 0.835 PSIR 0.167 IC 1.184

95% CI

p-value

(0.46–3.2) (0.93–1.43) (1.03–1.27) (0.87–1.57) (1.51–3.63) (1.28–4.49)

0.318 0.057 0.014 0.063 0.004 0.005

(0.99–1.47) (0.58–1.40) (0.49–1.03) (0.59–2.02) (0.81–2.61) (0.72–3.15) (0.46–1.50) (0.06–0.46) (0.68–2.07)

0.276 0.643 0.072 0.782 0.212 0.276 0.548 0.001 0.553

VFPS¼ Verbal Fluency and Processing Speed; CRSS ¼ Conceptual Reasoning and Set-Shifting; PSIR ¼ Processing Speed with Interference Resolution; IC ¼Inhibitory Control.

Not surprisingly, individuals with BD are less likely to be employed than controls with no psychiatric illness, but our findings indicate that patients with BD who maintained any employment, even if underemployed, exhibit greater cognitive dysfunction than their employed, non-affected healthy counterparts, consistent with recent findings that occupational recovery was associated with better cognitive performance (Bearden et al., 2011). Nearly 26% of the variance in our work status differentiation between diagnoses can be explained by cognitive factors. In particular, speed-related executive functioning measures appear to provide unique information in determining who falls into the employed diagnostic group. Bearden et al. (2011) found that working memory/attention and processing speed were associated with occupational recovery, though recovery was defined to include ratings of quantity and quality of work rather than a comparison of job position stability, which may account for differences in findings. In addition to cognitive factors predicting who may have BD and who may be employed within the BD group, cognitive factors is related to whether a BD patient is able to maintain employment,

regardless if they are underemployed. Employment was affected by emotional processing and executive functioning tasks that have set shifting or interference resolution components. Caution is warranted in over-interpreting these findings as our crosssectional study and analytical model cannot determine if cognitive functioning is a cause of occupational disability or underemployment or if those who work experience beneficial effects on their cognition as a result. It is tempting to speculate that our results could indicate that BD patients who exhibit difficulties on the job may require assessment of cognitive skills as these may be areas for targeted treatment, accommodations and/or remediation. It may be possible for those who can improve cognitive skills to retain and maintain employment. Need for accommodations or remediation can be identified early in the course of illness, with the opportunity for significantly increased productivity, quality, and job retention. Although not directly addressed in the literature, maintaining any type of employment may provide cognitive or other therapeutic benefits. This study has a number of strengths and a few limitations. Utilizing employment status as a functional outcome beyond using global assessment of functioning (GAF) scores provides more useful information to clinicians about a specific life domain in which specific treatments can be targeted. The GAF is heavily based on current symptomology and combines symptoms with overall level of functioning. While using employment status can be seen as a strength of this study, actual employment status does not capture quality or stability of work performance, a particularly important area in understanding functional recovery and warrants future investigation. There are also numerous situational and environmental factors that affect employment that are not routinely captured in our ongoing longitudinal study, but would important to include in future follow up studies (e.g., unemployment rates, access to transportation, regional employment opportunities). We included a large, well-characterized sample of euthymic BD and a well-matched comparison group. Number of mood episodes did not contribute to the prediction of work status despite the not working BD group having significantly more lifetime mood episodes (M ¼81.9) than the working BD group (M ¼30.9) yet equivalent years with the BD illness. This suggests that the not working group may be a sicker group in general and a closer examination of more specific illness severity factors should be examined further in future studies, Our not working

Please cite this article as: Ryan, K.A., et al., Emotion perception and executive functioning predict work status in euthymic bipolar disorder. Psychiatry Research (2013), http://dx.doi.org/10.1016/j.psychres.2013.06.031i

K.A. Ryan et al. / Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎

6

groups, especially the not working HC group was small and may limit some of the generalizations of these findings. However, we have found that including such a small not working HC group, no matter how small, is relatively rare in the literature. As is common among most naturalistic studies of this kind, it is difficult to control for the effects of medication, treatment adherence, and treatment optimization. We evaluated the impact of medication confounds, using a protocol often seen in the literature and accepted by BD researchers to assess total medication load (Almeida et al., 2009; Hassel et al., 2008; Sackeim, 2001) and did not show effects of most medication classes on employment status or a relationship between medication loading and cognition. Our cross-sectional design prevents examination of intraindividual differences in cognitive functioning over time. Future studies could investigate whether changes in cognitive functioning associated with disease progression might be a marker of employment likelihood. Furthermore, information about why or how long participants have been unemployed or disabled was not available to us and having this information might provide valuable insight into other factors that may impact employment status. 4.1. Implications Bipolar Disorder (BD) is a debilitating and chronic disorder with episodic symptoms. Despite clinical recovery, a substantial portion of BD patients are not working and this is well above the current levels of unemployment in the United States (Statistics, 2011), and a high number of BD patients are not working at their previous levels. Although the portion of unemployed BD patients is not substantially different from the control group, the cognitive factors related to unemployment status provide meaningful information. Treatment for BD generally focuses on clinical remission, but our results indicate that treatments should target areas of cognitive functioning as contributing factors to improve overall functional recovery, such as employability. These neuropsychological tools, with demonstrated ecological validity, can be used to determine the nature and responsiveness of treatment targets in an objective manner. A multidisciplinary approach to management of BD which integrates medication, psychosocial interventions and cognition would enhance functional recovery.

Acknowledgments We would like to acknowledge and thank Brennan Haase, Kathleen Hazlet, Nadia Huq, Lindsay Franti, Allison Kade, Michelle Kassel, Rachel Kay, Kortni Meyers, E. Michelle McFadden, and the rest of the staff of the Prechter Bipolar Research team for their contributions to this project. This research was supported by Heinz C. Prechter Bipolar Research Fund at the University of Michigan Depression Center (MGM, DM, ALW, KR, AV, MK, SAL).

Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.psychres.2013.06. 031.

References Altshuler, L., Tekell, J., Biswas, K., Kilbourne, A.M.P.M.P.H., Evans, D., Tang, D., Bauer, M.S.M.D., 2007. Executive function and employment status among veterans with bipolar disorder. Psychiatric Services 58, 1441–1447.

Almeida, J.R.C., Akkal, D., Hassel, S., Travis, M.J., Banihashemi, L., Kerr, N., Kupfer, D.J., Phillips, M.L., 2009. Reduced gray matter volume in ventral prefrontal cortex but not amygdala in bipolar disorder: significant effects of gender and trait anxiety. Psychiatry Research: Neuroimaging 171, 54–68. Armitage, S.G., 1946. An analysis of certain psychological tests used for the evaluation of brain injury. Psychological Monographs 60 i-48. Abbey, A., Abramis, D.J., Caplan, R.D., 1985. Effects of different sources of social support and social conflict on emotional well-being. Basic and Applied Social Psychology 6, 111–129. Bieling, P.J., MacQueen, G.M., Marriot, M.J., Robb, J.C., Begin, H., Joffe, R.T., Young, L.T., 2003. Longitudinal outcome in patients with bipolar disorder assessed by lifecharting is influenced by DSM-IV personality disorder symptoms. Bipolar Disorders 5, 14–21. Bearden, C.E., Shih, V.H., Green, M.F., Gitlin, M., Sokolski, K.N., Levander, E., Marusak, S., Hammen, C., Sugar, C.A., Altshuler, L.L., 2011. The impact of neurocognitive impairment on occupational recovery of clinically stable patients with bipolar disorder: a prospective study. Bipolar Disorders 13, 323–333. Benton, A.L., Hamsher, K.d., 1976. Multilingual Aphasia Examination. The University of Iowa, Iowa City, IA. Coryell, W., Scheftner, W., Keller, M., Endicott, J., Maser, J., Klerman, G.L., 1993. The enduring psychosocial consequences of mania and depression. The American Journal of Psychiatry 150, 720–727. Clark, L., Kempton, M.J., Scarnà, A., Grasby, P.M., Goodwin, G.M., 2005. Sustained attention-deficit confirmed in euthymic bipolar disorder but not in first-degree relatives of bipolar patients or euthymic unipolar depression. Biological Psychiatry 57, 183–187. Costa Jr., P.T., McCrae, R.R., 1992. NEO PI-R Professional Manual. Psychological Assessment Resources, Odessa, FL. Dickerson, F.B., Boronow, J.J., Stallings, C.R., Origoni, A.E., Cole, S., Yolken, R.H., 2004. Association between cognitive functioning and employment status of persons with bipolar disorder. Psychiatric Services 55, 54–58. Dixon, T., Kravariti, E., Frith, C., Murray, R.M., McGuire, P.K., 2004. Effect of symptoms on executive function in bipolar illness. Psychological Medicine 34, 811–821. Davis, J.M., Chen, N., 2004. Dose response and dose equivalence of antipsychotics. Journal of Clinical Psychopharmacology 24, 192–208. Delis, D., Kaplan, E., Kramer, J., Ober, B.A., 2000. California Verbal Learning Test-II. The Psychological Corporation, San Antonio, TX. El-Badri, S.M., Ashton, C.H., Moore, P.B., Marsh, V.R., Ferrier, I.N., 2001. Electrophysiological and cognitive function in young euthymic patients with bipolar affective disorder. Bipolar Disorders 3, 79–87. Fleck, D.E., Shear, P.K., Strakowski, S.M., 2005. Processing efficiency and sustained attention in bipolar disorder. Journal of the International Neuropsychological Society 11, 49–57. Goldberg, J.F., Harrow, M., Grossman, L.S., 1995a. Course and outcome in bipolar affective disorder: a longitudinal follow-up study. The American Journal of Psychiatry 152, 379–384. Goldberg, J.F., Harrow, M., Grossman, L.S., 1995b. Recurrent affective syndromes in bipolar and unipolar mood disorders at follow-up. The British Journal of Psychiatry 166, 382–385. Gutiérrez-Rojas, L., Jurado, D., Gurpegui, M., 2011. Factors associated with work, social life and family life disability in bipolar disorder patients. Psychiatry Research 186, 254–260. Green, P.W., Allen, L.M., 1997. The Emotional Perception Test. CogniSyst Inc, Durham, NC. Golden, C., 1978. Stroop Color and Word Test. Stoelting, Chicago, IL. Hamilton, M., 1967. Development of a rating scale for primary depressive illness. British Journal of Social Clinical Psychology 6, 278–296. Hassel, S., Almeida, J.R.C., Kerr, N., Nau, S., Ladouceur, C.D., Fissell, K., Kupfer, D.J., Phillips, M.L., 2008. Elevated striatal and decreased dorsolateral prefrontal cortical activity in response to emotional stimuli in euthymic bipolar disorder: no associations with psychotropic medication load. Bipolar Disorders 10, 916–927. Heaton, R.K., 1981. A Manual for the Wisconsin Card Sorting Test. Psychological Assessment Resources, Odessa, FL. Hosmer, D.W., Lemeshow, S., 2000. Applied Logistic Regression, 2nd ed. John Wiley & Sons, Inc, New York. Jaeger, J., Berns, S., Loftus, S., Gonzalez, C., Czobor, P., 2007. Neurocognitive test performance predicts functional recovery from acute exacerbation leading to hospitalization in bipolar disorder. Bipolar Disorders 9, 93–102. Keck Jr., P.E., 2006. Long-term management strategies to achieve optimal function in patients with bipolar disorder. The Journal of Clinical Psychiatry 67 (9), 19–24. (discussion 36–42). Leonardi, M., Sanchez-Moreno, J., Martinez-Aran, A., 2007. Detection of functional impairments in euthymic bipolar patients using the international classification of functioning, disability and health (ICF). In: 160th Annual Meeting of American Psychiatric Association, San Diego, CA. Lezak, M., 1995. Neuropsychological Assessment, 3rd ed. Oxford University Press, New York. Langenecker, S.A., Zubieta, J.K., Young, E.A., Akil, H., Nielson, K.A., 2007a. A task to manipulate attentional load, set-shifting, and inhibitory control: convergent validity and test-retest reliability of the parametric Go/No-Go test. Journal of Clinical and Experimental Neuropsychology 29, 842–853.

Please cite this article as: Ryan, K.A., et al., Emotion perception and executive functioning predict work status in euthymic bipolar disorder. Psychiatry Research (2013), http://dx.doi.org/10.1016/j.psychres.2013.06.031i

K.A. Ryan et al. / Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎ Langenecker, S.A., Saunders, E.F.H., Kade, A.M., Ransom, M.T., McInnis, M.G., 2010. Intermediate: cognitive phenotypes in bipolar disorder. Journal of Affective Disorders 122, 285–293. Langenecker, S.A., Caveney, A.F., Giordani, B., Young, E.A., Nielson, K.A., Rapport, L.J., Bieliauskas, L.A., Mordhorst, M.J., Marcus, S., Yodkovik, N., Kerber, K., Berent, S., Zubieta, J.-K., 2007b. The sensitivity and psychometric properties of a brief computer-based cognitive screening battery in a depression clinic. Psychiatry Research 152, 143–154. Murray, C.J.L., Lopez, A.D., 1997. Global mortality, disability, and the contribution of risk factors: global burden of disease study. The Lancet 349, 1436–1442. Medard, E., Dubertret, C., Peretti, C.S., Ades, J., D’Escatha, A., 2010. Descriptive study of the occupational outcome of bipolar patients. Journal of Occupational Rehabilitation 20, 293–298. Mann-Wrobel, M.C., Carreno, J.T., Dickinson, D., 2011. Meta-analysis of neuropsychological functioning in euthymic bipolar disorder: an update and investigation of moderator variables. Bipolar Disorders 13, 334–342. Martínez-Arán, A., Vieta, E., Colom, F., Torrent, C., Sánchez-Moreno, J., Reinares, M., Benabarre, A., Goikolea, J.M., Brugué, E., Daban, C., Salamero, M., 2004a. Cognitive impairment in euthymic bipolar patients: implications for clinical and functional outcome. Bipolar Disorders 6, 224–232. Martínez-Arán, A., Vieta, E., Reinares, M., Colom, F., Torrent, C., Sánchez-Moreno, J., Benabarre, A., Goikolea, J.M., Comes, M., Salamero, M., 2004b. Cognitive function across manic or hypomanic, depressed, and euthymic states in bipolar disorder. The American Journal of Psychiatry 161, 262–270. McGurk, S.R., Mueser, K.T., 2004. Cognitive functioning, symptoms, and work in supported employment: a review and heuristic model. Schizophrenia Research 70, 147–173. Meyers, J., Meyers, K., 1995. Rey Complex Figure and Recognition Trial: Professional Manual. Psychological Assessment Resources, Odessa, FL. Nurnberger Jr., J.I., Blehar, M.C., Kaufmann, C.A., York-Cooler, C., Simpson, S.G., Harkavy-Friedman, J., Severe, J.B., Malaspina, D., Reich, T., 1994. Diagnostic interview for genetic studies. Rationale, unique features, and training. NIMH Genetics Initiative. Archives of General Psychiatry 51, 849–859. (discussion 863–844). O’Shea, R., Poz, R., Michael, A., Berrios, G.E., Evans, J.J., Rubinsztein, J.S., 2010. Ecologically valid cognitive tests and everyday functioning in euthymic bipolar disorder patients. Journal of Affective Disorders 125, 336–340. Pope, M., Dudley, R., Scott, J., 2007. Determinants of social functioning in bipolar disorder. Bipolar Disorders 9, 38–44.

7

Rapport, L.J., Friedman, S.R., Tzelepis, A., Van Voorhis, A., 2002. Experienced emotion and affect recognition in adult attention-deficit hyperactivity disorder. Neuropsychology 16, 102–110. Ryan, K.A., Vederman, A.C., McFadden, E.M., Weldon, A.L., Kamali, M., Langenecker, S.A., McInnis, M.G., 2012. Differential executive functioning performance by phase of bipolar disorder. Bipolar Disorders 14, 527–536. Staner, L., Tracy, A., Dramaix, M., Genevrois, C., Vanderelst, M., Vilane, A., Bauwens, F., Pardoen, D., Mendlewicz, J., 1997. Clinical and psychosocial predictors of recurrence in recovered bipolar and unipolar depressives: a one-year controlled prospective study. Psychiatry Research 69, 39–51. Sanchez-Moreno, J., Martinez-Aran, A., Tabares-Seisdedos, R., Torrent, C., Vieta, E., Ayuso-Mateos, J.L., 2009. Functioning and disability in bipolar disorder: an extensive review. Psychother Psychosom 78, 285–297. Sackeim, H.A., 2001. The definition and meaning of treatment-resistant depression. The Journal of Clinical Psychiatry 62 (16), 10–17. Statistics, U.S.D.o.L.B.o.L., 2011. The Employment Situation – October 2011, Washington, DC. Tohen, M., Hennen, J., Zarate Jr., C.M., Baldessarini, R.J., Strakowski, S.M., Stoll, A.L., Faedda, G.L., Suppes, T., Gebre-Medhin, P., Cohen, B.M., 2000. Two-year syndromal and functional recovery in 219 cases of first-episode major affective disorder with psychotic features. The American Journal of Psychiatry 157, 220–228. Tabachnick, B., Fidell, L., 2007. Using Multivariate Statistics, 5th ed. Pearson Allyn & Bacon, Boston, MA. Vieta, E., Cieza, A., Stucki, G., Chatterji, S., Nieto, M., Sánchez-Moreno, J., Jaeger, J., Grunze, H., Ayuso-Mateos, J.L., 2007. Developing core sets for persons with bipolar disorder based on the international classification of functioning, disability and health. Bipolar Disorders 9, 16–24. Vinokur, A.D., Van Ryn, M., 1993. Social support and undermining in close relationships: their independent effects on the mental health of unemployed persons. Journal of Personality and Social Psychology 65, 350–359. Wechsler, D., 1997. Wechsler Adult Intelligence Scale-III. The Psychological Corporation, Cleveland, OH. Wechsler, D., 1999. Wechsler Abbreviated Scale of Intelligence (WASI) Manual. The Psychological Corporation, San Antonio, TX. Ware, J.E., 1996. The SF-36 health survey. In: Spilker, B. (Ed.), Quality of Life and Pharmacoeconomics in Clinical Trials, 2nd ed. Lippincott-Raven, Philadelphia, pp. 337–346. Young, R.C., Biggs, J.T., Ziegler, V.E., Meyer, D.A., 1978. A rating scale for mania: reliability, validity and sensitivity. The British Journal of Psychiatry 133, 429–435.

Please cite this article as: Ryan, K.A., et al., Emotion perception and executive functioning predict work status in euthymic bipolar disorder. Psychiatry Research (2013), http://dx.doi.org/10.1016/j.psychres.2013.06.031i