Enter Canada Site

Summary

  • A better understanding of the risk factors associated with future work disability will allow for the development of interventions to decrease worker absenteeism.
  • Employee disability and medical claims were analyzed to identify risk factors associated with future work disability leaves in patients with newly diagnosed depression (n = 9,954).
  • Patients most at risk for future work disability are those with high use of antidepressants, those with a history of comorbidities and previous work disability, and non-salaried and non-union employees.
  • This analysis suggests opportunities for interventions among patients with a new major depressive disorder diagnosis, including the reduction of time from diagnosis to the first psychotherapy visit.

Introduction

A recent estimate suggests that over half (54.6%) of household heads in the United States will report a work disability; among these, nearly 25% will report a severe work disability.1 Disabled workers lose earnings while struggling with the stigma associated with relying on disability benefits. Unplanned employee absences like disability leaves also cost employers an estimated 6% of payroll expenses.2 Identifying the risk factors associated with work disability and intervening on behalf of affected individuals will benefit the employer, patient, and payer.

Researchers have attempted to identify risk factors associated with future work disability leave. For example, in a large sample of Finnish, public sector employees, Airaksinen et al. found that much of the variability in worker disability incidence was explained by the following variables: age, body mass index, difficulty falling asleep, self-rated health, smoking, socioeconomic position, number of chronic illnesses, and number of sickness absences in the previous year.3 Unhealthy behaviors, such as heavy drinking and smoking, and stressful work conditions, such as excessive job demands and high job strain, were also identified as risk factors.3 However, few studies have attempted to identify risk factors for employee disability leaves due to mental health conditions – such as major depressive disorder (MDD) – exclusively. Given 7.4% of the global disability burden is attributable to behavioral and mental disorders, mental health conditions demand greater attention as the cause of worker disability leave.

For individuals with MDD, pharmacotherapy and psychotherapy are the primary treatment modalities. Burton et al. (2007) reported that employees with an antidepressant prescription were shown to have higher odds of a short-term disability leave when they failed to adhere to their treatment with antidepressants.4 Previous research also shows that patients with MDD frequently fail to adhere to their treatment when it is initiated, and their antidepressant adherence decreases with time.5,6 Many others never receive treatment; for example, Soria-Saucedo et al. reported that fewer than 20% of individuals diagnosed with MDD receive either pharmacotherapy or psychotherapy.7

In this research brief, employee disability and medical claims were analyzed to understand the risk factors associated with future work disability leaves in patients with newly diagnosed depression.

Methods

Data. This analysis used disability and medical claims from the IBM® Watson™ MarketScan® Commercial Claims and Encounters (CCAE) database and the MarketScan® Health and Productivity and Management database; claims made between 2007 and 2016 were included. Patients were only included if they had health care coverage for one year before and after their first reported MDD diagnosis. To help ensure a true MDD diagnosis, patients had to have at least two depression diagnoses noted in their records. Patients with a first MDD diagnosis following an inpatient visit were excluded because an inpatient stay suggests a confounding functional limitation. Since the treatment for MDD during breastfeeding and prenatal periods differs from the general population, patients were also excluded if they gave birth within one year of their first MDD diagnosis. To only evaluate predictors of work disability following a de novo depression diagnosis, patients could not have an antidepressant prescription, psychiatric visit, or temporary disability leave due to depression before their first depression diagnosis. Patients were also excluded if the following disorders, as defined by the multi-level Clinical Classification Software (CCS) grouper,8 were present in the year prior to the initial MDD diagnosis: bipolar disorders, delirium dementia and amnestic and other cognitive disorders, other central nervous system disorders, other and unspecified hereditary and degenerative nervous conditions, and schizophrenia and other psychotic disorders. 

Variables. The following demographic variables were taken from the CCAE database: age, sex, geographic location, employee’s job industry, and whether the employee was salaried or in a union. Coexisting conditions and comorbidities (such as anxiety disorders, diabetes, hypertension, and others) were included if they were noted in the year before the patient’s first MDD diagnosis. Adherence to antidepressant medication was measured as the proportion of days covered (PDC) during the acute phase of treatment, defined as 114 days following the first MDD diagnosis. Patients were considered adherent if they met the 80% (0.8) PDC threshold for adherence as suggested by the American Psychiatric Association’s Practice Guideline for the Treatment of Patients with Major Depressive Disorder.9

Statistical Analysis. Individual-level variables were tested for their association with time to work disability leave using Cox proportional hazards models. Variables were considered statistically significant at a p < 0.05 cutoff. Kaplan-Meier survival curves were also used to visualize the association between these variables and time to worker disability leave. For the multi-variable analysis, predictors were selected with the Least Absolute Shrinkage and Selection Operator (LASSO) procedure. Variables selected by the LASSO procedure were fit to the full data and reevaluated with a Cox proportional hazards model. All analyses were performed in the R statistical computing environment (version 3.4.3). 

Results

Demographics. A total of 10,129 patients with a new MDD diagnosis were included in this study population. Of these, 175 were later excluded because their disability leave occurred within the acute phase of depression treatment, before the efficacy of pharmacotherapy or psychotherapy could be evaluated, leaving a new total of 9,954 patients. Patients were followed for a median of 2.49 years before their MDD diagnosis and 3.45 years after their diagnosis. Over half (51%) of all diagnoses were made for moderate depression, while 22% and 27% of diagnoses were made for mild depression and severe depression, respectively. Among all patients, 2,558 (26%) had a work leave unrelated to pregnancy, and the remainder had no recorded work leave. The study population was almost evenly split between females (51%) and males (49%), and the median age at first MDD diagnosis was 38 years for all of the study population.

Single-Variable Analysis. Because of the large study population, many risk factors were shown to be significantly associated with time to work disability (Table 1). Females, hourly workers, those living in a rural area, and union members had a greater risk of work disability. Depression severity was negatively associated with time to disability (Figure 1A); patients with severe depression, as compared to mild and moderate depression, had the greatest risk of future work disability. The single-variable analysis also shows greater risk for patients that did not adhere to their prescribed antidepressant medication during the acute phase of depression treatment (Figure 1B). Notably, greater risk is shown for patients not receiving psychotherapy during the acute phase of depression treatment (Figure 1C). Lastly, patients with a prior disability leave show much greater risk for another disability than those without one (Figure 1D). About 15% of all patients considered in this study had a previous work disability leave.

Multi-Variable Analysis. The LASSO procedure identified a combination of 13 variables characterizing the groups most at risk for work disability leave (Table 2). Six of these were employer-related variables, including two variables indicating whether or not the employee belonged to a union (a risk factor), and whether or not the employee was salaried (a protective factor). The other four employer-related variables indicated the industry in which the employee was employed at the time of their first MDD diagnosis. For example, employees in the finance, insurance, or real estate industry at the time of their first MDD diagnosis had a greater risk of disability leave, whereas employees in the services industry at the time of their first MDD diagnosis had a lesser risk of disability leave. Patients covered by a health maintenance organization (HMO) insurance plan and patients with a psychotherapy visit within seven days of their first MDD diagnosis also had a lesser risk of disability leave.

Discussion

Employee disability and medical claims were used to identify risk factors associated with time to work disability leave for employees with a new MDD diagnosis. Measures of pharmacotherapy and psychotherapy adherence and utilization were shown to be associated with time to work disability leave. Demographic factors, comorbidities, and work-related factors were also revealed as risk factors. The final multi-variable analysis then identified a subset of variables that were most predictive of time to work disability.

The results from the single-variable analysis are consistent with previous studies analyzing the relationship between demographic factors and return-to-work (RTW) delays following a disability leave. For example, De Rijk et al. showed greater RTW delays among females, as compared to males.10 Other studies also show greater RTW delays among employees with previous claims.11 The lingering effects of a previous condition may explain the recurrence of disability leaves; these employees may also fear re-injury. The association between living in a rural area and disability incidence may similarly be explained by a lack of access to healthcare facilities and social support networks.

Six of the 13 variables selected in the multi-variable analysis were work related variables. These variables may act as surrogates for employee disability benefits. Salaried pay most likely accompanies other employee benefits, such as better health care coverage and employee wellness plans, which may explain the association shown here. Belonging to an employee union was also positively associated with future work leaves, but those in a union may benefit from more generous disability packages, which may provide them with an increased sense of financial security when considering the consequences of taking a disability leave.

The single- and multi-variable analyses revealed an association between antidepressant usage during the acute phase of treatment and future risk of work disability leave. Those with greater antidepressant usage had a greater risk of future work disability leaves as compared to patients who took no antidepressant medication. This finding is likely confounded by the severity of depression; individuals with more severe depression are more likely to be prescribed an antidepressant. Future work should consider the progression of depression severity through time, instead of just focusing on the initial diagnosis.

Perhaps the two most actionable and encouraging results from this study are as follows: 1) the decreased risk for patients with HMO insurance plans, and 2) the decreased risk for patients with a psychotherapy visit within seven days of their initial MDD diagnosis. Health maintenance organizations often provide integrated care and thus, MDD patients may be better able to afford health care services that can treat their depression, or any other condition that would cause a work disability. Our analysis also suggests that seeing a psychotherapist soon after an MDD diagnosis prevents future work disability. If these findings can be confirmed with further studies, then care coordinators should ensure proper, speedy referral to a mental health specialist for all patients diagnosed with MDD.

Conclusion

This analysis suggests a subset of patient-level variables that are most predictive of future work disability leave. Once these risk factors are identified, the appropriate interventions can be developed and implemented for these groups, to help prevent worker absenteeism. These early results also suggest two particularly actionable factors associated with decreased risk of future disability leave: 1) enrollment in a health management organization insurance plan, and 2) a psychotherapy visit with seven days of an initial MDD diagnosis.

About the Authors

Joshua Morrison is a second-year Master of Public Health student, with an Applied Biostatistics concentration, at the Colorado School of Public Health. He joined ReedGroup as a Health Informatics Researcher this year. His diverse research background includes quantifying larval insect movement and tracking the migration of Colorado bluebirds.

Dr. Fraser Gaspar is an Epidemiologist with MDGuidelines at ReedGroup since 2016. His research focuses on the factors that influence a patient’s successful return to activity and the use of evidence-based treatment guidelines in improving health outcomes. Dr. Gaspar completed his PhD and MPH at University of California Berkeley’s School of Public Health in Environmental Health Sciences.

Figures

figures charts

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Tables

 

table 2

 

References

  1. Laditka JN, Laditka SB. Work disability in the United States, 1968–2015: Prevalence, duration, recovery, and trends. SSM – Popul Heal. 2018;4(October 2017):126-134. doi:10.1016/j.ssmph.2017.12.006
  2. Mercer. The Total Financial Impact of Employee Absences Survey Highlights October 2008. 2008;(October). http://www.fmlainsights.com/wp-content/uploads/sites/311/2011/09/mercer-survey-highlights1.pdf.
  3. Airaksinen J, Jokela M, Virtanen M, et al. Development and validation of a risk prediction model for work disability: Multicohort study. Sci Rep. 2017;7(1):1-12. doi:10.1038/s41598-017-13892-1
  4. Burton WN, Chen C, Conti DJ, Schultz AB, Edington DW. The Association of Antidepressant Medication Adherence with Employee Disability Absences. Am J Manag Care. 2007;13(2):105-112.
  5. Keyloun KR, Hansen RN, Hepp Z, Gillard P, Thase ME, Devine EB. Adherence and Persistence Across Antidepressant Therapeutic Classes: A Retrospective Claims Analysis Among Insured US Patients with Major Depressive Disorder (MDD). CNS Drugs. 2017;31(5):421-432. doi:10.1007/s40263-017-0417-0
  6. Robinson R, Long S, Chang S, et al. Higher Costs and Therapeutic Factors Associated With Adherence to NCQA HEDIS Antidepressant Medication Management Measures: Analysis of Administrative Claims. J Manag Care Pharm. 2006;12(1):43-54.
  7. Soria-Saucedo R, Walter HJ, Cabral H, England MJ, Kazis LE. Receipt of Evidence-Based Pharmacotherapy and Psychotherapy Among Children and Adolescents With New Diagnoses of Depression. Psychiatr Serv. 2016;67(3):316-323. doi:10.1176/appi.ps.201500090
  8. Agency for Healthcare Research and Quality. Clinical Classifications for ICD-10 Data. www.ahrq.gov/research/data/hcup/icd10usrgd.html. Published 2018.
  9. American Psychiatric Association Practice Guidelines for the Treatment of Patients With Major Depressive Disorder. Am J Psychiatry. 2010;150(4):1-26. doi:10.1176/appi.books.9780890423387.654001
  10. De Rijk A, Janssen N, Alexanderson K, Nijhuis F. Gender differences in return to work patterns among sickness absentees and their associations with health: A prospective cohort study in the Netherlands. Int J Rehabil Res. 2008;31(4):327-336. doi:10.1097/MRR.0b013e3282fba37c
  11. Prang KH, Bohensky M, Smith P, Collie A. Return to work outcomes for workers with mental health conditions: A retrospective cohort study. Injury. 2016;47(1):257-265. doi:10.1016/j.injury.2015.09.011

Joshua Morrison

Joshua Morrison

Joshua Morrison is a second-year Master of Public Health student, with an Applied Biostatistics concentration, at the Colorado School of Public Health. He joined ReedGroup as a Health Informatics Researcher this year. His diverse research background includes quantifying larval insect movement and tracking the migration of Colorado bluebirds.