Background
Mental disorder is one of the leading causes of morbidity worldwide. Its cost and negative impact on productivity are substantial. Consequently, improving mental health-care system efficiency - especially service utilisation - is a priority. A systematic literature review reveals that prevalence rates at 12 months and lifetime are as follows: 10.6% and 16.6%, respectively, for anxiety disorders [
1]; 4.1% and 6.7% for major depressive disorders [
2]; 6.6% and 13.2% for alcohol use disorders; and 2.4% both in the case of drug use disorders [
3]. Mental disorders are frequently associated with alcohol or drug use disorders. The U.S. National Comorbidity Surveys evaluated that 42.7% of respondents with alcohol or drug disorder also had a mental disorder in the 12 previous months, and 14.7% a mental disorder along with alcohol or drug disorder [
4].
Risk factors and correlates to mental or substance use disorders have also been extensively investigated [
5‐
8]. Age, gender, income, and marital and employment status are the principal socio-demographic factors associated with the presence of mental disorder. Being female, middle-aged, widowed, separated or divorced and a low-income earner increases the risk of major depressive disorder [
6]. A systematic literature review showed that anxiety disorders were approximately twice as prevalent among females [
1]. For substance use disorders, studies reveal a generally greater prevalence among males and youths [
3].
Mental health-care service use has also been the subject of many epidemiological studies. The most frequently used model for identifying factors associated with service use is Andersen's behavioural model which classifies predictors of service use into three categories: predisposing, enabling, and needs-related factors [
9]. Predisposing factors are individual characteristics that existed prior to the illness such as age, gender, language, marital status, race/ethnicity, and country of birth. Several studies have found that people aged 25 to 44 [
10‐
12], females [
10‐
17], previously married [
12,
15,
16,
18,
19], highly educated [
18,
20], white [
11,
15,
21], and native-born [
22,
23] are most likely to use health-care services. Enabling factors refer to features that influence care delivery and attitudes toward care; they encompass variables such as income, social support, and geographical location. The most important enabling factor is income. People with more elevated socio-economic status tend to use psychiatric and psychological care more assiduously, even among individuals with the same insurance coverage [
20,
24‐
26]. Finally, needs-related factors include assessments of physical and mental health by patients and professionals, including diagnosis, severity of the disorder, and perceived needs. Depressive disorders [
20] and anxiety, particularly panic disorders [
27,
28] are strong predictors of health service use.
Utilisation of services has also been studied with regard to the use of primary mental health-care (e.g. general practitioners) or specialised mental health-care (e.g. psychiatrists). Individuals who use primary care are mainly female, older, more highly educated, live with a spouse or partner, and generally have anxiety or depressive disorders [
29]. Conversely, frequent users of psychiatric services (second- or third-line services) are generally male, young or middle-aged, unemployed, live alone, have low social support and, often, a dual diagnosis of mental and substance use disorder [
13,
30,
31]. Finally, youths and people with substance use disorders only use few health-care services [
32,
33].
DSM-IV, the most widely recognised mental health classification system, provides a detailed clinical profile of mental disorder; however, its use in forecasting needs or health-care service utilisation is limited [
34]. An alternative classification suggests that mental health-care users can be described in terms of clusters based on several characteristics. With the use of clusters, persons may be considered in broad terms; in addition, subgroups may be correlated with clinical and socio-demographic variables and patterns of service use [
34].
Cluster analysis has mainly been used to create typologies of patients with serious mental disorders [
34]. Studies have identified frequent users of in-patient services [
35,
36], patients with schizophrenia treated in the community [
37], patients with serious mental illness according to their level of functioning [
38], patients with dual diagnoses of serious mental disorders and substance use [
39] first-ever admitted psychiatric in-patients [
40]. Very few studies have explored subgroups of individuals with common mental disorders. An exception is a study by Mitchell and colleagues [
41] that used cluster analysis to classify adults with problematic online experiences and conventional problems (mental and physical health problems; family and/or other relationship problems; victimisation; aggressive behaviour) from a clinical perspective. Identifying individuals with common socio-demographic and clinical characteristics and patterns of service use, however, is essential to the efficacious planning of mental health-care delivery [
38].
In an effort to enhance knowledge of service needs profiling, this Canadian urban catchment area study (258,000 persons served by a psychiatric hospital) includes a typology of individuals diagnosed with mental disorder in a 12-month period based on their individual characteristics and use of services. Variables used in the cluster analysis are based on Andersen's behavioural model [
9], which considers that health-care service use is determined by predisposing, enabling, and needs-related factors.
Discussion
The study was designed to develop a typology of individuals, diagnosed with mental disorder during a 12-month period, based on individual characteristics and use of services. Its purpose is to generate knowledge on service needs profiling in support of efforts to facilitate mental health-care service planning. Mean prevalence of mental disorder in the last 12 months was 17%. According to epidemiological studies, the prevalence of mental disorder varies widely from country to country. In the International Consortium in Psychiatric Survey (ICPE), which focused on seven countries, the overall prevalence at 12 months was 29% in USA and 20% in Canada, as against 8% for Turkey [
43]. In a recent study based of a sample of more than 21,000 adults representative of the overall population in six European countries, Alonso and Lépine [
51] estimated the proportion of people affected by a mental disorder in the 12 previous months to be 11.5%. More recently, a meta-analysis of 27 studies estimated at 27% the proportion of European adults with at least one mental disorder in the 12 previous months [
52]. Differences in survey methods or instruments may account in part for these considerable variations [
5,
51,
53]. Some disorders (e.g. bipolar disorders and drug dependence in Western European countries) were not assessed in all the surveys [
53]. A greater reluctance toward participation or admission of mental illness in some countries and/or interviewer error are other possible biases that can explain the under- or over-estimation of the prevalence of mental disorders [
53,
54].
In our study, almost 50% of persons affected in the last 12 months by a mental disorder used health-care services for mental health reasons. In comparison to other studies, this number is high. According to the 2002 Canadian Community Health Survey Mental Health and Well-Being (CCHS 1.2), only 38.5% of Canadians used services for mental health reasons, when at least one mental disorder was present [
55]. In the 1997 Australian National Survey of Mental Health and Well-Being, only 35% of people with at least one mental disorder sought professional help [
56]. In our study, the proximity of a psychiatric hospital may account for the more assiduous use of mental health-care services. Globally, studies demonstrate that health-care services are underused by individuals with mental disorder. In Quebec and the rest of Canada, where public services are focused on the treatment of serious mental disorders and since psychological services are only partially available in the public health-care system, access to treatment for people with less severe disorders and without private insurance and/or with low income is limited. These situations explain in part the underutilisation of services for mental health reasons.
In our study, individuals who used services for mental health reasons received two services, on average, mainly from primary and specialised care providers, and close to 20% consulted at least four mental health-care professionals. The use of diverse and increasingly community-based professionals is perceived as a positive development by various authors [
57,
58]. Individuals who receive dual-modality treatment (e.g. psychopharmacology and psychotherapy) are less likely to abandon treatment than people who consult psychiatrists only [
58]. However, individuals with low income do not have easy access to psychologists (a service that often requires private insurance). A recent study has shown that cost is the main obstacle to psychotherapy access, especially for people with anxiety disorders [
59].
Cluster analysis yielded four user profiles including people with mainly anxiety disorders (Cluster 1), depressive disorders (Cluster 3), alcohol and/or drug disorders (Cluster 4), and multiple mental and dependence disorders (Cluster 2).
Two clusters (1 and 3) were more closely associated with females. In the other clusters (2 and 4), males were over-represented in comparison with the sample as a whole. It is interesting to note that these two clusters (where males are a majority) are linked to dependence disorders, regardless of association with mental disorders. The socio-demographic variables associated with the various clusters were consistent with previous studies on the prevalence and correlates of anxiety, mood, and dependence disorders in the general population. Our results show that anxiety [
1] and depressive disorders [
6] are more prevalent among females, and dependence disorders, principally alcohol dependence, and concurrent disorders more common among males [
3].
In three clusters, users aged 34 years or less were over-represented. Young adulthood is a critical life stage at which people leave the family home and may make far-reaching decisions with respect to education, career or parenthood [
60]. Generally, mental disorders begin during this period [
7]. Most general population surveys have found a marked prevalence of mental disorder in young adulthood. For example, in the National Comorbidity Survey Replication study (NCS-R), three-quarters of lifetime mental disorders emerged by age 24 [
60]. Youth is also associated with a greater risk of hospital readmission [
13,
61]. Conversely, age is a protective factor, with the risk of mental disorder decreasing as one gets older [
62].
Age of onset may account for the differences between Clusters 1 and 3. Cluster 1 included young women with anxiety disorders for the most part, though one half did experience a major depressive disorder. Conversely, Cluster 3 includes mainly middle-aged women who had only one major depressive episode without anxiety disorder in the last 12 months. It is possible that users in Cluster 3 successfully negotiated young adulthood and entered the job market or started a family without experiencing anxiety disorder or were successfully treated for it. Another explanation is that the major depressive episode occurred recently. If anxiety disorder tends to occur among younger individuals, mood disorder (including major depressive episodes) tends to occur among older individuals [
63]. According to one study [
6], lifetime rates and the probability of major depressive disorder are higher among baby-boomers than younger adults.
Co-morbidity with mental disorder appears to be the norm [
64,
65]. In three of four clusters, users were affected by more than one mental or dependence disorder. In Cluster 2, co-morbidity with mood disorder, anxiety, and dependence disorder was very frequent. Several studies indicate a significant correlation between alcohol dependence and depression [
66‐
68]. Intoxication by alcohol or drugs can induce symptoms similar to those of depressive disorder [
66]. Some drugs can increase stress and provoke panic attacks or other anxiety disorders [
4]. It is also known than several people with anxiety or mood disorder use alcohol or drugs as self-medication [
69]. Users with dual diagnoses present a challenge for mental health and addiction services and generally have worse treatment outcomes [
64,
67]. Co-morbid disorders are generally more chronic than pure mental disorders, and treatment is less effective [
4]. Cluster 4 was distinguished from Cluster 2 by the absence of mania, major depressive episodes, and anxiety. Cluster 4 was characterised by dependence disorder, combined with more marginal mental disorder, and exhibited a propensity for alcohol, rather than drug, use. Several studies have revealed that alcohol dependence is generally not as strongly associated with mental disorder as is drug dependence [
64,
67,
70,
71].
Clusters 1 and 2 were the most keenly affected by psychological distress. Conversely, Cluster 4, which encompassed only 1.1 mental or dependence disorders per subject, was not as greatly impacted by psychological distress as other clusters. These results seem to confirm that the number of mental disorders is associated with greater psychological distress [
64]. We may assume that multiple mental or dependence disorders can affect several domains essential to quality of life (work, daytime activities, social and intimate relationships, physical health, etc.).
Finally, Clusters 1 and 2 featured the greatest number of mental disorders per subject and the most frequent use of mental health-care services. According to several studies, needs factors are the prime predictors of service use [
9,
72,
73], and greater numbers of mental disorders result in more frequent use of health-care services [
64]. In addition, some authors have suggested that users with multiple mental and dependence disorders feel a greater impetus to seek treatment [
64,
74]. However, it is interesting to note that the mean number of health-care services used by Cluster 3, with only one mental disorder, is quite similar to that of Cluster 1. One possible explanation is that participants in Cluster 3, with the highest household income, make more frequent use of private psychologists to treat major depressive disorders. For Vasiliadis and colleagues, depression is the most significant predictor of service use [
20]. Gender may also account for similarities between Clusters 1 and 3 regarding service utilisation. It is known that females use more health-care services in general, mainly general practitioners and other primary care services. Age may also explain the differences: middle-aged persons are the peak users of mental health-care services [
75]. Conversely, younger people are less likely to perceive their need for treatment and often wish to solve problems on their own [
32]. Young adults are also more likely to drop out of treatment [
58]. Finally, participants in Cluster 4 use relatively few health-care services. This is the case for people affected only by substance disorders [
33]. They are usually less likely to think they need help than participants with co-morbid mental disorders [
74].
This study has some limitations. First, information was not available on participants' physical condition. Several epidemiological studies have reported that people with mental health disorders or dependence disorders often also have significant physical disorders, such as hypertension, diabetes or epilepsy [
76,
77]. The presence of a physical disease may account for more frequent use of health-care services. Second, our study did not include the full spectrum of psychiatric disorders, e.g. schizophrenia and other serious mental disorders, organic mental disorders, sexual disorders, eating disorders, personality disorders, and intellectual deficiencies. Several studies have reported a prevalence of personality disorders with anxiety, depression or substance-use disorders. Near half the people with a current mood or anxiety disorder have at least one personality disorder [
78]; identifying these disorders would have allowed us to refine our cluster analysis. Finally, the severity of mental disorder was not considered. Previous studies have reported that severe cases use more services than mildly severe or moderate ones [
33].
Conclusion
The study found considerable heterogeneity among socio-demographic characteristics, number of disorders, and number of health-care services used by individuals with mental or dependence disorders. Overall, there is significant underutilisation of mental health-care services with female consuming more services than men. When individuals sought services for mental health reasons, they generally saw more than one provider and used both primary and specialised mental health-care. As services are underutilised and mental disorders vary with regard to gender, age, and other characteristics, it is crucial to develop treatment modes or service programs that target specific mental disorder profiles. Our study reveals the existence of four subgroups of users with mental disorders: 'young females with anxiety disorders'; 'young low-income earners with multiple mental and dependence disorders'; 'middle-aged, high-income females with depressive disorders'; and 'young low-income earners with dependence disorders'. The second group exhibited the most socio-economic vulnerability and most frequent service utilisation.
Along with the need to target the four subgroups above with specific programs, our study highlights the relevance of focusing on younger individuals affected by multiple mental disorders or anxiety disorders concurrent with or without major depressive disorders. As concomitant problems are frequent among people with mental disorders, psychological services and/or addiction programs must also be taken into consideration as components of integrated programs or shared-care initiatives when planning treatment. In addition, as males seem to consult only when they suffer multiple mental and substance disorders, more outreach and promotion programs are needed to detect and facilitate mental health-care service access for them. Globally, public education on drug use and mental disorders and programs specially designed for youths and/or males may reduce this clientele's reticence to use health-care services. Integrating motivational and cognitive aspects of behavioural change in professional training may also lead to greater mental health-care utilisation. At last, enabling greater collaboration within the health-care system between professionals (including general practitioners) and programs, especially with regard to mental disorders and substance abuse, should lead to more timely and appropriate care.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
MJF, GG and MP designed the study. JMB carried out the statistic analyses with assistance from JC. MJF and GG wrote the article. All authors have read and approved the final manuscript.