Abstract
Objectives
To characterize the health-care utilization and economic burden associated with depression in Manitoba, Canada.
Methods
Patient-level data were retrieved from the Manitoba Centre for Health Policy administrative, clinical, and laboratory databases for the study period of January 1, 1996, through December 31, 2016. Patients were assigned to the depression cohort based on diagnoses recorded in hospitalizations and outpatient physician claims, as well as antidepressant prescription drug claims. A comparison cohort of nondepressed subjects, matched with replacement for age, gender, place of residence (urban vs. rural), and index date, was created. Demographics, comorbidities, intentional self-harm, mortality, health-care utilization, prescription drug utilization, and costs of health-care utilization and social services were compared between depressed patients and matched nondepressed patients, and incidence rate ratios and hazard ratios were reported.
Results
There were 190,065 patients in the depression cohort and 378,177 patients in the nondepression cohort. Comorbidities were 43% more prevalent among depressed patients. Intentional self-harm, all-cause mortality, and suicide mortality were higher among patients with depression than the nondepression cohort. Health-care utilization—including hospitalizations, physician visits, physician-provided psychotherapy, and prescription drugs—was higher in the depression than the nondepression cohort. Mean health-care utilization costs were 3.5 times higher among depressed patients than nondepressed patients ($10,064 and $2,832, respectively). Similarly, mean social services costs were 3 times higher ($1,522 and $510, respectively). Overall, depression adds a total average cost of $8,244 (SD = $40,542) per person per year.
Conclusions
Depression contributes significantly to health burden and per patient costs in Manitoba, Canada. Extrapolation of the results to the entire Canadian health-care system projects an excess of $12 billion annually in health system spending.
Keywords: depression, cost, economic burden, health-care utilization, mortality, suicide
Abstract
Objectifs
Caractériser l’utilisation des soins de santé et la charge économique associées à la dépression au Manitoba, Canada.
Méthodes
Les données sur les patients ont été extraites des bases de données administratives, cliniques et de laboratoire du Centre des politiques de santé du Manitoba pour la période de l’étude, du 1er janvier 1996 au 31 décembre 2016. Les patients ont été affectés à la cohorte de la dépression en fonction des diagnostics enregistrés lors des hospitalisations et des factures des médecins des services ambulatoires, ainsi que des demandes de remboursement des ordonnances de médicaments antidépresseurs. Une cohorte de comparaison de sujets non dépressifs, appariés en remplacement pour l’âge, le sexe, le lieu de résidence (urbain contre rural), et la date de début a été créée. Les données démographiques, les comorbidités, l’automutilation intentionnelle, la mortalité, l’utilisation des soins de santé, l’utilisation de médicaments sur ordonnance, et les coûts de l’utilisation des soins de santé et des services sociaux ont été comparés entre les patients déprimés et les patients appariés non déprimés, et les rapports des taux d’incidence ainsi que les rapports de risque ont été signalés.
Résultats
Il y avait 190 065 patients dans la cohorte de la dépression et 378 177 patients dans la cohorte de non-dépression. Les comorbidités étaient 43% plus prévalentes chez les patients déprimés. L’automutilation intentionnelle, la mortalité pour toutes causes, et la mortalité par suicide étaient plus élevées chez les patients souffrant de dépression que dans la cohorte de non-dépression. L’utilisation des soins de santé – notamment les hospitalisations, les visites au médecin, la psychothérapie donnée par un médecin, et les médicaments sur ordonnance – était plus élevée dans la cohorte de la dépression que dans celle de la non-dépression. Les coûts moyens de l’utilisation des soins de santé étaient 3,5 fois plus élevés chez les patients déprimés que chez ceux non déprimés (10 064 $ et 2 832 $, respectivement). De même, les coûts moyens des services sociaux étaient 3 fois plus élevés (1 522 $ et 510 $, respectivement). En tout, la dépression ajoute un coût moyen total de 8 244 $ (ET 40 542 $) par personne, par année.
Conclusions
La dépression contribue significativement à la charge de la santé et aux coûts par patient au Manitoba, Canada. L’extrapolation des résultats à tout le système de santé canadien projette un excès de 12 milliards de dollars annuellement aux dépenses du système de santé.
Introduction
Mental illness and addictions are a significant burden in Canada, exceeding that of other chronic illnesses such as cancer and infectious disease.1 This extensive health burden results in greater use of health-care resources and a substantial economic burden, estimated to be as large as $51 billion in Canada in 2003.2 In Ontario, Canada’s most populous province, high-cost mental health patients were found to incur 60% higher average health care costs than other high-cost patients who had no mental health–related costs.3,4 Similarly, when looking at non-mental health–related population health care spending, rates of mental illness and addiction increase across increasing cost groups, with the highest rates of mental illness observed for the highest costing patients in Ontario.5
Because of its high incidence, chronicity, and associated morbidity, depression exerts the greatest burden of all mental illnesses: It is a leading cause of disability and excess mortality in Canada.1,6–8 In addition to its psychological symptoms, depression is associated with chronic comorbidities including diabetes, heart disease, arthritis, asthma, back pain, chronic bronchitis, hypertension, and migraines.9,10 As well as understanding the health impact of depression, it is important to understand the corresponding economic burden in order to inform government spending and resource allocation.
While several studies have assessed the economic burden of mental illness overall, few studies have focused on the per patient cost of depression in Canada.11,12 In a population-based cohort study in Ontario, annual per capita direct health-care costs were more than 20% higher for patients with depression compared to a control cohort without depression or psychological distress ($3,210 and $2,629, respectively).11 In a separate Alberta study of patients who had at least one health-care visit for depression, costs associated with depression were highly variable, averaging to $550 per patient but reaching as high as $25,826 in the top 1% of the cohort.12 As costs reported in this study were depression-specific, costs associated with health services for other mental or physical illnesses were excluded; this is a major limitation of the study. Furthermore, prescription drug costs and indirect costs were not assessed in either study, producing conservative estimates of the burden of depression. A comprehensive assessment of depression costs in Canada is needed.
The Manitoba Centre for Health Policy (MCHP) houses administrative, clinical, and laboratory databases, allowing for record linkage and cohort generation for specific diagnoses, including major depressive disorder.13 The main objective of the current study was to perform a comprehensive evaluation of health-care utilization and economic burden of clinically significant depression (i.e., a depression requiring medical attention and treatment) in Manitoba, Canada, using patient-level data from the MCHP.
Methods
Study Design and Data Sources
This was a historical record linkage–matched cohort study. The enrollment period was January 1, 1996, through December 31, 2015, with the study period ending December 31, 2016, to ensure at least 1 year of follow-up data for all enrolled patients. This allowed for sufficient data collection to determine annual per patient health-care service utilization and costs. We gathered patient-level information from administrative, clinical, and social services databases housed at the MCHP, including Manitoba Health, the Winnipeg Regional Health Authority, Manitoba Families, and the Vital Statistics Agency. Manitoba Health provides publicly funded universal health insurance to the 1.3 million residents of Manitoba, with a participation rate >99%.14 Deidentified data are linked by an encrypted personal health information number with >95% linkage accuracy.13 A detailed description of the data collection is provided in Supplementary Materials.
Study Population
The inclusion criteria for this study were all Manitobans who were at least 13 years of age by the index date and were registered with Manitoba Health during the enrollment period. Patients younger than 13 years of age were not included as the onset of depression symptoms typically occurs during adolescence or young adulthood.15 The exclusion criterion was as follows: subject had <1 year of insurance coverage prior to the index date, with no minimum requirement post index date. For a member of the depression cohort, the index date was the date of the hospitalization during which the depression criteria were met or the date of first physician visit that met the depression definition. A member of the nondepression cohort was assigned the same index date as their match. Based on a previously validated algorithm,16 patients were assigned to the depression cohort if they met at least one of the following criteria: ≥1 hospitalization with one or more of the predetermined depression International Classification of Diseases (ICD) codes (Supplementary Table 2), or ≥5 physician claims with one or more of the predetermined depression ICD codes (Supplementary Table 2), or ≥1 physician claims and ≥7 prescription claims (Supplementary Table 2) in 2 consecutive years. The nondepression cohort was generated by matching each member of the depression cohort with up to two eligible persons who met inclusion and exclusion criteria but did not meet the depression criteria by the index date. Matching was performed based on age, gender, and place of residence (urban or rural). Urban refers to the cities of Winnipeg and Brandon, and rural refers to the rest of Manitoba. Members of the nondepression cohort were matched with replacement; therefore, each member could serve as a match for more than one member of the depression cohort.17
For both cohorts, follow-up time was measured from the index date to the earliest date of death, emigration, or the end of the study period (December 31, 2016). Additionally, follow-up for members of the nondepression cohort ended on the date of meeting criteria for depression, at which time they became members of the depression cohort.
Outcome Measures
Mortality (all cause and suicide), intentional self-harm, comorbidities, health-care utilization, prescription drug utilization, and costs of health-care utilization and social services were assessed for both cohorts. Intentional self-harm was defined using the following ICD codes: ICD-10 X6, X7, X80-84, Y97.0, and ICD-9 E95. Comorbidities, defined as any chronic disease, were determined based on previously validated algorithms using prespecified diagnosis, physician claim, and Drug Program Information Network (DPIN) codes (Supplementary Table 1).18,19 Costs were quantified for the following categories: hospital costs, physician costs, prescription drug costs, rent assist payments, and employment and income assistance (EIA). Hospital costs were calculated using standard costing measures from two values—a weight reflecting the relative intensity of resource use and a standard cost per weighted case. Inpatient stays and day procedures were assigned a resource intensity weight (RIW) and day procedure group weight (DPGWGT), respectively. Weights were calculated by the Canadian Institute for Health Information and were available in Manitoba Health’s Hospital Abstracts Database. For each year, the value of all hospital procedures in 2014 Canadian dollars (CAD) was calculated as (RIW + DPGWGT) × average cost of standard hospital stay (CSHS) for that year. The average CSHS for Manitoba in 2014/2015 was $6,152. Physician services costs were calculated as the sum of professional fees or tariffs paid to a physician for each service provided and recorded in Manitoba Health’s Medical Services Database. For prescription drugs, the cost of each drug and the dispensing fee was recorded in Manitoba Health’s DPIN database. When dispensing fees were not available, they were imputed using standard methods for Manitoba DPIN data.20 The total prescription drug cost was equal to the sum of drug ingredient cost and dispensing fee. Rent assist payments were calculated by tallying all rent assist payments made to a person. EIA was measured in months received and then converted to 2017 CAD by multiplying mean months with EIA per patient per year for each cohort by the mean monthly EIA payments documented for Manitoba, which were reported for the year 2017.21
For all categories except EIA, costs were measured in 2014 CAD, and final costs were converted to 2018 CAD using the inflation rate 6.55% from the all items Consumer Price Index (CPI).22 EIA costs were converted from 2017 CAD to 2018 CAD using the inflation rate 2.30% from the CPI. The cost differences reported here are primarily from the public payer perspective, with the exception of prescription drug claims. Information about drug claims in the Manitoba DPIN are captured regardless of type of insurance coverage and payer. Therefore, we have likely captured costs attributed to both public and private drug programs.
Data Analyses
All analyses were conducted for both cohorts—depression and matched nondepression comparators. Descriptive statistics, including mean, median, standard deviation (SD), first quartile (Q1) and third quartile (Q3), were used to summarize demographic, health-care utilization, prescription drug utilization, and cost of health-care utilization and social services. The prevalence, 95% confidence intervals (95% CI), incidence rate per 1,000 person-years, and age-adjusted incident rate ratios (IRRs) across the entire study period were calculated using standard time-to-event methods for the following outcomes: all-cause mortality, intentional self-harm, intentional self-harm mortality, health-care utilization, emergency department utilization, comorbidities, EIA, and rent assist. Time-to-event was measured to the occurrence of each outcome or to censoring. Multivariate Cox regression models were used to measure the hazard ratios for mortality due to all-causes and intentional self-harm. Hazard ratios were matching-adjusted by age, gender, place of residence, and index date. The prevalence of each of the following across the entire follow-up period was calculated: comorbidities, utilization of health-care resources, use of antidepressants. The mean (SD) utilization of each health-care resource and costs (2018 CAD) of health-care and social services were calculated per person per year based on the entire follow-up period.
Results
Cohort Demographics
A total of 190,065 patients were included in the depression cohort and 378,177 patients in the nondepression cohort. The mean follow-up periods in the depression and nondepression cohorts were 11.5 (Q1 = 5.9, Q3 = 17.0) and 10.3 (Q1 = 5.0, Q3 = 15.1) years, respectively. Patient demographics are summarized in Table 1. The mean age of both cohorts was 45 (Q1 = 29, Q3 = 57) years, respectively. The distribution of patients across age groups was similar in both cohorts. The majority of patients in both cohorts resided in urban areas (63.2% of depression cohort, 63.7% of nondepression cohort). When grouping patients according to income quintile, the proportion of patients in the higher income quintiles was lower in the depression than in the nondepression cohort (depression cohort: 18.7% in Q4, 16.9% in Q5; nondepression cohort: 19.7% in Q4, 18.2% in Q5).
Table 1.
Patient Demographics.
| Demographics | Depression | Nondepression |
|---|---|---|
| N | 190,065 | 378,177 |
| Follow-up time, years | ||
| Mean, median (Q1, Q3) | 11.5, 11.8 (5.9, 17.0) | 10.3, 9.5 (5.0, 15.1) |
| Male, N (%) | 64,135 (33.7%) | 127,616 (33.7%) |
| Age, years | ||
| Mean (SD) | 45 (20) | 45 (20) |
| Median (Q1, Q3) | 42 (29, 57) | 42 (29, 57) |
| 13 to 19 | 19,596 (10.3%) | 39,207 (10.4%) |
| 20 to 29 | 29,521 (15.5%) | 58,825 (15.6%) |
| 30 to 39 | 35,740 (18.8%) | 71,260 (18.8%) |
| 40 to 49 | 36,682 (19.3%) | 73,129 (19.3%) |
| 50 to 59 | 25,075 (13.2%) | 49,835 (13.2%) |
| 60 to 69 | 14,448 (7.6%) | 28,680 (7.6%) |
| 70 to 79 | 14,766 (7.8%) | 29,299 (7.7%) |
| 80 to 89 | 11,743 (6.2%) | 23,250 (6.1%) |
| 90+ | 2,494 (1.3%) | 4,692 (1.2%) |
| Residence locality | ||
| Urban | 120,090 (63.2%) | 240,902 (63.7%) |
| Rural | 66,450 (35.0%) | 131,825 (34.9%) |
| Unknown | 3,525 (1.9%) | 5,450 (1.4%) |
| Income quintile | ||
| Q1 (lowest) | 37,745 (19.9%) | 69,748 (18.4%) |
| Q2 | 41,050 (21.6%) | 80,180 (21.2%) |
| Q3 | 40,101 (21.1%) | 79,286 (21.0%) |
| Q4 | 35,618 (18.7%) | 74,522 (19.7%) |
| Q5 (highest) | 32,026 (16.9%) | 68,991 (18.2%) |
Health Outcomes
The overall prevalence and incidence of depression from 1996 to 2015 was 12.5% (95% CI, 12.5% to 12.6%) and 10.9 (per 1,000 person-years; 95% CI, 10.8 to 10.9), respectively. Across the entire follow-up period, the prevalence of any comorbidity was 43% higher among patients with depression than without depressions (53.0% vs. 37.0%; Table 2). Similarly, the intentional self-harm incidence rate was higher among patients with depression (1.5 per 1,000 patient-years; 95% CI, 1.4 to 1.5) than nondepressed patients (0.1 per 1,000 patient-years; 95% CI, 0.1 to 0.1), with an IRR of 19.5 (95% CI, 17.3 to 22.1). The incidence rate for suicide mortality was 0.4 (95% CI, 0.4 to 0.4) and <0.1 (95% CI, 0.0 to 0.0)) per 1,000 person-years for patients in the depression and nondepression cohorts, respectively (IRR 9.4; 95% CI, 7.9 to 11.2). All-cause mortality rates were also higher in the depression than the nondepression cohort. In the depression cohort, the all-cause mortality incidence rate per 1,000 person-years was 17.3 (95% CI, 17.1 to 17.4) compared to 10.4 (95% CI, 10.3 to 10.5) in the nondepression cohort. This corresponded to an IRR of 1.7 (95% CI, 1.6 to 1.7) for depression relative to nondepression.
Table 2.
The Prevalence and Incidence Rates of Health Outcomes during the Entire Follow-Up Period.
| Depression | Nondepression | |
|---|---|---|
| N | 190,065 | 378,177 |
| Comorbidities (any chronic disease), N (%) | 100,826 (53.0) | 140,050 (37.0) |
| Comorbidities (any chronic disease), incidence rate per 1,000 person-years (95% CI)a | 72.4 (72.0 to 72.9) | 49.6 (49.4 to 49.9) |
| Intentional self-harm, N (%) | 3,658 (1.9) | 291 (0.1) |
| Intentional self-harm, incidence rate per 1,000 person-years (95% CI)b | 1.5 (1.4 to 1.5) | 0.1 (0.1 to 0.1) |
| Intentional self-harm mortality, N (%)c | 835 (0.4) | 159 (<0.1) |
| Intentional self-harm mortality, incidence rate per 1,000 person-years (95% CI)d | 0.4 (0.4 to 0.4) | <0.1 (0.0 to 0.0) |
| All-cause mortality, N (%)e | 37,722 (19.8) | 40,302 (10.7) |
| All-cause mortality, incidence rate per 1,000 person-years (95% CI)f | 17.3 (17.1 to 17.4) | 10.4 (10.3 to 10.5) |
a IRR is 1.46 (95% CI, 1.45 to 1.47) for depression relative to nondepression.
b IRR is 19.53 (95% CI, 17.31 to 22.09) for depression relative to nondepression.
cMatching-adjusted hazard ratio is 9.28 (95% CI, 7.82 to 11.0) for depression relative to nondepression.
d IRR is 9.38 (95% CI, 7.90 to 11.19) for depression relative to nondepression.
eMatching-adjusted hazard ratio is 1.64 (95% CI, 1.61 to 1.66) for depression relative to nondepression.
f IRR is 1.66 (95% CI, 1.64 to 1.68) for depression relative to nondepression.
Note. Follow-up period: depression, 11.5 years (mean), 11.8 years (median); nondepression, 10.3 years (mean), 9.5 years (median). IRR = incidence rate ratio; CI = confidence interval.
Health-Care Utilization
Health-care utilization by patients with and without depression was reported for the entire duration of the study (Table 3) or as an annual average (Table 4). While the vast majority of patients in both cohorts had a physician visit during the study period, more depressed patients had at least one family doctor (99.6% vs. 96.3%), specialty (97.1% vs. 89.6%), or physician psychotherapy (79.3% vs. 38.3%) visit compared to nondepressed patients. Per year, each depression patient had more than double the mean number of family doctor (11.0 [SD = 15.0] vs. 5.0 [SD = 5.2]) and specialist (7.6 [SD = 19.4] vs. 3.5 [SD = 5.9]) visits compared to nondepressed patients. Notably, patients without depression attended a mean of 0.1 (SD = 0.5) physician psychotherapy visit per year compared to 1.7 (SD = 4.7) visits per patient per year in the depression cohort.
Table 3.
The Number of Patients Who Utilized Health-Care Services at Any Time during the Entire Follow-Up Period.
| Health-Care Service | Depression, N (%) | Nondepression, N (%) |
|---|---|---|
| Any physician visit | 189,954 (99.9) | 367,337 (97.1) |
| Family doctor visit | 189,325 (99.6) | 364,065 (96.3) |
| Specialist visit | 184,627 (97.1) | 338,991 (89.6) |
| Psychotherapy/consultationa | 150,665 (79.3) | 144,995 (38.3) |
| Hospitalizationb | 117,496 (61.8) | 155,890 (41.2) |
| Hospital readmission | 31,606 (16.6) | 30,105 (8.0) |
| ICU admission | 16,633 (8.8) | 18,594 (4.9) |
| Hospitalization due to self-harmc | 3,658 (1.9) | 291 (0.1) |
| Presented to emergency departmentd | 110,636 (58.2) | 150,210 (39.7) |
| Admitted to hospital from emergency departmente | 46,897 (24.7) | 47,501 (12.6) |
a See Supplementary Table 3 for codes defining psychotherapy.
b Incidence rate ratio (IRR) is 1.56 (95% CI, 1.55 to 1.57) for depression relative to nondepression.
c IRR is 19.53 (95% CI, 17.31 to 22.09) for depression relative to nondepression.
d IRR is 1.65 (95% CI, 1.63 to 1.66) for depression relative to nondepression.
e IRR is 1.82 (95% CI, 1.80 to 1.85) for depression relative to nondepression.
Table 4.
The Mean Utilization of Health-Care Services per Patient per Year during the Entire Follow-Up Period.
| Health-Care Service | Depression, Mean (SD) | Nondepression, Mean (SD) |
|---|---|---|
| Any physician visit | 18.6 (27.8) | 8.5 (8.8) |
| Family doctor visit | 11.0 (15.0) | 5.0 (5.2) |
| Specialist visit | 7.6 (19.4) | 3.5 (5.9) |
| Psychotherapy sessionsa | 1.7 (4.7) | 0.1 (0.5) |
| Hospitalization | 0.5 (4.1) | 0.1 (0.3) |
| Days in hospital | 8.3 (40.5) | 1.9 (8.3) |
| Days in intensive care unitb | 0.7 (0.5) | 0.4 (3.5) |
| Emergency department admission | 0.4 (2.6) | 0.1 (0.4) |
| Days receiving long-term care | 16.0 (61.2) | 4.2 (29.5) |
a See Supplementary Table 3 for codes defining psychotherapy.
b Mean for entire follow-up period, not mean per year.
The number of patients who had any hospitalization during the study period was 50% higher in the depressed compared to the nondepression cohort (61.8% vs. 41.2%; IRR 1.6; 95% CI, 1.6 to 1.6). Depressed patients spent a mean of 8.3 (SD = 40.5) days per year in the hospital compared to 1.9 (SD = 8.3) days per year for nondepressed patients. The majority of patients (58.2%) in the depression cohort presented to the emergency department at least once during the study period compared to 39.7% of the nondepression cohort (IRR 1.8; 95% CI, 1.8 to 1.9), and there was a mean of 0.4 (SD = 2.6) and 0.1 (SD = 0.4) emergency department admissions per patient per year in the depression and nondepression cohorts, respectively.
The use of antidepressant medications was significantly more common in the depression compared to the nondepression cohort (Table 5). The majority of depressed patients (80.4%) used selective serotonin reuptake inhibitors at least once during the study period, whereas only 4.3% of nondepressed patients used this class of drug. The use of other classes of antidepressants ranged from 21.8% to 40.3% and 1.5% to 9.0% among patients with and without depression, respectively.
Table 5.
The Number (%) of Ever-Users of Each Prescription Drug Category during the Entire Follow-Up Period.
| Drug Category | Depression, N (%) | Nondepression, N (%) |
|---|---|---|
| Tricyclic antidepressants | 62,407 (32.8) | 34,200 (9.0) |
| Norepinephrine reuptake inhibitors | 41,467 (21.8) | 11,831 (3.1) |
| Selective serotonin reuptake inhibitors | 152,802 (80.4) | 16,398 (4.3) |
| Serotonin and norepinephrine reuptake inhibitors | 76,581 (40.3) | 5,713 (1.5) |
| Other antidepressants | 58,984 (31.0) | 8,224 (2.2) |
| Benzodiazepine derivativea | 47,651 (25.1) | 6,233 (1.6) |
a Clonazepam, ATC code N03AE01.
Note. See Supplementary Table 4 for list of drugs that comprise each category.
Social Service Utilization
Over the duration of the study, social service utilization was significantly higher in the depression cohort, with nearly 3 times as many patients in the depression cohort receiving EIA compared to the nondepression cohort (17.8% vs. 6.1%). The crude incidence rates of receiving EIA were 10.7 (95% CI, 10.6 to 10.9) and 3.3 (95% CI, 3.2 to 3.4) per 1,000 person-years (IRR 3.3; 95% CI, 3.2 to 3.3) for depressed compared to nondepressed patients. Similarly, the number of patients receiving rent assist was more than 3 times higher in the depression than in the nondepression cohort (9.6% vs. 3.1%), with crude incidence rates of 8.0 (95% CI, 7.9 to 8.2) and 2.6 (95% CI, 3.2 to 3.4) per 1,000 person-years (IRR 3.3; 95% CI, 3.2 to 3.3), respectively.
Costs Associated with Health-Care and Social Service Utilization
The mean direct medical costs associated with depression were more than 3.5 times higher than that of the nondepression cohort ($10,064 [SD = $41,113] vs. $2,832 [SD = $7,601] per patient per year; Figure 1 and Supplementary Table 5). The direct costs consisted of hospital, physician, and prescription drug costs. The highest contributor to direct medical costs in both cohorts was hospital costs ($7,192 [SD = $38,761] vs. $1,701 [SD $6,623]). Prescription drug costs (depression $1,441 [SD = $2,962], nondepression $527 [SD = $2,101]) and physician costs (depression $1,431 [SD = $3,282], nondepression $605 [SD = $737]) were similar within each cohort. The mean rent assist payment was double in the depression compared to nondepression cohort ($19 [SD = $105] vs. $9 [SD = $87] per patient per year); similarly, mean EIA costs were 3 times higher for depression than for nondepressed patients ($1,503 [SD = $4,175] vs. $501 [SD = $2,505 per patient per year]).
Figure 1.
The mean (A) direct medical and (B) social service costs (2018 CAD) per patient per year during the entire follow-up period. CAD = Canadian dollar.
Discussion
Substantial differences in comorbidities, intentional self-harm, mortality, health-care utilization, and costs were observed in patients with depression compared to nondepressed patients, with all measures of health and economic burden being higher among those with depression. Intentional self-harm is the ninth leading cause of mortality in Canada, where intentional self-harm mortality rate was 0.1 per 1,000 population in 2016.23,24 Here, we reported that the depression and nondepression cohorts, respectively, had intentional self-harm mortality rates of 0.4 and 0.0 per 1,000 person-years, highlighting a significantly elevated risk of death by suicide among patients with depression.25 The higher intentional self-harm and all-cause mortality rates in the depression cohort were consistent with patterns previously reported in Ontario, Canada, where all-cause mortality rates for the depression and referent group were 14.2 and 7.9 per 1,000 person-years, respectively.6 Of note, the all-cause mortality rate reported for the nondepression group in the current study is higher than that reported for Manitoba by Statistics Canada (10.4 vs. 8.3 per 1,000 population).26 However, the Statistics Canada estimate represents all age groups, whereas our sample is restricted to ages 13 and older. Furthermore, 23% of our sample was aged 60 and above—the age-group with the highest mortality rates.
The number of physician visits, regardless of specialty, was higher in the depression cohort. Almost all depressed (99.9%) and nondepressed (97.1%) patients had at least one physician visit during the 10-year study period. This is consistent with the 12-month prevalence of physician visits for depressed patients in the United States who have and have not achieved remission (97.9% and 99.1%, respectively) reported by Dennehy et al.27 The mean numbers of physician visits per year for depressed and nondepressed patients in our study were 18.6 and 8.5 visits, respectively, similar to the number of physician visits per year reported by Dennehy et al. for U.S. patients who had or had not remitted from their depression (11.1 and 17.2, respectively).27 The average number of emergency department visits and days in hospital per year were also comparable between our depression cohort and the Dennehy nonremission group, and our nondepression cohort and the Dennehy remission group. Although the study designs and populations are not directly comparable, there are notable similarities in health-care resource utilization between our study and this study by Dennehy and colleagues. These findings suggest that achieving remission is associated with decreased health-care utilization to levels similar to patients with no depression diagnosis.
Although the use of antidepressant medications was considerably higher in the depression cohort than in the nondepression cohort, nearly 10% of the nondepression cohort was treated with at least one antidepressant during the study period. Most antidepressants are also prescribed for conditions other than depression. For example, in primary care, approximately 29% of antidepressants prescriptions are provided for indications such as anxiety, pain, insomnia, or nicotine dependence.28 Furthermore, antidepressant use in the nondepression cohort may be attributed to our conservative criteria for inclusion in the depression cohort, which was more likely to identify patients with moderate or severe depression. As a result, some patients with mild depression may have been included in the nondepression cohort.
Consistent with higher health-care resource utilization, costs associated with health-care and social services are significantly higher among patients with depression than those without. Costs associated with health-care and social services were $7,232 and $1,012 higher, respectively, per patient per year in the depression group. Using Statistics Canada population estimates for people aged 15 and older, and considering an annual prevalence of depression of 4.7%, this translates to an excess spending of over $400 million each year in in Manitoba and more than $12 billion per year across Canada.29,30 Not all costs were captured by the algorithm used in this study, including societal costs associated with reduced productivity, short- and long-term third-party disability, and impacts on caregivers, which are likely to be accrued among patients with depression. Moreover, federally provided social services and income assistance are not recorded in provincial databases; therefore, our estimates of social service utilization and associated costs underestimate the costs associated with depression. This is an area of future research.
The cost differences reported here are from the public payer perspective; however, not all public payer costs could be assessed in the current study, such as diagnostic, investigational, and screening services, home care, and ambulance services, and this is an area of future research. Conversely, Kellar et al. assessed private payer costs associated with treatment-resistant and non-treatment-resistant depression in Canada.31 Collectively, prescription drug, short-term disability, and long-term disability claims were $5,987 higher among employee claimants with treatment-resistant depression. Therefore, the economic burden of depression is evident for both public and private payers, emphasizing the need for better treatment approaches to help patients reach remission and recovery sooner.
There were several limitations to the design of this study that should be considered when interpreting the results. We used criteria to identify patients with clinically significant depression that required patients to have at least one hospitalization, or five physician visits for depression, or one physician visit plus seven or more prescriptions for antidepressants in 2 consecutive years. This resulted in the exclusion of less severe cases of depression from the depression cohort. Furthermore, patients in the nondepression cohort could enter the depression cohort if they later met criteria for depression during the study period. Therefore, some patients may have had undiagnosed depression when they were included in the nondepression cohort. It is also possible that a patient with depression prior to the study period may be initially classified in the nondepression cohort if they were already taking a medication but had not had a physician visit or hospitalization within the study period. However, it is unlikely that this patient would not have any depression-related physician visit and remain in the nondepression cohort for the duration for the study.
Additionally, health outcomes, health-care utilization, and costs may be higher in the nondepression cohort than in a typical healthy control population because our nondepression cohort had at least some contact with the Manitoba health-care system during the study period: 97.1% of the nondepressed patients had at least one physician visit during the study period. In Canada, only 73.8% of the population report having any contact with a physician in the past 12 months and our study was enriched for this treatment-seeking subset of the population.32 Thus, the difference in health outcomes, health-care utilization, and costs between patients with depression and without (i.e., healthy controls) would likely be larger if both healthy controls with and without contact with the health-care system could be considered. Furthermore, sensitivity analyses excluding patients who died during the study period were not performed, which is a limitation to the current study. The mortality rate in the depression cohort is considerably higher than the nondepression cohort. Earlier deaths may be associated with lower health care use and costs in the depression cohort. However, the last year of life is associated with higher health-care costs; therefore, this may increase the costs in the depression cohort.
Finally, the criteria used to derive the depression cohort included both the ICD-9 outpatient physician claim diagnostic code for Depressive disorder not elsewhere classified (311) and Episodic mood disorders (296). The 296 code does not differentiate between depression and bipolar disorder, which means the latter may have been included in the cohort, possibly inflating estimates of intentional self-harm.33 That said, bipolar disorder is much less common than depression, and likely 20% or less of the use of this diagnostic code will refer to a bipolar diagnosis.34 Additionally, we included the ICD-9 diagnostic code 298 for outpatient claims. Although the depression cohort may include a small number of patients with nondepressive psychosis due to this inclusion, this code was incorporated into our definition of depression as nearly 20% of subjects meeting criteria for a major depressive episode reported psychotic features in a study of the general population.35
Conclusions
The present study characterized the burden of depression in Manitoba, Canada. We have reported prevalence and incidence of depression to be 12.5 and 10.9, respectively. All measures of health and economic burden were highest among patients with depression than in matched nondepressed patients. The higher health-care and social service utilization among patients with depression were associated with 3.5 times greater costs than nondepressed patients. Based on our data, we project that each year depression costs the province of Manitoba and Canada an additional $400 million and $12 billion, respectively. Due to limitations of available data, this is likely an underestimation of the true economic burden of depression. Nonetheless, better services and treatment approaches for depression are needed to allow more patients to achieve remission and recovery sooner, resulting in improved overall health and a lower economic burden.
Supplemental Material
Supplemental Material, Mulsant_Supplementary_Materials for Economic Burden of Depression and Associated Resource Use in Manitoba, Canada by Julie-Anne Tanner, Jennifer Hensel, Paige E. Davies, Lisa C. Brown, Bryan M. Dechairo and Benoit H. Mulsant in The Canadian Journal of Psychiatry
Acknowledgments
We acknowledge the contributions of Salah Mahmud, Christiaan Righolt, Barret Monchka, and Geng Zhang.
Authors’ Note: The data used in this study were obtained from linked data sets contained in the Population Health Research Data Repository held at the Manitoba Centre for Health Policy (MCHP). All data are summarized in the Vaccine and Drug Evaluation Centre report, available publicly at https://vdec.ca/documents/Final-report-mood-disorders-v1.1.1.pdf. The conclusions are those of the authors, and no official endorsement by the MCHP, Manitoba Health, or other data providers is intended or should be inferred.
Declaration of Conflicting Interests: The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: JAT, PED, LCB, and BMD were employed by Assurex Health/Myriad at the time of the study and received stock options as part of their compensation. During the past 5 years, BHM has received research funding from Brain Canada, the CAMH Foundation, the Canadian Institutes of Health Research, and the U.S. National Institutes of Health (NIH); research support from Bristol-Myers Squibb (medications for a NIH-funded clinical trial), Eli-Lilly (medications for a NIH-funded clinical trial), Pfizer (medications for a NIH-funded clinical trial), Capital Solution Design LLC (software used in a study funded by CAMH Foundation), and HAPPYneuron (software used in a study funded by Brain Canada). The remaining authors declare no conflicts of interest.
Funding: The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Data analysis was funded by Assurex Health (Assurex Health is now Myriad Neuroscience). Funding during manuscript preparation was supported by Assurex Health and a Mitacs Elevate Postdoctoral Fellowship (JAT).
ORCID iD: Julie-Anne Tanner, PhD 
https://orcid.org/0000-0003-3460-9618
Supplemental Material: Supplemental material for this article is available online.
References
- 1. Ratnasingham S, Cairney J, Manson H, Rehm J, Lin E, Kurdyak P. The burden of mental illness and addiction in Ontario. Can J Psychiatry. 2013;58(9):529–537. [DOI] [PubMed] [Google Scholar]
- 2. Lim KL, Jacobs P, Ohinmaa A, Schopflocher D, Dewa CS. A new population-based measure of the economic burden of mental illness in Canada. Chronic Dis Can. 2008;28(3):92–98. [PubMed] [Google Scholar]
- 3. Claire de Oliveira JC, Rehm J, Kurdyak P. The role of mental health and addiction among high-cost patients: a population-based study. J Med Econ. 2018;12(4):348–355. [DOI] [PubMed] [Google Scholar]
- 4. de Oliveira C, Cheng J, Vigod S, Rehm J, Kurdyak P. Patients with high mental health costs incur over 30 percent more costs than other high-cost patients. Health Aff (Project Hope). 2016;35(1):36–43. [DOI] [PubMed] [Google Scholar]
- 5. Hensel JM, Taylor VH, Fung K, Vigod SN. Rates of mental illness and addiction among high-cost users of medical services in Ontario. Can J Psychiatry. 2016;61(6):358–366. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Chiu M, Vigod S, Rahman F, Wilton AS, Lebenbaum M, Kurdyak P. Mortality risk associated with psychological distress and major depression: a population-based cohort study. J Affect Disord. 2018;234:117–123. [DOI] [PubMed] [Google Scholar]
- 7. Institute for Health Metrics and Evaluation; [accessed Jan 21, 2019]. http://wwwhealthdataorg/canada.
- 8. Rizvi SJ, Cyriac A, Grima E, et al. Depression and employment status in primary and tertiary care settings. Can J Psychiatry. 2015;60(1):14–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Rotella F, Mannucci E. Depression as a risk factor for diabetes: a meta-analysis of longitudinal studies. J Clin Psychiatry. 2013;74(1):31–37. [DOI] [PubMed] [Google Scholar]
- 10. Patten SB, Williams JV, Lavorato DH, Modgill G, Jette N, Eliasziw M. Major depression as a risk factor for chronic disease incidence: longitudinal analyses in a general population cohort. Gen Hosp Psychiatry. 2008;30(5):407–413. [DOI] [PubMed] [Google Scholar]
- 11. Chiu M, Lebenbaum M, Cheng J, de Oliveira C, Kurdyak P. The direct healthcare costs associated with psychological distress and major depression: a population-based cohort study in Ontario, Canada . PLoS One. 2017;12(9):e0184268. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Slomp M, Jacobs P, Ohinmaa A, et al. The distribution of mental health service costs for depression in the Alberta population. Can J Psychiatry. 2012;57(9):564–569. [DOI] [PubMed] [Google Scholar]
- 13. Roos LL, Gupta S, Soodeen RA, Jebamani L. Data quality in an information-rich environment: Canada as an example. Can J Aging. 2005;24(Suppl 1):153–170. [DOI] [PubMed] [Google Scholar]
- 14. Martens P, Nickel M, Forget E, et al. The cost of smoking: A Manitoba Study; 2015. [accessed May 1, 2019]. http://mchp-appserv.cpe.umanitoba.ca/reference/Cost_Smoking_final.pdf.
- 15. Kessler RC, Bromet EJ. The epidemiology of depression across cultures. Annu Rev Public Health. 2013;34:119–138. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Marrie RA, Fisk JD, Yu BN, et al. Mental comorbidity and multiple sclerosis: validating administrative data to support population-based surveillance. BMC Neurol. 2013;13:16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Rothman KJ, Greenland S, Lash TL. Modern epidemiology. Vol. 3 Philadelphia, PA: Wolters Kluwer Health/Lippincott Williams & Wilkins; 2008. [Google Scholar]
- 18. Lix L, Yogendran M, Mann J. Defining and validating chronic diseases: an administrative data approach. Manitoba, Canada: Manitoba Centre for Health Policy; 2008. [Google Scholar]
- 19. Brownell MD, Yogendran MS. Attention-deficit hyperactivity disorder in Manitoba children: medical diagnosis and psychostimulant treatment rates. Can J Psychiatry. 2001;46(3):264–272. [DOI] [PubMed] [Google Scholar]
- 20. Metge C, Black C, Peterson S, Kozyrskyj AL. The population’s use of pharmaceuticals. Med Care. 1999;37(6 Suppl): JS42–JS59. [DOI] [PubMed] [Google Scholar]
- 21. Clark A. Income security to end poverty in Manitoba. Manitoba, Canada: Canadian Centre for Policy Alternatives; 2017. [Google Scholar]
- 22. Khan O, Blanco A, Conrad P, et al. Performance of Lynch syndrome predictive models in a multi-center US referral population. Am J Gastroenterol. 2011;106(10):1822–1827; quiz 1828. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Balmana J, Stockwell DH, Steyerberg EW, et al. Prediction of MLH1 and MSH2 mutations in Lynch syndrome. JAMA. 2006;296(12):1469–1478. [DOI] [PubMed] [Google Scholar]
- 24. Mercado RC, Hampel H, Kastrinos F, et al. Performance of PREMM(1,2,6), MMRpredict, and MMRpro in detecting Lynch syndrome among endometrial cancer cases . Genet Med. 2012;14(7):670–680. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Harris EC, Barraclough B. Suicide as an outcome for mental disorders. a meta-analysis. Br J Psychiatry. 1997;170:205–228. [DOI] [PubMed] [Google Scholar]
- 26. Jasperson KW, Tuohy TM, Neklason DW, Burt RW. Hereditary and familial colon cancer. Gastroenterology. 2010;138(6):2044–2058. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Dennehy EB, Robinson RL, Stephenson JJ, et al. Impact of non-remission of depression on costs and resource utilization: from the COmorbidities and symptoms of DEpression (CODE) study. Curr Med Res Opin. 2015;31(6):1165–1177. [DOI] [PubMed] [Google Scholar]
- 28. Wong J, Motulsky A, Abrahamowicz M, Eguale T, Buckeridge DL, Tamblyn R. Off-label indications for antidepressants in primary care: descriptive study of prescriptions from an indication based electronic prescribing system. BMJ. 2017;356:j603. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Freedman AN, Wideroff L, Olson L, et al. US physicians’ attitudes toward genetic testing for cancer susceptibility. Am J Med Genet A. 2003;120A(1):63–71. [DOI] [PubMed] [Google Scholar]
- 30. Patten SB, Williams JV, Lavorato DH, Wang JL, McDonald K, Bulloch AG. Major depression in Canada: what has changed over the past 10 years? Can J Psychiatry. 2016;61(2):80–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Kellar J, von Heymann C, Zingaro J, Kuriakose B, Li A, Mittmann N. Private-payer costs for employees with treatment-resistant depression. Am J Pharm Benefits. 2016;8(4):e67–e74. [DOI] [PubMed] [Google Scholar]
- 32. Sabatino SA, McCarthy EP, Phillips RS, Burns RB. Breast cancer risk assessment and management in primary care: provider attitudes, practices, and barriers. Cancer Detect Prev. 2007;31(5):375–383. [DOI] [PubMed] [Google Scholar]
- 33. Webb RT, Lichtenstein P, Larsson H, Geddes JR, Fazel S. Suicide, hospital-presenting suicide attempts, and criminality in bipolar disorder: examination of risk for multiple adverse outcomes. J Clin Psychiatry. 2014;75(8):e809–e816. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Moreno C, Hasin DS, Arango C, et al. Depression in bipolar disorder versus major depressive disorder: results from the National Epidemiologic Survey on alcohol and related conditions. Bipolar Disord. 2012;14(3):271–282. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Ohayon MM, Schatzberg AF. Prevalence of depressive episodes with psychotic features in the general population. Am J Psychiatry. 2002;159(11):1855–1861. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplemental Material, Mulsant_Supplementary_Materials for Economic Burden of Depression and Associated Resource Use in Manitoba, Canada by Julie-Anne Tanner, Jennifer Hensel, Paige E. Davies, Lisa C. Brown, Bryan M. Dechairo and Benoit H. Mulsant in The Canadian Journal of Psychiatry