Assessing the Costs and Benefits of Insuring Psychological Services as Part of Medicare for Depression in Canada

In North America, up to 20% of the population report having a common mental disorder, such as depression, in the course of their lifetime (1). Depression has been associated with decreased functional status and low perceived quality of life (2–4), mortality (5,6) and suicide rates that are four times higher than in the general population (2), and increased health system costs (7–9).

Although cost-effective psychological treatments for depression exist (10–12), data from the Canadian Community Health Survey (CCHS), conducted in 2012, showed that 39% of adults with major depression reported an unmet need for mental health care, with 71% reporting a need for counseling (13). A recent report by Quebec’s health commissioner on the performance of the health system clearly highlighted inequity in access to effective mental health services, such as psychotherapy (14). Although most patients prefer psychotherapy over antidepressants for depression (15), socioeconomic barriers prevent use of specialty providers of psychotherapy, such as psychologists, in primary care (16). The cost of psychological services is the primary concern (17).

In Canada, the family physician is the most frequently consulted professional for mental health reasons (18). Although most physicians are aware of evidence-based psychological interventions, they have reported not being comfortable in providing such interventions because of time constraints or inadequate training or because they preferred not to (19). The physicians cited cost as the largest barrier to referring patients to psychologists, because services are not covered by public health insurance.

Governments are currently struggling to decide whether to reimburse access to psychological services. To date, the United Kingdom and Australia have decided to reimburse psychological therapies in their health systems. In the United Kingdom, Layard and colleagues’ (20,21) cost-benefit analysis of the Improving Access to Psychological Therapies (IAPT) program showed that the program would pay for itself within five years. In France, a similar study estimated a cost-benefit ratio ranging between 1.14 and 1.95 (22).

In the United Kingdom, the economic modeling exercise (20) was helpful in the decision-making process involved in creating the IAPT program. Currently, Canada is examining how best to use the federal transfer funds for health in 2017 (23), and in the province of Quebec, the health technology assessment agency is working on a report on equitable access to psychological services. Therefore, the objective of this study was to evaluate the cost-effectiveness associated with rendering psychological services as part of Canadian Medicare for individuals with unmet mental health needs. This study provides important health system policy information for decision makers on how best to allocate resources for a more efficient public managed health care system.

Methods

Study Cohort

Using discrete event simulation (DES), we modeled the evolution of depression among patients over a 40-year period to assess the cost-effectiveness of increasing publicly funded access to psychotherapy in Canada, compared with the status quo. We included in the model incident cases of depression (2.9% of the population) among persons ages 20 to 85 and estimated patterns of past-year health service use for mental health reasons from the 2012 CCHS–Mental Health (24).

Model Overview

DES modeling with the ARENA–Rockwell software program was used to depict the disease progression of depression (25). As shown in studies by Le Lay and colleagues (26) and by Vataire and colleagues (27), this method is flexible and allows for an accurate depiction of the evolution of depression, a disease that is characterized by multiple episodes, with prognosis depending on disease history. The main advantage of DES is its flexibility in allowing individuals with different attributes to move from one event to another in sequential order while the model simultaneously takes into account important risk factors, such as age, disease history, and time to event (26). In DES, the model considers time-dependent events that trigger disease evolution and changes in health states (26,28). [More information about use of DES in this study is provided in an online supplement to this article.]

Model Parameters

We designed the model to predict the health outcomes and costs associated with different pathways of care received after an incident episode of depression. We modeled outcomes on the basis of type of health service recieved and the health events with which it was associated (Figure 1). The probability of events and model parameters were based on secondary analyses of data captured in the 2012 CCHS mental health survey carried out by Statistics Canada (24) (Table 1), as well as on epidemiologic, pharmacologic, cost (direct and indirect), and utility estimates, for the calculation of quality-adjusted life years, from the literature. [Tables in the online supplement present parameter estimates from the published literature.] The CCHS mental health data provided population estimates of major depression and prevalence of health service use for mental health reasons by community-dwelling adults ages 15 and older who were living in the ten provinces, representing 28.3 million Canadians. We also determined the past-year and lifetime prevalence estimates of major depression, hospitalizations, and suicide attempts.

Among respondents with past-year major depression (4.6% of the Canadian population), we assessed past-year health service use for mental health reasons (yes or no), by type of provider consulted and number of visits and antidepressant use. The following categories were studied: general practitioner or family physician (GP/FP) only, psychiatrist only, GP/FP and psychiatrist, other mental health specialist only (psychologist, psychotherapist or counselor, social worker, and nurse), GP/FP and other mental health specialist, psychiatrist and other mental health specialist, and GP/FP and psychiatrist and other mental health specialist. Average fees for each type of service were ascertained from national (29) and provincial billing manuals (30,31). The prevalence and annual costs of antidepressant use were also based on previous estimates (32).

Respondents who had used services in the past year were further categorized according to the CCHS data as receiving adequate mental health services (yes or no). Adequate mental health services were defined according to Canadian recommendations for the treatment of depression (12) as either 12 or more psychotherapy consultations with a psychiatrist, a GP/FP (at least 25 minutes), or another mental health specialist or use of an antidepressant with four visits to a GP/FP or a psychiatrist.

The Canadian population status quo (base-case scenario) was defined according to the results obtained from analyses of the CCHS data (Table 1). Briefly, among individuals who reported a major depressive episode, 64.3% had used services for mental health reasons, whereas 36.7% had not. Among these service users, 67.4% had received adequate care as defined above. Data on patterns of use of various provider types are presented in Table 1.

The alternative scenario of increased access to care was also based on analyses of 2012 CCHS data on unmet mental health needs, which showed that 17.0% and 12.7% of respondents with major depression reported a partially met or an unmet mental health need, respectively. Seventy-one percent of needs were for counseling (13), and in the increased-access scenario, this would result in 9% (12.7% × 71%) of past-year nonusers becoming users and 12.1% (17% × 71%) of past-year users who had not received adequate care receiving such care. Following the Australian GP gatekeeper model of psychological services (33), all these users were assigned to the GP/FP user group and to the group that used other mental health specialists. The population prevalence rates of health service use by provider categories in this alternative scenario are presented in Table 1. [Information about how estimates were drawn and the data sources used is included in the online supplement (15,23,24,27,32,34–64).]

Economic Analyses

Costs were assessed from a health system perspective and from a societal perspective; the latter included costs associated with productivity losses. Costs were then incorporated into an incremental cost-effectiveness analysis, which compared the status quo with coverage of psychotherapy as part of publicly funded psychological therapy services for incident cases of depression in which individuals had not received adequate care.

The improved-access scenario was based on results of the Better Access initiative in Australia (33) and the IAPT program in the United Kingdom (65,66), in which GP/FPs or psychiatrists follow and refer individuals to receive publicly funded psychological services. In those programs, adequate care was considered to consist, on average, of two visits with a GP/FP and eight sessions with a mental health professional; sensitivity analyses considered a lower estimate of two GP/FP visits and four sessions with a mental health professional and a higher estimate of four GP/FP visits and 12 sessions with a mental health professional. Medical, nonmedical, and productivity costs were tracked. Utility values for the depression and suicide-related health states (attempted suicide and suicide) were used to calculate quality-adjusted life years. Sensitivity analyses were also carried out to test the robustness of the results to changes in assumed data values with possible ranges of the cost data (lower and upper limits) and less conservative utility estimates.

Finally, to more accurately represent a dynamic adult population of incident and prevalent cases of depression, the ARENA program was run for a total of 105 years after which the last 40-year period, representing a steady-state picture of a population with depression, was retained in the cost-effectiveness analysis.

Results

The results of the 40-year model of the base-case scenario and the alternative scenario, in which there was increased access to psychological services, are presented in Table 2. A preliminary validation of the base-case scenario showed that the lifetime prevalence of suicide attempts and hospitalizations for mental health reasons among respondents with depression reached 14.6% and 30.2%, respectively. These results are similar to those from analyses of mental health data from the 2012 CCHS, which found that lifetime rates of attempted suicides and hospitalizations among respondents with major depression in the past year reached 14.1% and 29.9%, respectively.

In addition, in the base-case model, there were 250 suicides over the 40-year period, which translates into 9.1 suicides per 100,000 population per year. Earlier published results indicated that the percentage of suicides attributable to the population of individuals with mental or substance use disorders reached 74% (67). The suicide rate in the Canadian population for persons ages 20 to 85 was close to 12.1 per 100,000 in 2011 (68); it can be estimated that 8.9 suicides per 100,000 are attributable to affective disorders, such as depression, a rate similar to the rate that we observed in the model.

In the increased-access scenario, a significant decrease was noted in the number of days suicidal and depressed, with an increase in the number of days in a well or healthy state (Table 2). This translated into an average benefit of .17 quality-adjusted life years gained per person. The increased-access model also showed a decrease in health system and societal costs. The results indicated that by increasing access to psychotherapy, the alternative scenario dominated the status quo in that it was more effective and less costly (Table 3). The sensitivity analyses showed that the most important influence on the results was the use of the friction cost method and the human capital approach in assessing loss of productivity.

Discussion and Conclusions

This 40-year, prospective-value implementation study showed that rendering psychological services as a publicly funded treatment in Canada for adults with depression reporting an unmet mental health need leads to improved health-related quality of life and cost savings to the health system and society. Respondents with a reported unmet mental health need, mainly for counseling, who have not received adequate mental health care for their depression represent 8.1% of prevalent cases of depression in Canada. Increasing access and rendering publicly funded psychological services for this adult population would translate into additional costs, on average, of $123,212,872 CAD ($1,292 [two GP/FP visits plus eight psychotherapy sessions in a year] per individual multiplied by 95,366 individuals with unmet needs) (range $67,709,860–$190,922,732). The savings to society, calculated by using the human capital approach, would reach, on average, $246,997,940 CAD (range $120,733,356–$602,713,120).

These estimates need to be interpreted in light of the study’s limitations. First, the unit cost analyses were based on national data and the costs are for the province of Quebec, which may limit generalizability to other countries with public health care systems similar to Canada’s (69). Second, the results presented assume that eight psychotherapy sessions and two physician visits would, on average, be covered in the scenario of increased access to psychotherapy (20). Previous publications in Australia have reported that most people use six sessions, with improved outcomes on the ten-item Kessler Psychological Distress Scale and the 21-item Depression and Anxiety Stress Scale (33). Third, not all individuals with a need for mental health care would avail themselves of the services (20), and thus the average costs associated with increased access may be an overestimate. Inversely, one may argue that individuals without a significant need for mental health care may receive treatment because of the greater availability of services. Canadian population data indicate that 8.1% of those who do not meet criteria for a mental or substance use disorder report a need for mental health counseling and may want to access such care (13). A previous study showed that individuals who did not report a common mental or substance use disorder had consulted a psychologist in the previous year, with a median of four visits (16). How this would affect the estimates of cost and quality-adjusted life years is difficult to predict. However, in this subgroup of individuals without a clinical diagnosis who express a need for counseling, consulting a mental health professional may be associated with a decrease in psychological distress, which may be accompanied by improved quality of life and increased productivity, even if only marginal, and which may prevent an episode of major depression.

Fourth, the model did not consider factors, such as gender and general medical comorbidities. In regard to gender, females are usually more likely to seek services for mental health reasons (16), and thus the health outcomes may be overestimated for males. In addition, treatment of mental health problems with evidence-based psychological interventions can also significantly benefit general medical conditions. Therefore, the study may have underestimated the potential gain in health-related quality of life (70) and the decrease in health service use by individuals with depression and chronic general medical illnesses (71), as reported elsewhere. Fifth, the model did not account for any positive or negative lifetime events (for example, a new job or the loss of a job or bereavement after the death of a family member), which may also affect the incidence and persistence of episodes. Sixth, although the analyses included the costs and decrements in utility associated with the prevalence and persistence of antidepressant use, the direct effect that publicly funded psychotherapy may have on patterns of prescription—for example, incidence and persistence of antidepressant use—is difficult to estimate and was not considered. Finally, most mental health problems start in childhood (72), and the study may have underestimated the potential benefits of providing services to children with depression, including cost savings to society.

If one were to consider covering psychotherapy sessions for all Canadian adults with incident depression (2.9%), the program costs would reach $1.008 billion CAD. This estimate is similar to those in previous reports. In Australia, in the first three years of the Better Access initiative, close to 2,017,000 individuals received close to 11.1 million psychotherapy sessions (six sessions per individual) (73), for a total budget, in Canadian dollars, of $2.80 billion or $1,514 per individual. More than half of individuals had never consulted a professional for their mental health problems. In the United Kingdom, the cost of the IAPT program in 2006 was estimated at £600 million to treat 800,000 patients opting for psychological therapy (21). Recently, the National Health Service invested an additional £400 million in the IAPT program to offer services to children, adolescents, older adults, and persons with severe mental illness and chronic general medical problems (74). In France, the estimated yearly cost of covering 12 psychological sessions for persons with depression and those with anxiety disorders (close to 2.3% of the population) was $729 million (in Canadian dollars) (22). This estimate is lower than in our study, because a psychotherapy session in France costs €41 ($65 CAD), and in Canada it costs an average of $145 CAD.

Finally, our results indicate that covering psychological services as part of Medicare for individuals with major depression who do not receive adequate care would pay for itself. In Canada, every $1 invested in the program would yield on average $2 ($1.78–$3.15) in savings to society. In the United Kingdom, the return on investment for the IAPT program reached £1.75 for every £1 in health care expenditures (20,65,66). In France, spending €1 in psychotherapy for people with depression yielded savings of €1.95 (€1.30–€2.60) (23).

To put the results of this study into perspective for decision makers, public expenditures for mental health and addictions account for only 7.2% of the total health budget in Canada, amounting to $540 per capita (75). An additional investment of $3.42 per capita, representing .07% of the total health care budget, would lead to a more efficient publicly managed health care system that would produce cost savings for society and improved indicators of mental health in the population, including reductions in disability, in premature loss of life, and in loss of productivity. Major depression has been identified as a leading cause of disease burden (76), and health planners and decision makers should aim to develop mental health policies that reduce income barriers to accessing psychotherapy services in primary care (16,17) for individuals who are not otherwise able to access such care.

The health policy of covering psychological services as part of Medicare in a public managed health care system may have important implications. In Australia, the increasing costs of the Better Access program caused great concern, and many argued that the program was widening the socioeconomic gap in mental health service use (77). However, Harris and colleagues (78) showed that over 90% of Better Access users had a 12-month ICD-10 mental disorder or other relevant indication. Those without an affective or anxiety disorder who used services were not from higher socioeconomic groups, and those with a mental disorder received significantly more services than those without.

Finally, a successful health policy that is aimed at improving access to mental health services and that is relevant to a public health care system should consider the following: access (who gets treatment?), equity, type of services offered, effectiveness and cost-effectiveness (are health outcomes improved at a sustainable cost?), and patient-centered care (does the service meet individual mental health needs?) (79). Future research should also focus on modeling the impact of improved access to psychological services for children, adolescents, and clinically complex populations with multiple morbidities for whom the return on investment may be more important compared with the adult population (80).