Subjective life satisfaction and objective functional outcome in bipolar and unipolar mood disorders: A longitudinal analysis

Journal of Affective Disorders 89 (2005) 79 – 89 www.elsevier.com/locate/jad

Research report

Subjective life satisfaction and objective functional outcome in bipolar and unipolar mood disorders: A longitudinal analysis Joseph F. Goldberg a,b,*, Martin Harrow c b

a Affective Disorders Program, Silver Hill Hospital, New Canaan, CT, USA Department of Psychiatry Research, The Zucker Hillside Hospital, North Shore Long Island Jewish Health System, Glen Oaks, NY, USA c Department of Psychiatry, The University of Illinois College of Medicine, Chicago, IL, USA

Received 13 May 2004; accepted 8 August 2005 Available online 24 October 2005

Abstract Background: Quality of life (QOL) has gained increasing attention as an important yet underappreciated component of functional outcome in mood disorders. In particular, the relationship between subjective life satisfaction and objective measures of psychosocial adjustment has not been well-studied. The goal of the present study was to examine the longitudinal associations between subjective life satisfaction and objective functional outcome among individuals with bipolar and unipolar mood disorders. Method: One hundred fifty-seven mood disordered subjects were assessed at index hospitalization for bipolar mania (n = 35), unipolar psychotic depression (n = 27), or unipolar nonpsychotic depression (n = 95). All were prospectively followed up three times, at approximately 2, 4.5 and 7–8 years. Global outcome, work performance, social adjustment, recurrent depressive episodes, and dimensions of life satisfaction were assessed by semi-structured interviews using standardized ratings. Results: Subjective life satisfaction strongly paralleled global functioning, work performance and social adjustment at each follow-up for patients with unipolar nonpsychotic depression, but not bipolar disorder or unipolar psychotic depression. Depressive symptoms and objective functional impairment contributed to poor QOL in most domains, independent of illness chronicity, medication use, or affective disorder subtype. Limitations: Findings might have differed had a different QOL measure been used, although the present measure showed concurrent validity with a previously used instrument. Sample sizes for the bipolar and psychotic depression groups were sufficient to detect moderate, but not small, correlations between objective functioning and subjective QOL. Conclusions: Recurrent depression remains a substantial contributor to poor life satisfaction across affective disorder subtypes. Subjective QOL in bipolar and unipolar psychotic depression patients may not accurately reflect objective functional outcome status, potentially due to diminished insight, demoralization, or altered life expectations over time. D 2005 Elsevier B.V. All rights reserved. Keywords: Bipolar disorder; Quality of life; Depression; Affective recurrence; Outcome

* Corresponding author. Silver Hill Hospital, 208 Valley Road, New Canaan, CT 06840, USA. Tel.: +1 203 801 2363; fax: +1 203 966 9336. E-mail address: [email protected] (J.F. Goldberg). 0165-0327/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.jad.2005.08.008

80

J.F. Goldberg, M. Harrow / Journal of Affective Disorders 89 (2005) 79–89

1. Introduction Psychiatric outcomes research has paid increasing attention to the concept of quality of life (QOL). Among individuals with primary affective disorders, depression has been associated with frequent dissatisfaction in overall social, physical and emotional wellbeing (Broadhead et al., 1990; Cooke et al., 1996; Lehman, 1983; Pyne et al., 1997), as well as with traditionally studied areas of occupational disability and diminished work productivity. Although recent literature has drawn attention to subjective life satisfaction in major affective disorder patients (Arfen, 1997; Atkinson, 1997; Atkinson et al., 1997), little is known about the extent to which quality of life parallels objective measures of psychosocial functioning (Atkinson, 1997; Atkinson et al., 1997; Gill and Feinstein, 1994; Leplege and Hunt, 1997). It is also unknown whether different aspects of life satisfaction (such as work versus social satisfaction) are equally affected in bipolar and unipolar mood disorders. The impact of depression relapse on individual life satisfaction domains across mood disorder subtypes also is not well understood. QOL in patients with bipolar disorder appears poorer than in the general population (Arnold et al., 2000) or in patients with schizophrenia (Atkinson et al., 1997), and is comparable to or worse than that seen among the chronically medically ill (Arnold et al., 2000; Cooke et al., 1996; Robb et al., 1997, 1998). Though potentially linked with multiple depressive episodes (MacQueen et al., 2000), diminished life satisfaction in bipolar disorder persists even during periods of sustained euthymia (Cooke et al., 1996; Robb et al., 1997, 1998). Comparisons of life satisfaction between bipolar and unipolar mood disorders have not been extensively described. Russo et al. (1997) observed weaker inter-relationships among QOL dimensions for bipolar than unipolar patients, suggesting that life satisfaction may vary across domains among subtypes of mood disorders. Moreover, the extent to which self-report measures validly reflect life satisfaction also has been questioned in severe mood disorders (Atkinson et al., 1997). The foregoing observations prompted us to examine satisfaction in major life areas relative to objective functional outcome in work and social roles. We further wished to assess differences between life satis-

faction and depression relapse as it impacts global functioning (Goldberg et al., 1995a). Five components of life satisfaction were assessed relative to global outcome and objective measures of work and social adjustment across three successive follow-ups over 7– 8 years after index hospitalization. We hypothesized that bipolar and psychotic depression patients would have poorer life satisfaction as compared to patients originally hospitalized for unipolar depression, and that subjective life satisfaction ratings would reflect objective functional outcome more strongly among nonpsychotic unipolar depression patients than in more severe (i.e., bipolar and psychotic unipolar depression) patients.

2. Method 2.1. Subjects We evaluated 157 relatively early young patients initially as inpatients who were then prospectively followed over a 7–8-year period. Subjects were part of the Chicago Follow-up Study, a longitudinal outcome program studying features of thought disorder, positive and negative symptoms, psychosis, affective relapse, and psychosocial outcome in a large young adult cohort initially hospitalized for psychotic and mood-related disorders (Goldberg et al., 1995a,b, 2001; Goldberg and Harrow, 2004; Harrow and Miller, 1980; Harrow et al., 1988, 1989, 1997, 2000). Subjects included 35 patients originally hospitalized for bipolar I mania, 95 hospitalized for unipolar nonpsychotic depression, and 27 hospitalized for unipolar psychotic depression. All were diagnosed at the time of their index admission according to Research Diagnostic Criteria (RDC; Spitzer et al., 1978) using structured interviews. None of the unipolar patients had ever had a previous manic or hypomanic episode, and patients who met diagnostic criteria for schizophrenia, schizoaffective disorder, or primary substance use disorders at index were not involved in the study. The study group at baseline was representative of a typical general adult inpatient population. At index admission, subjects had a mean F S.D. age of 23.1 F 3.7 years. Seventy-six percent were Caucasian, 61% were female, 19% were currently married, and

J.F. Goldberg, M. Harrow / Journal of Affective Disorders 89 (2005) 79–89

the group had a mean F S.D. of 1.0 F 1.6 prior lifetime hospitalizations. Pharmacotherapies and other treatments during the follow-up period occurred under naturalistic conditions and were not a specific focus of the current investigation, but were recorded at the time of each assessment. 2.1.1. Follow-up assessments Follow-up evaluations, conducted by trained and experienced research assistants, included a series of standardized questionnaires and semi-structured interviews designed to assess varied aspects of psychopathology and functioning at each follow-up (Goldberg et al., 1995a,b, 2001; Goldberg and Harrow, 2004; Harrow et al., 1997). Global functioning in the year prior to each assessment was rated using an 8-point global outcome scale developed by Levenstein et al. (1966). This scale, used successfully in previous studies (Goldberg et al., 1995a,b; Harrow et al., 1997, 2000; Goldberg and Harrow, 2004; Levenstein et al., 1966), provides a comprehensive measure of psychosocial adjustment for the year preceding each follow-up, based on factors such as signs of psychiatric illness, rehospitalization, level of self-support, role performance, and social relationships. Inter-rater reliability for this 8point scale as used by our group is high (intraclass correlation coefficient = 0.92). Work performance in the year preceding each follow-up was rated using a 5-point index of work functioning developed by Strauss and Carpenter (1972). This scale accounts for the percentage of time in the preceding year in which patients were able to work effectively in their occupations or as a student or homemaker. Adequate inter-rater reliability has been established for this scale, which has been used previously by our group (Goldberg et al., 1995a,b; Goldberg and Harrow, 2004; Harrow et al., 1997, 2000). Psychotic symptoms and depressive episodes in the year preceding each follow-up were assessed using the Schedule for Affective Disorders and Schizophrenia (SADS; Endicott and Spitzer, 1978). Features of a depressive episode during the year preceding follow-up were rated from the SADS on a 5-point continuum, indicating no depressive syndrome (b1Q), mild depressive signs (b2Q), low-moderate subthreshold depression (b3Q), high-moderate sub-

81

threshold depression (b4Q), and full major depressive syndrome (b5Q). Psychotic symptoms at follow-up (i.e., delusions and/or hallucinations) were rated as bprobable or definiteQ versus babsent,Q based on SADS follow-up interviews. 2.1.1.1. Life satisfaction assessments. Five major areas of life satisfaction were rated based on a semistructured interview: 1) work satisfaction (among those who had worked in the year preceding followup), 2) feelings of economic security, 3) satisfaction with social activities and relationships, 4) satisfaction with living situation, and 5) self-perceived overall mental health. Subjects were asked the following questions: 1) bHow satisfied are you with your current work situation?Q 2) bHow secure do you feel economically?Q 3) bHow satisfied are you with your current social activities?Q and 4) bHow satisfied are you with your current living arrangements?Q Their responses were rated by interviewers on a scale ranging from 1 (bvery satisfiedQ) to 5 (bvery dissatisfiedQ). They were also asked, bHow would you rate your current overall mental health as compared to that of most other people?Q and their responses were again rated from 1 (bmuch better than the average person’sQ) to 5 (bmuch worse than the average person’sQ). High life satisfaction ratings were judged as scores of b1Q or b2Q on each of these 5-point rating systems. In order to assess the relationship between the five life satisfaction component dimensions and QOL as measured by a previously standardized QOL instrument, we calculated pairwise correlations between each of the five dimensions studied with corresponding items drawn from a 34-item QOL index developed by Ferrans and Powers (1985). This instrument has been validated and used in previous studies with large samples of psychiatric or medical patients (Ferrans and Powers, 1992). Before undertaking the present study, we rated the five life satisfaction measures described above in conjunction with the QOL scale by Ferrans and Powers in a group of 20 psychiatric outpatients with major mood disorders. Correlations between corresponding items from the present instrument and the QOL scale by Ferrans and Powers, respectively, were as follows: 1) subjective evaluation of mental health and satisfaction with mental health: r = 0.85, p b .001); 2) satisfaction with living arrangements and satisfaction with home/neighborhood:

82

J.F. Goldberg, M. Harrow / Journal of Affective Disorders 89 (2005) 79–89

Table 1 Sample characteristics Variable

Bipolar (N = 35)

Unipolar psychotic depression (N = 27)

Unipolar nonpsychotic depression (n = 95)

% % % %

49% 69% 23% 32%

63% 67% 11% 50%

65% 83% 20% 55%

54% 13.3 F 2.1 24.0 F 6.1

81% 13.0 F 1.5 21.9 F 4.2

89% 13.8 F 2.1 23.4 F 3.6

Female Caucasian Married at index with change in marital status over follow-up % with 0 or 1 prior hospitalizations Mean F S.D. years education at index Mean F S.D. age at index

r = 0.72, p b .001); 3) social satisfaction and satisfaction with leisure activities: r = 0.67, p b .05; 4) economic satisfaction and satisfaction with financial independence: r = 0.88, p b .01; and 5) work satisfaction with job satisfaction: r = 0.81, p b .04. Outcome data at the 7.5-year follow-ups were available for slightly over 80% of the original affectively disordered sample studied as inpatients. Subjects in the current sample did not differ significantly in age or education from a parallel sample of subjects who were assessed at the 2-year follow-ups, but not at the 7.5-year follow-ups. In addition, there were no significant differences at the 2-year follow-ups on global outcome or on the five QOL dimensions between the current sample and the subsample of

X2

F

df

3.045 6.216 2.653 5.297 20.016 1.792 1.860

p 2 4 4 2

.218 .184 .617 .071

2 2149 2146

b.001 .170 .159

patients studied at the 2-year but not the 7.5-year follow-ups. 2.2. Statistical methods Dichotomous variables were compared by chisquare analyses. Mean group differences across the three diagnostic groups were compared using one-way analyses of variance (ANOVAs). The strength of association between dependent variables (e.g., life satisfaction and overall outcome) was assessed by simple Pearson correlations. Changes over time in continuous variables involving the three diagnostic groups were analyzed by generalized linear model (GLM) repeated measures analyses of variance

Table 2 Naturalistic medication use at follow-up Medication grouping per follow-up Any medications 2-year 4.5-year 7–8-year Any antidepressant 2-year 4.5-year 7–8-year Any lithium 2-year 4.5-year 7–8-year Any antipsychotic 2-year 4.5-year 7–8-year

Bipolar

Unipolar psychotic depression

Unipolar nonpsychotic depression

X2

df

p

20/29 (69%) 22/34 (65%) 23/34 (68%)

11/23 (48%) 8/25 (32%) 8/24 (33%)

32/89 (36%) 29/91 (32%) 27/91 (30%)

9.753 11.861 15.392

2 2 2

.008 .003 b.001

4/29 (14%) 5/34 (15%) 4/34 (12%)

3/23 (13%) 1/25 (4%) 0/24 (0%)

12/88 (14%) 12/91 (13%) 12/91 (13%)

0.007 1.872 3.494

2 2 2

.997 .392 .174

14/28 (50%) 16/31 (52%) 20/33 (61%)

1/20 (5%) 1/23 (4%) 2/24 (8%)

10/87 (12%) 7/91 (8%) 7/91 (8%)

23.658 35.241 45.343

2 2 2

b.001 b.001 b.001

8/29 (28%) 10/34 (29%) 11/34 (32%)

7/23 (30%) 4/25 (16%) 6/24 (25%)

10/88 (11%) 8/91 (9%) 8/91 (9%)

6.882 8.452 11.226

2 2 2

.032 .015 .004

J.F. Goldberg, M. Harrow / Journal of Affective Disorders 89 (2005) 79–89

83

Table 3 Mean (FS.D.) life satisfaction ratings across diagnostic groups Follow-up and patient group

Dimensions of life satisfaction

2-year Bipolar (n = 35) Unipolar psychotic depression (n = 27) Unipolar nonpsychotic depression (n = 95) 4.5-year Bipolar (n = 35) Unipolar psychotic depression (n = 27) Unipolar nonpsychotic depression (n = 95) 7–8-Year Bipolar (n = 95) Unipolar psychotic depression (n = 27) Unipolar nonpsychotic depression (n = 95)

Work

Social

Economic

Living

Self-assessed mental health

1.8 F 1.1 2.3 F 1.2 2.5 F 1.2

2.9 F 1.4 2.8 F 1.4 2.7 F 1.3

3.0 F 1.7 2.8 F 1.3 2.7 F 1.4

2.3 F 1.4 2.6 F 1.3 2.2 F 1.3

2.1 F 0.7 2.5 F 1.1 2.4 F 1.0

2.2 F 1.1 2.3 F 1.0 2.2 F 1.2

2.9 F 1.3 2.4 F 1.3 2.8 F 1.2

3.0 F 1.6 2.7 F 1.2 2.6 F 1.4

2.6 F 1.5 2.3 F 1.4 2.2 F 1.3

2.2 F 0.9 1.9 F 0.7 2.1 + 0.9

2.3 F 1.0 2.0 F 0.8 2.1 F 0.9

2.7 F 1.2 2.6 F 1.4 2.7 F 1.3

2.5 F 1.3 2.9 F 1.3 2.6 F 1.2

2.1 F1.3 2.3 F 1.5 2.0 F 1.2

2.2 + 0.9 2.3 F 1.3 2.0 F 0.9

antipsychotic medications was higher in the bipolar than unipolar groups.

(ANOVAs). Multiple linear regression analyses were used to examine relationships between life satisfaction and recurrent depressive episodes, while controlling potential diagnostic differences. All statistical tests were two-tailed with an alpha level of 0.05. All analyses were conducted using SPSS for Windows, version 11.5 (Chicago, IL).

3.1. Domains of life satisfaction Tables 3 and 4, respectively, summarize mean scores for each of the five life satisfaction domains, with corresponding repeated measures ANOVAs, across the three diagnostic groups at each successive follow-up. In general, the majority of subjects in each group reported at least moderately high levels of life satisfaction in all domains at each time period. For those who worked in the year preceding a given follow-up, approximately three-quarters or more in all diagnostic groups rated work as bsatisfyingQ or bvery satisfying.Q Self-assessed mental health was rated as babove averageQ or bvery much above averageQ among two-thirds to three-quarters or more of all patients at all time periods. Repeated measures ANOVAs (Table 4)

3. Results Table 1 summarizes demographic and clinical characteristics for the three diagnostic groups. Notable was the significantly greater proportion of unipolar psychotic or nonpsychotic than bipolar subjects with one or no hospitalizations prior to the index admission. As shown in Table 2, naturalistic pharmacotherapy patterns were similar across affective disorder groups with regard to antidepressant use, but use of lithium or

Table 4 Longitudinal stability of life satisfaction domains: repeated measures analyses of variance Domain

Main effects Time

Social satisfaction Work satisfaction Economic security Living arrangements Self-assessed mental health

Time  Diagnosis

Diagnosis

F

df

p

F

df

p

F

df

p

0.464 0.385 0.778 0.488 2.425

2 2 2 2 2

.629 .682 .464 .615 .090

0.644 0.643 0.067 0.156 0.388

2 2 2 2 2

.527 .531 .936 .856 .679

0.637 1.480 0.488 0.787 3.729

4 4 4 4 4

.637 .216 .744 .535 .006

84

J.F. Goldberg, M. Harrow / Journal of Affective Disorders 89 (2005) 79–89

revealed no significant differences in any life satisfaction domains over time across diagnostic groups. For the total group of 157 subjects, significant associations were observed over time from the 2-year to 4.5-year follow-up and the 4.5-year to 7–8-year followup in several life satisfaction domains, including social satisfaction (r = 0.38, p b .001 and r = 0.28, p = .001, respectively); economic satisfaction (r = 0.59, p b .001 and r = 0.47, p b .001, respectively), and self-assessed mental health (r = 0.32, p b .001 and r = 0.44, p b .001). Satisfaction with living situations was consistent from the second to third follow-up (r = 0.26, p = .004) but not the first to second follow-up (r = 0.15, p = .117). Work satisfaction scores were not correlated from either the first to second (r = 0.05, p = .692) or second to third (r = 0.04, p = .741) assessment. 3.2. Interrelationships among life satisfaction domains Correlation matrices of life satisfaction domains for the total group of 157 subjects at each of the three follow-up assessments are presented in Table 5. In general, significant associations were seen more extensively at later than earlier follow-up assessments. Selfassessed mental health appeared to be the domain most consistently and significantly related to all other domains, at nearly every follow-up. 3.3. Associations between life satisfaction and global outcome Table 6 depicts the relationships among each area of life satisfaction and objectively measured global outcome at each follow-up. For the nonpsychotic depression patients, significant or highly significant correlations were observed between objective global functioning and each of the five component life satisfaction measures, at each of the three assessment periods. By contrast, associations between global outcome and life satisfaction domains were less robust, and mostly nonsignificant, for both the bipolar and psychotic depression groups. 3.4. Work performance and work satisfaction Although subjective work satisfaction ratings did not differ significantly among the three diagnostic

groups at any follow-up (Table 3), objective work functioning scores did. At the 2-year follow-up, work functioning scores were significantly lower (i.e., poorer) for the bipolar subjects (mean F S.D. = 2.4 F 1.7) than the psychotic depression (2.6 F 1.7) or nonpsychotic depression (3.2 F 1.3) groups. Similar findings emerged at the 4.5-year follow-up (mean F S.D. work functioning scores = 2.5 F 1.7 for bipolar patients, 3.2 F 1.1 for psychotic depression patients, 3.3 + 1.3 for nonpsychotic depression patients); and again at the 7–8-year follow-up (mean F S.D. work functioning scores = 2.6 F 1.6 for bipolar patients, 3.0 F 1.4 for psychotic depression patients, and 3.3 F 1.2 for nonpsychotic depression patients). A repeated measures ANOVA for mean work functioning scores yielded a highly significant main effect for Diagnosis ( F = 6.676, df = 2, p = .002) but not for Time ( F = 2.064, df = 2, p = .129) or a Diagnosis  Time interaction ( F = 0.846, df = 4, p = .497). Objective work functioning and subjective work satisfaction were not significantly associated with one another for the bipolar group at either the first (r = 0.42, p = .172), second (r = 0.01, p = .976) or third (r = 0.39, p = .091) follow-ups. Similarly, no significant associations were evident either for the unipolar psychotic depression group at any of the three follow-ups (r = 0.39, p = .272; r = 0.36, p = .190; r = 0.07, p = .816, respectively) or for the nonpsychotic depression group across the three assessments (r = 0.07, p = .635; r = 0.24, p = .077; r = 0.10, p = .444, respectively). 3.5. Social satisfaction and social functioning Objective social functioning scores did not differ significantly among the three diagnostic groups over time (repeated measures ANOVA: main effect for Diagnosis: F = 1.965, df = 2, p = .144; main effect for Time: F = 2.250, df = 2, p = .107; Diagnosis  Time interaction effect: F = 1.672, df = 4, p = .168). No significant correlations were observed between objective social functioning and subjective social satisfaction for the bipolar group (r = 0.21, p = .263; r = 0.01, p = .935; r = 0.03, p = .875 across three successive follow-ups, respectively). However, among the psychotic depression subjects, social functioning and social satisfaction were significantly related at the 2-year (r = 0.45, p = .030) and 7–8-

J.F. Goldberg, M. Harrow / Journal of Affective Disorders 89 (2005) 79–89

85

Table 5 Inter-relationships among life satisfaction domains Work

Social

Economic

Living situation

Self-assessed mental health

2-year follow-up Work Social Economic Living Self-assessed mental health

– .20 .50** .07 .18

– – .13 .24 .34***

– .21 .22

– .09

–

4.5-year follow-up Work Social Economic Living Self-assessed mental health

– .20 .06 .19 .38*

– – .27* .24** .32***

– .18 .48***

– .19*

–

7–8-year follow-up Work Social Economic Living Self-assessed mental health

– .30** .29* .18 .32**

– – .20 .22* .37***

– .27* .29*

– .33***

–

* p b .05. ** p b .01. *** p b .001.

year follow-ups (r = 0.51, p = .011) but not the 4.5year follow-up (r = 0.12, p = .567). Moreover, for the nonpsychotic depression subjects, this relationship

was very highly significant across all three followups (r = 0.41, p b .001; r = 0.40, p b .001; r = 0.42, p b .001, respectively).

Table 6 Correlations between dimensions of life satisfaction and objective global functioning Patient group

Bipolar (N = 35) 2-year follow-up 4.5-year follow-up 7–8-year follow-up Unipolar psychotic depression (N = 27) 2-year follow-up 4.5-year follow-up 7–8-year follow-up Unipolar nonpsychotic depressed (N = 95) 2-year follow-up 4.5-year follow-up 7–8-year follow-up * p b .05. ** p b .01. *** p b .001.

Dimensions of life satisfaction Work

Social

Economic

Living situation

Self-assessed mental health

.42 .07 .09

.38** .15 .31

.21 .35 .10

.29 .08 .10

.22 .21 .33*

.44 .22 .10

.29 .49** .15

.52 .22 .26

.03 .08 .32

.33 .14 .03

.27* .31** .29*

.42*** .40*** .32***

.25 .56*** .45***

.09 .21* .25**

.45*** .39*** .52***

86

J.F. Goldberg, M. Harrow / Journal of Affective Disorders 89 (2005) 79–89

tive functional outcome also significantly contributed to self-perceived mental health and satisfaction with patients’ living situations.

3.6. Relationship between recurrent depression and life satisfaction domains Prior to examining the impact of depressive symptoms at follow-up on life satisfaction, the three diagnostic groups were compared on their relative prevalence rates of depression in the year preceding the 7–8-year follow-up. Complete data on depressive syndromes were available for 30 of the bipolar, 17 of the unipolar psychotic, and 77 of the unipolar nonpsychotic groups. The presence of either bhigh-moderate subthresholdQ or bfull major syndromalQ depression was similar for the bipolar (9/30, or 30%), unipolar psychotic (6/17, or 35%) or unipolar nonpsychotic (24/77, or 31%) subjects. Using the full 5-point range of depression syndrome scores, there was no overall difference among the three groups ( F = 0.593, df = 2,121, p = .554). Table 7 presents results from a series of five separate multiple linear regression analyses calculated at the 7–8-year follow-up to examine associations between each of the five life satisfaction domains (i.e., each of the respective dependent variables) relative to the extent of past year recurrent depressive symptoms. Diagnostic grouping was controlled for as an independent variable in these analyses, along with other potential confounding variables (i.e., prior hospitalizations, medication use, and objective functional outcome scores). Significant associations were evident between depressive signs and nearly all life satisfaction domains, regardless of diagnosis. Objec-

4. Discussion Contrary to our initial hypothesis, overall dimensions of life satisfaction did not differ significantly among bipolar or unipolar depression patients, as measured across three successive follow-ups over a 7–8-year period. However, consistent with our second hypothesis, high concordance was seen between subjective life satisfaction and objective measures of global, work and social functioning among unipolar nonpsychotic depression patients; this relationship was less strong for bipolar or unipolar psychotic depression patients. These findings are consistent with those of Atkinson et al. (1997) in suggesting that among severely ill affective disorder patients, self-reported life satisfaction may not reliably reflect objective functioning. Although life satisfaction domains generally maintained consistency within themselves over time, correlations across specific domains were more variable at any given assessment. This could suggest that varied aspects of life satisfaction may arise independently of one another. Hence, patients who express high degrees of life satisfaction in one area may nevertheless feel low satisfaction in others. This would further indicate that clinicians should not pre-

Table 7 Associations between depressive syndromes at 7–8-year follow-up and life satisfaction domains Variable

Work satisfaction

Economic security

Living situation

Self-assessed mental health

N = 84

N = 123

N = 73

N = 123

N = 123

R 2 = .106

R 2 = .174

R 2 = .102

R 2 = .262

R 2 = .469

b Diagnosis Depressive syndrome 0,1 Prior hospitalizations Any medications at follow-up LKP functional outcome

Social satisfaction

t

p

b

t

p

b

t

p

b

t

p

b

t

p

.018 .231

0.164 .870 2.040 .045

.033 .295

0.356 .723 3.265 .001

.105 .098

0.791 .432 0.777 .440

.012 .233

0.143 2.726

.886 .134 .007 .312

1.818 4.309

.072 b.001

.043

0.394 .695

.068

0.710 .479

.040

0.307 .760

.086

0.947

.346 .040

.520

.604

.186

1.607 .112

.066

0.662 .509

.054

0.363 .718

.232

2.464

.015 .004

.046

.964

.017

.142 .887

.187

1.762 .081

.286

1.891 .063

.453

4.528 b.001 .487

5.732

b.001

J.F. Goldberg, M. Harrow / Journal of Affective Disorders 89 (2005) 79–89

sume different areas of life satisfaction (or dissatisfaction) necessarily to cluster homogeneously. A goal of this study was to examine component dimensions of life satisfaction, rather than to assess global QOL in bipolar illness using a multifactorial scale (e.g., Cooke et al., 1996; Robb et al., 1997, 1998). It is possible that the results reported herein might have differed had an alternative QOL instrument been used. However, Atkinson (1997) observed that disparities between subjective QOL and objective functional impairment did not vary with the use of different QOL instruments. Individuals with bipolar disorder or psychotic depressive disorders may be among those most vulnerable to chronic psychosocial impairment (Coryell et al., 1993), yet patients with these more severe disorders may have particular difficulty in acknowledging the impact of illness on functional well-being. Earlier findings by our group on self-monitoring in thought-disordered and psychotic schizophrenia patients would support the view that impaired selfmonitoring is common in patients with severe psychopathology (Harrow and Miller, 1980; Harrow et al., 1988, 1989). Disparities between subjective and objective functioning for the bipolar and unipolar psychotic (but not nonpsychotic) depression groups did not appear to be a function of recurrent depression, since depressive symptoms at follow-up were no less prevalent among the nonpsychotic unipolar depression group. Nor did chronicity of illness per se contribute directly to diminished life satisfaction in the regression models of QOL domains. Other factors would seem necessary to account for the discordance between subjective and objective psychosocial outcomes among the bipolar and psychotic depression subgroups. While numerous factors doubtless contribute to a phenomenon so multideterminate as life satisfaction, we would propose several interpretations of this finding: 1. Decreased awareness, or diminished insight. Some patients with severe affective disorders could be unaware of distressing reactions to illness-related disabilities. Such a phenomenon could involve the denial of illness and its consequences, and refusal to acknowledge unfavorable life circumstances. Lack of insight about the nature and consequences of psychopathology has been described as common

87

among patients with bipolar disorder (Dell’Osso et al., 2002; Ghaemi et al., 1995; Pallanti et al., 1999). The capacity to recognize functional deficits brought about by the disorder may in itself constitute an identifiable neurocognitive substrate (McGlynn and Schacter, 1997). Interestingly, data from the DSM-IV field trials revealed that insight is more impaired in bipolar than unipolar psychotic major depressive disorder, and more impaired in psychotic major depressive disorder than in nonpsychotic major depression (Amador et al., 1994). These earlier observations are consistent with the present findings. 2. Demoralization and desensitization to stress. It is possible that patients who have endured severe, recurrent bipolar or psychotic depressive episodes for many years after an index presentation may become resigned or desensitized to certain social, economic or occupational reversals, causing a blunted emotional reaction to chronic illness and repeated life disappointments. This type of bdesensitizationQ to life stresses over time is consistent with theories about kindling or behavioral sensitization relative to affective relapse (Post, 1992), insofar as episodes eventually may recur with increasing automaticity, independent from stressful life events. 3. Altered life expectations. A final consideration is that patients with severe, chronic, relapsing affective disorders may over time modify their expectations about life goals and acceptable psychosocial circumstances. Patients who fail to regain premorbid levels of adjustment may eventually adopt a new baseline or functional equilibrium state. If such a new homeostasis develops in the context of protracted psychosocial disability, life satisfaction may reflect an adaptation to more severe illness, with limited expectations. The current results are in accord with other reports suggesting that among severely ill affective disorder groups, objective functional disability may persist for years after symptomatic recovery from an index episode (Coryell et al., 1993; Gitlin et al., 1995; Tohen et al., 2000). Yet, the present findings also support the view that patients with extensive psychopathology may be more prone to over- or under-estimate their own disability or life satisfaction (Atkinson et al.,

88

J.F. Goldberg, M. Harrow / Journal of Affective Disorders 89 (2005) 79–89

1997). One implication of these data is that the service utilization needs of psychotic unipolar or bipolar patients should not necessarily be evaluated solely based on their subjective complaints. To the extent that individuals with severe mood disorders may have a diminished capacity to acknowledge and report occupational, social or other functional disabilities– or may, over time, become resigned to accept them without complaints–greater scrutiny may be needed by their clinicians, family members and others involved in their care to better assure that treatment needs become met. Recent observations from the National Institute of Mental Health (NIMH) Collaborative Depression Study document the persistence of affective (mainly depressive) symptoms among bipolar patients followed for up to 20 years (Judd et al., 2002). The present findings further underscore the adverse impact of depressive symptoms across most domains of life satisfaction, independent of their contribution to overt functional disability, as previously shown by our group (Goldberg et al., 1995a). It is noteworthy that the sample sizes of the bipolar and unipolar psychotic depression groups were smaller than that of the unipolar nonpsychotic depression group. Sufficient statistical power (80%) was available to detect moderate correlations in the bipolar (z 0.45) or unipolar nonpsychotic depression (z 0.50) groups, but correlations of smaller magnitude (z 0.30) among the unipolar nonpsychotic depression group. Importantly, however, correlations of at least moderate magnitude would be clinically more meaningful from the perspective of clinical significance. Thus, the current sample sizes were sufficiently large to identify clinically meaningful associations within each diagnostic group. In summary, the present findings suggest that most formerly hospitalized bipolar and unipolar patients describe comparable and at least moderately intact levels of life satisfaction across successive assessments over a 7–8-year follow-up period. However, while subjective life satisfaction closely parallels objective global, occupational and social adjustment in nonpsychotic unipolar depression, greater disparity is evident between subjective and objective outcomes among bipolar and unipolar psychotic depression patients. Future studies might usefully consider such distinctions when gauging the impact of disability and psy-

chosocial impairment after hospitalization in patients with bipolar and psychotic depressive disorders.

Acknowledgements Supported in part by NIMH K-23 Career Development Award MH-01936 (JFG) and NIMH grant MH26341 (MH), a NARSAD Young Investigator Award (JFG), and by the Stanley Bipolar Research Center of the Zucker Hillside Hospital. The authors wish to thank Robert N. Faull, BA, for his assistance with data analyses, and Barbara Napolitano, PhD, for her assistance with statistical design and interpretation.

References Amador, X.F., Flaum, M., Andreason, N.C., Strauss, D.H., Yale, S.A., Clark, S.C., Gorman, J.M., 1994. Awareness of illness in schizophrenia and schizoaffective and mood disorders. Arch. Gen. Psychiatry 51, 826 – 836. Arfen, C.L., 1997. Self-reported life satisfaction (letter). Am. J. Psychiatry 154, 1478. Arnold, L.M., Witzeman, K.A., Swank, M.L., McElroy, S.L., Keck, Jr., P.E., 2000. Health-related quality of life using the SF-36 in patients with bipolar disorder compared with patients with chronic back pain and the general population. J. Affect. Disord. 57, 235 – 259. Atkinson, M., 1997. Reply to Letter from Dr. Arfen. Am. J. Psychiatry 154, 1479. Atkinson, M., Zibin, S., Chuang, H., 1997. Characterizing quality of life among patients with chronic mental illness: a critical examination of the self-report methodology. Am. J. Psychiatry 154, 99 – 105. Broadhead, W.E., Blazer, D.G., George, L.K., Tse, C.K., 1990. Depression, disability days, and days lost from work in a prospective epidemiologic survey. JAMA 264, 2524 – 2528. Cooke, R.G., Robb, J.C., Young, L.T., Joffe, R.T., 1996. Wellbeing and functioning in patients with bipolar disorder assessed using the MOS-20 item short form (SF-20). J. Affect. Disord. 39, 93 – 97. Coryell, W., Scheftner, W., Keller, M., Endicott, J., Maser, J., Klerman, G.L., 1993. The enduring psychosocial consequences of mania and depression. Am. J. Psychiatry 150, 720 – 727. Dell’Osso, L., Pini, S., Cassano, G.B., Mastrocinque, C., Seckinger, R.A., Saettoni, R.A., Papasogli, A., Yale, S.A., Amador, X.F., 2002. Insight into illness in patients with mania, mixed mania, bipolar depression and major depression with psychotic features. Bipolar Disord. 4, 315 – 322. Endicott, J., Spitzer, R.L., 1978. A diagnostic interview: the schedule for affective disorders and schizophrenia. Arch. Gen. Psychiatry 35, 837 – 844.

J.F. Goldberg, M. Harrow / Journal of Affective Disorders 89 (2005) 79–89 Ferrans, C., Powers, M., 1985. Quality of life index: development and psychometric properties. Adv. Nurs. Sci. 8, 15 – 24. Ferrans, C., Powers, M., 1992. Psychometric assessment of the Quality of Life Index. Res. Nurs. Health 15, 111 – 119. Ghaemi, S.N., Stoll, A.L., Pope, Jr., H.G., 1995. Lack of insight in bipolar disorder. The acute manic episode. J. Nerv. Ment. Dis. 183, 464 – 467. Gill, T.M., Feinstein, A.R., 1994. A critical appraisal of the qualityof-life measurements. JAMA 72, 619 – 629. Gitlin, M.J., Swendsen, J., Heller, T.L., Hammen, C., 1995. Relapse and impairment in bipolar disorder. Am. J. Psychiatry 152, 1635 – 1640. Goldberg, J.F., Harrow, M., 2004. Consistency of remission and outcome in bipolar and unipolar mood disorders: a 10-year follow-up. J. Affect. Disord. 81, 123 – 131. Goldberg, J.F., Harrow, M., Grossman, L.S., 1995a. Recurrent affective syndromes in bipolar and unipolar mood disorders at follow-up. Br. J. Psychiatry 166, 382 – 385. Goldberg, J.F., Harrow, M., Grossman, L.S., 1995d. Course and outcome in bipolar affective disorder: a longitudinal follow-up study. Am. J. Psychiatry 152, 379 – 384. Goldberg, J.F., Harrow, M., Whiteside, J.E., 2001. Risk for bipolar illness in patients initially hospitalized for unipolar depression. Am. J. Psychiatry 158, 1265 – 1270. Harrow, M., Miller, J.G., 1980. Schizophrenic thought disorders and impaired perspective. J. Abnorm. Psychology 89, 717 – 727. Harrow, M., Rattenbury, F., Stoll, F., 1988. Schizophrenic delusions: an analysis of their persistence, of related premorbid ideas, and of three major dimensions. In: Oltmanns, Harrow , T., Maher, B. (Eds.), Delusional Beliefs. John Wiley and Sons, New York, pp. 184 – 211. Harrow, M., Lanin-Kettering, I., Miller, J.G., 1989. Impaired perspective and thought pathology in schizophrenic and psychotic disorders. Schizophr. Bull. 15, 605 – 623. Harrow, M., Sands, J.R., Silverstein, M.L., Goldberg, J.F., 1997. Course and outcome for schizophrenia versus other psychotic patients: a longitudinal study. Schizophr. Bull. 23, 287 – 303. Harrow, M., Grossman, L.S., Herbener, E.S., Davies, E.W., 2000. Ten-year outcome: patients with schizoaffective disorders, schizophrenia, affective disorders and mood-incongruent psychotic symptoms. Br. J. Psychiatry 177, 421 – 426. Judd, L.L., Akiskal, H.S., Schettler, P.J., Endicott, J., Maser, J., Solomon, D.A., Leon, A.C., Rice, J.A., Keller, M.B., 2002. The long-term natural history of the weekly symptomatic status of bipolar I disorder. Arch. Gen. Psychiatry 59, 530 – 537. Lehman, A., 1983. The well-being of chronic mental patients: assessing their quality of life. Arch. Gen. Psychiatry 40, 369 – 373.

89

Leplege, A., Hunt, S., 1997. The problem of quality of life in medicine. JAMA 278, 47 – 50. Levenstein, S., Klein, D.F., Pollack, M., 1966. Follow-up study of formerly hospitalized voluntary psychiatric patients: the first two years. Am. J. Psychiatry 122, 1102 – 1109. MacQueen, G.M., Young, L.T., Robb, J.C., Marriott, M., Cooke, R.G., Joffe, R.T., 2000. Effect of number of episodes on wellbeing and functioning of patients with bipolar disorder. Acta Psychiatr. Scand. 101, 374 – 381. McGlynn, S.M., Schacter, D.L., 1997. The neurobiology of insight: impaired awareness of deficits in a psychiatric context. Psychiatr. Ann. 27, 806 – 811. Pallanti, S., Querciolo, L., Pazzaglia, A., Rossi, A., Dell’Osso, L., Pini, S., Cassano, G.B., 1999. Awareness of illness and subjective experiences of cognitive complaints in patients with bipolar I and bipolar II disorder. Am. J. Psychiatry 156, 1094 – 1096. Post, R.M., 1992. Transduction of psychosocial stress into the neurobiology of recurrent affective disorder. Am. J. Psychiatry 148, 999 – 1010. Pyne, J.M., Patterson, T.L., Kaplan, R.M., Gillin, J.C., Koch, W.L., Grant, L., 1997. Assessment of the quality of life of patients with major depression. Psychiatr. Serv. 48, 224 – 230. Robb, J.C., Cooke, R.G., Devins, G.M., Young, L.T., Joffe, R.T., 1997. Quality of life and lifestyle disruption in euthymic bipolar disorder. J. Psychiatr. Res. 31, 509 – 517. Robb, J.C., Young, L.T., Cooke, R.G., Joffe, R.T., 1998. Gender differences in patients with bipolar disorder influence outcome in the medical outcomes survey (SF-20) subscale scores. J. Affect. Disord. 49, 189 – 193. Russo, J., Roy-Byrne, P., Reeder, D., Alexander, M., Dwyer-O’Connor, E., Dagadakis, C., Ries, R., Patrick, D., 1997. Longitudinal assessment of quality of life in acute psychiatric inpatients: reliability and validity. J. Nerv. Ment. Dis. 185, 166 – 175. Spitzer, R.L., Endicott, J., Robins, E., 1978. Research Diagnostic Criteria: rationale and reliability. Arch. Gen. Psychiatry 35, 773 – 782. Strauss, J.S., Carpenter, W.T., 1972. The prediction of outcome in schizophrenia: characteristics of outcome. Arch. Gen. Psychiatry 27, 739 – 746. Tohen, M., Hennen, J., Zarate, Jr., C.M., Baldessarini, R.J., Strakowski, S.M., Stoll, A.L., Faedda, G.L., Suppes, T., GebreMedhin, P., Cohen, B.M., 2000. Two-year syndromal and functional recovery in 219 cases of first-episode major affective disorder with psychotic features. Am. J. Psychiatry 157, 220 – 228.