Economic burden and antidepressant treatment patterns among patients with major depressive disorder in the United States

Abstract

BACKGROUND: The prevalence of major depressive disorder (MDD) continues to rise year over year, resulting in significant economic implications. However, when patients are treated with contemporary standard-of-care antidepressant pharmacotherapies, a suboptimal response is often attained, resulting in frequent treatment changes.

OBJECTIVE: To compare health care resource utilization (HCRU) and all-cause medical and pharmacy costs between commercially insured patients with an MDD diagnosis and matched non-MDD patients and explore treatment patterns among patients with MDD initiating antidepressant pharmacotherapy.

METHODS: This was a retrospective, observational analysis of IBM MarketScan US commercial claims data. Adults aged 18 years and older with continuous enrollment 12 or more months before and after the patient’s first MDD diagnosis from 2017 to 2018 were included in the analysis. HCRU and all-cause medical and pharmacy costs were compared among patients with MDD and a 1:1 exact-matched cohort of non-MDD patients during the same period (Objective 1). Treatment patterns (persistence, discontinuation, switch, combination, and augmentation) were analyzed for patients with MDD starting first-line antidepressant monotherapy for up to 12 months following their antidepressant initiation index date (Objective 2). Time to first treatment change or discontinuation was calculated and treatment patterns were graphically displayed in Sankey diagrams.

RESULTS: 625,272 patients with MDD were matched 1:1 to a cohort of non-MDD patients in Objective 1. Patients with MDD had statistically significantly greater all-cause medical (20.4 vs 9.4; P < 0.0001), outpatient (19.5 vs 9.0; P < 0.0001), emergency department (0.51 vs 0.23; P < 0.0001), inpatient (0.35 vs 0.11; P < 0.0001), and any mental health–related (7.7 vs 0.58; P < 0.0001) visits compared with non-MDD patients. Mean all-cause medical costs were $6,809 (P < 0.0001) higher among patients with MDD than among patients without MDD ($13,183 vs $6,374, respectively). In Objective 2, 44,485 patients with MDD who received antidepressant monotherapy as their first-line MDD treatment were examined. Among the first treatment patterns observed following initiation, 19.3% of patients persisted with their first-line therapy, 56.2% discontinued antidepressant therapy, 24.5% experienced a treatment change (switching, adding a second antidepressant, or augmenting their existing therapy). The median days until first treatment change were 65 days for those discontinuing and 47 days for those switching antidepressants. Among the 24.5% of patients with a treatment change, 50.0% experienced another change in therapy within 30 days.

CONCLUSIONS: The HCRU and costs accrued for patients with MDD is significantly greater than those for non-MDD patients. A large proportion of patients with MDD experienced treatment changes shortly after initiating their first-line antidepressant therapy. The results of this study highlight the need for reevaluation of the current MDD treatment paradigm.

DISCLOSURES: Drs Zhu and Namjoshi are employees of Biogen Inc. and may hold stock. Dr Ferries and Ms Suthoff are employees of Sage Therapeutics, Inc., and may hold stock and/or stock options. Dr Bera has no potential conflicts of interest to disclose. This research was funded by Sage Therapeutics and Biogen. Manuscript editorial services were provided by Boston Strategic Partners, Inc., funded by Sage Therapeutics and Biogen. This work was supported by Sage Therapeutics, Inc., and Biogen. The authors had full editorial control of the manuscript and provided final approval on all content.

Plain language summary

In this study, patients with major depressive disorder (MDD) had higher medical costs and utilization of health care resources than patients without a diagnosis of MDD. This study also examined the contemporary MDD standard-of-care treatment patterns. High rates of patients stopping and switching, as well as patterns of combining and adding additional therapies, were observed after patients started their antidepressant drug.

Implications for managed care

The results of this study indicate that medical costs and health care resource utilization are much greater for patients with MDD than for those without MDD and identified patterns of potential suboptimal treatment and prescribing patterns among patients treated with contemporary standard-of-care antidepressant therapies. This disparate economic burden calls for reevaluation of the contemporary MDD treatment paradigm and rethinking access to care policies.

Since the onset of the COVID-19 pandemic, the burden of major depressive disorder (MDD) and major depressive episodes has grown drastically in the United States and globally.^1-3 As of 2020, approximately 21 million adults in the United States reported experiencing at least 1 episode of depression within the past year.⁴ The impact of MDD has been disproportionate among young adults, with the prevalence of MDD among individuals between the ages of 18 and 34 years growing 54% from 2010 to 2018.⁵ Although the prevalence of MDD was already on the rise prior to 2020, the COVID-19 pandemic and accompanying stressors inflicted an exponential impact on the mental health of individuals around the world.⁶ As the second leading cause of disability in the United States, MDD contributes to a significant societal, clinical, and economic burden nationally.^6-10 Although the treatment of MDD has been associated with high direct medical costs, the accompanying indirect and workplace costs have also increased substantially over time.¹¹ A 2021 analysis from Greenburg et al found an approximately 37.9% increase in the incremental cost burden of MDD due to direct and indirect costs from 2010 to 2018.⁵

The American Psychiatric Association guidelines for treating MDD recommend psychotherapy and/or antidepressant medication as initial treatment and continuing antidepressant therapy for at least 4-9 months after remission.¹² Selective serotonin reuptake inhibitors (SSRIs) and serotonin-norepinephrine reuptake inhibitors (SNRIs) are generally recommended as first-line pharmacotherapy treatment options.¹³ However, MDD is generally considered an undertreated condition.^14,15 For patients who experience a major depressive episode, it is estimated that less than half will receive adequate or any pharmacotherapy.^16,17 Among those who do receive pharmacotherapy treatment with antidepressants, a suboptimal response is often attained, as studies have shown that approximately only one-third of patients experience remission in their depressive symptoms after using a first-line treatment.¹⁸ Contemporary existing antidepressant treatment options have been associated with delayed onset of response (ie, 4-8 weeks on average^12,19), low remission rates,^18,20 poor medication adherence and persistence,²¹ and high rates of side effects or adverse events,^22,23 all of which may result in poor treatment outcomes. Moreover, many patients who are treated with antidepressant pharmacotherapy experience a “trial and error” approach to treatment options, with high rates of treatment changes (ie, discontinuation, switching therapies, combining therapies, or augmenting),^24,25 highlighting a substantial unmet need for efficacious treatments. Uncontrolled or unresolved depression has been shown to increase the risk of developing or worsening comorbid conditions (eg, diabetes, coronary heart disease, and alcohol use disorders) and make chronic disease management more difficult.^26-31 Among patients with MDD, the incremental economic burden attributable to comorbid conditions accounts for the largest portion of the growing economic burden of MDD in the United States.⁵

To analyze the growing, substantial burden of MDD, the present study compared health care resource utilization (HCRU) and costs between commercially insured patients with an MDD diagnosis and matched patients without an MDD diagnosis from 2016 through 2019 (Objective 1). Given the growing clinical and economic burden of MDD in the United States, it is important to examine contemporary MDD standard-of-care prescribing patterns. Accordingly, this study also explored treatment patterns among patients with MDD initiating antidepressant pharmacotherapy (Objective 2). A prior retrospective assessment of claims by Gauthier et al evaluated treatment patterns in MDD and associated economic burden between 2003 and 2014.³² This study seeks to expand these findings by analyzing existing clinical practice treatment patterns as recent as 2019, as well as further investigating treatment patterns among patients with a treatment change following initiation of antidepressant monotherapy and graphically displaying the frequency of treatment changes.

Methods

DATA SOURCE

The study was conducted using real-world administrative claims data in the IBM MarketScan US claims database from January 1, 2016, to December 31, 2019. The MarketScan commercial database contains deidentified datasets that capture individual-level health care administrative claims for millions of commercially insured patients, including health care utilization data such as inpatient, outpatient, and prescription drug services (claims for mail order prescriptions and specialty pharmacies are included).³³ The commercial data are submitted to IBM by employer-sponsored private health insurance plans and contain employer-sponsored private health insurance data of more than 200 million employees, spouses, and dependents.

STUDY DESIGN

To establish baseline prevalence rates, for each year between January 1, 2016, and December 31, 2019, the number of patients with an International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) code diagnosis for MDD (ICD-10 codes: F32.0-F32.9, F33.0-F33.4, F33.8, F33.9, F41.2, F34.8, F34.9, F38.0, F38.1, F38.8, and F99) were identified among commercially insured patients aged 18 years and older. Patients included in the prevalence rates were required to have at least 1 month of continuous enrollment in each calendar year. Patients included in the health care resource utilization/cost analysis and treatment patterns analysis were required to have continuous enrollment 12 or more months before and after their index date. To confirm MDD diagnosis, patients in this study were required to have at least 2 ICD-10 claims for MDD in the outpatient (OP) setting on separate days less than 365 days apart or 1 or more ICD-10 claims for MDD in the inpatient (IP) or emergency department (ED) setting in the 12 months prior to or on the index date. Patients with MDD were excluded if a diagnosis for bipolar disorder, schizophrenia/schizoaffective disorder, or psychosis was identified in the claims data prior to or proceeding the index date (defined below).

OBJECTIVE 1

Design and Patient Cohorts. This study assessed Objective 1 using a retrospective, observational, matched cohort design that compared HCRU and health care costs between patients diagnosed with MDD and patients without an MDD diagnosis (controls) from January 1, 2017, to December 31, 2018. For the MDD cohort, the index date was defined as the first observed MDD diagnosis claim during the cohort identification period. A cohort of commercially insured, non-MDD controls were exact matched to MDD cases on a 1:1 basis based on the following factors: age, sex, Charlson Comorbidity Index (CCI) score category (score categories were classified as 1, 2, 3, and 4+), and geographic region at index date of the matched MDD case. Control patients’ index dates were randomly assigned dates in the same month and year as the index date for the matched case.

Outcomes Measures. Baseline descriptive demographic characteristics of patients with MDD and matched controls in the 12 months prior to the index date were evaluated. Mean, median and SDs, where appropriate, per-patient-per-year all-cause HCRU (claims) for IP admissions, ED visits, OP visits, and mental health-related HCRU and health care costs (total medical, IP, ED, OP, and pharmacy costs) in the 12 months post-index were evaluated during the follow-up period and compared between the matched cohorts. Health care costs were adjusted to 2020 US dollars using the medical care index of the Consumer Price Index.

OBJECTIVE 2

Design and Patient Cohort. Patients were included in the analysis for Objective 2 if they had 1 or more pharmacy claims for an antidepressant and an ICD-10 diagnosis for MDD prior to or on the index date. The index date was defined as the first observed antidepressant prescription fill date during the cohort identification period (January 1, 2017, to December 31, 2018). Patients who received any antidepressant prescription in the 6 months prior to their index date were excluded. Only patients who received antidepressant monotherapy as their first-line, index-date treatment were included in the analysis. Patients were excluded if their index antidepressant treatment was inconsistent with guideline-recommended first-line treatment (ie, combination of 2+ antidepressants, atypical antipsychotic, buspirone, amphetamine, modafinil, dopamine agonist [amantadine, pramipexole, ropinirole], anticonvulsants [lamotrigine, gabapentin, pregabalin], electroconvulsive therapy, and transcranial magnetic stimulation).¹³

Measurements. Cohort baseline demographic characteristics and antidepressant pharmacotherapy class initiated at index (SSRI, SNRI, atypical antidepressant [bupropion, maprotiline, mirtazapine, nefazodone, trazodone, vilazodone, and vortioxetine], or other) were reported using descriptive statistics. Treatment patterns following antidepressant therapy initiation were examined during the 12-month follow-up period and were classified into 5 utilization categories: (1) Discontinuation; (2) Switch; (3) Combination; (4) Augmentation; and (5) Persistence. The definitions for each treatment change are included in Table 1 and were based on descriptions outlined in Gauthier et al.²⁴ Patients were identified as persistent if no treatment changes were observed for the patient from the index date through the end of the follow-up period. The treatment patterns for patients with an observed treatment change following antidepressant initiation (discontinuation, switching, combining, or augmenting their antidepressant therapy) were further evaluated. Patients identified as having switched to a new antidepressant regimen were further analyzed to examine the drug regimens initiated in the 365 days following their first switch. Time (days) from the index treatment to the first treatment change or discontinuation were calculated and treatment patterns were visually displayed in Sankey diagrams, reflecting drug regimens at 90 days, 180 days, and 365 days post-index (for patients with treatment change or discontinuation) or post-first-switch (for patients who switched their treatment regimen). Sankey diagrams depict color nodes as treatments and the width of each link represents the number of patients transferred from one treatment to another. This diagram enables users to easily perceive the treatment transition trends over time.

TABLE 1.

Pharmacotherapy Treatment Pattern Definitions

Pharmacotherapy change	Definition
Discontinuation	Interruption of ≥ 42 consecutive days of the drug regimen initiated at the index treatment date24 Date of discontinuation is defined as the date of the discontinued medication’s last fill’s days supply
Switcha	Initiation of a new drug regimen defined by drug ingredient (antidepressant and/or atypical antipsychotic) within 42 days of the discontinuation of the drug regimen initiated at the index treatment date Switching from a branded to generic product with the same ingredient will not be defined as a switch
Combinationa	Treatment add-on resulting in the use of ≥ 2 antidepressants simultaneously
Augmentation	Treatment add-on resulting in the use of 1 antidepressant and an atypical antipsychotic simultaneously
Persistence	Absence of any of the treatment changes until the end of the follow-up period

^aPatients on combination therapy at the beginning of a line of therapy (post-index) who discontinued one treatment but remained on the other treatment were classified as having switched to a new drug regimen. Patients on combination therapy at index were not included in this analysis.

Results

Between January 1, 2016, and December 31, 2019, the overall prevalence rate of adult patients with an MDD diagnosis in the IBM MarketScan Commercial database increased from 3.8% (773,175/20,585,483) to 5.1% (951,654/18,521,821). Patients aged 18-34 years experienced the greatest increase in MDD diagnoses, with the prevalence rate increasing from 2.97% to 4.68% from 2016 to 2019.

OBJECTIVE 1

A total of 625,272 patients with MDD met all inclusion and exclusion criteria for the HCRU and costs comparison between January 1, 2017, and December 31, 2018, and were matched to patients in the non-MDD cohort. The mean (SD) age of the matched cohorts was 43 (13.5) years, 43.9% resided in the South, and 69.7% of the population was female (Table 2). Patients with MDD and the matched non-MDD cohort on average had a CCI score of less than 1 (0.52 ± 1.1 and 0.51 ± 1.1, respectively). The most prevalent comorbidities recorded in the MDD and non-MDD matched cohorts were chronic obstructive pulmonary disease (10.5% and 11.9%, respectively) and diabetes (10.6% and 10.0%, respectively).

TABLE 2.

Matched Cohort Baseline Descriptive Statistics

	MDD cohort (N = 625,272)	Matched non-MDD controls (N = 625,272)
Age at index date, mean ± SD [median], years	43.4 ± 13.5 [46]	43.4 ± 13.5 [46]
Sex, n (%)
Female	435,715 (69.7)	435,715 (69.7)
Male	189,557 (30.3)	189,557 (30.3)
Region, n (%)
North Central	147,204 (23.5)	147,204 (23.5)
Northeast	109,644 (17.5)	109,644 (17.5)
South	274,461 (43.9)	274,461 (43.9)
West	93,943 (15.0)	93,943 (15.0)
Baseline CCI score, mean ± SD [median]a	0.52 ± 1.1 [0]	0.51 ± 1.1 [0]
CCI category, n (%)
0	445,069 (71.2)	445,069 (71.2)
1	109,004 (17.4)	109,004 (17.4)
2	40,894 (6.5)	40,894 (6.5)
3	13,924 (2.2)	13,924 (2.2)
4+	16,381 (2.6)	16,381 (2.6)
CCI comorbidities, %b
Cancer	4.2	3.4
Cerebrovascular disease	1.6	2.0
Congestive heart failure	1.2	1.3
Chronic obstructive pulmonary disease	10.5	11.9
Diabetes	10.6	10.0
Diabetes with complications	2.5	2.5
Mild liver disease	3.1	3.5
Peripheral vascular disease	1.7	1.8
Renal disease	1.8	1.5
Rheumatologic disease	1.8	2.1

^a CCI score is a method of categorizing medical comorbidities of patients based on the International Classification of Diseases diagnosis codes found in administrative data.⁴⁴

^b Only those conditions with prevalence ≥ 1% are included in the table.

CCI = Charlson Comorbidity Index; MDD = major depressive disorder.

Over the course of the 12-month follow-up period, on average, patients with MDD experienced statistically significant greater total all-cause medical (20.4 vs 9.4; P < 0.0001), OP (19.5 vs 9.0; P < 0.0001), ED (0.51 vs 0.23; P < 0.0001), IP (0.35 vs 0.11; P < 0.0001), and any mental health–related (7.7 vs 0.58; P < 0.0001) visits compared with matched non-MDD patients (Table 3).

TABLE 3.

HCRU and Health Care Costs in an MDD Cohort Compared With Non-MDD Controls

	MDD cohort (N = 625,272)	Matched non-MDD controls (N = 625,272)	Differencea	P value
All-cause HCRU, mean ± SD [median]	20.4 ± 20.2 [14]	9.4 ± 12.8 [6]	11.0	<0.0001
Outpatient visits	19.5 ± 19.4 [14]	9.1 ± 12.4 [5]	10.4	<0.0001
Emergency department visits	0.51 ± 1.5 [0]	0.23 ± 0.77 [0]	0.28	<0.0001
Inpatient admissions	0.35 ± 1.7 [0]	0.11 ± 0.90 [0]	0.24	<0.0001
Number of prescription medications, mean ± SD [median]	14.8 ± 14.2 [11]	6.4 ± 8.5 [4]	8.4	<0.0001
Mental health care–related visits, mean ± SD [median]b	7.7 ± 12.3 [3]	0.58 ± 3.3 [0]	7.13	<0.0001
All-cause medical costs, mean ± SD [median] (2020)	13,183 ± 41,114 [3,631]	6,374 ± 28,722 [1,207]	6,809	<0.0001
Outpatient visit costs	8,277 ± 24,927 [2,886]	4,308 ± 20,347 [1,039]	3,969	<0.0001
Emergency department costs	1,128 ± 4,850 [0]	523 ± 3,504 [0]	605	<0.0001
Inpatient visit costs	3,778 ± 25,747 [0]	1,543 ± 15,995 [0]	2,235	<0.0001
Mental health–related medical costs, mean ± SD [median]	4,139 ± 17,762 [534]	386 ± 5,451 [0]	3,753	<0.0001
Total prescription medication costs, mean ± SD [median]	3,553 ± 13,884 [578]	2,006 ± 10,258 [121]	1,547	<0.0001

^a Difference = utilization in MDD cohort – utilization in matched non-MDD controls.

^b Any mental health–related visit (not MDD specific).

HCRU = health care resource utilization; MDD = major depressive disorder.

Overall, mean total all-cause medical costs were $6,809 higher ($13,183 vs $6,374; P < 0.0001) and total mental health-related medical costs were $3,753 higher ($4,139 vs $386; P < 0.0001) in patients with MDD compared with controls. These differences were largely driven by higher outpatient costs (difference: $3,969; P < 0.0001) and mental health-related inpatient costs (difference: $1,783; P < 0.0001).

OBJECTIVE 2

There were 44,485 patients with MDD identified who met inclusion and exclusion criteria and who received antidepressant monotherapy as their first-line MDD treatment (Figure 1). This cohort of patients, on average, was aged 38.9 years, was predominately female (67.1%), and resided in the Northeast/North Central or South (43.4% and 40.1%, respectively). Most patients initiated their antidepressant monotherapy with an SSRI (n = 27,972; 62.9%), followed by atypical antidepressants (n = 10,208; 22.9%), SNRIs (n = 5,013; 11.3%), and other antidepressants (n = 1,292; 2.9%; other included monoamine oxidase inhibitors [MAOIs] and tricyclic antidepressants).

FIGURE 1.

MDD Antidepressant Treatment Cohort Attrition Flowchart

Of the 44,485 patients with MDD initiating index antidepressant monotherapy, 19.3% (n = 8,578) persisted with their first-line therapy through the end of the follow-up period (Figure 2). Patients using SNRIs as their first-line antidepressant had a slightly higher rate of persistence (21.1% persistent, n = 1,058), followed by SSRI initiators (20.6% persistent, n = 5,755). Among the first treatment changes observed (discontinue, switch, combine, or augment cohorts; first change occurring at any time frame within the 12-month follow-up) following initiation of first-line therapy for patients with a treatment change, more than half of patients (n = 24,994, 56.2%) discontinued their initial therapy regimen, and 24.5% (n = 10,913) experienced a treatment change of switching to a different antidepressant (n = 5,301, 11.9%), adding a second antidepressant to their regimen (n = 4,405, 9.9%), or augmenting their existing therapy with an additional atypical antipsychotic (n=1,207, 2.7%) (Please see Table 1 for cohort definitions). The median days until first treatment change were as follows: 65 days for those patients discontinuing their index therapy (no further record of antidepressant use or interruption of ≥42 consecutive days before initiating another antidepressant); 47 days for those patients switching to a different antidepressant; 63 days for those patients adding on an additional therapy (use of a combination of ≥2 antidepressants simultaneously); and 29 days for those augmenting their regimen with an additional atypical antipsychotic.

FIGURE 2.

Patterns of Antidepressant Therapy Utilization in First Year of Treatment for Patients With a Treatment Change

Among the cohort of patients with a treatment change following antidepressant initiation (cohorts that switched, combined, or augmented their pharmacotherapy), 50.0% experienced another change in therapy within 30 days. Within the cohort of patients who discontinued their first-line pharmacotherapy (n = 24,994), 20.1% (n = 5,035) resumed use of an antidepressant therapy within 3 months and 33.3% (n = 8,318) within 6 months, with a median time to restarting pharmacotherapy of 92 days. By the end of the 1-year follow-up period, 74.7% of the total patients initiating an antidepressant (n = 33,217 of 44,485) had a treatment change, with most of the treatment change cohort (n = 24,772) discontinuing antidepressant pharmacotherapy altogether.

Within the cohort of patients who had a treatment change after initiating their first-line antidepressant pharmacotherapy (discontinuation, switch, combination, or augmentation; n = 35,907 of 44,485), changes were prevalent across all classes of antidepressants, and 53.8% had no record of being on an antidepressant 90 days after the index-treatment initiation. Among patients initiating an SSRI as their first-line monotherapy treatment who had a treatment change (n = 22,217 of 27,972), only 36.7% (n = 8,152) were still using an SSRI 90 days later, whereas the other 53.0% (n = 11,783) of patients had discontinued therapy and 10.3% experienced a treatment change (switch, combination, or augment; n = 2,282).

Similarly, only 13.1% (n = 1,125) of patients using atypical antidepressants at index (n = 8,558) were using an atypical antidepressant 1-year post-index. During this 1-year follow-up, 55.0% of this cohort (n = 619) displayed patterns of switching on-and-off atypical antidepressant therapy, whereas 18.0% (n = 202) switched between atypical antidepressants and other antidepressant classes throughout the course of the year.

Within index SNRIs users with a treatment change (n = 3,955 of 5,013), 69.1% (n = 2,733) of patients discontinued using the class within 90 days. Among those discontinuing use of SNRIs within the first 90 days after initiation, 2,113 (77.3%) patients had no pharmacy claim records of being on any antidepressant 90 days after initiation, and a small percentage switched to an SSRI, atypical, or other antidepressant (4.7%, 4.5%, and 13.5%, respectively).

Among patients using MAOIs and tricyclic antidepressants at index who experienced a treatment change (Other category; n = 1,177 of 1,292), 21.8% (n = 256) were still using an MAOI or tricyclic 6 months later, and 16.0% (n = 188) were still using them 1 year later. Almost half of patients (n = 563, 47.8%) discontinued their first-line MAOI or tricyclic within 90 days of initiating therapy and did not resume using any antidepressants over the course of the remaining 1-year follow-up.

A graphical display (Sankey diagram) of these treatment patterns for patients with a treatment change is presented in Figure 2. Sankey diagrams are visualizations that depict patient treatment journeys from therapy initiation through subsequent treatment transitions.

Treatment patterns for the subcohort of patients whose first treatment change post-index was initiating a new drug regimen (switch cohort, n = 5,301) (Figure 1) were further evaluated. Patients in the switch cohort were slightly younger when compared with the rest of the patients with MDD on antidepressant monotherapy at index (age: 36.8 vs 39.2 years; P < 0.001); however, there were no other statistically significant differences in the baseline characteristics. The median time for patients switching to a different antidepressant drug was 47 days, and the most common treatment pattern at 90 days following the patient’s first pharmacotherapy switch was absence of any therapy (n = 2,167, 40.8%). Throughout the 1 year of follow-up after the first treatment switch, continued patterns of cycling on and off treatments and switching to other antidepressants both within the same class and within different classes were observed (Figure 3). One-quarter of patients (24.4%, n = 1,293) switched on and/or off antidepressants within the same class from their index date through the 12-month follow-up period following their first treatment switch; however, most of these patients (66.7%, n = 863) had periods of discontinuation between treatment regimens. SSRIs remained the most common antidepressant class used across all time points. Among those initiating an SSRI at index (n = 3,292, 62.1%), the first treatment switch observed for 45.6% (n = 1,501) of the cohort was to another SSRI and 30.9% (n = 1,017) to an atypical antidepressant.

FIGURE 3.

Treatment Switching Patterns up to 1 Year After First Treatment Change

Discussion

This analysis highlights the extensive HCRU burden that MDD places on the US health care system, as well as the accompanying economic burden that potentially negatively impacts patients, their families, payers, and employers. While the discrepancy in health care utilization and costs between patients with MDD and non-MDD patients may indicate that MDD is suboptimally managed by contemporary standard-of-care pharmacotherapies, the antidepressant treatment pattern analysis affirms there are considerable challenges with existing therapeutic options. This analysis highlights a considerable heterogeneity in antidepressant use and patterns of multiple treatment changes, which may reflect what patients experience in the real world after initiation of antidepressant treatment, including cycling through multiple classes and episodes of discontinuing and resuming pharmacotherapy. These patterns may be indicative of unresolved depression, which may lead to a prolonged episode, functional impairment, and higher rates of relapse due to residual symptoms, all of which contribute to the overall burden of disease.^34-36

Other research has demonstrated that patients experiencing early improvement in their depressive symptoms within the first 2 weeks of symptom onset were more likely to achieve response to their antidepressant and were 6 times more likely to experience remission than patients without early improvements in depressive symptoms within the first 2 weeks of depressive symptom onset.³⁷ Additional research evaluating early treatment response found that shorter duration of depressive episodes and rapid treatment optimization within 2 weeks are associated with better symptomatic and functional outcomes in MDD.³⁸ However, the low prevalence of persistent patients throughout the follow-up period and the frequent treatment changes observed in this commercial patient population in a short time frame following antidepressant initiation may indicate that patients are not resolving their depression symptoms early in the treatment process, are having issues with their required cost sharing, or are experiencing undesirable treatment effects (ie, tolerability issues, side effects, or adverse effects). This research emphasizes the need for rapid resolution of depressive symptoms in a patient with MDD in the critical early weeks of antidepressant treatment initiation. Additionally, as demonstrated from this research, the frequency of treatment changes and the low persistence observed may also warrant rethinking the current chronic treatment paradigm and support the episodic treatment of MDD when appropriate. Additionally, although SSRIs were the most common antidepressant used for first-line therapy in this cohort, the results of this study reflect high rates of treatment changes in the early weeks following antidepressant initiation.

Previous research has found that when more treatment steps are required to treat MDD, patients experience lower acute remission rates¹⁸; however, it is still commonplace for payer policies to require a step therapy protocol for mental health disorders.^39-41 Step therapy protocols, also known as “step edits” or “fail first protocols,” require a patient to try or “fail” a first-line medication (ie, SSRI or SNRI) within a therapeutic class, prior to trying (and receiving coverage for) a second-line agent.⁴² A retrospective cohort study assessment of claims from more than 70,000 commercially insured and Medicare patients examined the economic burden associated with failed antidepressant therapy in the first 12 months following treatment initiation.⁴³ This research found that patients who did not respond to first-line treatment and required additional lines of antidepressant pharmacotherapy incurred significantly higher costs than those who completed treatment with a single line of antidepressant therapy. Many prescribers also follow general prescribing recommendations laid out in national treatment guidelines with recommendations for first-line pharmacotherapy treatment options.^12,13 Given that national treatment guidelines¹³ and step therapy protocols may require patients to initiate first-line antidepressant therapy with an SSRI and the importance of early improvement, it may be time to rethink policies that may impede access by overly generalizing treatment options. The economic burden associated with MDD outlined in this manuscript may warrant less interference between patients and their clinicians in choosing the treatment that they feel is best for the individual.

LIMITATIONS

This study has limitations that should be considered. First, in evaluating HCRU and health care costs, although patients with MDD were matched to non-MDD patients on CCI score and other demographic information, the analysis did not stratify patients by certain factors that were outside the scope of research that may have a variable effect on cost burden, such as MDD disease severity and certain sociodemographic factors such as race and ethnicity. For Objective 2, patients were required to have no previous fills for antidepressants in the 6 months prior to their index date; however, the patient’s experience with antidepressants prior to this 6-month period was unknown. Other factors that may influence the antidepressant the patient receives as first-line therapy and subsequent therapies were not accounted for in this study. Factors not included in this analysis that may influence treatment decisions include the severity of the patient’s MDD (ie, mild, moderate, or severe), past medical history, the patient’s health plan or insurance coverage of medications, or the patient’s prior use of antidepressant therapy more than 6 months prior to their study index date. Additionally, given that this was a claims analysis study, for patients discontinuing therapy throughout the study period, it was not possible to conclude whether the patient was in remission and no longer required antidepressant pharmacotherapy or paid for their prescription with case or discount card; however, it should be noted that most patients discontinuing therapy appeared to do so prior to recommended guidelines. Finally, this analysis was conducted within a commercially insured patient population; therefore, the results of this study may not be generalizable to other populations of patients with MDD.

Conclusions

These real-world claims analyses of patients with MDD emphasize much greater HCRU and health care costs for patients with MDD compared with a similar patient population without MDD. This study also highlights patterns of potential suboptimal treatment and prescribing patterns among patients with MDD who are treated with antidepressant therapies. Importantly, the economic burden associated with MDD calls for a reevaluation of the current MDD treatment paradigm. Further research and innovation in the quality and efficacy of MDD pharmacotherapy treatment options is critical.