Abstract
Aims: The current study examined the association between insufficient major depressive disorder (MDD) care and healthcare resource use (HCRU) and costs among patients with prior myocardial infarction (MI) or stroke.
Methods: This was a retrospective study conducted using the MarketScan Claims Database (2010–2015). The date of the first MI/stroke diagnosis was defined as the cardiovascular disease (CVD) index date and the first date of a subsequent MDD diagnosis was the index MDD date. Adequacy of MDD care was assessed during the 90 days following the index MDD date (profiling period) using 2 measures: dosage adequacy (average fluoxetine equivalent dose of ≥20 mg/day for nonelderly and ≥10 mg/day for elderly patients) and duration adequacy (measured as the proportion of days covered of 80% or higher for all MDD drugs). Study outcomes included all-cause and CVD-related HCRU and costs which were determined from the end of the profiling period until the end of study follow-up. Propensity-score adjusted generalized linear models (GLMs) were used to compare patients receiving adequate versus inadequate MDD care in terms of study outcomes.
Results: Of 1,568 CVD patients who were treated for MDD, 937 (59.8%) were categorized as receiving inadequate MDD care. Results from the GLMs suggested that patients receiving inadequate MDD care had 14% more all-cause hospitalizations, 4% more all-cause outpatient visits, 17% more CVD-related outpatient visits, 13% more CVD-related emergency room (ER) visits, higher per patient per year CVD-related hospitalization costs ($21,485 vs. $17,756), higher all-cause outpatient costs ($2,820 vs. $2,055), and higher CVD-related outpatient costs ($520 vs. $434) compared to patients receiving adequate MDD care.
Limitations: Clinical information such as depression severity and frailty, which are potential predictors of adverse CVD outcomes, could not be ascertained using administrative claims data.
Conclusions: Among post-MI and post-stroke patients, inadequate MDD care was associated with a significantly higher economic burden.
Introduction
Cardiovascular disease (CVD) is the leading cause of mortality in the United States and it is projected that about 40% of the population will be afflicted with some form of CVD by 2030Citation1. Depression is a well-known comorbid condition that coexists with CVDCitation2–4. It has been reported that about a third of stroke survivors experience depressionCitation5,Citation6. Depression among CVD patients can lead to increased mortality, a higher number of hospital admissions, lower medication adherence, poor follow-up with office visits, and overall lower compliance with recommended lifestyle modifications such as diet and exerciseCitation7–10.
The American Heart Association (AHA) reports the high prevalence of major depressive disorder (MDD) in patients with CVD and emphasizes the need for routine screening and treatment of MDD in this populationCitation11. A recent observational study among post-myocardial infarction (MI) and post-stroke patients showed that inadequate MDD treatment was associated with a higher risk of adverse CVD events such as stroke, angina, and a composite CVD outcome (which included stroke, angina, MI, and congestive heart failure), highlighting the importance of judicious treatment of depression among CVD patientsCitation12. In addition to adverse health outcomes, it has been shown that depression among CVD patients may lead to higher CVD-related healthcare costsCitation5,Citation13.
A recent statement by the AHA and American Stroke Association highlighted the importance of studies to evaluate the impact of treating post-stroke depression on subsequent healthcare utilizationCitation6. However, there still remains a lack of knowledge on the impact of MDD treatment on healthcare resource use (HCRU) and cost among CVD patients with comorbid depression. Therefore, the current study aimed to examine the real-world economic burden associated with inadequate MDD care among patients with MDD and a previous diagnosis of MI or stroke (referred to as the CVD cohort).
Methods
This retrospective, observational study analyzed de-identified data from the Truven Health MarketScan administrative claims database from 1 January 2010 to 31 December 2015. This database consists of commercial insurance claims data consolidated from more than 350 payers geographically distributed across the United StatesCitation14. The medical and pharmacy claims are linked to detailed patient and enrollment data, thereby providing patient demographics information, enrollment history, and service level and payment information for inpatient and outpatient treatments and care.
illustrates the study design. This study employed a retrospective, cross-sectional design. The baseline period (12-month period before CVD index date) was used to examine patients' baseline demographic and clinical characteristics. The immediate 3 months following the MDD index date were defined as the profiling period, which was used to ascertain the adequacy of depression treatment received. The reason behind using a 3-month period to assess MDD care adequacy was because the first 3 months post-MDD diagnosis comprise the acute phase of care. The American Psychiatric Association (APA) guidelines recommend treatment for MDD patients, irrespective of severity, during this phase. The period following the acute phase is known as the continuation and maintenance phase of care, during which treatment is recommended based on clinician assessment of response to acute-phase treatmentCitation15. For the current study, the time period following the profiling period until the end of the study was defined as the follow-up period, where the study outcomes (i.e. HCRU and costs) were assessed.
Figure 1. Study design. The date of the first MI/stroke diagnosis was defined as the CVD index date. The 12-month period prior to the CVD index date was defined as the baseline period for identifying patient demographics and clinical comorbidities. Following the CVD index date, the date of the first diagnosis of MDD was defined as the MDD index date. The 3-month period following the MDD index date was defined as the profiling period to evaluate the adequacy of depression care received. The time period following the end of the profiling period until the end of data cut was defined as the follow-up period for identification of the adverse CVD outcomes.
The study population included adult patients (≥18 years old) who were newly diagnosed with MDD and received treatment for their MDD following an initial diagnosis of MI/stroke. Receipt of MDD treatment was determined based on any prescription claims for selective serotonin reuptake inhibitors or serotonin and norepinephrine reuptake inhibitors. Patients were considered to have CVD if they had at least 1 claim with a relevant International Classification of Diseases, Ninth Revision, Clinical Modification diagnosis code for MI or stroke. The CVD index date was defined as the date of the first MI or stroke diagnosis. Likewise, the MDD index date was defined as the date of the first MDD diagnosis following the CVD index date. To identify newly diagnosed CVD and MDD, patients with any claims for CVD, MDD, or MDD medication use in the 12 months prior to the CVD index date (i.e. baseline period), were excluded from the study population. Patients with a claim of bipolar disorder or schizophrenia during any point in the study period were also excluded. Additionally, to ensure completeness of claims records, patients without continuous medical enrollment for at least 12 months before and 6 months after the CVD index date were also excluded.
Approval by an ethics review committee was not required because this study was a non-interventional, retrospective analysis conducted using de-identified insurance claims data. All results were reported only at the group level; no information was collected that could be used for participant identification.
Adequacy of depression treatment was evaluated during the 3-month profiling period based on the following 2 criteria: (1) antidepressant dosage adequacy and (2) antidepressant dosage duration. Patients who fulfilled both criteria received adequate depression treatment, otherwise they were categorized as receiving inadequate depression treatment. Such an algorithm has been previously used to identify patients with inadequate depression treatmentCitation12,Citation16. Antidepressant dosage adequacy was determined based on a patient’s average daily dose of antidepressants prescribed, expressed in fluoxetine-equivalent doses. A fluoxetine-equivalent dose of ≥20 mg/day for patients aged <65 years or ≥10 mg/day for patients aged ≥65 years was considered adequate. This dosage was calculated by first multiplying each antidepressant agent’s strength (in milligrams [mg]) by the quantity of drug dispensed, divided by the number of days supplied, and converting this to fluoxetine-equivalent doses. Following this, the average dose for each antidepressant agent prescribed was summed for each patient and divided by the number of agents prescribed. Antidepressant duration adequacy was assessed based on calculating the proportion of days covered during the 3-month profiling period, defined as the number of days between the first prescription date and the end date of the last prescription when a patient was in possession of the antidepressant. This measure considers early prescription refills and gaps in refills (up to 60 days in this study). A proportion of days covered value ≥80% was indicative of adequate treatment duration.
Study outcomes (i.e. HCRU and healthcare costs) were assessed over the follow-up period. HCRU included hospital admissions, emergency room (ER) visits, and outpatient office visits. We evaluated HCRU in 2 ways: (1) all-cause HCRU and (2) CVD-related HCRU. All-cause HCRU included all visits observed in the follow-up period while CVD-related HCRU only included visits that had a primary or secondary diagnosis code for a CVD event. Healthcare costs were estimated in terms of all-cause and CVD-related healthcare costs. The cost categories included hospital admission (inpatient) costs, ER costs, and outpatient office visits costs.
Descriptive statistics were used to report the baseline demographic and clinical characteristics of the study population as frequencies and percentages for categorical variables and means with standard deviations (SDs) for continuous variables. To account for confounding due to the baseline differences between patients receiving adequate versus inadequate depression treatment, a propensity score was generated predicting the probability of receiving adequate depression care given the baseline characteristics of each patient. This propensity score was then included in the subsequent analytical model to adjust for the baseline differences among patients receiving adequate and inadequate depression care. The c-statistic was used to assess the fit of the propensity score modelCitation17.
HCRU and costs were analyzed descriptively and reported as mean visits and cost per patient per year (PPPY). In addition, propensity-score adjusted generalized linear models with Poisson distribution were used to determine if all-cause and CVD-related HCRU differed between adequately and inadequately treated MDD patients with CVD. Generalized linear models with log link and gamma distribution were conducted for direct all-cause and CVD-related healthcare costs. Separate models were estimated for each cost category. Costs were adjusted for inflation to the 2016 US dollar. All statistical analyses were conducted using SAS 9.4 (SAS Institute Inc., Cary, NC).
Results
The study population comprised 1568 newly diagnosed CVD patients who were subsequently treated for MDD (). The mean age of the patients was 66 years and about half of them were female. Of these patients, 631 (40.2%) received adequate depression treatment and 937 (59.8%) received inadequate depression treatment according to our adequacy definitions. The mean follow-up time for both the adequately treated group and the inadequately treated group was 2 years. Compared to the inadequately treated group, the adequately treated group had fewer hospitalizations (p < .05) and less use of selective serotonin reuptake inhibitors (p < .001) at baseline. Other baseline sociodemographic and clinical characteristics between the groups are presented in . In order to account for these differences, all further analyses controlled for the propensity of receiving adequate MDD care. The c-statistic for the propensity model was 0.78.
Figure 2. Study cohort selection based on inclusion and exclusion criteria.
Table 1. Patient cohort characteristics at baseline.
Unadjusted average HCRU and healthcare costs PPPY during the follow-up period are presented in and , respectively. In the unadjusted analysis, all-cause and CVD-related hospital length of stay were significantly longer for patients receiving inadequate MDD care compared to those receiving adequate care. However, there were no other differences in HCRU and costs.
Table 2. Unadjusted average post-index all-cause and CVD-related HCRU (per patient per year [PPPY]).
Table 3. Unadjusted average post-index all-cause and CVD-related healthcare costs (per patient per year [PPPY]).
and present the adjusted HCRU and costs results from the generalized linear models. Patients receiving inadequate MDD care had 14% more (incident rate ratio [IRR]: 1.14; 95% confidence interval [CI]: 1.01–1.30; p = .036) and 21% longer (IRR: 1.21; 95% CI: 1.15–1.27; p < .001) all-cause hospitalizations, 4% more all-cause outpatient visits (IRR: 1.04; 95% CI: 1.02–1.06; p < .001), 24% longer CVD-related hospitalizations (IRR: 1.24; 95% CI: 1.17–1.31; p < .001), 13% more CVD-related ER visits (IRR: 1.13; 95% CI: 1.03–1.24; p = .006) and 17% more CVD-related outpatient visits (IRR: 1.17; 95% CI: 1.13–1.21; p < .001) compared to patients receiving adequate MDD care. CVD-related hospitalizations were not statistically different between both groups.
Table 4. Adjusted estimates of all-cause and CVD-related HCRU from generalized linear models.
Table 5. Adjusted estimates of all-cause and CVD-related healthcare costs (per patient per year [PPPY]) from generalized linear models.
As shown in , adjusted mean all-cause outpatient costs ($2,820 vs. $2,055 PPPY; p < .001), CVD-related hospitalization costs ($21,485 vs. $17,756 PPPY; p = .0435), and CVD-related outpatient costs ($520 vs. $434 PPPY; p = .0312) were significantly higher for CVD patients receiving inadequate depression treatment versus those receiving adequate treatment. Costs incurred from all-cause hospitalizations, all-cause ER visits, and CVD-related ER visits were not statistically different between both treatment groups.
Figure 3. Direct all-cause and CVD-related costs (per patient per year) among CVD patients receiving adequate versus inadequate MDD care. (A) Adjusted all-cause costs. (B) Adjusted CVD-related costs. Panel (A) shows all-cause adjusted cost estimates from generalized linear models comparing CVD patients receiving adequate versus inadequate MDD care; Panel (B) shows CVD-related adjusted cost estimates from generalized linear models comparing CVD patients receiving adequate versus inadequate MDD care. Abbreviation. CVD, cardiovascular disease; MDD, major depressive disorder; ER, emergency room.
Discussion
This is the first study to examine the impact of the adequacy of depression treatment on HCRU and healthcare costs among CVD patients newly diagnosed with depression. Specifically, the study findings demonstrated that those who received adequate treatment had significantly fewer all-cause hospitalizations, outpatient visits, CVD-related ER visits, and CVD-related outpatient office visits compared to the patients who received inadequate MDD care. With respect to healthcare costs incurred, CVD patients who were deemed adequately treated also had significantly lower all-cause outpatient costs and CVD-related outpatient and hospitalization costs.
Existing studies have observed an increase in HCRU and healthcare costs among CVD patients attributed to the diagnosis of depressionCitation5,Citation13,Citation18–20. For example, a study by Jia et al. found that about half of veterans with acute stroke had post-stroke depression, and those with post-stroke depression were observed to have more hospitalizations, longer hospital length of stay, and more outpatient visits compared to those with no post-stroke depressionCitation19. Although these studies have shown that depression increases healthcare costs and HCRU among CVD patients, no studies have empirically assessed the impact of adequacy of depression treatment on these economic outcomes.
Previous studies have shown the benefit of adequately treating CVD patients with MDD on clinical outcomes such as MI, stroke, congestive heart failure, and anginaCitation12. Treatment with antidepressants such as selective serotonin reuptake inhibitors reduces platelet aggregation and therefore potentially could have a cardioprotective impact among post-MI patientsCitation21,Citation22. Thus, adequate depression care, characterized by appropriate dosage and duration of treatment, resulted in a lower risk of adverse CVD outcomes in previous studies. These studies did not examine the impact of adequate depression care on HCRU and cost outcomes. The current study adds to this previous literature by suggesting that in addition to improvement in adverse CVD outcomes, adequately treating depression among patients with previous MI or stroke also reduces overall and CVD-related HCRU and increases healthcare cost savings. Hence, adequate treatment of depression among CVD patients diagnosed with depression can circumvent some of these problems or undesirable outcomes.
By demonstrating that adequate treatment of depression in CVD patients lowers HCRU and healthcare costs, the findings from this study address an important gap in the literature and have implications for clinicians as well as payers. From a clinical standpoint, it is important to recognize and screen for depression among patients with CVD, especially since clinicians were observed to have different attitudes and practices towards screening and treating depression among patients with CVDCitation23. It is already known that under-recognition and under-treatment of depression can lead to higher healthcare costsCitation24,Citation25. Therefore, from a payer’s perspective, it is critical to recognize the cost-effectiveness of adequate depression treatment. Proper allocation of resources to facilitate access to adequate depression treatment will help reduce HCRU and healthcare costs, especially in this critically ill population of patients with prior MI or stroke.
Results from this study should be interpreted in consideration of few limitations. First, the study analyzed data from a commercial insurance database, which primarily included people with employer-sponsored insurance. Therefore, the findings may not necessarily generalize to all CVD patients with depression. Second, as with most administrative claims data, clinical information such as depression symptomology and severity of depression that may impact MDD care adequacy could not be ascertained. Frailty, which is a potential predictor of adverse CVD outcomes, could not be controlled in the analysis as this information is not captured in administrative claims data. Nevertheless, the study adjusted for total comorbidity burden using the Charlson Comorbidity Index (CCI) for each patient in all multivariable analyses. Third, in the assessment of treatment adequacy, the study used prescription claims data, which allow for the determination of a drug being dispensed and thereby reflect drug availability to the patient, but not necessarily actual drug-taking behaviorCitation26. Fourth, this study did not include drug/pharmaceutical costs in the calculation of total healthcare expenditure. These costs were excluded because any pharmaceuticals administered to the patient in an inpatient setting would not be captured in MarketScan's administrative claims data. The cost of any pharmaceuticals administered to the patient during hospitalization was captured as a part of the total cost for the hospital stay. The MarketScan data only provides aggregated costs for inpatient claims whereby costs of all inpatient services (including pharmaceuticals) utilized by a patient from the date of admission until discharge are summed up to calculate the total cost of that particular hospitalization episode. However, it was not possible to obtain itemized costs for individual services/pharmaceuticals that the patient may have received during the inpatient stay. For drugs administered outside a hospital setting, the dataset does not collect information on any discounts, rebates, and/or coupons that manufacturers may provide to patients at point of sale. Also, only outpatient dispensations covered by commercial, Medicaid, or Medicare insurance appeared in the database. The database does not record prescriptions for which a patient may have paid out of pocket. Fifth, although psychotherapy (particularly cognitive behavioral therapy) has been shown to be effective for the treatment of MDD, the current analysis only considered MDD patients who were treated with any pharmacotherapy to determine treatment adequacy. Psychotherapy was not considered for this study. Also, there was no way of identifying the type of psychotherapy received (i.e. cognitive behavioral therapy) by the patient and guidelines report varying efficacy by typeCitation27. Future studies should consider incorporating psychotherapy into the definition of MDD care adequacy. Finally, the APA guidelines for the treatment of the acute phase of MDD state that pharmacotherapy is recommended as the initial treatment choice for patients with mild to moderate MDD and definitely should be provided for those with severe MDDCitation15.
Conclusions
This is the first study that provides empirical economic evidence supporting the use of adequate depression treatment among CVD patients newly diagnosed with depression. The study results suggest that among CVD patients newly diagnosed with depression, the receipt of adequate depression treatment leads to lower HCRU and healthcare costs. The results are also timely given the recent AHA and American Stroke Association statement emphasizing the need for future studies to evaluate the effect of depression treatment post-CVD on subsequent outcomesCitation6.
Transparency
Declaration of funding
The study was funded by Pfizer Inc.
Declaration of financial/other interests
Ruchit Shah, Xin Gao, and Jennifer Stephens are employees of Pharmerit International, LP, which received research funding from Pfizer Inc. for this study. Chinmay Deshpande was an employee of Pharmerit International, LP, when the study was conducted. Ahmed Shelbaya, Elizabeth Pappadopulos, and Patricia Schepman are employees of Pfizer Inc. and own Pfizer stock. Rita Prieto works for Pfizer GEP S.L.U. Spain and owns Pfizer stock.
JME peer reviewers on this manuscript have no relevant financial or other relationships to disclose.
Author contributions
The study was designed by Sripal Bangalore, Ruchit Shah, Xin Gao, Elizabeth Pappadopulos, Chinmay Deshpande, Ahmed Shelbaya, Rita Prieto, Jennifer Stephens, Patricia Schepman, and Roger McIntyre. The data were analyzed by Ruchit Shah and Chinmay Deshpande. Sripal Bangalore, Ruchit Shah, Xin Gao, Elizabeth Pappadopulos, Chinmay Deshpande, Ahmed Shelbaya, Rita Prieto, Jennifer Stephens, Patricia Schepman, and Roger McIntyre wrote the paper. All authors read and approved the final version of the manuscript.
Previous presentations
An earlier version of this analysis was presented as a poster at the American Managed Care and Specialty Pharmacy Annual Meeting 2018 (April 24–26, 2018).
Acknowledgements
The authors would like to thank the Pharmerit project team members Rachel Shah and Zhiyong Chen for their contributions to the research, interpretation, and critical review.
References
- Heidenreich PA, Trogdon JG, Khavjou OA. Forecasting the future of cardiovascular disease in the United States. Circulation. 2011;123(8):933.
- Fan AZ, Strine TW, Jiles R, et al. Depression and anxiety associated with cardiovascular disease among persons aged 45 years and older in 38 states of the United States, 2006. Prev Med. 2008;46(5):445–450.
- Ha J-H, Wong C-K. Pharmacologic treatment of depression in patients with myocardial infarction. J Geriatr Cardiol. 2011;8(2):121–126.
- Bradley SM, Rumsfeld JS. Depression and cardiovascular disease. Trends Cardiovasc Med. 2015;25(7):614–622.
- Rutledge T, Vaccarino V, Johnson BD, et al. Depression and cardiovascular health care costs among women with suspected myocardial ischemia: prospective results from the WISE (Women’s Ischemia Syndrome Evaluation) study. J Am Coll Cardiol. 2009;53(2):176–183.
- Towfighi A, Ovbiagele B, El Husseini N, et al. Poststroke depression: a scientific statement for healthcare professionals from the American Heart Association/American Stroke Association. Stroke. 2017;48(2):e30–e43.
- Chang L, Liu N. The safety, efficacy, and tolerability of pharmacological treatment of depression in patients with cardiovascular disease: a look at antidepressants and integrative approaches. Heart and Mind. 2017;1(1):8–16.
- Myers V, Gerber Y, Benyamini Y, et al. Post-myocardial infarction depression: increased hospital admissions and reduced adoption of secondary prevention measures – a longitudinal study. J Psychosom Res. 2012;72(1):5–10.
- Chung ML, Dekker RL, Lennie TA, et al. Antidepressants do not improve event-free survival in patients with heart failure when depressive symptoms remain. Heart Lung. 2013;42(2):85–91.
- Frasure-Smith N, Lespérance F, Talajic M. Depression and 18-month prognosis after myocardial infarction. Circulation. 1995;91(4):999–1005.
- Lichtman JH, Bigger JT, Blumenthal JA, et al. Depression and coronary heart disease: recommendations for screening, referral, and treatment: a science advisory from the American Heart Association Prevention Committee of the Council on Cardiovascular Nursing, Council on Clinical Cardiology, Council on Epidemiology and Prevention, and Interdisciplinary Council on Quality of Care and Outcomes Research: endorsed by the American Psychiatric Association. Circulation. 2008;118(17):1768–1775.
- Bangalore S, Shah R, Pappadopulos E, et al. Cardiovascular hazards of insufficient treatment of depression among patients with known cardiovascular disease: a propensity score adjusted analysis. Euro Heart J Qual Care Clin Outc. 2018;4(4):258–266.
- Husaini B, Levine R, Sharp L, et al. Depression increases stroke hospitalization cost: an analysis of 17,010 stroke patients in 2008 by race and gender. Stroke Res Treat. 2013;2013:846732.
- IBM Watson Health. IBM MarketScan Research Databases for Health Services Researchers (White Paper). 2018 [cited 2019 Nov 4]. Available from: https://www.ibm.com/downloads/cas/6KNYVVQ2
- Gelenberg A, Freeman M, Markowitz J, et al. Practice guideline for the treatment of patients with Major Depressive Disorder: third edition. San Fransisco (CA): American Psychiatric Association; 2010.
- Charbonneau A, Rosen AK, Owen RR, et al. Monitoring depression care: in search of an accurate quality indicator. Med Care. 2004;42(6):522–531.
- Westreich D, Cole SR, Funk MJ, et al. The role of the c-statistic in variable selection for propensity score models. Pharmacoepidem Drug Safe. 2011;20(3):317–320.
- Sullivan M, Simon G, Spertus J, et al. Depression-related costs in heart failure care. Arch Intern Med. 2002;162(16):1860–1866.
- Jia H, Damush TM, Qin H, et al. The impact of poststroke depression on healthcare use by veterans with acute stroke. Stroke. 2006;37(11):2796–2801.
- Ghose SS, Williams LS, Swindle RW. Depression and other mental health diagnoses after stroke increase inpatient and outpatient medical utilization three years poststroke. Med Care. 2005;43(12):1259–1264.
- McFarlane A, Kamath MV, Fallen EL, et al. Effect of sertraline on the recovery rate of cardiac autonomic function in depressed patients after acute myocardial infarction. Am Heart. 2001;142(4):617–623.
- Yekehtaz H, Farokhnia M, Akhondzadeh S. Cardiovascular considerations in antidepressant therapy: an evidence-based review. J Tehran Heart Cent. 2013;8(4):169.
- Feinstein RE, Blumenfield M, Orlowski B, et al. A national survey of cardiovascular physicians’ beliefs and clinical care practices when diagnosing and treating depression in patients with cardiovascular disease. Cardio Rev. 2006;14(4):164–169.
- Panzarino PJ, Jr. The costs of depression: direct and indirect; treatment versus nontreatment J Clin Psychiatry. 1998;59(20):11–14.
- Davidson JR, Meltzer-Brody SE. The underrecognition and undertreatment of depression: what is the breadth and depth of the problem? J Clin Psychiatry. 1999;60(7):4–9.
- Crystal S, Akincigil A, Bilder S, et al. Studying prescription drug use and outcomes with Medicaid claims data strengths, limitations, and strategies. Med Care. 2007;45(2):S58–S65.
- Kennedy SH, Lam RW, McIntyre RS, et al. Canadian network for mood and anxiety treatments (CANMAT) 2016 clinical guidelines for the management of adults with major depressive disorder: section 3. Pharmacological treatments. Can J Psychiatry. 2016;61(9):540–560.