Original Article

Sustained unemployment in psychiatricoutpatients with bipolar disorder: frequencyand association with demographic variablesand comorbid disorders

Zimmerman M, Galione JN, Chelminski I, Young D, Dalrymple K,Ruggero CJ. Sustained unemployment in psychiatric outpatients withbipolar disorder: frequency and association with demographicvariables and comorbid disorders.Bipolar Disord 2010: 12: 720–726. ª 2010 The Authors.Journal compilation ª 2010 John Wiley & Sons A ⁄S.

Objectives: The negative impact of bipolar disorder on occupationalfunctioning is well established. However, few studies have examinedthe persistence of unemployment, and no studies have examined theassociation between diagnostic comorbidity and sustainedunemployment. In the present report from the Rhode Island Methodsto Improve Diagnostic Assessment and Services (MIDAS) project, wedescribed the amount of time unemployed in the five years beforethe evaluation in a large cohort of outpatients diagnosed with bipolardisorder, and determined the demographic and clinical correlatesof sustained unemployment.

Methods: A total of 206 patients diagnosed with DSM-IV bipolar I orbipolar II disorder were interviewed with semi-structured interviewsassessing comorbid Axis I and Axis II disorders, demographic andclinical variables. The interview included an assessment of the amountof time missed from work due to psychiatric reasons during the pastfive years. Persistent unemployment was defined as missing up to twoyears or more from work.

Results: Less than 20% of the patients reported not missing any timefrom work due to psychiatric reasons, and more than one-third missedup to two years or more from work. Prolonged unemployment wasassociated with increased rates of current panic disorder and a lifetimehistory of alcohol abuse or dependence. Patients with prolongedunemployment were older and experienced more episodes of depression.

Conclusions: Most patients presenting for the treatment of bipolardisorder have missed some time from work due to psychiatric reasons,and the persistence of employment problems is considerable. Comorbidpsychiatric disorders are a potentially treatable risk factor for sustainedunemployment. It is therefore of public health significance to determineif current treatments are effective in bipolar disorder patients withcurrent panic disorder, and if not, to attempt to develop treatmentsthat are effective.

Mark Zimmerman, Janine N Galione,Iwona Chelminski, Diane Young,Kristy Dalrymple andCamilo J Ruggero

Department of Psychiatry and Human Behavior,

Brown Medical School, Providence, RI, USA

doi: 10.1111/j.1399-5618.2010.00869.x

Key words: anxiety disorders – bipolar disorder –

comorbidity – unemployment

Received 17 February 2010, revised and accepted

for publication 20 August 2010

Corresponding author:

Mark Zimmerman, M.D.

Bayside Medical Center

235 Plain Street

Providence, RI 02905, USA

Fax: 401-441-7107

E-mail: mzimmerman@lifespan.org

The authors of this paper do not have any competing commercial or financial interests to disclose in connection with this manuscript.

Bipolar Disorders 2010: 12: 720–726 ª 2010 John Wiley and Sons A/S

BIPOLAR DISORDERS

720

Bipolar disorder is a serious illness resulting insignificant psychosocial morbidity, reducedhealth-related quality of life, and excess mortality.Recent research has suggested that bipolar disor-der, when defined to include milder variants suchas bipolar II disorder and bipolar disorder nototherwise specified, is more prevalent than hadbeen previously reported (1, 2). Bipolar disorderincurs disproportionately higher costs on patients,the mental health care system, and society thando most other mental illnesses, and is ranked asone of the leading causes of disability worldwide(3). The disorder costs significantly more thanmost other mental illnesses to treat (4, 5) andplaces a stark burden on patients, includingincreased risk of suicide (6, 7) and profounddisruptions in work and social functioning (8–10).Over half of the patients with bipolar I disorderwill have made an attempt to end their own life(8), with many patients eventually succeeding(6, 7).Bipolar disorder disrupts all domains of func-

tioning including occupational functioning. In areview of the literature, Dean and colleagues (11)found that compared to patients with otherpsychiatric disorders, bipolar disorder was associ-ated with elevated rates of sustained unemploy-ment, absenteeism, and poor work performance.The most consistent correlates of functionalimpairment in bipolar disorder patients are thepresence and persistence of depressive symptoms(12, 13). Little research, however, has examinedwhether unemployment was associated with diag-nostic comorbidity.Epidemiologic and clinical studies have found

that individuals with bipolar disorder have highrates of anxiety, substance use, impulse control,and personality disorders (1, 14, 15). In patientswith bipolar disorder, comorbidity has been foundto be associated with greater chronicity at the timeof the evaluation, poorer longitudinal course, andinferior response to treatment (16–18). Althoughcomorbidity has been associated with symptompersistence, and symptom persistence has beenassociated with occupational impairment, we areaware of only one study of the association betweencomorbidity and employment status. McElroy andcolleagues (14) reported that compared to patientswith no current comorbid disorders, patients withone or more current comorbid axis I disorders hada significantly higher rate of ‘‘limitations of occu-pational functioning.’’ If comorbidity is associatedwith more frequent and persistent unemploymentin bipolar disorder patients, this has clinical andpublic health significance because comorbidity ispotentially treatable.

Many studies of occupational functioning inbipolar disorder have relied on scales to dimen-sionally characterize the degree of functionalimpairment (12, 13, 17) rather than examiningunemployment rates. In fact, we are not aware ofany studies that have examined the amount of timepatients with bipolar disorder report missing workbecause of psychiatric illness. Nor are we aware ofany studies that have attempted to identify theclinical characteristics of patients with bipolardisorder who have lengthy periods of unemploy-ment. In the present report from the Rhode IslandMethods to Improve Diagnostic Assessment andServices (MIDAS) project, we described theamount of time patients report being unemployedin the five years before the evaluation, and deter-mined the demographic and clinical correlates ofsustained unemployment. We hypothesized thatprolonged unemployment would be associatedwith greater symptom chronicity, poorer socialsupport, and a higher frequency of comorbidpsychiatric disorders.

Methods

The Rhode Island MIDAS project represents anintegration of research methodology into a com-munity-based outpatient practice affiliated with anacademic medical center (19). A comprehensivediagnostic evaluation is conducted upon presenta-tion for treatment. This private practice grouppredominantly treats individuals with medicalinsurance (including Medicare but not Medicaid)on a fee-for-service basis, and it is distinct from thehospital�s outpatient residency training clinic thatpredominantly serves lower income, uninsured,and medical assistance patients. Data on referralsource was recorded for the last 1,200 patientsenrolled in the study. Patients were most frequentlyreferred from primary care physicians (31.6%),psychotherapists (15.8%), and family members orfriends (17.6%).The present report is based on 206 patients

diagnosed with DSM-IV bipolar I or bipolar IIdisorder. Twenty additional patients diagnosedwith bipolar disorder were excluded from theanalysis because they were not expected to workdue to being students, retired, physically disabled,or primarily responsible for running the household.The patients were interviewed by a diagnostic raterwho administered a modified version of the Struc-tured Clinical Interview for DSM-IV (SCID) (20)and the Structured Interview for DSM-IV Person-ality (SIDP-IV) (21). During the course of theMIDAS project the assessment battery has beenmodified at times. The assessment of all DSM-IV

Sustained unemployment in outpatients with bipolar disorder

721

personality disorders was not introduced until thestudy was well underway and the proceduraldetails of incorporating research interviews intoour clinical practice had been well established,though we had introduced the assessment ofborderline and antisocial personality disorder atthe beginning of the study. In June, 2008 westopped administering the full SIDP-IV and con-tinued to only administer the borderline personal-ity disorder module. We therefore only includedborderline personality disorder in the analysisbecause it was the only personality disorderassessed in all patients.The interview included items from the Schedule

for Affective Disorders and Schizophrenia(SADS) (22) on best level of social functioningduring the past five years and during adolescence.These items were rated on a 6- or 7-point scalewhich was dichotomized as good or better func-tioning (good, very good, superior) and fair orworse functioning (fair, poor, very poor, grosslyinadequate). We also included from the SADS theitem assessing the amount of time missed fromwork due to psychiatric reasons during the pastfive years. On the SADS, the amount of timemissed from work is categorized as follows:0 = did not work at all because was not expectedto work (e.g., retired, student, housewife, physi-cally ill, or some other reason not related topsychopathology); 1 = virtually no time at all outof work or absenteeism unrelated to psychopa-thology; 2 = only a few days to 1 month; 3 = upto 6 months; 4 = up to 1 year; 5 = up to2 years; 6 = up to 3 years; 7 = up to 4 years;8 = up to almost 5 years; 9 = worked none, orpractically none of the time because of reasonsrelated to psychopathology. At the beginning ofthe study we only scored the SADS item regard-ing the amount of time missed from work.Approximately midway through the project webegan recording the actual number of dayspatients reported that they missed from work inorder to get a more precise estimate of thenumber of days unemployed. This informationwas collected for 85 of the 206 patients.The SCID ⁄SADS interview included assess-

ments of prior psychiatric hospitalizations, lifetimehistory of suicide attempts, number of episodes ofdepression, and duration of the current episode.The Clinical Global Index of depression severity(23) was rated on all patients.As described elsewhere, the diagnostic raters

were highly trained and monitored throughout theproject to minimize rater drift (19). Reliability wasexamined in 65 patients. A joint-interview designwas used in which one rater observed another

conducting the interview, and both raters indepen-dently made their ratings. For disorders diagnosedin at least two patients by at least one of the tworaters, the Kappa coefficients were: major depres-sive disorder (j = 0.90), dysthymic disorder(j = 0.88), bipolar disorder (j = 0.75), panicdisorder (j = 0.95), social phobia (j = 0.84),obsessive compulsive disorder (j = 1.0), specificphobia (j = 0.93), generalized anxiety disorder(j = 0.85), posttraumatic stress disorder (j =0.87), alcohol abuse ⁄dependence (j = 0.64), drugabuse ⁄dependence (j = 0.64), any somatoformdisorder (j = 1.0), and borderline personalitydisorder (j = 1.0). The intraclass correla-tion coefficient of reliability for the SADSitem �time missed from work� was also high(ICC = 0.96).We compared the demographic, clinical, and

diagnostic characteristics of patients with bipolardisorder who were and were not unemployed forup to two years or more during the five years priorto the evaluation. T-tests were used to compare thetwo groups on continuously distributed variables.Categorical variables were compared by the chi-square statistic, or by Fisher�s Exact Test if theexpected value in any cell of a 2 · 2 table was lessthan 5. After the univariate analyses, we conducteda stepwise logistic regression analysis to determinewhich of the variables were independently associ-ated with prolonged unemployment. Only thosevariables that were significant (p < 0.05) in theunivariate analyses were included in the regressionanalysis.

Results

The data in Table 1 show the demographic char-acteristics of the 206 patients. The majority of thesubjects were white, female, and married or single.More patients were diagnosed with DSM-IVbipolar II disorder (56.3%, n = 116) than bipolarI disorder (43.7%, n = 90). Because this is atreatment seeking sample, it is not surprising thatthe majority of the patients were in episode at thetime of the evaluation (68.4%, n = 141). Almostone-quarter of the patients were in partial remis-sion (24.3%, n = 50), and less than 10% were inremission (7.3%, n = 15).Only 36 (17.5%) patients reported not missing

any time from work due to psychiatric reasons(Table 1). A total of 83 (40.3%) patients reportedmissing only a few days up to six months of workduring the past five years, and approximately one-third (34.5%, n = 71) missed up to two years ormore. We defined prolonged unemployment asmissing up to two years or more of work. For the

Zimmerman et al.

722

entire group the median number of weeks missedfrom work during the five years prior to theevaluation was 15.2 weeks.

Prolonged unemployment was not significantlyassociated with either remission status or bipolarsubtype, therefore, the data was analyzed for theentire group. Prolonged unemployment was asso-ciated with significantly older age (39.1 ± 9.5versus 34.1 ± 11.5, t = 3.1, p < 0.01), but notwith gender, education, or race. Patients who didnot work for psychiatric reasons for up to twoyears or more were diagnosed with significantlymore current Axis I disorders (2.2 ± 1.7 versus1.6 ± 1.5, t = 2.5, p < 0.05). The data in Table 2show that prolonged unemployment was associ-ated with increased rates of current panic disorder,posttraumatic stress disorder, and borderline per-sonality disorder. The results were the same whenlifetime rates of diagnosis were examined exceptthat the difference in rate of an alcohol use disorderwas now significant [60.6% versus 45.9%, oddsratio (OR) = 1.8, 95% confidence interval (CI):1.0–3.2, p < 0.05].In contrast to bipolar disorder patients who had

not been unemployed for more than two years ofthe prior five years, the patients with prolongedunemployment reported poorer adolescent andcurrent social functioning (Table 3). Patients withprolonged unemployment experienced more epi-sodes of depression and were more likely to havebeen hospitalized for psychiatric reasons.We conducted a regression analysis in order to

examine the independent correlates of prolongedunemployment. Only diagnostic and psychosocialvariables that were significantly different betweenthe two groups were entered into the analysis.

Table 1. Demographic characteristics and time missed from work due topsychopathology in 206 psychiatric outpatients with bipolar disorder

Characteristic

Gender, n (%)Female 122 (59.2)Male 84 (40.8)

Education, n (%)Less than high school 13 (6.3)Graduated high school 136 (66.0)Graduated college or greater 57 (27.7)

Marital status, n (%)Married 72 (35.0)Living with someone 11 (5.3)Widowed 1 (0.5)Separated 16 (7.8)Divorced 33 (16.0)Never married 73 (35.4)

Race, n (%)White 188 (91.3)Black 4 (1.9)Hispanic 3 (1.5)Asian 1 (0.5)Other 10 (4.9)

Age, years, mean (SD) 35.8 (11.1)Time missed from work, n (%)

Almost none 36 (17.5)Only a few days up to 1 month 43 (20.9)Up to 6 months 40 (19.4)Up to 1 year 16 (7.8)Up to 2 years 25 (12.1)Up to 3 years 10 (4.9)Up to 4 years 7 (3.4)Up to 5 years 29 (14.0)

Table 2. Univariate analyses of current DSM-IV disorders in psychiatric outpatients with bipolar disorder who were and were not unemployed for at least twoyears in the five years before presenting for treatment

Prolongedunemployment(n = 71)

No prolongedunemployment(n = 135)

Oddsratio 95% CI p-value

Anxiety disorders, n (%)Panic disorder 27 (38.0) 26 (19.3) 2.6 1.4–4.9 <0.01Agoraphobia without history of panic 2 (2.8) 1 (0.7) 3.9 0.3–43.6 n.s.Social phobia 31 (43.7) 47 (34.8) 1.5 0.8–2.6 n.s.Specific phobia 15 (21.1) 16 (11.9) 2.0 0.9–4.3 n.s.Posttraumatic stress disorder 26 (36.6) 19 (14.1) 3.5 1.8–7.0 <0.001Generalized anxiety disorder 11 (15.5) 37 (27.4) 0.5 0.2–1.0 n.s.Obsessive-compulsive disorder 14 (19.7) 19 (14.1) 1.5 0.7–3.2 n.s.Any anxiety disorder 53 (74.6) 86 (63.7) 1.7 0.9–3.2 n.s.

Substance use disorders, n (%)Alcohol abuse ⁄ dependence 6 (8.5) 12 (8.9) 0.9 0.3–2.6 n.s.Drug abuse ⁄ dependence 2 (2.8) 10 (7.4) 0.4 0.1–1.7 n.s.Any substance use disorder 8 (11.3) 20 (14.8) 0.7 0.3–1.8 n.s.

Any eating disorder, n (%) 9 (12.7) 12 (8.9) 1.5 0.6–3.7 n.s.Any somatoform disorder, n (%) 7 (9.9) 8 (5.9) 1.7 0.6–5.0 n.s.Any impulse control disorder, n (%) 4 (5.6) 10 (7.4) 0.7 0.2–2.5 n.s.Borderline personality disorder, n (%) 26 (36.6) 28 (20.7) 2.2 1.2–4.2 <0.05

n.s. = non significant.

Sustained unemployment in outpatients with bipolar disorder

723

Sustained unemployment was significantly associ-ated with current panic disorder (b = 0.8,OR = 2.3, p < 0.05), a lifetime history of alcoholuse disorder (b = 0.5, OR = 1.6, p < 0.05), olderage (b = 0.1, OR = 1.1, p < 0.01), and greaternumber of depressive episodes (b = 0.1,OR = 1.1, p < 0.01).

Discussion

Chronic illnesses such as bipolar disorder havemultiple effects on occupational functioning, rang-ing from reduced work efficiency to occasionalmissed days from work to sustained periods ofunemployment to permanent disability. In oursample of psychiatric outpatients with bipolardisorder presenting for treatment, we found thatmost patients reported missing some work in thepast five years due to psychiatric symptoms, andthat brief interludes of missing work were morecommon than long periods of unemployment.Because the total number of days missed fromwork was much greater in the group with pro-longed unemployment, we focused on the demo-graphic and clinical variables characterizing thepatients with sustained unemployment.Approximately one-third of the patients had

been unable to work for up to two years and longerin the five years preceding the evaluation. It isdifficult to compare the prevalence and amount ofmissed work in the current study with other studiesbecause various definitions of sustained unemploy-ment have been used. For example, Morgan andcolleagues (10) found that half of the patients they

interviewed were ‘‘mainly unemployed over theprevious 12 months’’ and that two-thirds wereunemployed at the time of the evaluation. In theStanley Center Bipolar Disorder Registry, half ofthe participants were not working (24). In the first1,000 patients enrolled in the Systematic TreatmentEnhancement Program for Bipolar Disorder(STEP-BD), the rate of unemployment was lowerthan these other studies (22.0% unemployed and15.3% disabled); however, half of the patients werein recovery at the time of the baseline evaluation(25). In the Collaborative Depression Study theunemployment rate in patients with bipolar dis-order was 39.5% in the last year of a five-yearfollow-up (26). The enduring nature of occupa-tional role impairment associated with bipolardisorder was demonstrated in the Iowa 500 studywhich found that 24% of the bipolar disorderpatients were ‘‘occupationally incapacitated due tomental illness’’ 30 to 40 years after their indexhospitalization (27). Regardless of the definition,all studies have found a high rate of unemploymentin patients with bipolar disorder.Reviews of the impact of bipolar disorder on

employment have focused on such clinical vari-ables as age of onset, presence and persistence ofmood symptoms, cognitive functioning, psychoticfeatures, and social support (28). We are not awareof any prior studies that have examined the impactof the breadth of diagnostic comorbidity onunemployment in patients with bipolar disorder.Because almost all psychiatric disorders have beenassociated with impairment in occupational func-tioning, we expected comorbidity to be associated

Table 3. Univariate analyses of psychosocial characteristics of psychiatric outpatients with bipolar disorder who were and were not unemployed for at least twoyears in the five years before presenting for treatment

Prolongedunemployment(n = 71)

No prolongedunemployment(n = 135) t or v2 p-value

Fair-poor past social functioning, n (%) 39 (54.9) 30 (22.2) 22.3 <0.001Fair-poor current social functioning, n (%) 40 (56.3) 40 (29.6) 14.0 <0.001Age of onset, mean (SD) 9.3 (17.2) 9.9 (17.1) 0.1 n.s.No. of depressive episodes, mean (SD)a 8.5 (11.8) 7.6 (7.6) 3.5 0.001Current episode direction, weeks, medianb 27.0 31.0 )1.3c n.s.Chronic depression > 2 years duration, n (%)b 22 (39.3) 26 (26.5) 3.7 n.s.Clinical Global Index of Severity, mean (SD) 1.4 (2.6) 1.3 (2.2) 2.0 <0.05Severity of current suicidal ideation, mean (SD) 1.5 (1.3) 1.3 (1.1) 1.3 n.s.History of suicide attempt, n (%) 33 (46.5) 49 (36.3) 2.0 n.s.History of psychiatric hospitalization, n (%) 45 (63.4) 63 (46.7) 5.2 <0.05

aMaximum number of episodes coded was 20.bData was missing for two subjects. Episode duration refers to duration of depression and this was not rated for 50 subjects who werenot in an episode of depression at the time of the evaluation, leaving a final sample of 56 for prolonged unemployment and 98 for noprolonged unemployment.cThis value is for z.n.s. = not significant.

Zimmerman et al.

724

with poorer functioning. While several disorderswere associated with sustained unemployment inthe univariate analyses, in the logistic regressionanalysis only current panic disorder and lifetimealcohol use disorder were found to be independentcorrelates of sustained unemployment.These results are consistent with other studies

that have found that comorbid panic disorder isassociated with poorer quality of life, greaterpsychosocial impairment, and poorer outcome inpatients with bipolar disorder. For example, Feskeand colleagues (29) found that a history of panicattacks predicted poorer response to acute treat-ment of patients with bipolar I disorder. In anotherreport from the Pittsburgh group, Frank andcolleagues (18) found that elevated scores on thePanic-Agoraphobic Spectrum–Self Report scalepredicted higher levels of depressive symptomsduring acute phase treatment of patients withbipolar I disorder and longer time to achieveremission. Other studies have found that otheranxiety disorders have been associated with fewerdays well, poorer psychosocial functioning andquality of life, delayed time to recovery, quickertime to relapse, and lower likelihood of achievingeuthymia (17, 30).There is thus accumulating evidence of the

clinical and public health significance of comorbidanxiety disorders in bipolar disorder. This high-lights the importance of examining the efficacy ofestablished treatments in patients with this combi-nation of disorders, and the development of newtreatment strategies for this group. Recent reviewsof the treatment of bipolar disorder with comorbidanxiety disorders have focused on the efficacy ofmood stabilizers for anxiety disorders, and theefficacy and possible risks of anti-anxiety medica-tions such as antidepressants in patients withbipolar disorder (15, 31). Noteworthy was the lackof a single study of the treatment of patients withboth bipolar disorder and an anxiety disorder.Consequently, treatment recommendations forpatients with bipolar disorder and comorbid anx-iety disorders have been based on extrapolationsfrom studies of the treatment of each disorderindividually (15, 31).A lifetime history of alcohol use disorder also

predicted persistent unemployment. In markedcontrast to the lack of empirical study of thetreatment of patients with comorbid bipolar andanxiety disorder, there are a small number ofcontrolled and uncontrolled studies of pharmaco-therapy and psychotherapy of bipolar disordercomorbid with substance use disorder (31).Some limitations of this study should be

considered. The most obvious limitation is that

the duration of unemployment was based onpatients� retrospective reports rather than pro-spective observation or reviewing patients�employment records. It is possible that patientsoverestimated the amount of time that they wereunemployed or the amount of time unemploy-ment was attributable to psychiatric illness.Although studies comparing self-reported absen-teeism with employment records have found highcorrelations between the assessments (32, 33), nostudies have examined accuracy of self-reportsover a five-year period. It is also possible thatcertain forms of psychopathology are associatedwith biased reporting of either time missed fromwork, or attributing time unemployed to psychi-atric illness. We are not aware of any studiesexamining this issue.The assessment of unemployment due to psy-

chiatric illness did not attempt to differentiatewhich types of pathology might have been respon-sible for the unemployment. Thus, we cannotattribute the unemployment entirely to the symp-toms of bipolar disorder. Also, we did not assesscognitive functioning, which has been found to beassociated with functional capacity (34).The study was conducted in a single outpatient

practice in which the majority of patients werewhite, female, and had health insurance. Replica-tion of the results in samples with differentdemographic characteristics is warranted. Wewould hypothesize that in clinics treating unin-sured patients the rates of sustained unemploymentwill be higher.In conclusion, the results of the present study

indicate that not only do a large percentage ofpatients presenting for the treatment of bipolardisorder miss some time from work due to psychi-atric reasons, but that the persistence of employ-ment problems is considerable. Repeated episodesof depression and comorbid panic and alcohol usedisorder were independent correlates of persistentunemployment. It is therefore of public healthsignificance to determine if current treatments areeffective in bipolar disorder patients with currentanxiety or alcohol use disorders, and if not, toattempt to develop treatments that are effective.

References

1. Merikangas KR, Akiskal HS, Angst J et al. Lifetime and12-month prevalence of bipolar spectrum disorder in theNational Comorbidity Survey replication. Arch Gen Psy-chiatry 2007; 64: 543–552.

2. Hantouche EG, Akiskal HS, Lancrenon S et al. Systematicclinical methodology for validating bipolar-II disorder:data in mid-stream from a French national multi-site study(EPIDEP). J Affect Disord 1998; 50: 163–173.

Sustained unemployment in outpatients with bipolar disorder

725

3. Murray C, Lopez A. Alternative projections of mortalityand disability by cause 1990–2020: Global Burden ofDisease Study. Lancet 1997; 349: 1498–1504.

4. Bryant-Comstock L, Stender M, Devercelli G. Health careutilization and costs among privately insured patients withbipolar I disorder. Bipolar Disord 2002; 4: 398–405.

5. Kent S, Fogarty M, Yellowlees P. A review of studies ofheavy users of psychiatric services. Psychiatr Serv 1995; 46:1247–1253.

6. Isometsa ET, Henriksson MM, Aro HM, Lonnqvist JK.Suicide in bipolar disorder in Finland. Am J Psychiatry1994; 151: 1020–1024.

7. Angst F, Stassen HH, Clayton PJ, Angst J. Mortality ofpatients with mood disorders: follow-up over 34–38 years.J Affect Disord 2002; 68: 167–181.

8. Ruggero CJ, Chelminski I, Young D, Zimmerman M.Psychosocial impairment associated with bipolar II disor-der. J Affect Disord 2007; 104: 53–60.

9. Kessler RC, Akiskal HS, Ames M et al. Prevalence andeffects of mood disorders on work performance in anationally representative sample of U.S. workers. Am JPsychiatry 2006; 163: 1561–1568.

10. Morgan VA, Mitchell PB, Jablensky AV. The epidemiol-ogy of bipolar disorder: sociodemographic, disability andservice utilization data from the Australian National Studyof Low Prevalence (Psychotic) Disorders. Bipolar Disord2005; 7: 326–337.

11. Dean BB, Gerner D, Gerner RH. A systematic reviewevaluating health-related quality of life, work impairment,and healthcare costs and utilization in bipolar disorder.Curr Med Res Opin 2004; 20: 139–154.

12. Simon GE, Bauer MS, Ludman EJ, Operskalski BH,Unutzer J. Mood symptoms, functional impairment, anddisability in people with bipolar disorder: specific effects ofmania and depression. J Clin Psychiatry 2007; 68: 1237–1245.

13. Bauer MS, Kirk GF, Gavin C, Williford WO. Determi-nants of functional outcome and healthcare costs inbipolar disorder: a high-intensity follow-up study. J AffectDisord 2001; 65: 231–241.

14. McElroy S, Altshuler L, Suppes T et al. Axis I psychiatriccomorbidity and its relationship to historical illness vari-ables in 288 patients with bipolar disorder. Am J Psychi-atry 2001; 158: 420–426.

15. Freeman MP, Freeman SA, McElroy SL. The comorbidityof bipolar and anxiety disorders: prevalence, psychobiol-ogy, and treatment issues. J Affect Disord 2002; 68: 1–23.

16. Henry C, Van den Bulke D, Bellivier F, Etain B, RouillonF, Leboyer M. Anxiety disorders in 318 bipolar patients:prevalence and impact on illness severity and response tomood stabilizer. J Clin Psychiatry 2003; 64: 331–335.

17. Otto MW, Simon NM, Wisniewski SR et al. Prospective12-month course of bipolar disorder in out-patients withand without comorbid anxiety disorders. Br J Psychiatry2006; 189: 20–25.

18. Frank E, Cyranowski JM, Rucci P et al. Clinical signifi-cance of lifetime panic spectrum symptoms in the treat-ment of patients with bipolar I disorder. Arch GenPsychiatry 2002; 59: 905–911.

19. Zimmerman M. Integrating the assessment methods ofresearchers in routine clinical practice: The Rhode Island

Methods to Improve Diagnostic Assessment and Services(MIDAS) project. In: First M ed. Standardized Evaluationin Clinical Practice, Vol. 22. Washington, DC: AmericanPsychiatric Publishing, Inc., 2003: 29–74.

20. First MB, Spitzer RL, Gibbon M, Williams JBW. Struc-tured Clinical Interview for DSM-IV Axis I Disorders –Patient edition (SCID-I ⁄P, version 2.0). New York:Biometrics Research Department, New York State Psy-chiatric Institute, 1995.

21. Pfohl B, Blum N, Zimmerman M. Structured Interview forDSM-IV Personality. Washington, DC: American Psychi-atric Press, Inc., 1997.

22. Endicott J, Spitzer RL. A diagnostic interview: TheSchedule for Affective Disorders and Schizophrenia. ArchGen Psychiatry 1978; 35: 837–844.

23. Guy W. ECDEU Assessment Manual for Psychopharma-cology. Rockville: National Institute of Mental Health,1976.

24. Elinson L, Houck P, Pincus HA. Working, receivingdisability benefits, and access to mental health care inindividuals with bipolar disorder. Bipolar Disord 2007; 9:158–165.

25. Kogan JN, Otto MW, Bauer MS et al. Demographic anddiagnostic characteristics of the first 1000 patients enrolledin the Systematic Treatment Enhancement Program forBipolar Disorder (STEP-BD). Bipolar Disord 2004; 6: 460–469.

26. Coryell W, Scheftner W, Keller M, Endicott J, Maser J,Klerman G. The enduring psychosocial consequences ofmania and depression. Am J Psychiatry 1993; 150: 720–727.

27. Tsuang MT, Woolson RF, Fleming JA. Long-termoutcome of major psychoses. I. Schizophrenia and affectivedisorders compared with psychiatrically symptom-freesurgical conditions. Arch Gen Psychiatry 1979; 36: 1295–1301.

28. Huxley N, Baldessarini RJ. Disability and its treatment inbipolar disorder patients. Bipolar Disord 2007; 9: 183–196.

29. Feske U, Frank E, Mallinger AG et al. Anxiety as acorrelate of response to the acute treatment of bipolar Idisorder. Am J Psychiatry 2000; 157: 956–962.

30. Boylan KR, Bieling PJ, Marriott M, Begin H, Young LT,MacQueen GM. Impact of comorbid anxiety disorders onoutcome in a cohort of patients with bipolar disorder.J Clin Psychiatry 2004; 65: 1106–1113.

31. Goodwin F, Jamison K. Manic-Depressive Illness: BipolarDisorders and Recurrent Depression, 2nd edn. New York:Oxford University Press, 2007.

32. Ferrie JE, Kivimaki M, Head J, Shipley MJ, Vahtera J,Marmot MG. A comparison of self-reported sicknessabsence with absences recorded in employers� registers:evidence from the Whitehall II study. Occup Environ Med2005; 62: 74–79.

33. Kessler RC, Barber C, Beck A et al. The World HealthOrganization Health and Work Performance Question-naire (HPQ). J Occup Environ Med 2003; 45: 156–174.

34. Gilbert AM, Olino TM, Houck P, Fagiolini A, Kupfer DJ,Frank E. Self-reported cognitive problems predict employ-ment trajectory in patients with bipolar I disorder. J AffectDisord 2010; 124: 324–328.

Zimmerman et al.

726