February 2010  << Back  

  Can't open the PDF? Click here for help.

 

 RESEARCH ARTICLE

Health-related quality of life in depression: A STAR*D report

Ella J. Daly, MB, MRCPsych*

Department of Psychiatry, University of Texas Southwestern Medical Center, Dallas, TX, USA

*Ella J. Daly, MB, MRCPsych, is currently an employee of Johnson & Johnson PRD. At the time this work was completed, she was an assistant professor at the University of Texas Southwestern Medical School at Dallas, where she continues as an adjunct faculty member.

Madhukar H. Trivedi, MD

Department of Psychiatry, University of Texas Southwestern Medical Center, Dallas, TX, USA

Stephen R. Wisniewski, PhD

Epidemiology Data Center, Graduate School of Public Health, University of Pittsburgh, Pittsburgh, PA, USA

Andrew A. Nierenberg, MD

Massachusetts General Hospital, Depression Clinical and Research Program, Boston, MA, USA

Bradley N. Gaynes, MD, MPH

Department of Psychiatry, University of North Carolina School of Medicine, Chapel Hill, NC, USA

Diane Warden, PhD, MBA

Department of Psychiatry, University of Texas Southwestern Medical Center, Dallas, TX, USA

David W. Morris, PhD

Department of Psychiatry, University of Texas Southwestern Medical Center, Dallas, TX, USA

James F. Luther, MA

Epidemiology Data Center, Graduate School of Public Health, University of Pittsburgh, Pittsburgh, PA, USA

Amy Farabaugh, PhD

Massachusetts General Hospital, Depression Clinical and Research Program, Boston, MA, USA

Ian Cook, MD

UCLA Semel Institute for Neuroscience and Human Behavior, Department of Psychiatry and Biobehavioral Sciences, Los Angeles, CA, USA

A. John Rush, MD

Department of Psychiatry, University of Texas Southwestern Medical Center, Dallas, TX, USA

BACKGROUND: Although major depressive disorder (MDD) is associated with significant impairments in health-related quality of life (HRQOL), few studies have evaluated HRQOL dysfunction in multiple domains. This report examined the psychological, physical, and social domains in a large sample of outpatients who entered the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) trial.

METHODS: The relationship of HRQOL and baseline sociodemographic and clinical features, including depressive severity, was evaluated. We assessed HRQOL with the 12-item Short Form Health Survey, the 5-item Work and Social Adjustment Scale, and the 16-item Quality of Life Enjoyment and Satisfaction Questionnaire.

RESULTS: Among 2307 participants, greater depressive symptom severity was associated with poorer HRQOL. After controlling for age and depression severity, lower HRQOL was related independently to being African American or Hispanic, less educated, unemployed, divorced or separated, having public medical insurance, and to having more general medical disorders. We found impairments across all 3 domains, with low correlations between the 3 measures of HRQOL chosen, suggesting that they evaluate different and nonoverlapping aspects of function.

CONCLUSION: Sociodemographically disadvantaged patients with greater general medical and depressive illness burden are at greatest risk for poorer quality of life. Distinct impairments are seen in the 3 domains of HRQOL.

KEYWORDS: Health-related quality of life, major depressive disorder, domains of quality of life

ANNALS OF CLINICAL PSYCHIATRY 2010;22(1):43–55

  INTRODUCTION

Depression is not only a common, often chronic, and recurrent disorder, but it is cardinally associated with significant impairment in work and daily social and psychological well-being.1 Major depressive disorder (MDD) is the fourth leading cause of disability worldwide2 and is predicted to become the second leading cause by the year 2020.3 The Medical Outcomes Study4 found that depressed individuals have comparable or worse physical, psychosocial, and role functioning than those who have chronic medical conditions. Similarly, the World Health Organization (WHO) Collaborative Study on Psychological Problems in General Health Care5 reported increased functional disability, even after controlling for physical disease severity among patients with depression. Furthermore, the longer a patient remains symptomatic, the lower the chances of a complete recovery,6 and thus greater dysfunction.

Health-related quality of life (HRQOL) in depression involves at least 3 specific domains of health—physical, psychological, and social—and each can be measured by either objective assessment or subjective perception. Furthermore, there are many components within each domain (eg, symptoms, ability to function, and disability). Whereas disease-specific instruments focus on the domains most relevant to the disease or condition under study, both generic and disease-specific instruments of HRQOL have often been used in general populations to assess a wide range of domains applicable to health states, conditions, and diseases.7,8 Currently no widely accepted single disease-specific measure of HRQOL evaluating all 3 domains is available for depression.

We previously reported on clinical and sociodemographic factors associated with HRQOL using baseline data from the first 1500 consecutively enrolled participants in the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) trial.9 HRQOL was assessed using the 12-item Short Form Health Survey (SF-12),10 the Work and Social Adjustment Scale (WSAS),11 and the Quality of Life Enjoyment and Satisfaction Questionnaire (Q-LES-Q).12 These specific measures offer a complementary assessment of different HRQOL domains.

Based on 1397 evaluable participants, we found that greater depressive symptom severity was associated with poorer HRQOL on all measures, confirming that the associated impairments in HRQOL are reflected in the 3 somewhat distinct domains of functioning.9 This finding has implications for the treatment of MDD in clinical practice settings, since these domains might be treatment targets and/or provide a means to assess outcomes. All 3 measures of HRQOL used in STAR*D were also independently correlated with each other.

The current report examines baseline data from the subsequent and distinct group of 2307 STAR*D enrollees and adds to the growing body of evidence suggesting that function and quality of life are important areas of focus for treatment of patients with MDD.

  METHODS

Study description

The methods of STAR*D are detailed elsewhere.13,14 STAR*D was designed to prospectively define which of several treatments are most effective for outpatients with nonpsychotic major depressive disorder (MDD) who report an unsatisfactory clinical outcome to an initial treatment and, if necessary, subsequent treatment(s). The diagnosis of nonpsychotic MDD was confirmed with the DSM-IV-TR15 checklist. Initially, eligible and consenting participants were treated with a selective serotonin reuptake inhibitor (SSRI). After up to 14 weeks of treatment, participants with an adequate clinical response could enter a 12-month naturalistic follow-up phase, while those with partial or no response could enter one or more subsequent randomized clinical trials. Clinical research coordinators (CRCs) at each clinical site worked closely with the participants and clinicians. Research outcome assessors (ROAs), masked to treatment assignment, collected depressive symptom ratings through telephone interviews. Additional research outcomes were collected by an automated interactive voice response (IVR) telephone interview.16

Study population

Of the second group (n = 2541) of consecutive outpatients who provided written informed consent and were enrolled in STAR*D, a total of 2307 completed baseline IVR assessments. Participating patients were enrolled from both primary care (N = 18) and specialty care (N = 23) clinics in either the public or private sector. Potentially eligible participants were age 18 to 75 and diagnosed with nonpsychotic MDD. Participants with general medical conditions (GMCs) were eligible if their GMCs did not contraindicate one or more of the protocol treatments. Participants with currently active substance abuse were also eligible, unless the participant required inpatient detoxification. Ineligible participants included those with bipolar disorder; schizophrenia; schizoaffective disorder; MDD with psychotic features (lifetime); or a current primary diagnosis of anorexia nervosa, bulimia nervosa, or obsessive-compulsive disorder. Participants who were pregnant or breastfeeding were also ineligible.

Baseline assessments

At the baseline visit, CRCs collected clinical and demographic information as well as completing the 17-item Hamilton Rating Scale for Depression (HRSD17)17 to establish inclusion/exclusion criteria, and the 16-item Quick Inventory of Depressive Symptomatology–Clinician-rated (QIDS C16),18 which assesses the 9 symptom domains to diagnose MDD as defined by the DSM-IV-TR. Participants completed the self-report Psychiatric Diagnostic Screening Questionnaire (PDSQ)19 to establish other current Axis I disorders. The HRSD17 and the 30-item Inventory of Depressive Symptomatology–Clinician-rated (IDS-C30)20,21 were collected with 72 hours of the baseline visit.

Health-related quality of life measures

Many instruments have been developed to quantify HRQOL in patients with depression and other chronic medical conditions.22-26 Three distinct measures were selected for use in STAR*D because they assess different domains of HRQOL. The Short Form Health Survey (SF-12),10 the Work and Social Adjustment Scale (WSAS),11 and the Quality of Life Enjoyment and Satisfaction Questionnaire (Q-LES-Q)12 were administered using the IVR system within 72 hours of the baseline visit.

The 12-item Short Form Health Survey (SF-12) is a self-report measure of perceived health and functioning, consisting of 12 questions ranging from general health status to specific physical limitations.10 The 12 items selected from the original 36-item Short Form Health Survey (SF-36)8,27 include questions about the patient’s general health and limitations in work or other activities due to health or emotional problems (eg, feeling depressed or anxious). Some items are scored as absent/present, while other items are scored on a Likert scale. Two subscale scores are generated: a physical health factor score and a mental health factor score. Each subscale has a possible range of 0 to 100, with a higher score indicating higher functioning.

The norm for each subscale is a score of 50. The SF-12 has been shown to be valid and to have acceptable test-retest reliability.28 Scores on the SF-12 Mental (r=–0.33 to –0.48) and the SF-12 Physical (r=–0.19 to –0.29) are modestly associated with overall depressive symptom severity.

The Work and Social Adjustment Scale (WSAS) is a 5-item, self-report scale that assesses the ability to work, to manage home and social affairs, and to form and maintain close relationships.11 Each question is rated on a 0-to-8 Likert scale, with 0 indicating no impairment and 8 indicating very severe impairment. The range of the total score is 0 to 40, with lower scores indicating higher functioning. A WSAS score >20 suggests at least moderately severe functional impairment. Scores between 10 and 20 indicate measurable functional impairment but less severe clinical symptomatology. The WSAS has demonstrated adequate internal consistency (Cronbach alpha ranges from 0.70 to 0.94) and test-retest reliability (r=0.73).11 Depressive symptom severity on standardized scales is strongly associated with the score on the WSAS (r=0.73), with significant differences between mildly-to-moderately depressed and moderately-to-severely depressed individuals.11

The Quality of Life Enjoyment and Satisfaction Questionnaire (Q-LES-Q) is a self-report instrument designed to measure satisfaction and enjoyment in various domains of functioning (eg, physical health, work, and household duties).12 The short version used in this study has 16 items, of which the first 14 assess discrete domains such as social relationships, living or housing situation, and physical health. Item 15 concerns respondents’ satisfaction with medication they are taking, if applicable. Item 16 is a global rating in which respondents are asked to rate their “overall life satisfaction and contentment.” Each item is scored on a 5-point Likert scale that indicates the degree of enjoyment or satisfaction achieved during the past week (1=very poor, 5=very good). Higher scores indicate better function.

Items 1 through 14 of the Q-LES-Q can be summed into a total score, with higher scores representing greater life enjoyment and satisfaction. For the present analyses, we used only the total score generated by the sum of the first 14 items. This scale has a Cronbach alpha of 0.90 and a test-retest reliability of 0.74 (n = 54).25 The construct validity of the Q-LES-Q is supported by moderately negative correlations with the Clinical Global Impressions–Severity of Illness scale (CGI-S) (r=–0.62 for the summary scale)25 and the HRSD17 (r=–0.61 for the summary scale).25

Medical comorbidity

The current burden due to general medical conditions (GMCs) was assessed using the Cumulative Illness Rating Scale (CIRS).29,30 The 14-item CIRS, completed by the clinician or CRC using a manual to guide scoring,31 gauged the number and severity/morbidity of GMCs relevant to different organ systems. Each condition was scored from 0 (no problem) to 4 (extremely severe/immediate treatment required/end organ failure/severe impairment in function), and the scores were summed across conditions to generate a total score.

Analytic methods

Pearson correlation coefficients were used to estimate the association among the HRQOL measures. All subsequent analyses were conducted separately for each measure of HRQOL. Bivariate analyses between the HRQOL measures and clinical and sociodemographic characteristics were assessed using a Pearson correlation coefficient when the clinical features and sociodemographic characteristics were continuous. For binary measures (eg, gender), a 2-group test was used to compare mean HRQOL scores. For discrete measures with >2 levels (eg, marital status), analyses of variance models were used to compare the HRQOL scores across the groups. If differences were detected, post hoc tests were conducted.

Regression analyses employing the backward elimination procedure were used to identify factors independently associated with HRQOL after controlling for the effects of age and depressive symptom severity (as measured by the HRSD17 total score). In the backward elimination analyses, the F statistics associated with the full set of independent variables are calculated and the variable contributing least to the model is eliminated. This process is repeated until all remaining independent variables are significant at the P < .10 level.

We controlled for age and severity of depression because these variables are consistently associated with HRQOL. Since we sought to identify additional factors independent of age and depressive symptom severity, age and severity of depression were forced into each of the regression models. Also, the variables for monthly household income and duration of index episode were transformed (natural logarithm) to meet the assumptions of the models.

  RESULTS

Baseline characteristics

TABLE 1 shows baseline demographic and clinical characteristics for the study sample. Of the 2307 participants, 63.2% were female, with the ethnic/racial composition of the sample reflective of the US Census numbers,32 except for a higher representation of African Americans (17% in this study vs 14% in the US population). Participants reported moderate-to-severe depressive symptom severity, and medical comorbidity was common. The sample reported moderately good physical function but low mental function on the SF-12. Similarly, the WSAS and Q-LES-Q revealed moderate impairment in work and social adjustment as well as life enjoyment and satisfaction.


TABLE 1

Baseline sociodemographic and clinical characteristics (N = 2307)

CHARACTERISTIC n (% N) Mean ± SD
Age   40.7 ± 13.3
Male gender 850 (36.8)  
Race
  Caucasian 1752 (76.1)  
  African American 392 (17.0)  
  Other 159 (6.9)  
Hispanic 332 (14.4)  
Highest educational degree
  None 311 (13.5)  
  High school/associate 1422 (61.7)  
  College/advanced 570 (24.8)  
Monthly household income ($)   2412 ± 3204
Employment status
  Employed 1283 (55.7)  
  Unemployed 890 (38.6)  
  Retired 130 (5.6)  
Medical insurance
  Private 1100 (49.2)  
  Public 344 (15.4)  
  None 792 (35.4)  
Marital status
  Single 706 (30.7)  
  Married/cohabiting 944 (41.0)  
  Divorced/separated 572 (24.8)  
  Widowed 81 (3.5)  
Age at first episode   25.6 ± 14.7
Duration of illness (y)   15.2 ± 13.1
No. episodes   6.0 ± 12.3
Duration of current episode (mo)   24.2 ± 50.8
CIRS Severity index   1.2 ± 0.7
CIRS—No. of categories endorsed   3.1 ± 2.4
HRSD17   19.6 ± 6.4
IDS-C30   35.3 ± 11.3
QIDS-SR16   15.4 ± 4.4
SF-12—Physical   49.2 ± 12.1
SF-12—Mental   26.9 ± 8.8
WSAS   23.5 ± 9.3
Q-LES-Q   41.7 ± 15.4
CIRS: Cumulative Illness Rating Scale; HRSD17: Hamilton Rating Scale for Depression; IDS-C30: Inventory of Depressive Symptomatology–Clinician-rated; QIDS-SR16: Quick Inventory of Depressive Symptomatology–Self-reported; Q-LES-Q: Quality of Life Enjoyment and Satisfaction Questionnaire; SF-12: Short Form Health Survey; WSAS: Work and Social Adjustment Scale.
Do the HRQOL measures assess distinct domains?

For the most part, the HRQOL measures used in the study have minimal overlap in their assessment of function, quality of life, and satisfaction with life based on the Pearson correlation coefficients. For example, the SF-12 mental subscale is negatively correlated with both the SF-12 physical subscale (r=–0.4) and the WSAS (r=–0.44). Similarly, the SF-12 physical subscale is negatively correlated with the WSAS (r=–0.24). Finally, the Q-LES-Q has minimal overlap with either the SF-12 mental subscale (r=0.37) or the SF-12 physical subscale (r=0.36). Of note, unlike the other HRQOL measures where there is minimal overlap, the Q-LES-Q is moderately (negatively) correlated with the WSAS (r=–0.68) (TABLE 2).


TABLE 2

Intercorrelations among HRQOL measures

HRQOL MEASURE Q-LES-Q SF-12 MENTAL SF-12 PHYSICAL
SF-12 Mental 0.36874    
SF-12 Physical 0.36060 –0.39993  
WSAS –0.68139 –0.43528 –0.24256
HRQOL, health-related quality of life; Q-LES-Q: Quality of Life Enjoyment and Satisfaction Questionnaire; SF-12: Short Form Health Survey; WSAS: Work and Social Adjustment Scale.
All P values <.0001.
What sociodemographic features are associated with HRQOL?

TABLE 3 presents the associations between participant sociodemographic features and various measures of HRQOL, without adjustments for age or depressive severity.


TABLE 3

Associations of sociodemographic characteristics with HRQOL

CHARACTERISTIC SF-12 PHYSICALa SF-12 MENTALb WSASc Q-LES-Qd
MEAN ± SD P MEAN ± SD P MEAN ± SD P MEAN ± SD P
Gender   .1721   .0027   .9641   .5663
  Male 48.7 ± 12.6   27.6 ± 8.4   23.5 ± 8.9   41.9 ± 15.0  
  Female 49.4 ± 11.9   26.5 ± 9.1   23.5 ± 9.6   41.6 ± 15.6  
Race   <.0001   <.0001   <.0033   .0893
  Caucasian 50.0 ± 11.9   26.5 ± 8.6   23.2 ± 9.2   42.1 ± 15.4  
  African American 44.9 ± 12.2   28.6 ± 9.8   24.9 ± 10.0   40.2 ± 15.6  
  Other 50.5 ± 12.0   27.1 ± 8.9   23.4 ± 8.8   41.8 ± 14.1  
Hispanic   <.0001   .0037   .9833   .0045
  Yes 45.8 ± 11.2   28.2 ± 8.3   23.5 ± 10.1   39.3 ± 16.8  
  No 49.7 ± 12.2   26.7 ± 8.9   23.5 ± 9.2   42.1 ± 15.1  
Highest educational degree   <.0001   <.0001   .0020   <.0001
  None 43.1 ± 11.2   30.3 ± 9.4   23.2 ± 10.7   40.1 ± 17.0  
  High school/associate 48.5 ± 12.1   26.7 ± 8.7   24.0 ± 9.3   40.7 ± 15.4  
  College/advanced 54.1 ± 10.7   25.6 ± 8.6   22.4 ± 8.6   45.0 ± 13.9  
Employment status   <.0001   <.0001   <.0001   <.0001
  Employed 52.6 ± 10.4   26.0 ± 8.5   22.3 ± 8.8   44.1 ± 14.4  
  Unemployed 45.4 ± 12.8   27.4 ± 9.0   25.5 ± 9.6   38.0 ± 16.1  
  Retired 41.1 ± 11.7   32.0 ± 9.3   21.2 ± 9.8   43.7 ± 14.8  
Medical insurance   <.0001   <.0001   <.0001   <.0001
  Private 51.6 ± 11.3   26.6 ± 8.7   22.5 ± 9.3   44.4 ± 14.9  
  Public 42.2 ± 11.5   28.9 ± 9.3   23.6 ± 9.7   37.8 ± 15.9  
  None 48.8 ± 12.4   26.5 ± 8.8   24.7 ± 9.1   39.9 ± 15.1  
Marital status   <.0001   <.0001   <.0001   <.0001
  Single 52.3 ± 11.2   25.5 ± 8.4   23.3 ± 8.9   43.0 ± 14.7  
  Married/cohabiting 48.8 ± 11.9   28.0 ± 9.2   22.4 ± 9.6   43.0 ± 15.5  
  Divorced/separated 46.9 ± 12.8   26.4 ± 8.6   25.6 ± 9.1   37.8 ± 15.4  
  None 43.1 ± 11.2   29.5 ± 9.1   23.3 ± 8.8   42.8 ± 14.0  
  r P r P r P r P
Age –0.352 <.0001 0.235 <.0001 –0.004 .8460 –0.036 .0858
Monthly household income 0.214 <.0001 –0.038 .0913 –0.128 <.0001 0.179 <.0001
HRQOL, health-related quality of life; Q-LES-Q: Quality of Life Enjoyment and Satisfaction Questionnaire; r: correlation coefficient; SF-12: Short Form Health Survey; WSAS: Work and Social Adjustment Scale.
aPairwise tests significant at P < .05: blacks < whites and others; college/advanced > high school/associates > no degree; employed > unemployed > retired; private insurance > no insurance > public insurance; singles > married/cohabiting > divorced/separated > widowed.
bPairwise tests significant at P < .05: blacks > whites; college/advanced < high school/associates < no degree; employed < unemployed < retired; public > private and no insurance; married/cohabiting and widowed > single and divorced/separated.
cPairwise tests significant at P < .05: blacks > whites; college/advanced < high school/associates; unemployed > employed and retired; no insurance > private; divorced/separated > single and married/cohabiting.
dPairwise tests significant at P < .05: college/advanced > high school/associates and no degree; employed < unemployed and retired; private > no > public insurance; divorced/separated < single, married/cohabiting, and widowed.
Statistical significance is designated by bold type.
What clinical features are associated with HRQOL?

TABLE 4 presents the associations between participant clinical features and the 3 measures of HRQOL, without adjustments for age or depressive severity.


TABLE 4

Associations of clinical characteristics with HRQOL

CHARACTERISTIC SF-12 PHYSICAL SF-12 MENTAL WSAS Q-LES-Q
r P r P r P r P
Age at first episode –0.187 <.0001 0.172 <.0001 –0.033 .1160 0.033 .1158
Duration of illness (y) –0.088 <.0001 –0.012 .5624 0.063 .0026 –0.075 .0004
No. episodes 0.034 .1361 –0.067 .0031 0.026 .2565 –0.015 .5010
Duration of current episode (mo) –0.183 <.0001 0.089 <.0001 0.055 .0083 –0.056 .0071
CIRS Severity index –0.451 <.0001 0.209 <.0001 0.075 .0003 –0.141 <.0001
HRSD17 –0.295 <.0001 –0.249 <.0001 0.445 <.0001 –0.544 <.0001
IDS-C30 –0.282 <.0001 –0.305 <.0001 0.507 <.0001 –0.606 <.0001
QIDS-SR16 –0.146 <.0001 –0.388 <.0001 0.519 <.0001 –0.550 <.0001
CIRS: Cumulative Illness Rating Scale; HRSD17: Hamilton Rating Scale for Depression; HRQOL, health-related quality of life; IDS-C30: Inventory of Depressive Symptomatology–Clinician-rated; QIDS-SR16: Quick Inventory of Depressive Symptomatology–Self-reported; Q-LES-Q: Quality of Life Enjoyment and Satisfaction Questionnaire; SF-12: Short Form Health Survey; WSAS: Work and Social Adjustment Scale.
Statistical significance is designated by bold type.
What sociodemographic and clinical features are independently associated with HRQOL?

TABLE 5 presents those factors that were independently associated with measures of HRQOL after controlling for age and symptom severity.

African Americans reported worse physical function on the SF-12 compared with Caucasians. Other factors associated with worse reported physical function included being unemployed, having public medical insurance, having increased GMCs, being of Hispanic ethnicity, and having more severe depression. In contrast, having more education was associated with better physical function. On the SF-12 mental function subscale, better mental function was associated with African American race and employment status (with unemployed participants reporting better mental function than those who were employed). Men also reported slightly better mental function than women. Interestingly, increased GMCs appeared to be associated with better mental function. In contrast, worse reported mental function was associated with having more years of education, more severe depression, and being divorced or separated.

Worse reported work and social function, as assessed by the WSAS, was associated with greater severity of depression, more years of education, being unemployed, and being divorced or separated (compared with being married). Finally, worse reported quality of life and satisfaction, as measured by the Q-LES-Q, was associated with greater severity of depression, being unemployed, having a high school education or greater, and being divorced or separated.


TABLE 5

Characteristics independently associated with HRQOL*

CHARACTERISTIC SF-12 PHYSICAL SF-12 MENTAL WSAS Q-LES-Q
B P sr2 B P sr2 B P sr2 B P sr2
Agea –3.407 .0523 0.1263 3.098 .0276 0.0602 –0.003 .8704 0.0002 –0.041 .1664 0.0008
HRSD17a –0.369 <.0001 0.0887 –0.412 <.0001 0.0641 0.643 <.0001 0.2030 –1.265 <.0001 0.3151
Male gender       0.844 .0416 0.0004            
Race (ref: Caucasian)
  African American –2.244 .0014 0.0109 2.049 .0003 0.0084            
  Other 0.709 .4822 0.0004 0.903 .2661 0.0022            
Hispanic –1.484 .0408 0.0117 1.276 .0282 0.0101       –1.547 .0883 0.0010
Highest educational degree (ref: none)
  High school/associate 0.834 .3103 0.0080 –1.759 .0077 0.0026 1.926 .0046 0.0002 –2.787 .0078 0.0013
  College/advanced 3.561 .0002 0.0284 –3.645 <.0001 0.0225 2.688 .0008 0.0028 –2.680 .0280 0.0002
Employment status (ref: employed)
  Unemployed –3.620 <.0001 0.0263 1.111 .0103 0.0036 1.818 .0002 0.0089 –2.912 <.0001 0.0103
  Retired –3.033 .0124 0.0054 1.681 .0824 0.0028 –0.278 .7884 0.0001 0.257 .8684 1.91-5
Monthly household income 0.000 .0379 0.0037       –0.000 .0136 0.0059 0.000 .0181 0.0051
Medical insurance (ref: private)
  Public –2.343 .0040 0.0067       –1.569 .0235 0.0024      
  None 0.162 .7840 9.91-7       0.023 .9639 0.0001      
Marital status (ref: married/cohabiting)
  Single       –0.740 .1432 0.0002 1.255 .0213 0.0007 –1.940 .0176 0.0005
  Divorced/separated       –1.911 .0002 0.0081 1.934 .0005 0.0047 –3.535 <.0001 0.0080
  Widowed       –1.170 .3252 0.0006 2.997 .0181 0.0028 –0.059 .9752 4.08-7
Age at first episode 3.181 .0695 0.0001 –2.965 .0347 0.0021            
Duration of illness (y) 3.219   0.0027 –3.018 .0324 0.0031            
Duration of current episode (mo) –0.013 .0172 0.0037                  
CIRS Severity index –3.918 <.0001 0.0369 1.857 <.0001 0.0159            
ß: beta coefficient; CIRS: Cumulative Illness Rating Scale; HRQOL, health-related quality of life; HRSD17: Hamilton Rating Scale for Depression; Q-LES-Q: Quality of Life Enjoyment and Satisfaction Questionnaire; SF-12: Short Form Health Survey; sr2: squared, semipartial correlation coefficient (amount of variance in Y explained by X); WSAS: Work and Social Adjustment Scale.
*Characteristics retained by multivariate backward elimination regression model.
aInclusion forced into each model.
Statistical significance is designated by bold type.

  DISCUSSION

The current results replicate our earlier finding of associations between greater depressive symptom severity and reduced HRQOL on all measures,9 as well as independent associations between worse HRQOL and race, ethnicity, marital status, education level, employment status, health insurance status, and presence of current GMCs. These data also further support the growing literature linking mood disorders to functional disabilities in the social, emotional, and physical domains of life.33

In the past, there has been criticism of the routine inclusion of quality of life (QOL) evaluation in clinical trials,34 with concerns raised about the rationale for its use and the clinical utility of such measures. Quality of life is, however, increasingly recognized as an important outcome measure in the care of patients with severe mental illness.35-37 In this article, we examined more closely the specific domains of HRQOL that were found to be impaired in this large sample of depressed patients at baseline. We found low correlations in the assessment of function, quality of life, and satisfaction with life based on the Pearson correlation coefficients for psychological, physical, and social domains (TABLE 2), suggesting that these 3 measures of HRQOL are evaluating different and nonoverlapping aspects of function.

A closer inspection of the specific measures chosen reveals that some emphasize certain domains more than others (eg, the physical subscale of the SF-12 covers predominantly the physical domain, with some overlap with the social domain). Similarly, the mental subscale of the SF-12 predominantly covers the psychological domain with some overlap with the social domain and symptoms of MDD. In contrast, the WSAS focuses more on the social domain with some physical components. Similarly, the Q-LES-Q also focuses more on the social domain, with only a few items related to the impact of mood and physical health. Our findings suggest that these 3 measures of HRQOL offer a complementary assessment of different HRQOL domains.

Looking at the physical domain (largely covered by the SF-12 physical subscale), increased comorbid medical illness, fewer years of education, having public health insurance, being African American or Hispanic, and being unemployed were all associated with worse reported physical function. The association with these particular sociodemographic factors replicates previous findings in terms of medical comorbidity as well as race and ethnicity, as discussed in more detail below.

In terms of the psychological domain (largely covered by the SF-12 mental subscale), better mental function was associated with African American race, being unemployed, and having fewer years of education. Interestingly, increased comorbid medical illness also appeared to be independently associated with better mental function. This result, however, should be interpreted with caution, given that in the analysis the investigators controlled for depressive severity. Furthermore, increased levels of GMCs are known to have a negative impact on MDD.4,38-40 After correction for depression severity in the analysis, this may give the appearance of lower levels of depression in people with these conditions.

Finally, in the social domain (largely covered by the WSAS and Q-LES-Q scales), being unemployed, having more years of education, and being divorced or separated were associated with poorer reported social function. For the purpose of the following discussion, we will focus on those associations that were replicated from our original study (P ≤ .05 level).9

As mentioned earlier, a body of evidence suggests that comorbidity of depression with GMCs is associated with worse HRQOL.4,39 Our findings indicating that increased general medical comorbidity is associated with decreased function in the physical domain on self-report replicates our earlier results9 and supports prior findings.4 This finding has particular relevance for primary care practice, where the majority of patients with depression are treated, further emphasizing the importance of assessment and monitoring of HRQOL in depressed patients with comorbid medical illness.

Similarly, as in our earlier findings, African American participants reported worse SF-12 physical function but better SF-12 mental function than their Caucasian counterparts. These findings are in general agreement with data from the sample of patients with depression in the Medical Outcomes Study, which showed that although African Americans were the least likely to report suicidal ideation, they reported the poorest quality of life,41 suggesting that racial differences in symptom presentation and HRQOL could have clinical and social consequences in general medical settings. Also, when screening for depression, clinicians need to be more aware of the possibility of underreporting of depressive symptoms in this patient population and may need to probe for evidence of other signs of functional impairment.

Finally, as with our earlier findings, we found that employment status was independently related to multiple measures of HRQOL. Patients who were unemployed reported worse physical function, worse work and social function, and worse quality of and satisfaction with life. A recent study also found evidence that increased financial hardship is associated with worse HRQOL and that those who are depressed have higher rates of unemployment.42 Similarly 2 large, community-based European studies—the Study of the Epidemiology of Mental Disorders (ESEMeD) project and the Netherlands Mental Health Survey and Incidence Study (NEMESIS)—reported higher rates of mental disorders43 and disability33 in those who were unemployed.

Some of the current results were counterintuitive. For example, having more years of education was again associated with worse mental functioning on the SF-12 mental function scale. Possible explanations include a possible association between being more educated and more pessimistic because of increased awareness of the impact of the disorder or because being more educated may influence the individual’s self-reporting of how their depression affects their life.

The amount of variance accounted for by the variables in this analysis is small, and the absence of normative data precludes understanding how these associations are different among depressed patients than among normal controls. Also, because of the large sample size, minor differences in the mean scores may have potentially resulted in statistical significance without any meaningful clinical significance. Only 2 variables, age and baseline severity of depression, individually account for more than 5% of the variance in some of the models, which implies that while the other variables may be statistically significantly associated with HRQOL, they do not account for much of the variation. Other limitations of the current study include: (1) all measures of function were in the form of subjective self-report, and any association with depression could simply represent depressed individuals’ negative self-assessment rather than truly poorer function; (2) the HRQOL data obtained were from a cross-sectional sample, thus precluding inference of cause and effect; (3) the study lacked a non-MDD comparison group or normative community sample; and (4) as with our earlier study,9 results may not be generalizable to different patient populations.

Further complicating interpretation of this data, as Ormel and colleagues point out in a recent publication,44 it is not clear how much psychosocial disability during a major depressive episode (MDE) was present before the MDE, and how much remains after the episode. The authors suggest the relationship between depression and HRQOL is complex, probably including reverse causality and shared causes. In NEMESIS, the large prospective Dutch psychiatric population-based survey, the investigators looked at 4 domains of functioning: spouse/partner, work/employment, housekeeping, and leisure time activities. The investigators found that patients having a first MDE during the study period had higher disability scores (self-report) long before their episode (trait effect), compared with those who never had an MDE.44 They also found that during the MDE, disability further increased in those patients experiencing their first or a recurrent MDE (state effect), returning in most cases to premorbid levels after remission, except in those subjects who experienced a severe recurrent episode (scar effect).

One strength of this study is that it provides a strong replication of our earlier finding in a second group of STAR*D enrollees. The literature contains many papers in which different factors are shown to be associated with HRQOL. However, the variables that are associated with HRQOL are not consistent across studies. For example, gender may be associated in one study but not in another. Since these studies have different designs in different populations, one cannot say that the studies show different associations because of the study design, population, or other reasons. Interestingly, a recent article discussed the high rate of nonreplication (lack of confirmation) of research discoveries,45 suggesting that the probability that a research claim is true may depend on study power and bias, the number of studies testing the same question, and the ratio of true to no relationships. Given this situation, a research finding is less likely to be true when the study sample is small, when effect sizes are small, when there is less preselection of tested relationships, and when designs and outcomes are more flexible. In contrast, we were able to conduct a replication study in a second large sample, in the same study with the same design and same sites, thus removing a significant amount of variability that exists in the literature.

Even though it is likely that the relationship between depression and HRQOL is complex, our findings may have implications for the treatment of MDD in clinical practice settings. The individual domains of HRQOL (ie, psychological, physical, and social) may represent important treatment targets (state) in addition to symptoms of depression, and progress in these areas can be objectively measured. Given that in our community-based sample, there appears to be significant impairment in both physical and social domains at baseline, we suggest that clinicians treating patients who have MDD should monitor these areas throughout treatment as well as consider additional interventions targeting both areas. It is possible that some aspects of HRQOL improve with treatment (pharmacologic and/or psychotherapy) and that partial response may be associated with the presence of ongoing, specific HRQOL deficits. Implementing programs such as those focused on disease self-management, as well as educational interventions, may potentially improve outcomes when added to pharmacologic treatment.

The difficulty of interpreting HRQOL data in this population also highlights the need to develop a single, clinically useful HRQOL measure that is specific to MDD and captures all 3 domains. In the meantime, based on our findings to date, we know that those measures used in STAR*D are easy to administer and have minimal overlap in their assessment of function, quality of life, and satisfaction with life.

  CONCLUSION

The findings from the current analysis replicate many of our previous findings in a second large participant sample and further emphasize the importance of assessing these 3 different but complementary domains of HRQOL in patients with MDD in clinical practice settings. Factors such as race, ethnicity, monthly income, marital status, employment status, education level, health insurance status, medical comorbidity, and age at first onset of MDD were again found to be independently associated with HRQOL at baseline, emphasizing the importance of measuring functional status as well as depressive symptoms in this patient population, both to identify areas in need of improvement and to monitor treatment progress. Additionally, as mentioned in the discussion, in our large, community-based sample, there appears to be significant impairment in both physical and social domains at baseline, suggesting that clinicians treating patients with MDD should monitor these specific domains throughout treatment as well as consider additional interventions targeting both areas. For example, specific interventions such as disease self-management may be a useful adjunctive treatment for those patients who continue to demonstrate functional impairment despite optimal pharmacologic treatment. Such an approach may aid clinicians in the treatment of patients, allowing them to better tailor treatment to the individual patient.

ACKNOWLEDGEMENTS: This project was funded by the National Institute of Mental Health under Contract N01MH90003 to UT Southwestern Medical Center at Dallas (PI: A.J. Rush). The content of this publication does not necessarily reflect the views or policies of the US Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the US government.

We appreciate the support of Bristol-Myers Squibb, Forest Laboratories, GlaxoSmithKline, King Pharmaceuticals, Organon, Pfizer Inc, and Wyeth Pharmaceuticals for providing medications at no cost for this trial. We would also like to acknowledge the editorial support of Jon Kilner, MS, MA.

DISCLOSURES: Ella Daly, MB, MRC Psych, is currently an employee of Johnson & Johnson PRD, however, at the time this work was completed, she was an assistant professor at the University of Texas Southwestern Medical School, Dallas, where she continues as an adjunct faculty member.

Madhukar H. Trivedi, MD, is a consultant for AstraZeneca; Bristol-Myers Squibb; Cephalon; Eli Lilly and Company; Evotec; Fabre-Kramer Pharmaceuticals, Inc.; Forest Pharmaceuticals; GlaxoSmithKline; Janssen, L.P.; Johnson & Johnson PRD; Medtronic; Neuronetics; Otsuka Pharmaceuticals; Pfizer Inc; Shire Development Inc.; and Wyeth-Ayerst Laboratories. He serves on speakers bureaus for AstraZeneca; Bristol-Myers Squibb; Forest Pharmaceuticals; and Otsuka Pharmaceuticals. He receives grant support from the Agency for Health-care Research and Quality; the National Institute of Mental Health; the National Institute on Drug Abuse; and Targacept.

Stephen R. Wisniewski, PhD, is a consultant for Bristol-Myers Squibb (2007-2008), Case Western University (2007), Cyberonics, Inc. (2005-2009), ImaRx Therapeutics, Inc. (2006), Organon (2007), and Singapore Clinical Research Institute (2009).

Andrew A. Nierenberg, MD, is a full-time employee of the Massachusetts General Hospital (MGH). He has received grant/research support through MGH from the National Institute of Mental Health, PamLab, Pfizer Inc, and Shire. He received honoraria from American Drug Utilization Review, the American Society for Clinical Psychopharmacology, Baystate Medical Center, Belvoir Publications, Columbia University, Hillside Hospital, Imedex, MBL Publishing, MJ Consulting, New York State, Physicians Postgraduate Press, SUNY Buffalo, the University of Pisa, the University of Texas Southwestern Medical Center at Dallas, and the University of Wisconsin. Dr. Nierenberg is a presenter for the MGH Psychiatry Academy (MGHPA), whose education programs were supported through independent medical education (IME) grants from AstraZeneca, Bristol-Myers Squibb, Eli Lilly and Company in 2009. He provides advisory/consulting services to the American Psychiatric Association (only travel expenses paid), Appliance Computing, Inc. (Mindsite), BrainCells Inc., Brandeis University, Eli Lilly and Company, Novartis, PGxHealth, Schering-Plough, Shire, Takeda, and Targacept. Dr. Nierenberg owns stock options in Appliance Computing, Inc., and BrainCells Inc. Through MGH, he is named for copyrights to the Clinical Positive Affect Scale and the MGH Structured Clinical Interview for the Montgomery Asberg Depression Scale exclusively licensed to the MGH Clinical Trials Network and Institute (CTNI). Also through MGH, Dr. Nierenberg has a patent extension application for the combination of buspirone, bupropion, and melatonin for the treatment of depression.

Bradley N. Gaynes, MD, MPH, receives grant/research support from the Agency for Healthcare Research and Quality, M-3 Information, and the National Institute of Mental Health.

Amy Farabaugh, PhD, receives grant/research support from the National Alliance for Research in Schizophrenia and Depression (NARSAD) and the National Institute of Mental Health (K23).

Ian Cook, MD, receives research support from Aspect Medical Systems, the National Institutes of Health, Neuronetics, and Sepracor. He provides advisory/consulting services to Bristol-Myers Squibb, Neuronetics, Scale Venture Partners, and USDOJ, and serves on speakers bureaus for the Medical Education Speakers Network, Neuronetics, and Wyeth Pharmaceuticals. Patents on biomedical inventions are assigned to and owned by the University of California.

A. John Rush, MD, is a speaker for Bristol-Myers Squibb and Otsuka Pharmaceuticals.

Diane Warden, PhD, MBA, David W. Morris, PhD, and James F. Luther, MA, report no competing interests or disclosures.

    REFERENCES

  1. Kessler RC, Berglund P, Demler O, et al, And the National Comorbidity Survey Replication. The epidemiology of major depressive disorder: results from the National Comorbidity Survey Replication (NCS-R). JAMA. 2003;289:3095–3105.
  2. Ustun TB, Ayuso-Mateos JL, Chatterji S, et al. Global burden of depressive disorders in the year 2000. Br J Psychiatry. 2004;184:386–392.
  3. Murray CJL, Lopez AD. The global burden of disease. Boston, MA: Harvard University Press; 1996.
  4. Wells KB, Stewart A, Hays RD, et al. The functioning and well-being of depressed patients. Results from the Medical Outcomes Study. JAMA. 1989;262:914–919.
  5. Ormel J, VonKorff M, Ustun TB, et al. Common mental disorders and disability across cultures. Results from the WHO Collaborative Study on Psychological Problems in General Health Care. JAMA. 1994;272:1741–1748.
  6. Keller MB, Lavori PW, Mueller TI, et al. Time to recovery, chronicity, and levels of psychopathology in major depression. A 5-year prospective follow-up of 431 subjects. Arch Gen Psychiatry. 1992;49:809–816.
  7. Bergner M, Bobbitt RA, Carter WB, et al. The Sickness Impact Profile: development and final revision of a health-status measure. Med Care. 1981;19:787–805.
  8. Ware JE, Sherbourne CD. The MOS 36-Item Short-Form Health Survey (SF-36). 1. Conceptual framework and item selection. Med Care. 1992;30:473–483.
  9. Trivedi MH, Rush AJ, Wisniewski SR, et al. Factors associated with health-related quality of life among outpatients with major depressive disorder: a STAR*D report. J Clin Psychiatry. 2006;67:185–195.
  10. Ware JE, Kosinski M, Keller SD. A 12-item Short-Form Health Survey: construction of scales and preliminary tests of reliability and validity. Med Care. 1996;34:220–233.
  11. Mundt JC, Marks IM, Shear MK, et al. The Work and Social Adjustment Scale: a simple measure of impairment in functioning. Br J Psychiatry. 2002;180:461–464.
  12. Endicott J, Nee J, Harrison W, et al. Quality of Life Enjoyment and Satisfaction Questionnaire: a new measure. Psychopharmacol Bull. 1993;29:321–326.
  13. Fava M, Rush AJ, Trivedi MH, et al. Background and rationale for the sequenced treatment alternatives to relieve depression (STAR*D) study. Psychiatr Clin North Am. 2003;26:457–494, x.
  14. Rush AJ, Fava M, Wisniewski SR, et al. Sequenced treatment alternatives to relieve depression (STAR*D): rationale and design. Control Clin Trials. 2004;25: 119–142.
  15.  Diagnostic and statistical manual of mental disorders, 4th ed, text rev. Washington, DC: American Psychiatric Association; 2000.
  16. Kobak KA, Taylor LV, Dottl SL, et al. A computer-administered telephone interview to identify mental disorders. JAMA. 1997;278:905–910.
  17. Hamilton M. A rating scale for depression. J Neurol Neurosurg Psychiatry. 1960;23:56–62.
  18. Rush AJ, Trivedi MH, Ibrahim HM, et al. The 16-item Quick Inventory of Depressive Symptomatology (QIDS), clinician rating (QIDS-C), and self-report (QIDS-SR): a psychometric evaluation in patients with chronic major depression. Biol Psychiatry. 2003;54:573–583.
  19. Zimmerman M, Mattia JI. The reliability and validity of a screening questionnaire for 13 DSM-IV Axis I disorders (the Psychiatric Diagnostic Screening Questionnaire) in psychiatric outpatients. J Clin Psychiatry. 1999;60:677–683.
  20. Rush AJ, Giles DE, Schlesser MA, et al. The Inventory for Depressive Symptomatology (IDS): preliminary findings. Psychiatry Res. 1986;18:65–87.
  21. Rush AJ, Gullion CM, Basco MR, et al. The Inventory of Depressive Symptomatology (IDS): psychometric properties. Psychol Med. 1996;26:477–486.
  22. Trivedi MH. Sensitizing clinicians and patients to the social and functional aspects of remission. J Clin Psychiatry. 2001;62:32–35.
  23. Ware JE Jr. Conceptualization and measurement of health-related quality of life: comments on an evolving field. Arch Phys Med Rehabil. 2003;84:S43–S51.
  24. Papakostas GI, Petersen T, Mahal Y, et al. Quality of life assessments in major depressive disorder: a review of the literature. Gen Hosp Psychiatry. 2004;26:13–17.
  25. Endicott J, Nee J, Harrison W, et al. Quality of life enjoyment and satisfaction questionnaire: a new measure. Psychopharmacol Bull. 1993;29:321–326.
  26. Andresen EM, Meyers AR. Health-related quality of life outcomes measures. Arch Phys Med Rehabil. 2000;81:S30–S45.
  27. Ware JE, Snow KK, Kosinski M, et al. SF-36 health survey: manual and interpretation guide. Boston, MA: The Health Institute; 1993.
  28. Gandek B, Ware JE, Aaronson NK, et al. Cross-validation of item selection and scoring for the SF-12 Health Survey in nine countries: results from the IQOLA Project. International Quality of Life Assessment. J Clin Epidemiol. 1998;51:1171–1178.
  29. Miller MD, Paradis CF, Houck PR, et al. Rating chronic medical illness burden in geropsychiatric practice and research: application of the Cumulative Illness Rating Scale. Psychiatry Res. 1992;41:237–248.
  30. Linn BS, Linn MW, Gurel L. Cumulative illness rating scale. J Am Geriatr Soc. 1968;16:622–626.
  31. Miller M, Tower A. A manual of guidelines for scoring the Cumulative Illness Rating Scale for geriatrics (CIRS-G). Pittsburgh, PA: University of Pittsburgh; 1991.
  32.  US Census Bureau. American community survey. http://www.census.gov/acs/www/. Accessed August 20, 2009.
  33. Bijl RV, Ravelli A. Current and residual functional disability associated with psychopathology: findings from the Netherlands Mental Health Survey and Incidence Study (NEMESIS). Psychol Med. 2000;30:657–668.
  34.  Quality of life and clinical trials [editorial]. Lancet. 1995;346:1–2.
  35. Sim K, Mahendran R, Chong SA. Health-related quality of life and psychiatric comorbidity in first episode psychosis. Compr Psychiatry. 2005;46:278–283.
  36. Lehman AF. Measures of quality of life among persons with severe and persistent mental disorders. Soc Psychiatry Psychiatr Epidemiol. 1996;31:78–88.
  37. Linzer M, Spitzer R, Kroenke K, et al. Gender, quality of life, and mental disorders in primary care: results from the PRIME-MD 1000 study. Am J Med. 1996;101:526–533.
  38. Katon WJ. Clinical and health services relationships between major depression, depressive symptoms, and general medical illness. Biol Psychiatry. 2003;54:216–226.
  39. Evans DL, Charney DS. Mood disorders and medical illness: a major public health problem. Biol Psychiatry. 2003;54:177–180.
  40. Kessler RC, Nelson CB, McGonagle KA, et al. Comorbidity of DSM-III-R major depressive disorder in the general population: results from the US National Comorbidity Survey. Br J Psychiatry Suppl. 1996;168:17–30.
  41. Jackson-Triche ME, Greer Sullivan J, Wells KB, et al. Depression and health-related quality of life in ethnic minorities seeking care in general medical settings. J Affect Disord. 2000;58:89–97.
  42. Avis NE, Assmann SF, Kravitz HM, et al. Quality of life in diverse groups of midlife women: assessing the influence of menopause, health status and psychosocial and demographic factors. Qual Life Res. 2004;13:933–946.
  43. Alonso J, Angermeyer MC, Bernert S, et al. Prevalence of mental disorders in Europe: results from the European Study of the Epidemiology of Mental Disorders (ESEMeD) project. Acta Psychiatr Scand Suppl. 2004;21–27.
  44. Ormel J, Oldehinkel AJ, Nolen WA, et al. Psychosocial disability before, during, and after a major depressive episode: a 3-wave population-based study of state, scar, and trait effects. Arch Gen Psychiatry. 2004;61:387–392.
  45. Ioannidis JP. Why most published research findings are false. PLoS Med. 2005;2:696–701.

CORRESPONDENCE: Madhukar H. Trivedi, MD, Professor of Psychiatry, Lydia Bryant Professorship in Psychiatric Research, Director, Mood Disorders Program, University of Texas, Southwestern Medical Center at Dallas, 5323 Harry Hines Blvd. Dallas, TX 75390-9119 USA E-MAIL: madhukar.trivedi@utsouthwestern.edu