DEPRESSION

AND

ANXIETY 27 : 7889 (2010)

Research Article
EMPLOYER BURDEN OF MILD, MODERATE, AND SEVERE MAJOR DEPRESSIVE DISORDER: MENTAL HEALTH SERVICES UTILIZATION AND COSTS, AND WORK PERFORMANCE
Howard G. Birnbaum, Ph.D.,1 Ronald C. Kessler, Ph.D.,2 David Kelley, B.S.,1 Rym Ben-Hamadi, M.Sc.,1 Vijay N. Joish, Ph.D.,3 and Paul E. Greenberg, M.A. M.S.1

Background: Treatment utilization/costs and work performance for persons with major depressive disorder (MDD) by severity of illness is not well documented. Methods: Using National Comorbidity Survey-Replication (20012002) data, US workforce respondents (n 5 4,465) were classified by clinical severity (not clinically depressed, mild, moderate, severe) using a standard self-rating scale [Quick Inventory of Depressive Symptomatology SelfReport (QIDS-SR)]. Outcomes included 12-month prevalence of medical services/medications use/costs and workplace performance. Treatment costs (employers perspective) were estimated by weighing utilization measures by unit costs obtained for similar services used by MDD patients in claims data. Descriptive analysis across three severity groups generated v2 results. Results: Using a sample of 539 US workforce respondents with MDD, 13.8% were classified mild, 38.5% moderate, and 47.7% severe cases. Mental health services usage, including antidepressants, increased significantly with severity, with average treatment costs substantially higher for severe than for mild cases both regarding mental health services ($697 vs. $388, v2 5 4.4, P 5 .019) and antidepressants ($256 vs. $88, v2 5 9.0, P 5 .001). Prevalence rates of unemployment/disability increased significantly (v2 5 11.7, P 5 .003) with MDD severity (15.7, 23.3, and 31.3% for mild, moderate, and severe cases). Severely and moderately depressed workers missed more work than nondepressed workers; the monthly salary-equivalent lost performance of $199 (severely depressed) and $188 (moderately depressed) was significantly higher than for nondepressed workers (v2 5 10.3, Po.001). Projected to the US workforce, monthly depression-related worker productivity losses had human

Analysis Group, Inc., Boston, Massachusetts Department of Health Care Policy, Harvard Medical School, Boston, Massachusetts 3 sanofi-aventis, Bridgewater, New Jersey
2

Funding for this research was provided by sanofi aventis. National Comorbidity Survey Replication (NCS-R) is supported by the National Institute of Mental Health (NIMH; U01-MH60220) with supplemental support from the National Institute of Drug Abuse (NIDA), the Substance Abuse and Mental Health Services Administration (SAMHSA), the Robert Wood Johnson Foundation (RWJF; Grand 044780), and the John W. Alden Trust.
Correspondence to: Howard Birnbaum, Ph.D., Analysis Group, Inc, 111 Huntington Avenue 10th floor, Boston, MA 02199. E-mail: hbirnbaum@analysisgroup.com

Received for publication 25 February 2009; Revised 15 April 2009; Accepted 17 April 2009 DOI 10.1002/da.20580 Published online 30 June 2009 in Wiley InterScience (www. interscience.wiley.com).

r 2009 Wiley-Liss, Inc.

Research Article: Employer Burden of Depression by Severity

79

capital costs of nearly $2 billion. Conclusions: MDD severity is significantly associated with increased treatment usage/costs, treatment adequacy, unemployment, and disability and with reduced work performance. Depression and r 2009 Wiley-Liss, Inc. Anxiety 27:7889, 2010. Key words: healthcare cost; work loss; presenteeism; clinical severity depression

affect 18.1 million people in 2000 in the United States, and to have a lifetime prevalence of 16.2%[1,2] and an annual prevalence of 6.6%.[2] Depression is a chronic disease with poor remission rates.[3] Patients with depression tend to have more comorbid conditions, both other mental illnesses and chronic physical diseases,[4] than those without MDD. The economic consequences of MDD are substantial. This is true not only because treating the disease itself, as well as its accompanying comorbidities, requires substantial health care resources,[1,5] but even more so because depression is associated with reduced productivity both at paid and unpaid work, increased work loss and disability, and suicidality. Greenberg et al. estimated the annual societal costs of clinical depression at $83.1 billion in 2000.[1] Kessler et al. estimated that MDD resulted in 27.2 lost workdays per ill worker each year.[6] Kessler et al. showed that the major workplace impact of depression is likely to be captured as presenteesim, which can be several times the costs of work loss time alone.[6] While the substantial workplace costs of MDD are well documented, less is known about the impact of MDD severity on workplace burden. The literature on the economic consequences of treatment failure for patients with depression indicates that the direct and indirect costs of severe depression can be substantial.[710] In particular, depressed patients who do not respond to treatment have more than twice the resource use and costs of patients who are not treatmentresistant as well as higher indirect costs.[8,10] However, the literature on costs of severe MDD that categorize patients using actual clinical measures is sparse. Previous research by Kessler et al. used preliminary data from the National Comorbidity Survey Replication (NCS-R) and classified survey respondents with 12-month MDD into different severity groups (mild, moderate, and severe) based on transformational rules for the Quick Inventory of Depressive Symptomatology Self-Report (QIDS-SR).[11] That study, based on preliminary NCS-R data, showed that MDD severity was associated with increased number of days out of role (i.e., days where respondents were unable to work or carry on usual activities because of depression) in the year prior to the interview. Although that study compared the resource utilization of respondents by

Major Depressive Disorder (MDD) was estimated to

INTRODUCTION

severity of MDD, it did not estimate costs associated with MDD severity. The objective of the current report is to document in a descriptive analysis the employer burden of MDD by severity of illness in terms of lost work performance and increased treatment costs. By way of background and to provide a context, descriptive information on employment status is also provided. We use data from the full NCS-R sample, interviewed in 20012002, and a privately insured administrative claims database to estimate the employer costs associated with MDD by severity.

MATERIALS AND METHODS

DATA SOURCE AND SAMPLE
The NCS-R is a nationally representative US survey that was conducted between February 2001 and December 2002 of Englishspeaking household residents aged 18 or older (regarding respondents experiences in the prior 12 months). Part 1 of the interview included a core diagnostic assessment administered to all respondents (N 5 9,282). Part 2 included assessments of risk factors, consequences, and services of the core disorders. Part 2 was administered to 5,692 of the 9,282 NCS-R respondents, over-sampling those with clinically significant psychopathology. The Part 2 respondents who were not sampled with certainty were weighted by the inverse of their probability of selection to reconstruct the Part 1 distribution of the sample. Finally, the full Part 2 sample was poststratified to approximate the distribution of the 2000 Census on a range of socio-demographic variables. More details on NCS-R sampling and weighing are reported elsewhere.[12] The subset of Part 2 respondents who were in the workforce (n 5 4,465) was the focus of this analysis. They comprised respondents who were either employed (currently working for a salary or hourly wage) at least 20 hr per week, unemployed (looking for work), or disabled (unable to work currently or in the next six months due to their disability). Respondents who met CIDI/Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (CIDI/DSM-IV) criteria for 12-month MDD were self-administered a truncated version of the Quick Inventory of Depressive Symptomatology Self-Report (QIDS-SR)[13] to assess symptom severity in the worst month of the past year. Respondents included both treated and untreated MDD patients. The QIDS-SR is a fully structured measure that is strongly related both to the clinician-administered Inventory of Depressive Symptomatology Clinician Rating (IDS-C) and to the Hamilton Rating Scale of Depression (HRSD).[14,15] Transformation rules developed for the QIDS-SR were used to convert scores into clinical severity categories mapped to conventional HRSD ranges of none (i.e., not clinically depressed), mild, moderate, and severe.[2] Depression and Anxiety

80

Birnbaum et al.

MEASURES
All respondents to Part 2 of the NCS-R were asked about receiving treatment for emotional problems and use of psychotropic medications in the past 12 months.[2,11] Outcomes included the prevalence of mental health medical services and medication use over this recall period, the average number of visits and pill days on medication over this period, and the prevalence of treatment adequacy. Mental health services include hospital (inpatient treatment), psychiatrist (any treatment from a psychiatrist), other mental health (nonpsychiatrist mental health specialty treatment, e.g., psychologist, therapist, etc.), general medical (all other mental health treatment not classified above), and any healthcare (either of mental health specialty treatment or general medical treatment as well as human services and complementary medicine). Measures of MDD treatment adequacy follow previous research, where treatment adequacy was defined as receiving either (1) at least four outpatient visits with any type of physician for pharmacotherapy that included use of either an antidepressant or mood stabilizer for a minimum of 30 days or (2) at least eight outpatient visits with any professional in the specialty mental health sector for psychotherapy lasting a mean of at least 30 min.[2]

WORKPLACE PERFORMANCE
We also considered the following workplace self-reported outcomes, based on survey questions about the previous months work experience: employment status (employed, disabled, and unemployed), workplace performance (hours worked, self-rated performance, and days of missed work), and workplace burden (estimated by multiplying work hours lost by self-reported hourly income). Following previous work by Kessler et al., linear regressions adjusting for age, age squared, race, and sex were used to model adjusted average hours, self-rating, and reduced work and performance.[2,16,17] The self-rating model, based on the respondents rating of his own performance, was adjusted for the average rating of similar workers and similarly, the reduced work and performance model, based on the total hours worked in the month prior to the survey, were adjusted for expected hours worked. Logistic regressions were used to model the likelihood of missing one or more days of work in the past month. The adjusted average monthly hours equal the expected number of hours worked per month that the respondent worked plus extra hours worked and minus the number of hours missed in the month prior to the survey. Reduced work and performance was calculated by adding hours missed to the product of self-rated reduced work performance and the average monthly hours worked. The number of hours of reduced work and performance was multiplied by the hourly income to estimate the indirect costs to an employer of reduced performance. Total monthly US burden of reduced work and performance was estimated through extrapolation using government workforce statistics. The fringe benefit rate (37.8%, calculated as benefits divided by wages/salaries using Bureau of Labor Statistics data from March 2001)[18] was included in calculating the monthly employer economic burden because the employer pays for fringe benefits. This value was multiplied by the civilian workforce and the prevalence rate of MDD among employed to extrapolate the total monthly US burden of reduced work and performance.

array of industries and job classifications. Costs were computed from the perspective of employers as payments to providers by insurers (not including copayments or deductibles). Unit costs were estimated using data for the survey period (i.e., 20002001). The claims data contain un-identified information on patients demographics (e.g., age and gender), medical and pharmacy claims, and monthly eligibility files. Specifically, patients utilization of medical services was recorded with date of service, actual payments to providers, associated diagnoses (ICD-9), and performed procedures (CPT). Patients pharmacy claims contain prescribed medications (NDC), dates of prescription fills, days of supply, strength, quantity, and payment. To estimate unit costs representative of MDD costs comparable to the NCS-R survey workforce respondents, patients were selected for the analysis if they had a diagnosis of MDD (ICD-9 codes of either 296.2 or 296.3) before 2001 and were between the ages of 1864. Claims from the sample identified were grouped into treatment sectors measured by the NCS-R data. All mental health-related claims, based on ICD-9 codes, were compiled for the sample claimants for the years 2000 and 2001, corresponding to the years in which the survey was taken. Costs for any mental healthcare and any healthcare are weighted averages of the distinct treatment sectors from the claims database. To account for differences in patients hospital length of stay, the cost per day for inpatient stays was determined using the best-fitted line of the average costs per unique episode. The total cost of antidepressant use is based on the average number of antidepressants days supply and a claims based per day unit cost by antidepressant class. Indirect costs were calculated based on NCS-R survey questions. The following measures were calculated: adjusted average monthly hours worked, work performance self-rating, and the prevalence of missing one or more days of work in the past month. On the basis of the days missed and the productivity rating multiplied by the average number of monthly hours, an hourly estimate of monthly reduced work and performance was calculated. The incremental difference of monthly reduced work and performance between the nondepressed and each severity group was multiplied by each severity groups hourly wage to produce incremental indirect cost to the employer.

STATISTICAL ANALYSES
Descriptive measures including demographics and treatment prevalence were compared across three depression severity groups using Wald w2 statistics. Resource utilization comparisons used regression models that generated F statistics adjusted for demographics. Multiple regression analysis was used to predict 12-month prevalence and usage, with controls for sociodemographic variables and substance use disorders, following previous NCS-R studies.[6] We used a Generalized Linear Model to estimate the effect of severity on the use of mental health services and work performance.

RESULTS
DEMOGRAPHICS Using a sample of 4,465 workforce respondents, the prevalence of MDD among the US workforce was 7.6%. Using a weighted subsample of 539 MDD respondents, the prevalence of mild, moderate, and severe depression was 13.8, 38.5, and 47.7% in the US workforce (Po.05) (Table 1). Age, sex, race, and marital status were associated significantly with MDD, but were not associated with MDD severity: the prevalence of depression is highest among younger respondents (8.5% for respondents aged 1844 vs. 3.0% for

COST ANALYSIS
To estimate the cost of treatment, unit costs were applied to relevant (survey) utilization measures. Costs were estimated by weighing utilization measures by unit costs obtained for similar services used by depressed patients in a US privately insured claims database of 31 US employers (Ingenix Employer Solutions), covering 2.3 million beneficiaries. These companies have operations nationwide in a broad Depression and Anxiety

TABLE 1. Sociodemographic correlates of depression and depression severity NCS-R: N 5 4,465 Part 2 respondents in the workforce
Prevalence of depression Mild
b

Distribution of depression severitya Moderate Prevalence

e

Severe Prevalencee (95% CI)

Demographic variable Prevalence 7.6 8.5 8.5 7.0 3.0 3.0 3.0 2.4 1.0 1.9 1.0 0.8 0.7 1.4 1.0 2.0 1.7 1.0 1.3 1.0 1.2 1.0 1.0 0.9 1.1 1.0 (0.7, (0.6, (0.8, (1.0, 1.6) 1.3) 1.6) 1.0) 12.2 14.6 10.0 19.7 (0.9, (0.7, (0.9, (1.0, 1.9) 1.4) 1.6) 1.0) 8.4 18.8 13.9 11.3 (4.1, (12.7, (9.9, (6.8, (8.0, (8.5, (7.1, (11.7, (1.5, 2.7) (1.4, 2.2) (1.0, 1.0) 10.4 16.7 13.6 (6.1, 17.3) (10.2, 26.2) (9.1, 20.0) 16.5) 27.0) 19.3) 18.4) 18.2) 24.0) 14.0) 31.3) 36.6 34.9 42.6 31.8 35.7 42.4 41.4 45.5 28.5 41.2 42.2 (0.6, (0.5, (0.9, (1.0, 1.2) 0.9) 2.1) 1.0) 15.3 21.4 7.2 13.2 (7.6, (12.2, (2.5, (9.9, 28.5) 34.8) 19.2) 17.2) 48.0 30.9 36.0 38.4 (35.6, (18.8, (17.0, (32.9, 9.2 (1.5, 2.3) (1.0, 1.0) 12.0 17.0 (9.3, 15.3) (11.9, 23.8) 37.8 39.7 50.3 43.3 60.7) 46.3) 60.5) 44.2) (26.6, 48.0) (25.1, 46.2) (35.2, 50.2) 5.5 42.1) 45.1) 48.8) 51.7) 52.9) 36.3) 50.7) 55.3) 36.7 47.7 56.8 48.4 53.0 48.4 43.8 59.8 45.5 43.6 47.3 42.4 56.9 48.8 38.1 (2.0, (1.9, (1.6, (1.0, 4.5) 4.7) 3.8) 1.0) 9.9 5.5 38.1 (4.6, (4.1, (7.2, (7.1, 10.57 (8.6, 13.4) (8.0, 11.1) (4.8, 6.8) 27.58 (6.8, (5.8, (7.0, (5.5, 3.7 8.0 6.8 8.5 7.8 2.9 (6.3, (5.4, (7.2, (5.7, 10.0) 8.4) 10.1) 10.5) 11.8) 8.4) 9.7) 8.8) 8.9) 7.5) 15.2) 9.1) 6.5 5.6 10.5 8.0 10.8 9.5 5.7 9.0 7.0 8.3 7.0 (8.6, 11.3) (4.7, 6.5) (32.6, 43.2) (33.2, 46.7) 4.4 (6.8, (7.3, (7.4, (5.6, (2.0, 31.13 8.6) 9.9) 9.9) 8.7) 4.5) 13.8 15.3 14.1 11.0 23.6 (11.0, (9.2, (9.8, (6.3, (11.3, 17.3) 24.3) 20.0) 18.3) 42.7) 38.5 38.5 35.6 42.7 34.8 (34.1, (29.6, (28.6, (31.2, (16.2, 5.1 43.1) 48.4) 43.3) 55.0) 59.6) 47.7 46.1 50.2 46.4 41.6 (95% CI) Odds ratio (95% CI) Prevalence (95% CI) (95% CI)

Subgroup

Total samplef Age

All respondents 18rager29 30rager44 45rager59 60rage w2: 3 df test 1, 6 df test 2

(43.1, (38.3, (42.4, (34.6, (19.7,

52.3) 54.2) 58.1) 58.6) 67.4) (44.9, 55.7) (36.7, 50.1) (23.0, (33.1, (34.0, (42.8, 52.9) 62.7) 77.1) 54.1) (41.2, 64.5) (39.6, 57.2) (36.9, 51.0)

Sex

Female Male w2: 1 df test 1, 2 df test 2

Race

Hispanic Black Other White w2: 3 df test 1, 6 df test 2

Marital status

Separated/widowed/divorced Never married Married/cohabitating w2: 2 df test 1, 4 df test 2

Education

011 years 12 years 1315 years 161 years w2: 3 df test 1, 6 df test 2

(23.0, (27.3, (36.3, (31.7, 11.7 (38.2, (21.8, (32.3, (30.2, 23.2

(49.7, (34.9, (36.6, (38.8, (33.1, (48.8, (41.3, (30.0,

69.2) 56.5) 51.0) 56.0) 52.2) 64.6) 56.3) 46.8)

Research Article: Employer Burden of Depression by Severity

Region

Midwest South West Northeast w2: 3 df test 1, 6 df test 2

Only among respondents who are depressed. The prevalence rates for each depression severity group are only reported among the depressed in each demographic subgroup. c Total prevalence is percent with depression among the row subgroup. d Odds ratios are the odds of depression within the row subgroup in comparison to the contrast category. e Mild, moderate, and severe prevalences are the percents among depressed respondents in the row subgroup. f The total sample of depressed respondents (N 5 539) consisted of mildly (N 5 78), moderately (N 5 207), and severely (N 5 254) depressed respondents. These are weighted sample sizes. People with depression were over-sampled in Part 2. As a result, the ratio of cases with depression to the total sample size (539/4465) is larger than the weighted prevalence estimate of 7.6%. All prevalence estimates and odds-ratios are based on weighted data to correct for the over-sampling of cases into the Part 2 sample. Po0.05.

81

Depression and Anxiety

82

Birnbaum et al.

respondents 601 years); females have a higher prevalence than males (9.9 vs. 5.5%). Education was not significantly associated with either MDD or its severity. RESOURCE USE AND TREATMENT ADEQUACY There was a significant association between MDD severity and the use of mental health services. Usage of psychiatrist services increased with MDD severity, where 9.6% of patients with mild MDD used this type of service, compared to 15.6% for patients with moderate MDD and 23.8% for patients with severe MDD (P 5.034). Similarly, usage of other mental health services increased with MDD severity, where 17.1% (mild), 18.5% (moderate), and 28.5% (severe) patients used these other mental health services (P 5.026). Overall, 34.8% of mild, 48.0% of moderate, and 58.7% of severely MDD patients (P 5.004) used any type of healthcare treatment (Table 2). The use of antidepressants also increased with MDD severity. Patients in the severe group were more likely to use antidepressants than patients in the mild group (39.5 vs. 21.1%, P 5.006). They also had a higher number of days on medication (108.5 vs. 37.1 days, P 5.001). This finding holds for most antidepressant drug classes (except modified cyclics and SNRIs) (Table 2). Similarly, the adequacy of mental health services increased with MDD severity: psychiatrist treatment adequacy was statistically significantly higher for patients with severe MDD (3.3, 4.5, and 12.8%, P 5.004). The adequacy of other mental health treatment increased with MDD severity, but the association was not statistically significant (P 5.104). Overall, any mental health, general medical, any healthcare, and any treatment adequacy increased with MDD severity (all Po.02) (Table 2). COSTS The average 12-month costs per MDD patient in the workforce were substantially higher for patients with severe MDD vs. mild MDD. Costs for patients with severe MDD were generally at least twice those for persons with mild MDD (hospital: $91 vs. $33; other mental health: $468 vs. $203; any mental health: $697 vs. $388; general medical: $138 vs. $53; any healthcare: $834 vs. $442) (Table 3). Psychiatrist costs were similar for mild and severely depressed patients ($152 vs. $138). Costs for patients with moderate MDD were similar to those with severe MDD. EMPLOYMENT STATUS Compared to nondepressed respondents in the workforce, depressed patients have a significantly higher prevalence of disability and unemployment. Among workforce respondents, depressed respondents were 3.7 times more likely to be disabled than
Depression and Anxiety

nonrespondents (Po.001). Among the nondisabled workforce respondents, the depressed were 1.6 times more likely to be unemployed than the nondepressed (P 5.027). Disability increased from 10.8% to 16.4% to 20.1% for mild, moderate, and severe MDD, respectively. Similarly, unemployment among the nondisabled workforce respondents increased from 5.4% for mild, to 8.2% for moderate, and 14.1% for severe MDD (P 5.007, Table 4). Combining these mutually exclusive groups (disabled and unemployed) creates an even more substantial association. The prevalence of disabled or unemployed among workforce respondents was 12.7% for nondepressed and 26.1% for depressed respondents. The prevalence rates of disability/unemployment increased with MDD severity. Of the severely depressed respondents in the workforce, 31.3% were unemployed or disabled, which is significantly higher than the prevalence for mild (15.7%) or moderate MDD (23.3%) (P 5.003, Table 4).

WORKPLACE PERFORMANCE There were 3,740 employed respondents who worked in the past month. Depressed respondents worked significantly fewer adjusted monthly hours than nondepressed respondents: mildly depressed workers are estimated to miss 13.7 additional hours per month (P 5.016), while moderate and severely depressed respondents did not show any significant decrease in monthly hours worked (Table 5). Depression also had a significant impact on self-rated work performance. Moderately and severely depressed respondents in the workforce were 4 to 5% less productive than the mildly depressed and nondepressed (P 5.004 and Po.001, respectively). Severely depressed workers were also more likely to miss one or more days of work when compared to moderately, mildly, and nondepressed respondents. Compared to nondepressed respondents, mildly depressed respondents had reduced work performance by three hours (although not statistically significant), compared with 12.0 hr for moderately depressed and 14.8 hr for severely depressed respondents (P 5.001) (Table 5). This significant association between severity and work performance translates into significant costs for moderately and severely depressed respondents: the monthly cost of lower workplace performance was $44 for mildly depressed respondents compared to nondepressed, $188 per moderately depressed respondents, and $199 per severely depressed respondent (Po.001) (Table 5). Taking fringe benefits into account, the total monthly employer economic burden of reduced work and performance amounted to $60 for mild, $259 for moderate, and $274 for severely depressed respondents. Projected to the entire US workforce, these monthly MDD-related losses in worker productivity have a human capital cost of nearly $2.0 billion (with

TABLE 2. 12-Month treatment and medication usage of depression and depression severity NCS-R: N 5 4,465 Part 2 respondents in the workforce
Prevalence for depressed respondents Mild Prevalence (95% CI) Prevalence (95% CI) Prevalence (95% CI) (95% CI)
b b b

Distribution by depression severitya Moderate Prevalenceb Waldc w2 Severe

Prevalenced Sector of treatment 0.9 18.7 23.1 32.3 32.5 51.3 10.6 9.0 56.3 32.3 4.3 5.0 22.8 3.4 6.4 8.2 11.2 15.6 5.0 18.4 2.3 19.6 (6.1, 11.1) (8.8, 14.2) (12.8, 18.8) (3.1, 8.0) (15.2, 22.1) (1.2, 4.3) (16.4, 23.3) 3.3 6.2 6.2 2.5 8.6 2.6 11.2 (0.7, 14.0) (2.2, 16.1) (2.2, 16.1) (0.6, 9.5) (3.7, 18.7) (0.6, 10.5) (5.3, 22.2) 4.5 8.8 11.8 2.6 13.6 1.5 13.6 (27.8, (2.8, (3.4, (19.0, (2.3, (4.3, 37.1) 6.3) 7.2) 27.0) 5.0) 9.4) 21.1 3.9 3.7 9.9 3.7 4.1 (13.3, 32.0) (1.2, 12.5) (1.1, 11.3) (4.7, 19.7) (0.9, 13.7) (0.8, 18.0) 27.3 1.7 1.5 21.6 1.7 4.6 (21.1, (0.6, (0.5, (15.5, (0.7, (2.4, 34.6) 5.1) 4.8) 29.2) 4.4) 8.4) (2.3, 8.5) (5.3, 14.1) (7.6, 17.9) (1.1, 6.2) (9.0, 20.0) (0.5, 5.0) (9.0, 20.0) (50.8, 61.6) 35.9 (25.9, 47.3) 53.2 (46.9, 59.4) 64.7 39.5 6.4 8.1 27.4 4.6 8.5 12.8 14.7 21.3 7.8 25.1 2.9 26.9 (56.1, 72.4) (32.4, (3.8, (5.7, (21.3, (2.7, (5.7, (9.2, (10.9, (17.1, (4.6, (20.5, (1.4, (22.1, 47.0) 10.5) 11.4) 34.6) 7.8) 12.7) 17.5) 19.4) 26.3) 12.8) 30.4) 6.2) 32.4) (0.4, 2.4) (15.0, 23.0) (19.4, 27.2) (28.7, 36.2) (27.2, 38.2) (46.0, 56.5) (8.4, 13.3) (6.1, 13.2) 1.4 9.6 17.1 19.1 20.6 34.8 7.5 5.8 (0.2, 10.2) (3.5, 23.8) (9.9, 27.9) (11.7, 29.7) (10.6, 36.1) (24.1, 47.3) (3.4, 16.1) (2.5, 13.0) 0.4 15.6 18.5 27.2 32.6 48.0 11.7 9.0 (0.1, 2.7) (10.4, 22.9) (13.3, 25.1) (21.7, 33.6) (26.3, 39.7) (41.6, 54.4) (7.8, 17.2) (4.7, 16.6) 1.2 23.8 28.5 40.3 35.8 58.7 10.7 9.9 (0.4, (18.4, (23.2, (34.4, (28.1, (51.0, (7.5, (6.2, 4.1) 30.3) 34.4) 46.4) 44.3) 65.9) 14.9) 15.6)

Hospital Psychiatrist Other mental health Any mental healthe General medical Any healthcaree Human service Complementary alternative medicine Any treatmente

1.4 6.8 7.3 14.8 2.3 11.0 0.9 1.2 18.4 10.3 5.7 9.7 7.5 5.1 3.9 11.3 4.5 9.3 8.3 12.4 0.8 13.4

Medication

Antidepressantse Buproprion Modified cyclic SSRI SNRI Other antidepressant

Treatment adequacy

Psychiatrist Other mental health Any mental healthe General medical Any healthcaree Human service Any treatment

Research Article: Employer Burden of Depression by Severity

Prevalence

(95% CI)

Prevalence

(95% CI)

Prevalence

(95% CI)

Prevalence

(95% CI)

Wald F f Statistic

Number of visits and pill days on medicationd Sector of treatmentg Hospital Psychiatrist Other mental health Any mental health General medical Any healthcaree Human service Complementary alternative medicine Any treatment 0.1 1.5 3.8 5.4 1.0 6.4 0.9 2.4 9.6 (7.0, 12.2) (0.0, (0.9, (2.7, (4.1, (0.8, (5.0, (0.4, (0.8, 0.3) 2.0) 4.9) 6.7) 1.3) 7.9) 1.3) 4.0)

0.1 2.1 2.2 4.4 0.5 4.9 1.8 0.6 7.2

(0.0, (0.0, (0.4, (0.7, (0.2, (1.2, (0.6, (0.0,

0.2) 5.7) 4.0) 8.1) 0.8) 8.5) 3.0) 1.2) (3.1, 11.4)

0.1 0.7 2.8 3.5 0.9 4.5 0.5 1.4 6.3

(0.0, (0.3, (1.7, (2.4, (0.5, (3.1, (0.2, (0.0,

0.2) 1.1) 3.8) 4.7) 1.3) 5.8) 0.8) 2.7) (4.4, 8.2)

0.2 1.9 5.1 7.2 1.3 8.5 0.9 3.8 12.9

(0.0, (1.3, (3.2, (5.0, (0.9, (6.1, (0.1, (0.7,

0.5) 2.5) 6.9) 9.4) 1.7) 10.8) 1.7) 6.8) (8.3, 17.5)

0.5 7.0 2.6 4.4 3.2 4.8 2.5 2.8 3.7

83

Depression and Anxiety

84

TABLE 2. Continued
Prevalence for depressed respondents Mild Prevalenceb (95% CI) (64.3, (2.7, (4.6, (36.1, (2.4, (7.3, 94.8) 10.3) 14.9) 51.4) 10.7) 18.7) 37.1 0.7 3.2 24.8 5.9 2.5 (14.2, 60.0) (0.0, 1.9) (0.0, 8.3) (3.7, 45.9) (0.0, 13.4) (0.0, 7.1) 58.8 2.6 2.0 41.2 2.1 11.0 (41.6, 76.1) (0.0, 7.0) (0.0, 5.0) (27.5, 54.8) (0.0, 5.4) (4.9, 17.1) 108.5 11.4 17.9 51.3 10.3 17.6 (82.7, (4.0, (8.7, (40.1, (1.9, (7.8, 134.3) 18.8) 27.2) 62.4) 18.8) 27.4) (95% CI) (95% CI) (95% CI) 79.5 6.5 9.7 43.7 6.6 13.0 Prevalenceb Prevalenceb Prevalenceb Waldc w2 9.0 3.6 6.5 2.0 1.9 6.4 Moderate Severe Distribution by depression severitya

Depression and Anxiety Cost by depression severitya Mild (95% CI) Mean (95% CI) Mean Moderate (95% CI) Mean Severe (95% CI) Wald F P-value $24 $51 $254 $329 $100 $429 $88 ($34, $142) $139 ($0, $72) ($22, $80) ($157, $352) ($223, $426) ($56, $143) ($297, $534) ($98, $180) $91 $138 $468 $697 $138 $834 $256 ($0, $229) ($94, $182) ($295, $640) ($475, $892) ($94, $180) ($592, $1050) ($195, $317) 0.5 7.0 2.6 4.4 3.2 4.8 9.0 ($0, $129) ($66, $146) ($250, $448) ($388, $670) ($82, $140) ($483, $791) ($152, $224) $33 $152 $203 $388 $53 $442 ($0, $105) ($0, $409) ($33, $374) ($129, $757) ($17, $88) ($187, $819)

Medicationg

Antidepressants Buproprion Modified cyclic SSRI SNRI Other antidepressant

Only among respondents who are depressed. For Part 1, total prevalence is the percent of respondents with depression seeking treatment in the row sector. For Part 2, total means are among depressed respondents. c Wald w2 is a 2 df test of depression severity on the likelihood of treatment in the row sector for Part 1. Logistic regression model used to generate w2 test was adjusted for age, race, and sex for Part 1. d Mild, moderate, and severe prevalences are the percents of respondents within those sub-categories that are seeking treatment in the row sector. Means are among respondents in those subcategories. e Any mental health treatment includes hospital, psychiatrist, and other mental health treatment sectors. The combination of any mental health and general medical treatment yields the any healthcare treatment sector. Any treatment is the aggregate of all sectors of treatment. Similarly antidepressant medication is the aggregate of all the listed drug classes. f Wald F is a 2 df test of the effect of depression severity on the number of visits or days on medication for Part 2. Linear regression model used to generate F test was adjusted for age, sex, and race for part 2. g Only respondents who have made at least one visit to at least one sector of treatment or had at least one antidepressant prescription fill are included in Part 2. Po0.05.

Birnbaum et al.

TABLE 3. 12-Month treatment and medication cost by sector of treatment and by severity group NCS-R: N 5 4,465 Part 2 respondents in the workforce

Cost for depressed respondents

Mean

Sector of treatment Hospital Psychiatrist Other mental health Any mental health General medical Any healthcare

$57 $106 $350 $513 $111 $624

Medication Antidepressants

$188

Total cost computed by multiplying average number of visits and pill days on medication (Table 2) by unit costs computed from AGs employer claims data (20012002) for MDD patients. Po0.05.

Research Article: Employer Burden of Depression by Severity

85

Notes: The prevalence represents the percent of people among workforce respondents who are disabled, unemployed (among the nondisabled workforce respondents), and unemployed or disabled among workforce respondents. The odds ratio is the likelihood for depressed respondents compared to nondepressed respondents to be in the row category. Total prevalence is the percent of depressed respondents with employment status equal to the row category. Mild, moderate, and severe prevalences are the percents of respondents in those sub-categories with employment status equal to the row category. Wald w2 comparing depressed to nondepressed respondents is based on a model including 4,465 Part 2 respondents in the workforce. Wald w2 comparing mild, moderate, and severe depression is based on a model including 539 Part 2 respondents in the workforce with depression. Logistic regression models used to generate w2 tests were adjusted for age, race, and sex.

Statistic P-value

0.062

0.007

0.003

Wald w2

severely depressed respondents accounting for 55% of that cost).

11.7

5.6

9.9

(14.9, 26.5)

(24.5, 39.0)

(9.3, 20.8)

(95% CI)

DISCUSSION
This study estimated a 7.6% prevalence of MDD, which is similar to estimates in other research.[2,19,20] While the average MDD respondent had 79.5 days on antidepressant medication in the year before interview, this average is based on a sample in which only 1/3 of all MDD respondents took antidepressants. Hospital inpatient visits were not common among MDD respondents, with only 0.9% of MDD respondents having a hospital inpatient stay, resulting in a very low mean number of nights spent in a hospital (0.1 per year) among all MDD respondents. Consistent with these utilization measures, only 20% of all MDD respondents received treatment that could be considered even minimally adequate in terms of established MDD treatment guidelines. Respondents with more severe MDD, though, were more likely than others to receive adequate treatment. Regarding workplace performance, when examining different severity groups, there was a consistent trend: respondents with more severe MDD were more likely to be disabled or unemployed. Comparing mildly, moderately, and severely depressed respondents showed a trend that increased MDD severity was associated with increased disability and greater unemployment among nondisabled workforce participants. As a result, the employer burden may underestimate the societal cost of MDD. Similarly, increased MDD severity was associated with worse performance at work. There was a significant association between MDD severity and adjusted average monthly hours worked, work performance self-rating, the probability to miss one or more work days in past month, and monthly reduced work and performance. Note, however, that while increased MDD severity was not associated with reduced hours worked, it was associated with reductions in other measures of workplace performance. The reason we present the adjusted average monthly hours worked is to calculate the work loss of each severity group; this measure is descriptive, and lower adjusted hours are not necessarily a consequence of the severity of depression. For example, it may be that severely depressed individuals work more hours (160.4 hr/month) than those who are mildly depressed (155.4 hr/month), perhaps because it takes them longer to get their tasks completed. Alternatively, people may be severely depressed because they work the most hours. In general, this descriptive analysis of workplace performance is simply meant to provide context for the employer burden calculation and should not be interpreted as indicating a causal relationship between employment status and depression. For a more detailed analysis on the relationship between
Depression and Anxiety

Severe

TABLE 4. Employment status for nondepressed and depressed respondents NCS-R: N 5 4,465 Part 2 respondents in the workforce

By depression severity

Prevalence

20.1

14.1 (5.3, 12.5) 8.2 (2.0, 13.7) 5.4 0.027 4.9 (1.0, 2.5) 1.6 (7.5, 14.5) 10.5 6.6 (5.5, 7.8)

Moderate

Estimate (95% CI) Statistic P-value Prevalence (95% CI) Prevalence

(10.7, 24.4)

16.4

Mild

(5.2, 21.4)

10.9

Depressed vs. nondepressed

Wald w

0.000

82.1

Odds ratio

(2.8, 4.9)

3.7

Depressed respondents

Prevalence

(14.1, 21.3)

17.4

Nondepressed respondents

Prevalence

6.6

Disabled among workforce respondents Unemployed among nondisabled workforce respondents Unemployed or disabled among workforce respondents

12.7

(11.1, 14.4)

(95% CI)

(5.4, 7.9)

26.1

(21.2, 31.6)

(95% CI)

2.6

(2.0, 3.5)

47.6

0.000

15.7

(8.6, 26.9)

23.3

(16.4, 31.9)

(95% CI)

31.3

86

TABLE 5. Workplace burden for employed respondents NCSR: N 5 3,740 employed respondents who worked in the past month
Mean or prevalence Respondent subgroup Mildly depressed Moderately depressed Severely depressed Nondepressed Wald w2 3 df 2.92 85.2 79.4 79.1 84.7 8.13 34.0 34.0 40.8 27.5 12.34 37.0 47.4 49.8 34.4 10.26 14.5 15.6 13.4 16.9 (11.7, (13.3, (11.6, (16.0, 17.3) 17.9) 15.2) 17.8) 43.9 188.0 199.1 0.0 60.5 259.0 274.4 0.0 530.2 358.3 1,065.0 0.0 (83.9, 171.6) (83.3, 292.7) (101.5, 296.8) (0.0, 0.0) (27.0, (40.1, (42.7, (32.6, 46.9) 54.7) 57.0) 36.3) 3.0 12.0 14.8 0.0 (5.8, 11.8) (5.3, 18.8) (7.6, 22.1) (0.0, 0.0) 0.7 3.6 4.1 0.492 0.001 0.000 (24.7, (24.5, (34.3, (25.1, 44.6) 44.9) 47.7) 30.0) 1.2 1.2 1.6 1.0 (0.8, (0.8, (1.2, (1.0, 1.9) 1.9) 2.2) 1.0) 0.8 0.9 3.1 0.426 0.378 0.004 (81.5, (76.6, (76.7, (83.9, 89.0) 82.2) 81.5) 85.5) 0.7 (4.2) (4.5) 0.0 (2.6, 4.1) (6.9, 1.4) (6.9, 2.1) (0.0, 0.0) 0.4 3.0 3.8 0.666 0.004 0.000 Mildly depressed Moderately depressed Severely depressed Nondepressed Wald w2 3 df Mildly depressed Moderately depressed Severely depressed Nondepressed Wald w2 3 df Mildly depressed Moderately depressed Severely depressed Nondepressed Wald w2 3 df Mildly depressed Moderately depressed Severely depressed Nondepressed Mildly depressed Moderately depressed Severely depressed Nondepressed Mildly depressed Moderately depressed Severely depressed Nondepressed Mildly depressed Moderately depressed Severely depressed Nondepressed 155.4 169.4 160.4 173.7 (142.5, (158.7, (150.4, (171.0, 168.3) 180.1) 170.4) 176.5) (13.7) (0.6) (9.4) 0.0 (24.7, 2.7) (11.9, 10.7) (19.0, 0.1) (0.0, 0.0) 2.5 0.1 2.0 Estimate (95% CI) Estimate (95% CI) T-test P-value 0.016 0.918 0.053 b or odds ratio

Depression and Anxiety

Adjusted average monthly hours worked

Work performance self-rating: 100pt scale

Miss one or more work days in past month

Monthly reduced work and performance (hours)

Birnbaum et al.

Hourly income ($)

Monthly cost of reduced work and performance ($)

Total Monthly Employer Economic Burden of Reduced Work and Performance ($)

Total Monthly US Economic Burden of Reduced Work and Performance ($ millions)

Notes: Linear regression was used to model adjusted average hours, self-rating, and reduced work and performance. The models were adjusted for age, age squared, race, and sex. The self-rating model was adjusted for the average rating of similar workers. The reduced work and performance model was adjusted for average hours worked. Logistic regression was used to model the likelihood of missing one or more days of work in the past month. The model was adjusted for age, race, and sex. Adjusted average monthly hours equals average monthly hours plus extra hours worked minus the number of work hours missed. Monthly reduced work and performance equals work hours missed plus the reduction in work performance multiplied by the average monthly hours worked. Monthly cost of reduced work and performance is equal to the income mean multiplied by the reduced work and performance regression estimates. Total monthly employer economic burden is calculated by multiplying the monthly cost of reduced work and performance by the fringe benefit rate. Total monthly US economic burden is calculated by multiplying the total monthly employer economic burden by the US workforce, and the prevalence rate of depression among the employed. US civilian workforce for 4th quarter 2001 equaled 134,308,000 according to the Bureau of Labor Statistics. The BLS reported, as of March 2001, that fringe benefits were 37.8% of wages and salaries.

Research Article: Employer Burden of Depression by Severity

87

mental health, potential confounders, and labor market outcomes, see Chatterji (2007).[21] These factors taken together lead to an association between greater employer economic burden and US economic burden with increased MDD severity. Among MDD respondents in the workforce, there was a significant association between MDD severity and treatment usage and costs. The cost estimates are consistent with prior research. For example, annualizing the $2 billion monthly incremental workplace burden (i.e., excess costs of reduced absenteeism/ presenteeism, compared to non-MDD employees) yields an estimate of $24 billion. For a similar period, 20012002, Stewart estimated annual excess work absence and reduced performance costs of US workers with depression at $31 billion.[22] Greenberg estimated workplace costs (absenteeism/presenteeism) in 2000 at $36 billion; however, that estimate included other forms of depression (e.g., bipolar disorder) as well as short- and long-term disability leave, which were not included here.

As discussed elsewhere, workplace costs account for most of the costs of depression. Table 6 organizes some key study results regarding the prevalence and adequacy of treatment, as well as healthcare treatment costs and work performance costs (annualized from monthly cost) for a 10,000 employee firm, of which 760 employees have MDD (using a 7.6% depression prevalence rate). The severe and moderate MDD groups account for 96.5% of the employers costs due to reduced work performance. While the per employee costs due to reduced work performance of patients with moderate MDD are similar to those with severe MDD ($2,256 vs. $2,389), patients with severe MDD account for a disproportionate share of treatment costs (63.8% of costs vs. 26.5% for moderately severe). This cost pattern is consistent with treatment patterns; patients with severe MDD have a higher treatment rate compared to patients in the moderate MDD group (58.7 vs. 48.0%), a higher treatment adequacy rate (25.1 vs. 13.6%), and, therefore, a higher proportion of treated patients who are treated adequately (42.8 vs.

TABLE 6. Treatment and work performance costs of depression in a firm of 10,000 employees
Severity of depression Mild [A] Prevalence of depression by severitya Number of depressed employeesb Treatment ratec Number of treated employees Percent of depressed receiving adequate treatmentc Number of employees treated adequately Treatment adequacy as a proportion of all treatment Annualized depression treatment cost per depressed employeed Annualized depression treatment cost of all depressed employees Percent of depression treatment cost by severity Annualized work performance cost per depressed employeee Annualized work performance cost of all depressed employees Percent of work performance cost by severity Annualized treatment cost 1 work performance cost per depressed employee Annualized treatment cost 1 work performance cost of all depressed employees Percent of treatment cost 1 work performance cost by severity
a b

Moderate [B] 38.5% 293 48.0% 140 13.6% 40 28.3% $429 $125,525 26.5% $2,256 $660,106 41.7% $2,685 $785,631 38.2%

Severe [C] 47.7% 363 58.7% 213 25.1% 91 42.8% $834 $302,342 63.8% $2,389 $866,060 54.8% $3,223 $1,168,402 56.8%

Total cost (or employees) [D] 5 [A]1[B]1[C]

13.8% 105 34.8% 36 8.6% 9 24.7% $442 $46,357 9.8% $527 $55,272 3.5% $969 $101,629 4.9%

(760) (390) (140)

$474,224

$1,581,438

$2,055,662

See Table 1 The number of depressed employees is calculated by multiplying the 10,000 employees in the firm by the overall depression prevalence rate (7.6%) and the prevalence rate of depression by severity. c See Table 2. d See Table 3. e See Table 5; the monthly work performance cost per depressed employee from Table 5 was annualized. Depression and Anxiety

88

Birnbaum et al.

28.3%). This pattern suggests a potential opportunity to improve health outcomes among MDD patients by targeting the treatment rate and treatment adequacy rate for patients with moderate MDD that is now found among those who have severe MDD. LIMITATIONS The results are based on data from a subsample of persons in the workforce among respondents in the most recent nationally representative survey of mental disorders in the United States (the NCS-R), which was conducted in 20012002. The results rely on 12-month recall of mental health medical services and medication utilization except for the workplace performance data, which asks about performance and hours worked in the previous 30-day period. The indirect (workplace performance) costs are incremental as they represent the additional cost of MDD over that of a non-MDD worker, and therefore do not represent absolute costs. The medical/drug utilization and cost data are limited to mental health services only. This paper discusses only the gross effects of depression and does not address the endogeneity of other comorbidities common among depressed patients (e.g., pain, sleep disorders, and headaches).[23] It is difficult to discern if depression causes these comorbidities or vice versa. Thus, the results represent the cost associated with MDD employees without addressing the underlying causal relationship with other comorbidities. To investigate the issue of causality was beyond the scope of this paper and future research should explore this relationship.

CONCLUSIONS
Among MDD respondents in the workforce, there was a significant association between MDD severity and treatment usage and costs, as well as between treatment adequacy and severity. Rates of unemployment and disability increased with MDD severity, whereas work performance decreased with MDD severity. The direct and indirect costs of MDD by severity group raise the question of how to cost-effectively treat MDD. The prevalence of mild depression and its potential to become more severe may suggest focusing treatment on the mildly depressed to head off more serious illness. Alternatively, because depression treatment is proactive and treatment quality is so variable and problematic, some studies have found that it is not cost-effective to screen for depression.[24] This study was not designed to explore intervention strategies or resource allocation and to do so would require addressing the issue of causality. Generally, the results regarding treatment adequacy imply a need for overall improvement of treatment quality, specifically in regard to MDD severity. Better management of MDD with attention to severity levels may offer direction for improving care and better management of employers direct and indirect costs.
Depression and Anxiety

Acknowledgments. Funding for this research was provided by sanofi aventis. Howard Birnbaum, David Kelley, and Paul Greenberg are employed by Analysis Group, Inc., an independent consultancy providing research to various pharmaceutical, government, and nongovernment organizations. Ms. Rym Ben-Hamadi is a consultant to Analysis Group. Dr. Kessler is a professor at Harvard Medical School and has been a consultant for GlaxoSmithKline Inc., Kaiser Permanente, Pfizer Inc., sanofi aventis, Shire Pharmaceuticals, and Wyeth-Ayerst; has served on advisory boards for Eli Lilly & Company and Wyeth-Ayerst; and has had research support for his epidemiological studies from Bristol-Myers Squibb, Eli Lilly & Company, GlaxoSmithKline, Johnson & Johnson Pharmaceuticals, Ortho-McNeil Pharmaceuticals Inc., Pfizer Inc., and sanofi aventis. Vijay N. Joish is an employee of sanofi aventis. All authors were responsible for the design and analysis of the research. Results and conclusions are wholly the work of the authors. The authors thank Mark Lane, Matthew Hsieh, and Evan Kantor for their helpful research support. Collaborating NCS-R investigators include Ronald C. Kessler (Principal Investigator, Harvard Medical School), Kathleen Merikangas (Co-Principal Investigator, NIMH), James Anthony (Michigan State University), William Eaton (The Johns Hopkins University), Meyer Glantz (NIDA), Doreen Koretz (Harvard University), Jane McLeod (Indiana University), Mark Olfson (New York State Psychiatric Institute, College of Physicians and Surgeons of Columbia University), Harold Pincus (University of Pittsburgh), Greg Simon (Group Health Cooperative), Michael Von Korff (Group Health Cooperative), Philip S. Wang (NIMH), Kenneth Wells (UCLA), Elaine Wethington (Cornell University), and Hans-Ulrich Wittchen (Max Planck Institute of Psychiatry; Technical University of Dresden). The views and opinions expressed in this report are those of the authors and should not be construed to represent the views of any of the sponsoring organizations, agencies, or US Government. A complete list of NCS publications and the full text of all NCS-R instruments can be found at http://www.hcp.med.harvard.edu/ncs. Send correspondence to ncs@hcp.med.harvard.edu. The NCS-R is carried out in conjunction with the World Health Organization World Mental Health (WMH) Survey Initiative. We thank the staff of the WMH Data Collection and Data Analysis Coordination Centres for assistance with instrumentation, fieldwork, and consultation on data analysis. These activities were supported by the National Institute of Mental Health (R01 MH070884), the John D. and Catherine T. MacArthur Foundation, the Pfizer Foundation, the US Public Health Service (R13-MH066849, R01MH069864, R01-DA016558), the Fogarty International Center (FIRCA R03-TW006481), the Pan American Health Organization, Eli Lilly and Company, Ortho-McNeil Pharmaceutical, Inc.,

Research Article: Employer Burden of Depression by Severity

89

GlaxoSmithKline, and Bristol-Myers Squibb. A complete list of WMH publications can be found at http:// www.hcp.med.harvard.edu/wmh/.

12.

REFERENCES
1. Greenberg PE, Kessler RC, Birnbaum HG, Leong SA, Lowe SW, Berglund PA, Corey-Lisle PK. The economic burden of depression in the United States: how did it change between 1990 and 2000? J Clin Psychiatry 2003;64:14651475. 2. Kessler RC, Berglund P, Demler O, Jin R, Koretz D, Merikangas KR, Rush AJ, Walters EE, Wang PS. National comorbidity survey replication. The epidemiology of major depressive disorder: results from the National Comorbidity Survey Replication (NCS-R). JAMA 2003;289;30953105. 3. Rush A, Trivedi M, Wisniewski S et al. Acute and longer-term outcomes in depressed outpatients requiring one or several treatment steps: a STARD report. Am J Psychiatry 2006;163: 19051917. 4. Bloom BS. Prevalence and economic effects of depression. Manag Care 2004;13:916. 5. Greenberg PE, Stiglin LE, Finkelstein SN, Berndt ER. The economic burden of depression in 1990. J Clin Psychiatry 1993;54:405418. 6. Kessler RC, Akiskal HS, Ames M et al. Prevalence and effects of mood disorders on work performance in a nationally representative sample of U.S. workers. Am J Psychiatry 2006;163:15611568. 7. Crown WH, Finkelstein S, Berndt ER et al. The impact of treatment-resistant depression on health care utilization and costs. J Clin Psychiatry 2002;63:963971. 8. Corey-Lisle PK, Birnbaum H, Greenberg PEP, Marychenko M, Claxton AJ. Identification of a claims data signature and economic consequences for treatment-resistant depression. J Clin Psychiatry 2002;63:717726. 9. Russell JM, Hawkins K, Ozminkowski RJ et al. The cost consequences of treatment-resistant depression. J Clin Psychiatry 2004;65:341347. 10. Greenberg PE, Corey-Lisle PK, Birnbaum H, Marynchenko M, Claxton A. Economic implications of treatment-resistant depression among employees. Pharmacoeconomics 2004;22:363373. 11. 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

13. 14.

15. 16.

17.

18.

19.

20.

21.

22.

23.

24.

in patients with chronic major depression. Biol Psychiatry 2003;54:573583. Kessler RC, Berglund P, Chiu WT et al. The US national comorbidity survey replication (NCS-R): design and field procedures. Int J Methods Psychiatr Res 2004;13:6992. Rush AJ, Carmody T, Reimitz P-E. The inventory of depressive symptomatology (IDS). Int J Methods Psychiatr Res 2000;9:4559. Rush AJ, Gullion CM, Basco MR, Jarrett RB, Trivedi MH. The inventory of depressive symptomatology (IDS). Psychol Med 1996;26:477486. Hamilton M. A rating scale for depression. J Neurol Neurosur Psychiatry 1960;23:5662. Kessler RC, Barber C, Beck AL et al. The World Health Organization Health and Work Performance Questionnaire (HPQ). J Occup Environ Med 2003;45:156174. Kessler RC, Ames M, Hymel PA et al. Using the WHO health and work performance questionnaire (HPQ) to evaluate the indirect workplace costs of illness. J Occup Environ Med 2004;46:S23S37. Bureau of Labor Statistics. Employer Costs for Employee Compensation March 2001. United States Department of Labor. June 29, 2001. Hasin DS, Goodwin RD, Stinson FS, Grant BF. Epidemiology of major depressive disorder: results from the national epidemiologic survey on alcoholism and related conditions. Arch Gen Psychiatry 2005;62:10971106. Stinson FS, Dawson DA, Patricia Chou S et al. The epidemiology of DSM-IV specific phobia in the USA: results from the national epidemiology survey on alcohol and related conditions. Psychol Med 2007;37:10471059. a M, Lu M, Takeuchi D. Psychiatric disorders Chatterji P, Alegr and labor market outcomes: evidence from the National Latino and Asian American Study. Health Econ 2007;16:10691090. Stewart WF, Ricci JA, Chee E, Hahn SR, Morganstein D. Costs of lost productive worktime among US workers with depression. JAMA 2003;289:31353144. Kessler R, White LA, Birnbaum H et al. Comparative and interactive effects of depression relative to other health problems on work performance in the workforce of a large employer. J Occup Environ Med 2008;50:809816. Valenstein M, Vijan S, Zeber JE, Boehm K, Buttar A. The costutility of screening for depression in primary care. Ann Intern Med 2001;134:345360.

Depression and Anxiety

Copyright of Depression & Anxiety (1091-4269) is the property of John Wiley & Sons Inc. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.