Beyond Symptomatic Improvement:Assessing Real-World Outcomes in Patients With Major Depressive Disorder

Abstract

Objective:

To quantify the negative impact that major depressive disorder (MDD) has on quality of life, disability, and work, family, and overall psychosocial functioning. Available scales that assess these areas of impairment as they relate to patients with MDD are described.

Data Sources:

PubMed searches were conducted using the following terms: (MDD OR major depressive disorder) AND (absenteeism OR absente*); AND (quality of life OR QOL); AND (psychosocial function*); AND (presente* OR presenteeism); AND (health care cost* OR [health care] cost*); AND (health outcome*); AND (functional outcome*); AND (family life); AND (disabil* OR disability); AND (work function*); AND (unemployment OR unemploy*). The literature search was conducted in July 2008 and was restricted to English language articles. There were no limits set on the dates of the search.

Study Selection:

Two hundred twenty potential articles were identified. Among these studies, 48 presented primary data directly demonstrating the effect of MDD on quality of life, disability, and work, family, and overall psychosocial functioning.

Data Extraction:

Primary data were compiled from these studies and are summarily described. Available scales that assess quality of life, disability, and work, family, and overall psychosocial functioning are also described.

Data Synthesis:

MDD was found to be associated with significant disability and declines in functioning and quality of life. The Sheehan Disability Scale, the 36-item Short-Form Health Survey, and the Work Limitations Questionnaire were the most commonly used scales according to this review of the literature, but the majority of studies used direct and indirect disability measures, such as health care and other disability-related costs.

Conclusions:

In addition to assessing symptomatic outcomes, physicians should routinely assess their depressed patients on “real-world” outcomes. The development of a concise functional outcome measure specific to MDD is necessary for busy clinical practices.

According to the National Comorbidity Survey (NCS), major depressive disorder (MDD) has an estimated lifetime prevalence of 16%. NCS data also reveal that the majority of participants that met MDD criteria were currently employed and during the past year experienced some form of role impairment in work, family, or social functioning. The degree of impairment was commensurate with the severity of depressive symptomatology.¹ Furthermore, the World Health Organization has estimated that, among the 10 most disabling diseases (eg, diabetes, tuberculosis, hepatitis, sexually transmitted diseases), MDD is responsible for 5% of the total global disease burden associated with these disorders² and is ranked fourth among the leading causes of disease burden.³

Despite the overwhelming evidence demonstrating that MDD exerts a significantly negative effect on functioning and quality of life (Table 1), when depressed patients are assessed, either in a clinical or research setting, the focus primarily remains on the severity of depressive symptoms, with significantly less attention being paid to the impact that MDD has on these “real-world” outcomes. Other areas of medicine (eg, cardiovascular medicine,^4,5 diabetes⁶) have begun to expand the assessment of treatment outcomes from simply measuring symptomatic improvement to include a broader range of outcomes, thereby providing a useful example of how assessing these outcomes can help in the decision-making process of the various health care stakeholders (ie, physicians, employers, and payors/managed care plans) involved in the treatment of depressed patients.

Table 1.

Summary of Key Studies Assessing the Disability Associated With Major Depressive Disorder (MDD)

Reference	Study Description	Primary Results
Arnow et al, 200695	Patients with and without MDD and chronic pain were evaluated with the SF-8 to assess the impact of MDD on health-related quality of life	MDD/disabling pain: 31.3; P < .0001 vs all groups
		MDD/nondisabling pain: 27.9; P < .0001 vs controls
		MDD alone: 26.7; P < .0001 vs controls
		Neither MDD nor pain: 16.8
Barefoot et al, 199616	Patients with coronary artery disease with and without depression (n = 1,250) were followed for a median of 15 y	Risk of cardiac death
		χ21 = 9.25; P = .002 vs nondepressed patients
		Total mortality
		χ21 = 14.67; P < .001 vs nondepressed patients
		51% of moderate-to-severe depression group died of cardiac causes
Baune et al, 200744	Patients with common medical disorders with and without comorbid MDD were assessed to determine the impact of MDD on functioning and health outcomes	Risk of experiencing disability days (mean risk vs those without comorbid MDD)
		Overall: 1.34; P = .012
		Neurologic: 2.62; P = .0005
		Gastrointestinal: 1.38; P = .079
		Cardiovascular: 1.16; P = .359
		Endocrine: 1.70; P = .047
		Allergic: 1.06; P = .399
		Respiratory: 2.02; P = .069
Beekman et al, 200245	Adults aged 55–85 y (n = 2,200) were followed for 3 y to determine the impact of MDD on use of health services	Risk associated with the onset of MDD, OR (95% CI)
		Disability days: 2.10 (1.38–3.20)
		Hospital admission: 1.80 (1.15–2.81)
		Use of paramedical services: 1.88 (1.15–3.08)
		Impaired general satisfaction: 4.11 (2.65–6.38)
		Risk associated with persistent MDD, OR (95% CI)
		Fewer physician visits: 0.38 (0.19–0.75)
		Impaired general satisfaction: 3.73 (2.04–6.83)
		Dissatisfaction with services: 3.08 (1.39–6.83)
Birnbaum et al, 20038	The disability costs associated with depression were analyzed by gender	Total costs
		Males with MDD: $8,502 (work absence: 42%; prescription drugs: 11%; medical: 48%)
		Males without MDD: $3,458 (work absence: 39%; prescription drugs: 13%; medical: 47%)
		Females with MDD: $9,265 (work absence: 50%; prescription drugs: 11%; medical: 40%)
		Females without MDD: $5,091 (work absence: 53%; prescription drugs: 11%; medical: 36%)
Breslau et al, 200396	Patients with migraines, patients with severe headaches, and nonheadache controls were followed for 2 y to assess the relationship between MDD and headaches	Risk for migraine, OR (95% CI)
		Controls with MDD: 3.4 (1.4–8.7); P < .01 vs controls without MDD
		Risk for MDD, OR (95% CI)
		Migraine without MDD: 5.8 (2.7–12.3); P < .0001 vs controls without headache or MDD
Breslin et al, 200697	Adults aged 18–60 y were followed for 4 y and assessed to determine the effect of MDD on activity limitation	Risk for activity limitation associated with MDD, OR
		Home
		Female: 3.8
		Male: 4.2
		Work
		Female: 3.4
		Male: 3.4
		Other
		Female: 3.5
		Male: 5.7
Burton et al, 200498	Depressed employees (n = 1,491) from a large financial services company were assessed with the WLQ	Risk for low score on WLQ scales, OR (95% CI)
		Time: 2.05 (1.83–2.30); P < .05
		Physical: 1.49 (1.32–1.68); P < .05
		Mental: 2.46 (2.20–2.76); P < .05
		Output: 2.24 (2.00–2.50); P < .05
Carta et al, 200399	Direct health care costs of patients with depression, patients with a chronic somatic illness, and healthy controls were compared	Hospital days, mean
		Depressed: 5.8 d (P < .0001 vs somatic and healthy)
		Antidepressant treatment: 2.0 d (P < .04 vs no antidepressant treatment)
		No antidepressant treatment: 7.4 d
		Somatic: 4.3 d
		Healthy: 1.1 d
		Daily drug expenditures, mean
		Depressed: €0.53 (P < .0001 vs somatic and healthy)
		Antidepressant treatment: €0.65 (P = .28 vs no antidepressant treatment)
		No antidepressant treatment: €0.51
		Somatic: €0.52
		Healthy: €0.21
		Drug and hospitalization expenditures during previous year
		Depressed: €2,289.41 (P < .0001 vs somatic and healthy)
		Antidepressant treatment: €1,715.81 (P < .04 vs no antidepressant treatment)
		No antidepressant treatment: €2,528.40
		Somatic: €1,750.37
		Healthy: €474.11
Cramer et al, 2003100	Epilepsy patients with mild-moderate (n = 74) or severe (n = 166) depression and those without depression (n = 443) were assessed with a seizure severity scale	Patients with mild-moderate and severe depression experienced more severe, bothersome, and frequent seizures that were more difficult to recover from than nondepressed patients
Dorenlot et al, 200546	Dementia outpatients were followed for 3 y to assess the impact of MDD on rates of institutionalization	Risk for institutionalization, OR (95% CI)
		MDD: 1.6 (1.1–2.5); P = .043 vs non-MDD controls
Druss et al, 200010	Employees (n = 15,153) of a large US corporation who filed health claims were assessed on health care and disability costs	Depressive disorder (MDD, dysthymia, depressive disorder not otherwise specified)
		Sick days: 9.9; P < .05 vs diabetes, heart disease, hypertension, back problems, all other
		Per capita health/disability costs: $5,415; P < .01 vs hypertension, all other
		Diabetes, heart disease, hypertension, or back problem with comorbid depressive disorder
		Health care costs: $7,407; P < .001 vs any disorder without comorbid depressive disorder
		Sick days: 13.5; P < .01 vs any disorder without comorbid depressive disorder
		Per capita health care costs: $7,906; P < .001 vs any disorder without comorbid depressive disorder
Druss et al, 1999101	Health care costs of patients with MDD (n = 1,081) were compared with nondepressed controls	Increase in outpatient care costs: $1,326; P < .001 vs nondepressed controls
		Increase in inpatient care costs: $1,581; P < .001 vs nondepressed controls
Egede et al, 2004102	Patients with MDD and diabetes, MDD without diabetes, and diabetes only and nondepressed, nondiabetic controls were assessed on their physical functioning	Prevalence of overall functional disability
		MDD/diabetes: 77.8%; P < .0001 vs controls
		Diabetes only: 58.1%; P < .0001 vs controls
		MDD only: 51.3%; P < .0001 vs controls
		No MDD/no diabetes: 24.5%; P < .0001 vs controls
		Risk of functional disability, OR (95% CI)
		MDD: 3.02 (2.66–3.44)
		Diabetes: 2.46 (2.15–2.82)
		MDD and diabetes: 6.15 (3.86–9.80)
Egede, 200747	The 12-mo prevalence of MDD in patients with a number of CMDs was assessed to determine the effect of MDD on health care utilization	Risk, OR (95% CI)
		≥ 1 day absent from work
		CMD: 0.98 (0.80–1.21)
		CMD/MDD: 1.22 (0.88–1.68)
		≥ 1 day spent in bed
		CMD: 0.97 (0.85–1.10)
		CMD/MDD: 1.60 (1.28–2.00)
		Functional disability: yes
		CMD: 1.06 (0.91–1.24)
		CMD/MDD: 2.48 (1.96–3.15)
Ford et al, 2004103	The effect of depression on health care utilization was assessed	Patients with depression are significantly more likely to be high (14%; OR [95% CI]: 2.2 [1.2–3.9]) vs midrange (7%) health care utilizers
Frasure-Smith et al, 1993104	Patients who met criteria for a myocardial infarction and screened positively for MDD (n = 35) were followed for 6 mo to compare mortality rates with nondepressed controls (n = 187)	Risk for mortality, OR (95% CI)
		5.74 (4.61–6.87); P < .0006 vs nondepressed controls
Greenberg et al, 20037	The economic burden of depression for the year 2000 was calculated from a variety of sources presenting US data	Inpatient: $8,883 (million)
		Outpatient: $6,083 (million)
		Pharmaceutical: $10,400 (million)
		Absenteeism: $36,248 (million)
		Presenteeism: $15,195 (million)
Haarasilta et al, 2005105	Adolescents and young adults (n = 942) were followed for 1 y to assess the relationship between MDD and health	Risk of poor health, OR (95% CI)
		Chronic illness: 1.59 (0.93–2.70)
		Diagnosed chronic illness: 1.76 (1.03–3.01)
		Respiratory allergies: 2.71 (1.14–6.45)
		Other allergies: 1.31 (0.65–2.66)
		Musculoskeletal: 1.51 (0.60–3.79)
		Neurologic: 1.85 (0.53–6.45)
		Migraine: 3.67 (0.96–14.0)
		Disabling chronic illness: 1.35 (0.55–3.31)
		Poor self-perceived health: 2.56 (1.23–5.34)
		≥ 3 sick days: 2.01 (1.23–3.29)
Hoge et al, 2002106	Health care utilization of active military personnel between 1990 and 1999 was calculated to identify the effect that a psychiatric diagnosis has on this metric	MDD (1990–1999)
		Hospitalizations: 11,264
Ambulatory visits: 100,866
Janssens et al, 2003107	Patients with multiple sclerosis (n = 101) were screened for depression to assess its relationship to SF-36 scores	SF-36 physical health
		Physical functioning: β = −.45; P < .001
		Role-physical functioning: β = −.53; P < .001
		Bodily pain: β = −.43; P < .001
		General health: β = −.41; P < .001
		SF-36 mental health
		Vitality: β = −.51; P < .001
		Social functioning: β = −.57; P < .001
		Role-emotional functioning: β = −.45; P < .001
		Mental health: β = −.64; P < .001
Katon et al, 200313	The medical costs of older adults diagnosed with MDD were compared with nondepressed controls	Cost ratios associated with MDD, OR (95% CI)
		Total: 1.49 (1.28–1.72)
		Total outpatient: 1.47 (1.36–1.56)
		Outpatient depression: 1.78 (1.42–2.24)
		Outpatient nondepression: 1.36 (1.18–1.56)
Kaufmann et al, 199917	Patients with myocardial infarction (n = 331) were followed for 12 mo to assess the effect of MDD (27%) on mortality in these patients	Mortality
		MDD: 18.7%; OR = 2.33 (95% CI = 1.16–4.65); P = .015
		Non-MDD: 9.0%
Keenan-Miller et al, 2007108	Adolescents were followed for 5 y to assess the effect of MDD on health outcomes in young adulthood	MDD at age 15 y, β (95% CI)
		Interviewer-rated health: .16 (0.04–0.29); P = .01
		SF-36 self-rated health: 1.10 (0.17–2.02); P = .02
		SF-36 physical limitations: .62 (−0.22–1.45); P = .15
		Visit to medical professional: 1.26 (0.61–1.90); P = .001
		SF-36 work role impairment: .38 (0.15–0.60); P = .001
		Chronic illness at age 20 y, OR (95% CI): 1.62 (0.98–2.67); P = .06
Kessler et al, 1999109	Work disability data from 2 nationally representative populations were assessed to quantify the effect of 30 d of experiencing MDD	Short-term work disability: 45.9% (P < .05 vs non-MDD workers)
		Work disability days: 7.6 (P < .05 vs non-MDD workers)
		Salary-equivalent disability costs: $267
Kessler et al, 2003110	The disability data from a nationally representative sample diagnosed with hypertension, arthritis, asthma, and ulcer with MDD were compared	Difference in number of role impairment days
		Hypertension: 1.6 (P = NS vs no comorbid MDD)
		Arthritis: 2.2 (P = NS vs no comorbid MDD)
		Asthma: 2.4 (P < .05 vs no comorbid MDD)
		Ulcer: 3.1 (P < .05 vs no comorbid MDD)
Kessler et al, 20069	Using NCS data, an analysis of performance was conducted in workers with MDD	Aggregated (total US population) impact of MDD
		Absenteeism
		Days per year (million): 72.2; P < .05
		Dollars per year (million): 11,742; P < .05
		Presenteeism
		Days per year (million): 150.5; P < .05
		Dollars per year (million): 24,482; P < .05
		Total
		Days per year (million): 225.0; P < .05
		Dollars per year (million): 36,602; P < .05
Kouzis and Eaton, 1994111	The predictors of emotional disability days were assessed	Disability associated with MDD
		≥ 1 disability day: 44%
		Risk for disability day: OR = 27.8 (95% CI = 6.93–108.96)
Lerner et al, 200423	Work outcomes in employees with MDD (n = 75) were compared with healthy controls (n = 169) over 6 mo	WLQ scales (change from baseline to 6 mo)
		Physical
		MDD: –3.3; P = NS
		Controls: –0.8
		Time
		MDD: –3.9; P = .014
		Controls: –0.2
		Mental
		MDD: –9.4; P < .001
		Controls: –1.4
		Output
		MDD: –12.0; P < .001
		Controls: –1.4
Lerner et al, 2004112	The effect of MDD on work productivity outcomes was assessed with the WLQ	WLQ scales
		Mental-interpersonal: β = 50.8; P < .001
		Physical: β = 12.9; P = .004
		Time: β = 46.4; P < .001
		Output: β = 59.7; P < .001
		Days missed: 2.2; P < .001
Lespérance et al, 2002113	The association between 5-y risk of cardiac mortality and BDI severity score was assessed in patients with myocardial infarction (n = 879)	Cardiac mortality, OR (95% CI)
		BDI score 5–9 vs < 5: 1.76 (0.98–3.17); P = .059
		BDI score 10–18 vs < 5: 3.17 (1.79–5.60); P < .001
		BDI score ≥ 19 vs < 5: 3.13 (1.56–6.27); P = .001
Luber et al, 2000114	The health care utilization of internal medicine outpatients with a diagnosis of MDD was compared to nondepressed controls over 1 y	Health care visits
		MDD: 5.3; P < .001 vs controls
		Controls: 2.9
		Total health care costs
		MDD: $2,808; P = .001 vs controls
		Controls: $1,891
McIntyre et al, 2008115	The impact of MDD on work functioning was assessed in a large, nationally representative sample	Risk associated with MDD, OR (95% CI)
		≥ 1 disability day in past 2 wk: 5.6 (4.1–7.7); P < .05 vs reference group
		Good job security: 0.7 (0.6–0.8); P < .05 vs reference group
McQuaid et al, 1999116	A population of primary care patients was used to assess the impact of MDD on work functioning	Missed work days
		MDD: 58.1% (χ21 = 15.10; P < .001)
		No MDD: 28.7%
		Cut down on activities
		MDD: 77.0% (χ21 = 9.03; P < .01)
		No MDD: 54.3%
Muchmore et al, 2003117	A predictor analysis was conducted in arthritis patients to assess the impact of arthritis and associated conditions, including MDD, on disability	Risk associated with comorbid depression, OR (95% CI)
		Short-term disability: 1.01 (1.00–1.02); P < .0001
		Long-term disability: 2.23 (1.49–3.32); P < .0001
		Worker's compensation: 1.45 (1.25–1.70); P < .0001
Papapetropoulos et al, 200648	PD patients with and without MDD were assessed with PD rating scales	PD severity ratings
		Unified Parkinson's Disease Rating Scale, mean (SD)118
		PD/MDD: 58.1 (32.3); P = .004
		PD: 37.3 (31.1)
		Hoehn and Yahr, mean (SD)119
		PD/MDD: 2.7 (1.0); P = .07
		PD: 2.2 (0.9)
		Schwab and England, activities of daily living performed, % (SD)120
		PD/MDD: 69.4% (22.1); P = .03
		PD: 78.4% (22.3)
Penninx et al, 2001121	Mortality rates of patients diagnosed with and without CD and with and without MDD (n = 2,847) were compared	CD mortality, OR (95% CI)
		No CD/MDD: 3.8 (1.4–10.6)
		CD/No MDD: 3.4 (2.4–4.9)
		CD/MDD: 10.5 (4.1–26.7)
		IHD mortality, OR (95% CI)
		No CD/MDD: 5.1 (1.6–16.9)
		CD/No MDD: 4.5 (2.8–7.1)
		CD/MDD: 17.7 (6.0–51.9)
Rovner, 1993122	Nursing home patients were followed for 1 y to determine the effects of MDD on mortality rates	Risk for mortality in MDD, OR (95% CI): 1.59 (1.02–2.51)
Rumsfeld et al, 2003123	Veterans’ affairs patients with a history of acute coronary syndrome with MDD (n = 1,431) were compared with nondepressed controls (n = 526) on a variety of health and quality of life outcomes	MDD vs controls, OR (95% CI)
		Higher angina frequency: 2.40 (1.86–3.10); P < .001
		Greater physical limitations: 2.89 (2.17–3.86); P < .001
		Worse quality of life: 2.84 (2.16–3.72); P < .001
Saarijärvi et al, 200236	SF-36 data from patients with MDD and nondepressed controls were compared to determine the impact of MDD on quality of life	Relationship between SF-36 and BDI in MDD patients
		Physical functioning (P ≤ .01)
		Role functioning-physical (P < .001)
		Role functioning-emotional (P < .01)
		Energy (P = .0001)
		Emotional well-being (P = .0001)
		Social functioning (P = .0001)
		Bodily pain (P = .0001)
		General health perception (P = .0001)
Simon et al, 1995124	Primary care patients with and without depression were compared on health care costs over 1 y following diagnosis	Annual total direct health care costs (eg, outpatient and inpatient mental health care, outpatient primary care, inpatient medical)
		Depressed: $4,246
		Not depressed: $2,371
Simon et al, 200049	A 2-y follow-up of patients beginning antidepressant therapy assessed patients on a variety of work outcomes according to treatment response (ie, persistent [n = 35], improved [n = 137], or remitted [n = 118] depression)	Total health care costs
		Persistent: $4,082
		Improved: $3,459
		Remitted: $2,816
		Employed, % (F2,262 = 5.88, P = .003)
		Persistent: 70.1
		Improved: 83.8
		Remitted: 85.4
		Days of work missed (F2,226 = 10.62, P < .001)
		Persistent: 16.80
		Improved: 10.37
		Remitted: 6.29
Sobocki et al, 2006125	Direct and indirect health costs (in year 2004 € million) collected from published studies conducted in 28 European countries were calculated	Total costs: €117,851
		Total direct costs: €41,688
		Hospitalization costs: €10,424
		Drug costs: €9,013
		Outpatient care: €22,252
		Total indirect costs: €76,163
		Morbidity: €72,189
		Mortality: €3,974
Spitzer et al, 199535	Primary care patients diagnosed with MDD (n = 115) were compared with controls on the subscales of the SF-20 to assess the impact of MDD on health-related quality of life	MDD vs nonpsychiatrically diagnosed patients
		Physical functioning: –21.8; P < .001
		Bodily pain: –22.7; P < .001
		Role functioning: –45.5; P < .001
		General health: –30.4; P < .001
		Social functioning: –31.7; P < .001
		Mental health: –36.5; P < .001
Stewart et al, 200328	Depressed (n = 219) and nondepressed (n = 908) employees were compared on the cost of lost productive time	Lost productive time (hr/wk)
		Absenteeism: 1.2
		Presenteeism: 7.2
		Total lost productive time: 8.4
		Cost of lost productive time ($ billion per year)
		Absenteeism: 3.18
		Presenteeism: 18.18
		Total lost productive time: 21.36
Sullivan et al, 1997126	The physical functioning of patients with coronary artery disease was compared by baseline HDRS17 score and severity of depression over a 12-mo follow-up period	Relationship between functioning and baseline HDRS17 score
		Physical function score at 12 mo: r = –0.27; P < .001
		Activity interference at 12 mo: r = 0.23; P < .01
		A significant association (ANOVA) between baseline depressive severity and physical functioning was observed at baseline (P < .001) and 12 mo (P = .01)
Sullivan et al, 2000127	The physical functioning of patients with coronary artery disease was compared by baseline HDRS17 score and severity of depression over a 5-y follow-up period	Relationship between disability and HDRS17 score
		SF-36 scales
		Physical function: r = 0.26; P < .01
		Physical role: r = 0.36; P < .001
		Pain: r = 0.28; P < .01
		Social function: r = 0.31; P < .001
		Mental health: r = 0.21; P < .05
		Emotional role: r = 0.21; P < .05
		Vitality: r = 0.36; P < .0001
		General health: r = 0.33; P < .0001
Unützer et al, 1997128	A 4-y prospective study of Medicare enrollees ≥ 65 y who were screened for depression; health care costs were compared between those with depression (n = 353) and those without (n = 2,165)	Median costs 1 y after baseline
		Depressed: $2,147
		Not depressed: $1,461
		Median costs 4 y after baseline
		Depressed: $15,423
		Not depressed: $10,152
Wells et al, 198930	Data from 11,242 participants of the Medical Outcomes Study receiving treatment from general medical providers were subdivided into NCC and MDD groups	Physical functioning
		NCC: 85.4; P < .05
		MDD: 81.3
		Social functioning
		NCC: 91.7; P < .0001
		MDD: 83.3
		Role functioning
		NCC: 87.0; P < .0001
		MDD: 74.6

Abbreviations: ANOVA = analysis of variance, BDI = Beck Depression Inventory, CD = cardiac disease, CMD = chronic medical disorder, HDRS₁₇ = 17-item Hamilton Depression Rating Scale, IHD = ischemic heart disease, NCC = no chronic condition, NCS = National Comorbidity Survey, NS = not significant, OR = odds ratio, PD = Parkinson's disease, SF-8 = 8-item Short-Form Health Survey, SF-20 = 20-item Short-Form Health Survey, SF-36 = 36-item Short-Form Health Survey, WLQ = Work Limitations Questionnaire.

The objective of this systematic review was to quantify the negative impact that MDD has on quality of life, disability, and work, family, and overall psychosocial functioning. Available scales that assess these areas of impairment as they relate to patients with MDD are also described.

METHOD

PubMed searches were conducted using the following terms: (MDD OR major depressive disorder) AND (absenteeism OR absente*); AND (quality of life OR QOL); AND (psychosocial function*); AND (presente*OR presenteeism); AND (health care cost* OR [health care] cost*); AND (health outcome*); AND (functional outcome*); AND (family life); AND (disabil* OR disability); AND (work function*); AND (unemployment OR unemploy*). The literature search was conducted in July 2008 and was restricted to English language articles. There were no limits set on the dates of the search.

Clinical Points

• ♦ Major depressive disorder is associated with significant declines in functioning and quality of life.

• ♦ In addition to measuring the severity of depressive symptoms, clinicians and researchers should assess depressed patients’ level of functioning and quality of life.

• ♦ A number of scales are currently available that are valid and reliable measures of functioning and quality of life.

Two hundred twenty potential articles were identified. Among these studies, 48 presented primary data directly demonstrating the effect of MDD on these outcomes. Primary data were compiled from these studies and are summarily described. Available scales that assess these outcomes are also described.

COSTS ASSOCIATED WITH MDD

The substantial direct and indirect health care costs associated with depression alone^7–9 and comorbid with other conditions^10–13 have been quantified by a number of large-scale evaluations. Correspondingly, depression also has been shown to negatively impact an individual's overall health and can lead to increased costs for treating any co-occurring disorders. For instance, in 2 studies that assessed diabetic patients with and without comorbid depression, significantly higher health care costs¹⁴ and worse clinical outcomes, including higher mortality rates,¹⁵ were observed in those with a co-occurring depressive disorder.

In a large-scale assessment of employees of a major corporation,¹⁰ the per-capita health and disability costs (including direct health care expenditures and sick/disability days) associated with depression were $5,415 annually, equivalent to diabetes ($5,472), heart disease ($5,523), and back problems ($4,388), but significantly greater than hypertension ($3,372; P = .002) and a category containing “all other” reasons for filing a health claim ($1,292; P < .001). In addition, when MDD co-occurred with one of these general medical conditions, the associated health care costs increased nearly 2-fold compared to patients with the medical disorder alone. The resulting cost to the corporation was $2.2 million.^10,11 In addition, other studies have demonstrated a significant association between MDD, cardiac death,^16,17 and total mortality.¹⁶ Beyond medical comorbidities, MDD is also commonly comorbid with other psychiatric disorders. In particular, MDD commonly co-occurs with generalized anxiety disorder, which results in significant disability.¹⁸

IMPAIRED FUNCTIONING AND QUALITY OF LIFE ASSOCIATED WITH MDD

The terms functioning and quality of life are often used interchangeably; however, the disability associated with each and the means in which they are assessed are multidimensional and can differ significantly. Assessments of functioning generally include performance-based metrics, such as one's ability to engage in expected or usual responsibilities at work or home, and can be assessed using objective measures, such as days of work missed, or subjectively, using patient-rated assessments. However, functioning does not distinctly involve work performance; functioning in other areas, such as in familial, societal, and marital roles, also can be adversely affected by MDD. Quality-of-life measures, on the other hand, generally describe the subjective quality of an individual's day-to-day experiences, which involve enjoyment and satisfaction with one's life, but also can be related to the patient's performance in his or her expected role.

Work Functioning

A number of studies have empirically demonstrated that MDD not only negatively impacts those who are currently employed^19–22 but also is associated with an increased risk for job loss.²³ One large-scale analysis of NCS data suggested that those in the middle of their careers are more likely than their younger counterparts to lose their jobs as the result of a major depressive episode.²⁴ The most obvious way that MDD impacts work performance is the elevated rates of absenteeism among depressed employees.^9,10 However, the negative effect that MDD has on work performance extends beyond whether an employee is present to perform his or her job.

Presenteeism describes the impaired functioning that employees experience when they attend work but suboptimally perform their daily activities, in this case, due to their depressive symptoms. Depressed employees have been shown to be particularly prone to detriments in mental-interpersonal and time management tasks, as well as overall job performance, compared to nondepressed employees.^25–28 In 1 study conducted in a large company, the lost productive time associated with MDD was reported to be nearly 6 hours per week, at a cost of $44 billion annually.²⁸ In addition, less severe forms of depression (ie, subthreshold depressive symptoms,^29,30 minor depression,^31–33 and dysthymia³⁴) also have been shown to negatively impact work functioning.

In 1 large-scale work performance study, NCS participants currently experiencing a depressive episode were compared to those with episodes that resolved > 1 to 6 months ago, > 6 to 12 months ago, and > 12 months ago. These patients had responded to treatment but continued to experience significant residual symptoms that were more pronounced than the level seen in the euthymic controls. Patients who were currently experiencing a depressive episode were significantly more likely to experience lost days than healthy controls and those with residual symptoms; however, compared to healthy controls, depressed patients and even those with residual symptoms were significantly more likely to experience difficult days and cutback days (in which the depressed patient did not get as much done as usual) than healthy controls.^29,30

A growing body of data has demonstrated that depressive symptoms include impairments in cognitive functioning that can result in declining work performance. A recently published study that assessed untreated MDD patients using work performance measures following an error or receipt of negative feedback demonstrates that dysfunctional information processing associated with depression can lead to declines in productivity.¹⁹ This study and a number of other neuroimaging studies have suggested that these maladaptive cognitive processes involve areas of the brain that are thought to be associated with the symptoms of MDD (eg, the prefrontal cortex, anterior cingulate cortex, hippocampus).^19–22

Psychosocial Functioning

The components of psychosocial functioning vary between scales and sometimes overlap with assessments of work performance but are mainly differentiated by the assessment of other areas of functioning beyond the workplace. Several studies have been conducted that support the claim that MDD has a negative impact on overall social functioning.^35,36 Results from an analysis of patients experiencing different levels of depressive symptomatology (ie, subthreshold depressive symptoms, minor depression/dysthymia, and MDD) have suggested that psychosocial disability increases correspondingly with the severity of depressive symptoms.³⁷ In a more recently conducted study, which assessed patients with MDD and bipolar I and II disorder, those experiencing MDD were found to experience significant psychosocial impairment during the majority of follow-up visits.³⁸

Quality of Life

Quality of life is difficult to concisely define due to its subjective nature and the overlap between assessing functional outcomes; therefore, measuring quality of life is more complex than the other outcomes described. For example, assessing the quality of a person's life without examining how the individual functions in his or her expected roles, in either family life or work performance, is difficult.

The impact of a major depressive episode on health-related quality of life in individuals with^39–48 or without^39–43 a general medical condition has been well established in the literature, and treating MDD to a full remission of symptoms can improve quality of life to a greater extent than when significant residual symptoms are present.^49,50

TREATMENT OF THE FUNCTIONAL DISABILITY ASSOCIATED WITH MDD

A number of empirically supported treatment options have demonstrated effectiveness in alleviating the disability caused by MDD,^30,51–56 as well as in lowering health care costs,^49,57–65 improving quality of life,^50,66,67 and decreasing absenteeism in depressed patients,^49,68 particularly in those who receive early and adequate treatment.^69,70 For example, the Sequenced Treatment Alternatives to Relieve Depression trial, a large-scale assessment of an MDD treatment algorithm conducted in a real-world clinical setting, demonstrated that following sequential treatment alternatives for patients not reaching remission with first-line treatment can lead to significant improvements in functioning and quality of life.^53,71

Despite increased usage of the multiple antidepressant medications available to patients and improvements in empirically supported treatment guidelines and algorithms, data are available that demonstrate that a disproportionately large number of patients continue to receive inadequate treatment for their depressive episodes. In a cross-sectional analysis of medical records from 2 cohorts of depressed patients (ie, 1993 to 1994 and 2003 to 2004), an increase in the use of adequate antidepressant doses was observed; however, the use of sequential antidepressant treatment options and psychotherapy remained low.⁷²

An effective way of lowering the costs associated with MDD is to encourage physicians to use guideline-derived forms of treatment and utilize enhanced treatment options, such as incorporating care managers for monitoring the patient's symptoms, adverse events, and adherence.⁷³ In addition, 1 study demonstrated that by taking steps to improve employee access to effective depression treatment—for example, by lowering copayments and using a selective contracting network and a mental health destigmatization program—the likelihood of initiating treatment and having more mental health visits increased.⁷⁴

The relationship between MDD, functional disability, and impaired quality of life has been suggested to be bidirectional. In addition to MDD negatively impacting functioning and quality of life, the presence of these impairments at baseline has been linked to poor antidepressant treatment response.^53,75,76 It also has been suggested that functional impairment and a lower quality of life are associated with an elevated risk for the recurrence of a major depressive episode.^77–79

ASSESSING DECLINES IN FUNCTIONING AND QUALITY OF LIFE ASSOCIATED WITH MDD

Table 2 describes some commonly used functional outcome measures and assessments of quality of life. The 36-item Short-Form Health Survey (SF-36)⁸⁰ was the most widely used scale in the studies identified during our systematic review of the literature. Its length and complexity have been impediments to its regular use in clinical settings; however, a shortened 12-item version (SF-12) is also available that may be better suited for use in busy practice settings.⁸¹ In addition, the Sheehan Disability Scale (SDS)⁸² and the Work Limitations Questionnaire (WLQ)⁸³ were commonly used. The SDS is widely used in research settings because it has been shown to be sensitive to treatment effects, and its concise means of assessing the overall level of functioning make it desirable for use in clinical settings as well.

Table 2.

Commonly Used Assessments of Functioning and Quality of Life

Scale	Number of Items	Areas of Assessment	Rating Scale
36-Item Short-Form Health Survey80	36 patient-rated items	8 subscales: physical functioning, role-physical, bodily pain, general health perception, vitality, social functioning, role-emotional, mental health	Items are rated numerically either with a Likert scale (1 to 5 and 1 to 6) or with yes (1) or no (2) responses
Endicott Work Productivity Scale87	25 patient-rated items	Assesses how medical disorders impact work performance across a range of activities (eg, pace of work, organization, prioritizing, not performing expected tasks)	Items are generally rated on a 5-point scale (0 [never] to 4 [almost always]); there are also open-ended items that ask patients the number of hours worked and to provide a reason why fewer hours than expected have been worked
Global Assessment of Functioning (GAF), Social and Occupational Functioning Assessment Scale (SOFAS), Global Assessment of Relational Functioning (GARF)89	Each scale provides 1 score that assesses functioning	GAF: Measures social, occupational, and psychological functioning in relation to psychiatric symptom severity	GAF: 10 categorical ratings each within a 10-point range (ie, 100 to 91: superior functioning; 10 to 1: nonfunctioning [ie, violent, suicidal, persistent lack of hygiene])
		SOFAS: Measures social and occupational functioning independent of symptom severity	SOFAS: 10 categorical ratings each within a 10-point range (ie, 100 to 91: superior functioning; 90 to 81: good functioning with minimal, socially appropriate symptomatology; 80 to 71: transient and minimal impairment)
		GARF: 3 major content areas associated with familial and other social relationships: problem solving, organization, emotional climate	GARF: 5 categories with scores ranging from 100 to 81 (relational unit is functioning satisfactorily) to 20 to 1 (relational unit has become too dysfunctional to retain continuity of contact and attachment)
Quality of Life Enjoyment and Satisfaction Questionnaire86	93 items	5 subscales: physical health, subjective feelings, leisure time activities, social relationships, general activities	Items are rated on a 5-point scale (0 [never] to 4 [almost always]) and generally follow the same format: During the past week, how often have you been pleased with your work accomplishments, kept your room/apartment/house cleaned to your satisfaction, shopped for food or other household items to your satisfaction, and joked or laughed with other people?
		3 scales that are appropriate for each patient: work, household duties, school/course work
Sheehan Disability Scale82	3 patient-rated items	Work/school, social life, family life functioning	Combination of visual, verbal, and numerical anchors (0 [not at all] to 10 [extremely]) ≥ 5 on any item is indicative of impaired functioning
Social Adjustment Scale–Self-Report85	42 items divided into 4 categories: performance, friction, finer aspects of relationships, inner feelings/satisfaction	Work, social/leisure, extended family, and marital, parental, and familial relationships	5-point Likert scale with higher scores indicating a greater level of impairment; total score can be obtained by adding the individual item scores then dividing by the number of items that have actually been scored
Work Limitations Questionnaire83	25 items	4 subscales: time demands (5 items), physical demands (6 items), mental-interpersonal demands (9 items), output demands (5 items)	Scores range from 0 (limited none of the time) to 100 (limited all of the time) and represent the reported amount of time in the prior 2 wk respondents experienced work limitations; additionally, using an algorithm, Work Limitations Questionnaire scores can be converted into an estimate of productivity loss
Work Productivity and Activity Impairment Questionnaire88	6 items	Assesses patients on the number of hours lost during the last days due to the patient's “problem” or any other reason, the number of hours worked, and the effect the patient's “problem” has on productivity at work and on other activities	An overall work productivity score and a quantification of overall work productivity and impairment in regular activities are obtained
World Health Organization 5-Item Well-Being Index84	5 patient-rated, depression-specific items	Cheerful/good spirits, calm/relaxed, active/vigorous, wake feeling fresh/rested, life is full of things that interest me	Rated on a 0 (at no time) to 5 (all of the time) Likert scale
			A score of 25 represents the best possible quality of life, and a score < 13 is indicative of a diagnosis of major depressive disorder

The World Health Organization 5-item Well-Being Index,⁸⁴ Social Adjustment Scale–Self-Report (SAS-SR),⁸⁵ and Quality of Life Enjoyment and Satisfaction Questionnaire (Q-LES-Q)⁸⁶ are commonly used to assess psychosocial functioning and quality of life, whereas the Endicott Work Productivity Scale,⁸⁷ Work Productivity and Activity Impairment Questionnaire,⁸⁸ and WLQ⁸³ are commonly used assessments of work dysfunction.

Perhaps the most familiar means of assessing the functional declines associated with psychiatric disorders is Axis V of the multiaxial diagnostic methodology used in the fourth edition of the American Psychiatric Association's Diagnostic and Statistical Manual of Mental Disorders.⁸⁹ The Global Assessment of Functioning (GAF) asks the clinician to rate the level of functioning in relation to symptom severity. This combination of 2 distinct outcomes on 1 axis has been the primary criticism of the GAF, as a separation between these outcomes is important to obtaining an accurate level of functional disability independent of symptom severity. In addition, the Global Assessment of Relational Functioning (GARF) and Social and Occupational Functioning Assessment Scale (SOFAS) were designed as supplemental assessments of functioning to be used in conjunction with the GAF.⁸⁹ The GARF was specifically designed to assess familial and other long-term relationships, whereas the SOFAS was designed to assess social and occupational functioning independent of symptom severity.⁸⁹

ISSUES ASSOCIATED WITH USING “REAL-WORLD” OUTCOME MEASURES

Despite the benefits of assessing functioning and quality of life to the various stakeholders involved with the treatment of MDD, the potential issues associated with using such scales must be noted. Some of the commonly used assessment tools described above have some unmet needs. Namely, they can be cumbersome for use in the clinical setting because they are generally too lengthy to administer in the ever-lessening duration of clinical visits. As was previously mentioned, a number of these scales, such as the SF-36, WLQ, SAS-SR, and Q-LES-Q, are thought to be too lengthy for use during clinical visits. In addition, there is some ambiguity as to what these scales actually measure, mainly due to the lack of consistency in the items they use and the lack of a clear gold standard. On the other hand, some scales are too specific. The assessments of work functioning, for example, focus too narrowly on the area of work dysfunction to be used alone in a clinical setting wherein a level of overall functional status is desired.

When performing follow-up assessments using these scales, it is important to note that a lag time in improvements in functional impairment has been observed in relation to improvements in depressive symptomatology. When assessing functioning, > 8 weeks may be required before improvements are observed, whereas improvements in depressive symptoms generally occur sooner.^90,91 A secondary analysis of data from a randomized trial investigating selective serotonin reuptake inhibitor treatment,⁹² which calls for a broader definition of depression remission that expands beyond symptom severity, demonstrated that depressive symptoms improve in synchrony and are correlated with work functioning, even though depressive symptoms improved to a greater degree. It is also important to note that the majority of the scales described above are not designed to diagnose MDD and are meant only to be used for screening and to assess and monitor changes in disability. If the results are suggestive of a depressive disorder, then a validated diagnostic assessment tool should be used to make a proper diagnosis.

CONCLUSIONS

The development of an assessment tool that can address the issues described above would be a benefit to the various parties involved with the treatment of MDD (eg, researchers, patients, health care plans). The SDS has many positive attributes but has not yet been widely accepted as the gold standard in assessing disability in patients with MDD. The limitations of the SDS include its assessment of patients on only 3 domains of functioning and the lack of a depression-specific focus. The SF-12, a shortened version of the SF-36, is another valuable assessment tool due to its more detailed assessment of disability yet compact enough length for use in clinical practice. Perhaps creating a “hybrid” scale adopting the best of various options that can become a standard measure for assessing the disability associated with MDD will become necessary. Along with assessing the impact on the emotional and physical symptoms of MDD, assessing functional disability and declines in quality of life should continue to become a routine part of clinical care and outcome measures in clinical trials that assess the efficacy of antidepressant treatment.^43,78,93,94 As calls for the measurement-based care of MDD continue, particularly for primary care physicians, a thorough assessment of the impact that MDD has on a patient's life will become a necessity.

Author contributions: Drs Langlieb and Guico-Pabia equally contributed to the study by analyzing and interpreting data, drafting and revising the manuscript, and obtaining funding.

Potential conflicts of interest: Dr Langlieb serves on the advisory boards for Eli Lilly and Pfizer Inc. Dr Guico-Pabia is an employee of Pfizer Inc, formerly Wyeth Research.

Funding/support: This analysis was supported by Wyeth Research, Collegeville, Pennsylvania, which was acquired by Pfizer Inc in October 2009.