Private-payer Costs for Employees with Treatment-resistant Depression
Individuals with treatment-resistant depression do not achieve and sustain remission after multiple pharmacotherapies, with an unknown impact on employment. This study investigates the expenditures of TRD from a private-payer perspective in Canada.
Objective: Individuals with treatment-resistant depression (TRD) do not achieve and sustain remission after multiple pharmacotherapies. The impact on employment is unknown. Our objective was to investigate the expenditures of TRD from a private-payer perspective in Canada.
Methods: An employer-sponsored benefits plan database (2011-2012) was used to define a cohort of non-TRD and TRD employee claimants. The employees’ covered family members were not included. TRD employee claimants were defined as those employees with previous experience with a number of antidepressants. The costs of prescription medication utilization, short-term disability (STD), and long-term disability (LTD) benefits were calculated (2011 and 2012 $CAN) for both non-TRD and TRD groups. Descriptive statistics were used to characterize the cohort of employee claimants and the resources and costs for employee claims.
Results: There were 55,324 and 61,028 employee claimants in 2011 and 2012, respectively. Approximately 10% of employee claimants (9.9% in 2011, 9.7% in 2012) were treated for depression, and 1.3% of employee claimants were classified as TRD employee claimants. The medication costs for treating depression were approximately $780 per TRD employee claimant compared with ~$300 per non-TRD employee claimant. Weighted average STD costs per employee disability claimant were approximately $7300 for TRD and $5000 for non-TRD. Weighted average LTD costs per employee disability claimant were approximately $13,800 for TRD and $12,600 for non-TRD.
Conclusions: Employee claimants identified with TRD had higher medication, STD, and LTD costs than those with non-TRD. Limitations include lack of diagnostic information for employee claimants and small sample sizes for STD and LTD subgroups.
Am J Pharm Benefits. 2016;8(4):e67-e74
Depression is a common mental health disorder, affecting an estimated 350 million people worldwide.1 More than 1 million Canadians suffer from some form of depressive illness.2 One of the most significant depressive disorders, major depressive disorder (MDD), is associated with a high rate of nonrecovery and recurrence.3
The estimated 1-year prevalence of MDD in the United States is approximately 7%-10%,3-6 and according to the Global Burden of Disease Study, MDD is one of the world’s leading causes of disability.5,7 It profoundly impacts occupational functioning, social interactions, family dynamics, and physical health.2 At its worst, it can lead to suicide, with approximately 1 million lives lost annually across the globe.1
MDD involves the occurrence of 1 or more major depressive episodes. Each episode is defined as a period lasting at least 2 weeks that is characterized either by depressed mood (most of the day and/or nearly every day) and/or markedly diminished interest or pleasure in all or almost all activities.6 In addition to these core features, during the same 2-week period an individual must also have 5 of the following: weight loss or gain; insomnia or hypersomnia; psychomotor agitation or retardation; fatigue; feelings of worthlessness or excessive guilt; diminished ability to concentrate; recurrent thoughts of death; suicidal ideation or suicide attempt.6
Unfortunately, most adults with MDD fail to achieve remission with first-line pharmacological treatment strategies.3 In fact, at least half will not achieve and sustain remission following multiple antidepressant pharmacologic approaches.3 The individuals of this latter group are said to have treatment-resistant depression (TRD). TRD is the clinical term used to define this subgroup of individuals with MDD who fail to respond to conventional antidepressant treatment approaches.8
In defining TRD, McIntyre and colleagues propose that failure to respond to an adequate trial of 2 or more mechanistically dissimilar antidepressants would identify a treatment-resistant population.2 Fava and colleagues further define an adequate antidepressant trial as one in which the patient receives standard doses (ie, doses superior to placebo in randomized, double-blind studies) administered continuously for a minimum duration of 6 weeks.9
TRD is receiving increasing attention because it is thought to be a major contributor to the burden of illness associated with MDD.3,7 A conservative estimate for 1-year prevalence of TRD in the United States is 1%-4%, equal to or greater than that of schizophrenia, obsessive-compulsive disorder, or Alzheimer’s dementia.4
One reason the burden is considered particularly substantial is that age of onset is less than 30 years in most affected individuals.3 As such, the potential for significant impact on social and occupational functioning is tremendous at both the individual and population levels.3
In the United States alone, the estimated annual costs attributable to MDD are approximately $83 billion, with indirect costs due to decreased psychosocial functioning, most notably, workforce performance, being a major contributor (>60%).3,10 The World Health Organization (WHO) launched the Global Burden of Disease Study, in 1992, to quantify disease burden. At that time, MDD was ranked 4th among all diseases despite its relatively low mortality rate.7
Although both the WHO disabilities and cost figures noted above are for MDD, which encompass all types and severities, the prevailing impression is that those with TRD are the most disabled of those with MDD, and hence are those who likely drive these burden findings.7 Kubitz and colleagues showed that the greater occurrence of comorbidities in TRD patients adds to the higher utilization of medical resources, hence the greater economic burden associated with TRD.11
Greenberg and colleagues also showed that individuals with a diagnosis of TRD used more than twice as many medical services compared with depressed claimants without TRD (non-TRD) and incurred significantly greater work-loss costs.12 Given the working age of the population affected by TRD and the functional impairment thought to be associated with the condition, there is significant impact on work functioning.
Other studies illustrate that patients with TRD contribute a disproportionately high burden of illness compared with patients with MDD who respond to treatment.5
At this time, no Canadian study has attempted to quantify the costs associated with TRD from a private-payer perspective, namely, medication utilization and employment changes. The purpose of this pilot study is to begin to understand the burden of TRD in Canada from a private-payer perspective, which includes costs related to prescription drugs and potential disability benefits for employees, specifically. Covered family members are not included in the analysis at this time.
An employer-sponsored benefits plan database was used to define a cohort of employee claimants with depression, both with and without TRD. The database contained claims of 8 national employers across 3 benefit lines: prescription medications, short-term disability, and long-term disability.
The claims data were assessed over a 3-year study period. There was a 6-month screening period to exclude employee claimants who were no longer covered or who no longer had a prescription claim for a medication with an indication for depression during this screening period. This excluded employee claimants not meeting eligibility criteria, and analyses were performed on the 2011-2012 data.
Since diagnosis information was not available from the benefits plan database, 2011 and 2012 prescription claim data were analyzed to generate a “depressed” employee cohort. Claimants not receiving a drug with an indication for depression were excluded.
Claimants who received any drug indicated for depression were included in the cohort, unless they received a claim for antipsychotics prior to receiving antidepressant therapy, a combination frequently used for the depressed mood component of other psychotic disorders. Claimants who received lithium monotherapy were excluded since lithium is often used to treat or prevent bipolar depression.
Claimants were also excluded if they were receiving monotherapy with tricyclic antidepressants (TCAs) as their first and only antidepressant exposure. Since TCAs are considered third-line therapy for depression treatment, we hypothesized that most claimants receiving monotherapy with these agents, with no previous antidepressant exposure, are more likely being treated for sleep and/or pain disorders.2
Claimants were also excluded if they were being treated with trazodone in doses less than 100 mg daily, as this dose is subtherapeutic to treat depression, and more likely used to treat insomnia. Claimants meeting the exclusion criteria are referred to as “all other employees” (ie, non-depressed).
The “depressed” employee cohort was further divided into the TRD and non-TRD groups. Using the definition proposed by McIntyre and colleagues, TRD claimants were defined as any of the following: (a) individuals on at least their 3rd antidepressant monotherapy trial; (b) individuals receiving antidepressant augmentation therapy (antidepressant plus one of these four options: atypical antipsychotic or lithium at a daily dose ≤600 mg; or thyroid hormone; or buspirone; or modafinil); (c) individuals on combination antidepressant therapy (2 or more antidepressants at the same time for at least 100 days); (d) individuals receiving monotherapy with phenelzine and/or tranylcypromine; (e) individuals on quetiapine monotherapy with a history of anti-depressant use in the dose range of 150 mg to 300 mg. Non-TRD represented the remaining depressed patients.3
The costs of prescription medication utilization, short-term disability (STD), and long-term disability (LTD) benefits for employee claimants were calculated (2011 and 2012 $CAN) for both the TRD and non-TRD groups. Drug costs were calculated using the total amount eligible per claim (ie, up to a maximum eligible mark-up on ingredient costs and maximum eligible dispensing fee as determined during claim adjudication), which includes amount paid by the plan, the claimant (deductibles, co-payments), or other plans through coordination of benefits.
Drug costs were stratified into depression drugs and non-depression drugs. This cost did not include over-the-counter medications or other medications not adjudicated through the drug plan. Disability costs for TRD and non-TRD were based on STD and LTD claim costs, which represented the employer payouts and are a product of individual plan design (the level of payout may not be consistent across employers).
Descriptive statistics were used to characterize the cohort of employee claimants, as well as utilization and costs for employees. Data were analyzed by year of claim.
There were 55,324 and 61,028 employee claimants in 2011 and 2012, respectively. Nearly 10% of employee claimants were treated for depression each year with approximately 1.3% classified as TRD claimants, based on their medication claims history (717 and 798 defined as TRD claimants in 2011 and 2012, respectively). In comparison, 4744 (8.6%) and 5137 (8.4%) were defined as non-TRD claimants in 2011 and 2012, respectively. Figure 1 is the consort diagram for employee claimants in 2012.
The mean ages of employee claimants with treated depression (non-TRD and TRD claimants) were approximately 43.4 and 43.3 years, and 57.5% and 58.1% were male, in 2011 and 2012, respectively (Tables 1 and 2).
The average annual per capita drug plan costs for the non-TRD and TRD employee claimant groups are summarized in Table 3. Total prescription drug expenditures were more than 2 times higher in employees with TRD when compared with those in the non-TRD depressed cohort. Furthermore, per capita drug plan expenditures were nearly 4 times higher in the TRD employee cohort compared with the non-depressed employee population (eAppendix Figure, available at www.ajpb.com).
A higher proportion of employees defined as TRD made an STD and/or LTD claim compared with the non-TRD depressed employees in 2011 and 2012. In 2012, 5.0% and 4.1% of all TRD employees had STD and LTD claims for depression, respectively, compared with 1.0% and 0.3% of non-TRD depressed employees (Table 3).
The number of days lost per STD claim per year and mean annual costs paid for disability were also higher in the TRD cohort compared with the non-TRD cohort (Table 3). In 2012, TRD employees had longer STD claim durations than did non-TRD depressed employees, losing an average of 108.7 days per STD claim compared with 76.7 days for non-TRD depressed employees.
Considering the overall burden of illness (BOI) of depression, the combined costs of prescription drugs, STD claims, and LTD claims are more than 4 times higher in the TRD cohort compared with the non-TRD depressed cohort. In 2012, the average per capita plan spending was $1762.48 for TRD employees versus $375.18 for non-TRD depressed employees (Table 4).
The change in per capita plan spending from 2011 to 2012 was driven by disability claims. In 2012, disability claims represented only 22% of the overall BOI for non-TRD depressed employees versus 55% of the overall BOI for TRD employees (Table 4).
This is the first study to attempt to examine the cost impact of TRD on private payers based on employee claimants. Covered family members are not included in this analysis. This is also the first study of its kind to explore cost differences between individuals with TRD versus non-TRD.
Of the many therapeutic options, antidepressant therapy remains the most common initial choice as it is the most studied and the best-evidenced treatment for MDD.2 Other areas of active research in the treatment of MDD include the use of ketamine, an N-methyl-D-aspartate receptor antagonist that has been shown to have rapid antidepressant effects in study patients; and deep brain stimulation, an experimental intervention shown to be effective in patients with TRD.2,13
While the number of treatment options for depression is significant, the mainstay of treatment remains pharmacotherapy due to patient preference, prescriber comfort, and availability/access issues with many nonpharmacological modalities. The main focus of this paper is not the treatment of MDD or TRD, but rather the impact of MDD versus TRD from a private-payer perspective.
In this private-payer analysis, per capita spending for prescription drug, STD, and LTD claims were as great as 4 times higher in TRD employee claimants versus non-TRD depressed claimants. The change in per capita plan spending from 2011 to 2012 was driven by disability claims (representing 22% of total costs in non-TRD depressed employees and 55% in TRD employees in 2012).
These data are consistent with a number of studies that evaluated the impact of depression on employment and lost productivity.14 Our work showed that there are employment differences associated with TRD. Literature shows that even minor levels of depressive symptoms have been associated with decreased work functioning.15 Severe depression was associated with decreased probability of working, and patients had more days off work due to their illness.14-17
Although we had a predominantly male claimant cohort based on the plans included in the analysis, there is some evidence that income loss may be gender related.18 Sixty percent of the claimant pool was employed in the construction and labor sectors, which have higher male distributions. Seventy percent of the employee claimants were male. Female claimants represent about 30% of this employee claimant sample (Table 1).
Further examination of the claimants, including family members of employees, may provide additional information on the prevalence of TRD, which could allow for closer alignment with the literature. When our sample was compared with the Canadian workforce, overall, here almost half of women were employed in the labor workforce (2014) and were employed in sales and service, business, finance, and administration occupations. In contrast, men were most likely employed in trades; as transport and equipment operators and in related occupations; in sales and service occupations; and in management occupations.19
Our study is subject to limitations. First, as previously noted, no consensus definition of TRD exists.3,4. Several approaches to staging TRD have been developed, including the Thase-Rush, Massachusetts General Hospital, European Staging, and Maudsley models; however, no single system has been universally adopted by clinicians and/or researchers.3,4
Second, we used antidepressant utilization to identify our cohorts since antidepressant medications are the mainstay of treatment for MDD, and antidepressant augmentation and/or combination therapy is the mainstay for TRD.4 The definition of the cohorts for this study was based on prescription drug claims data as the sole mechanism to categorize patients into non-TRD versus TRD groups.
This classification method was selected since there are limited diagnostic codes within the private-payer dataset. Specifically, there are no diagnostic classifications until an individual’s disability claim is accepted. This poses significant challenges when defining groups, as in psychiatry it is common to use antidepressants and other psychotropic medications to treat multiple conditions; hence, prescription drug claims alone cannot always accurately imply a diagnosis. For example, antidepressant monotherapy is used as first-line treatment for anxiety disorders and numerous pain disorders.
Thus, it is plausible that some claimants in the sample were inaccurately categorized into the non-TRD group. However, our estimates of prevalence are in line with those of other studies.3 Another limitation of defining cohorts based on medication use is that claimants being treated with non—medication-based treatments (ie, psychotherapy or electroconvulsive therapy) may have been missed. Also, claims for prescription medication do not confirm that individuals were adherent with the medication. Hence, it is plausible that individuals were non-adherent, which may impact changes in therapy and progression to disability.
Other study limitations include: (a) the study does not capture casual absenteeism or presenteeism, only private-payer expenditures, as absenteeism and presenteeism records would need to be captured by the employers and these data are not readily available; (b) this analysis was cross-sectional in nature, namely each claimant cohort was distinct. Claimants may or may not be in each of the years; (c) the 2-year time horizon of this study is quite short, meaning this work did not allow for longitudinal analyses of claimants or examine the impact of time on expenditures; (d) our analysis was limited to a subset of employee data with 3 benefit lines, which means we only reflected a limited insured population, which may not be representative of a general treated population. Furthermore, the number of disability claims was low, so the results may not be generalizable to the overall treated population; (e) because of male predominance in the overall employee claimant cohort, these results may not be generalizable to the entire private payer market; (f) we did not have enough data to evaluate with comparative statistics or time-based statistics; (g) the overall sample size for the cohort was small; and (h) we did not have information on the primary reason for receiving STD and LTD. As this was a preliminary analysis, only descriptive analyses were performed.
Finally, our results do not represent the total cost burden to employers or employees, which may include casual absenteeism, allied health professional activities, other treatment procedures, and family claimants.
Our work shows advantages and disadvantages of using private claims data to understand burden of diseases, specifically, in this case, burden of TRD to employers. In terms of advantages, we fill some of the gaps in the knowledge about private-payer expenditures. Having access to more treatment data from another funding source would improve the understanding of multifaceted treatment care and provide a fuller evaluation of utilization across this disease and others, should the methods be extrapolated to other diseases.
Moreover, other benefits (eg, allied health personnel) covered by private insurance could provide further analyses. In term of disadvantages, it is important to highlight that the limitations outlined above (eg, definitions, timing, representativeness) are all typical of analyses conducted with claims data. Future research will build on this preliminary work in order to improve methods and representativeness of the data.
This work has generated the construction of an initial cohort of individuals with TRD in the private-payer space. We have shown that TRD is associated with higher costs than non-TRD. Future work is needed to examine the utility and flexibility of private-payer data for analyses of burden of disease in the nonpublic space with the possibility to link to outcomes databases as well.
The authors thank collaborating members at The Great West Life Assurance Company for providing their transactional-level claims dataset in support of this project. They also thank Kevin O’Connor, director of Federal and Private Healthcare at Janssen, for his efforts in data acquisition, study planning, and data review.
Author Affiliations: The Centre for Addiction and Mental Health (JK), Toronto, ON; Leslie Dan Faculty of Pharmacy, University of Toronto (JK), ON; Cubic Health Inc (CVH, JZ), Toronto, ON; Janssen Inc (BK, AL), Toronto, ON; Sunnybrook Health Research Institute, Toronto, ON (NM)
Source of Funding: Unrestricted funding from Janssen Inc.
Author Disclosures: Authors received funding from Janssen to support the analysis.
Authorship Information: Concept and design (JK, CVH, JZ, BK, AL, NM); acquisition of data (CVH, NM); analysis and interpretation of data (JK, CVH, JZ, BK, AL, NM); drafting of the manuscript (JK, NM); critical revision of the manuscript for important intellectual content (JK, CVH, JZ, BK, AL, NM); statistical analysis (JZ, NM); obtaining funding (BK); administrative, technical, or logistic support (BK, AL, NM), supervision (CVH, BK, NM)
Address correspondence to: Aileen Li, BSc Phm, Janssen Inc, 19 Green Belt Dr., Toronto, ON, Canada, M3C 1L9. E-mail: firstname.lastname@example.org.
- Depression: fact sheet. World Health Organization website. http://www.who.int/mediacentre/factsheets/fs369/en/index.html. Published October 2012. Accessed August 11, 2014.
- Kennedy SH, Milev R, Giacobbe P, et al; Canadian Network for Mood and Anxiety Treatments (CANMAT). Canadian Network for Mood and Anxiety Treatments (CANMAT). Clinical guidelines for the management of major depressive disorder in adults. J Affect Disord. 2009;117(suppl 1):S44-S53. doi: 10.1016/j.jad.2009.06.039.
- McIntyre RS, Filteau MJ, Martin L, et al. Treatment-resistant depression: definitions, review of the evidence, and algorithmic approach. J Affect Disord. 2014;156:1-7. doi: 10.1016/j.jad.2013.10.043.
- Holtzheimer PE. Advances in the management of treatment-resistant depression. Focus (Am Psychiatr Publ). 2010;8(4):488-500.
- Mrazek DA, Hornberger JC, Altar CA, Degtiar I. A review of the clinical, economic, and societal burden of treatment-resistant depression: 1996-2013. Psychiatr Serv. 2014;65(8):977-987. doi: 10.1176/appi.ps.201300059.
- Diagnostic criteria and codes. In: American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 5th ed. Arlington, VA: American Psychiatric Association; 2013:155-168.
- World Health Organization. The global burden of disease: 2004 update. http://www.who.int/healthinfo/global_burden_disease/GBD_report_2004update_full.pdf. Published 2008. Accessed August 11, 2014.
- Greden JF. The burden of disease for treatment-resistant depression. J Clin Psychiatry. 2001;62(suppl 16):26-31. Review.
- Fava M, Davidson KG. Definition and epidemiology of treatment-resistant depression. Psychiatr Clin North Am. 1996;19(2):179-200.
- 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(9):1561-1568.
- Kubitz N, Mehra M, Potluri RC, Garg N, Cossrow N. Characterization of treatment resistant depression episodes in a cohort of patients from a US commercial claims database. http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0076882. Published October 18, 2013. Accessed August 11, 2014.
- Greenberg P, Corey-Lisle PK, Birnbaum H, Marynchenko M, Claxton A. Economic implications of treatment-resistant depression among employees. Pharmacoecomonics. 2004;22(6):363-373.
- Murrough JW, Iosifescu DV, Chang LC, et al. Antidepressant efficacy of ketamine in treatment-resistant major depression: a two-site randomized controlled trial. Am J Psychiatry. 2013;170(10):1134-1142. doi: 10.1176/appi.ajp.2013.13030392.
- Greenburg PE, Stiglin LE, Finkelstein SN, Berndt ER. The economic burden of depression in 1990. J Clin Psychiatry. 1993;54(11):405-418.
- Beck A, Crain AL, Solberg LI, et al. Severity of depression and magnitude of productivity loss. Ann Fam Med. 2011;9(4):305-311. doi: 10.1370/afm.1260.
- Simon GE, Ludman EJ, Unützer J, Operskalski BH, Bauer MS. Severity of mood symptoms and work productivity in people treated for bipolar disorder. Bipolar Disord. 2008;10(6):718-725. doi: 10.1111/j.1399-5618.2008.00581.x.
- Backenstrass M, Frank A, Joest K, Hingmann S, Mundt C, Kronmüller KT. A comparative study of nonspecific depressive symptoms and minor depression regarding functional impairment and associated characteristics in primary care. Compr Psychiatry. 2006;47(1):35-41.
- Marcotte DE, Wilcox-Gok V, Redmon DP. The labor market effects of mental illness: the case of affective disorders. In: Salkever D, Sorkin A. The economics of disability. Amsterdam, Netherlands: JAI Press, 2000:181-210.
- Statistics Canada. Figure 3. Distribution of the employed population by major occupational groups, men and women aged 15 years and over, May 2011.
- https://www12.statcan.gc.ca/nhs-enm/2011/as-sa/99-012-x/2011002/c-g/c-g03-eng.cfm. Published May 2011. Accessed August 11, 2014.