Gender differences in the use of psychiatric outpatient specialist services in Tromsø, Norway are dependent on age: a population-based cross-sectional survey

Population-based health surveys have been conducted in Tromsø since 1974 [25, 26]. The cross-sectional sixth Tromsø Study (Tromsø 6), consisting of two comprehensive self-administered questionnaires, clinical examinations and laboratory tests, was conducted from October 2007 to December 2008. Four groups were invited; every resident aged 40–42 or 60–87 years (n = 12,578), a 10 % random sample of individuals aged 30–39 (n = 1,056), a 40 % random sample of people aged 43–59 (n = 5,787) and all subjects who had attended the second visit of the fourth Tromsø Study, if not already included in the other three groups (n = 341). The sampling reflected the need for repeated measurements and follow-up, and the need to enrol new participants for new and ongoing projects.

Our data was collected from the two self-administered questionnaires. The first questionnaire was mailed with the invitation about two weeks ahead of the suggested appointment time. Participants were invited to attend whenever suitable within the survey opening hours (between 09:00 and 18:00). Non-respondents were given one reminder. The second questionnaire was handed out upon attendance, and most participants completed it while waiting for the clinical examination. The comprehensive Tromsø 6 data include self-reported demographic and socio-economic characteristics, information about symptoms and diseases, health status, and use of medicines and health services. Since residents with severe mental disorders are unlikely to participate in population-based surveys like Tromsø 6 [27] the study mainly includes persons without psychotic symptomatology.

In the present study we excluded participants who failed to inform about psychiatric outpatient visits (n = 1,639), anxiety/depression (n = 869), or both (n = 221). The final general population sample consisted of 10,253 participants (Fig. 1). For further analyses of participants with moderate or severe anxiety/depression we excluded those who reported no anxiety/depression (n = 8,485). This final sample consisted of 1,768 participants (Fig. 1).

The use of psychiatric outpatient services at least once during the previous year (yes/no) was designated as the dependent variable. This included visits to private and public providers, and also visits to hospital staff like nurses and social workers, supervised by psychiatrists and psychologists.

Gender was designated key independent variable. Adjustment independent variables were age in 10-year groups, marital status, income, education, self-reported degree of anxiety/depression, and GP visits last year (yes/no). For marital status we used the original response options: married/cohabitant or single. The income variable referred to the household’s total gross income in the previous year. Eight original response categories were merged into low income (< NOK 200,000), low middle income (NOK 201,000-400,000), high middle income (NOK 401,000-700,000) and high income (> NOK 700,000). We defined three education response categories from the original five: low (primary and part of secondary school), middle (high school) and high education (college or university). The self-reported anxiety/depression variable was obtained from the Euro Quality of Life Group five Dimensions (EQ-5D) score question with the three response options “I am not/moderately/extremely anxious/depressed” [28].

Data was analysed by means of descriptive statistics and logistic regressions. Differences between men and women were investigated by applying univariable as well as multivariable logistic regression models adjusted for age, marital status, income, education, self-reported degree of anxiety/depression, and GP visits last year. Analyses were performed for both genders as well as for men and women separately. The adjustment variables were introduced collectively into the regression models. All analyses were obtained as dummy- and trend analyses. Since some of the groups in the anxiety/depression sample were small, and there was no significant lack of linearity, we chose to report the trend analyses exclusively. First-order interactions were tested by introducing interaction terms in the regression models, and by stratifications.

The sixth Tromsø Study and this particular study have been approved by the Regional Committee for Medical and Health Research Ethics (REK 2009/2536 and 2014/1667/REK nord).

Table 1 gives the characteristics of the general population and the population with anxiety/depression. In total, 17.3 % of the general population reported anxiety/depression, 21.5 % of women and 12.3 % of men. Visits to psychiatric outpatient services during a year were reported by 3.9 %, 4.6 % among women and 3.3 % among men (Table 1).

Univarable logistic regressions showed that the likelihood of visiting psychiatric outpatient services was lower among men than women (OR 0.71, CI 0.58–0.87, p = 0.001). These gender differences were not maintained after adjustment for age, marital status, income, education, self-reported degree of anxiety/depression, and GP visits last year (Table 2), thus confirming our hypothesis. The relation between the use of services and gender was, however, modified by age, with an increased probability of use among women compared to men in younger ages (interaction term gender x age OR 1.54, CI 1.27–1.87, p < 0.001). Analyses stratified by age confirmed that men used psychiatric outpatient services significantly less than women in ages 30–49 years (OR 0.58, CI 0.39–0.84, p = 0.004), but slightly more than women in ages 50 years and over (OR 1.36, CI 1.00–1.83, p = 0.047). The likelihood of a visit to psychiatric outpatient services decreased by age in women aged 30–69 years, whereas it increased by age in men aged 30–59 years (Fig. 2). Overall, the use of services decreased significantly with higher age (Table 2).

Other factors associated with the use of psychiatric outpatient services in the general population were lower income, higher education, more severe anxiety/depression, and GP visits the last year (Table 2). We found the same for women when analysing genders separately, whereas for men anxiety/depression and GP visits during the last year were the only statistically significant variables associated with use (Table 2).

The likelihood of visiting psychiatric outpatient services decreased with age in women with anxiety/depression, whereas men’s probability of visiting peaked in ages 40-59 years (Fig. 3). Only 12.4 % with anxiety/depression reported psychiatric outpatient visits during the last year, 13.5 % of women and 10.5 % of men (Table 1). Among those with severe anxiety/depression, 52.9 % of women and 30 % of men reported visits. Corresponding numbers among those with moderate anxiety/depression were 12.3 % and 10.1 %, respectively.

In line with our hypothesis, there were no statistically significant gender difference among those with self-reported anxiety/depression; neither in univariable analysis (OR 0.75, CI 0.55–1.02, p = 0.068) nor in multivariable analysis adjusted for age, marital status, income, education, self-reported degree of anxiety/depression, and GP visits last year (Table 3). Use was significantly reduced by higher age (Table 3).

Among women with anxiety/depression we found younger age, higher education, more severe degree of anxiety/depression, and GP visits the last year to be associated with use, whereas in men only a more severe anxiety/depression was associated with psychiatric outpatient visits (Table 3).

The main finding of the current study is that in a general population the probability of a psychiatric outpatient visit was significantly lower among men compared to women in ages 30–49 years, whereas the trend was the opposite for ages 50 years and over. Only 3.9 % of the general population visited psychiatric outpatient services in a year. In both genders the probability of visiting was strongly associated with symptoms of anxiety/depression, and GP visits during the last year. In women the likelihood of a visit was also associated with lower age, lower income, and higher education.

Among those with self-reported anxiety/depression only12.4 % visited psychiatric outpatient services in a year. There were no statistically significant gender differences in use of services in this subgroup. The probability of visiting was strongly associated with more severe symptoms of anxiety/depression. In women the likelihood of a visit was additionally associated with higher education and GP visits the last year.

It is well documented throughout the world that men use mental health services less than women [29‐32], but the picture has been ambiguous for higher age groups [10‐14]. Our finding that gender differences in psychiatric outpatient specialist utilisation are dependent on age in a general population has, to our knowledge, hardly been reported in previous studies. Olivier et al. found that men in a general GP population in Somerset England were less likely to say that they would seek any kind of help for mental health problems, and that help-seeking increased steeper by age in men compared to women [33]. This does not underpin our finding that women’s use of psychiatric specialists decreased by age, but substantiates our study regarding the relative use between genders since utilisation rates for men and women approached in higher ages. In line with our study, Mosier et al. found that Canadian women’s probability of service use for mental health reasons decreased with age [30]. We found no significant trend in men’s utilisation by age, whereas Mosier et al. found an age dependent increase in men’s service utilisation [16, 30]. There is, however, evidence that the prevalence of anxiety and depressive disorders decrease by age throughout a lifespan both in men and women [19, 30, 34], strongly indicating that other factors than prevalence of disease are of great importance in understanding patterns of psychiatric services utilisation, particularly in men. Possible explanations for the opposite gender differences in ages 30-49 and 50 years and over could be that men might not recognize and accept mental health problems to the same extent as women during their 30s and 40s [35], which in turn may relate to requirements for social manifestations of masculinity in a traditional sense [36‐38]. These notions might be intertwined with the phenomenon of stigma, leading to patients preferring not to receive care at all or not to receive care in specialist psychiatric settings [10, 39, 40]. Pattyn et al. found that men were more likely than women to be recommended self-care for mental health problems, both by women and by other men. This report also states that men report more “negative attitudes towards professional help-seeking, but that also women (re)construct masculinity norms by giving different treatment advice to a man compared to another woman” [40]. This might suggest that men should be encouraged to perceive, recognize and accept mental problems to a larger extent, and that the acceptance of such problems among men, in particular among younger men (<50 years), should increase in society and health services.

In the subpopulation with self-reported anxiety/depression we found no statistically significant gender differences in use of outpatient psychiatric services. This is in line with previous studies reporting that genders do not differ in use of services once mental problems are perceived and recognised [9, 15, 16, 30, 35].

Our finding that use was generally lower in higher ages (Tables 2 and 3) is in accordance with studies from Norway [41] and other countries [7, 10, 12, 42, 43]. It is consistently reported that older adults (55–65 years and over) of both genders are less likely than younger and middle aged to perceive a need for care, to receive referrals from primary care, and to use specialist psychiatric services [10, 43]. In older adults visits to the GP for mental health problems seem to be more common than specialist visits [13]. This might implicate that health care providers and services should be aware of a possible under-treatment of older adults compared to younger adults when treatment in specialist services is concerned.

Only 12.4 % of those who reported some degree of anxiety/depression had visited psychiatric specialist services during the previous year. Wang et al. found that among patients with mood, anxiety and/or substance disorders in 10 different high-income countries 37.6 % (New Zealand) to 52.2 % (Spain) received specialist care [18]. We found that even among patients with severe symptoms less than half (46.8 %) had visited specialist services, whereas Wang et al. found that 24 % (Japan) to 61 % (Belgium) had visited specialists. A Canadian survey showed that 54 % of respondents meeting criteria for major depression had consulted health care services for mental health reasons in the previous year (with no significant differences between genders) [9]. Hämäläinen et al. even reported that about half of those suffering from severe anxiety/depression in Finland did not use any health care services for these symptoms [42], whereas specialist level services accounted for 64 % of use associated with major depressive disorders, and 71 % of use associated with anxiety disorders [39]. It has been estimated that around half of those who use services for mental health problems in high-income countries use specialty services [18]. Thus, it is likely that some of our study participants, who did not visit psychiatric specialist services, received GP treatment for their mental health symptoms. However, Olivier et al. found that the preferred source of help for mental health problems was relatives and friends, and that one in four said that they would not seek help anywhere [33]. One in six would not seek help from their GP [33]. Our low visit rates adds to a solid documentation that the use of mental health services both in general populations and in people with depression and anxiety disorders in high income countries is limited, indicating that these symptoms are undertreated [18, 34, 44]. Treatment rates are even lower in low and middle income countries [18].

In line with others’ research, our study shows that the main determinant for seeking help for mental health problems was the severity of the symptoms [8, 33, 35, 42, 45]. Furthermore, it is not surprising that GP visits also turned out to be an important indicator of specialist utilisation in both genders, since referrals are required for specialist treatment in Norway.

The likelihood of psychiatric specialist use in a general population increased with higher education but decreased with higher income. This finding is in accordance with research from Norway [46] and the US [7], and has been discussed elsewhere [3]. However, we revealed an effect modification for gender and education, indicating even higher use among women compared to men as education increases. Schomerus et al. reported an association of higher education and help-seeking for depression, but did not observe any gender-specific effect [47]. Our finding might be due to an increased perception of mental problems in highly educated women combined with knowledge that effective treatment exists, and a larger willingness to seek care compared to highly educated men.

Particular strengths of this study were the large sample size, the high response rate, and the comprehensive coverage of information about health, disease, and socio-economic status in the questionnaires.

Nevertheless, the study should be interpreted in light of some limitations. Despite a high response rate, our sample may not be entirely representative of the general population, as it is well known that women, married/cohabitants, healthier persons, and higher socio-economic groups are more likely to participate in population surveys [48]. In Tromsø 6, attendees were older, and the proportions of married/cohabitants and women were higher than in non-attendees [25, 26]. In the second Tromsø Study (1979–80) the participation of men and women with psychiatric morbidity was approximately 20 % lower than for those without such morbidity [27], and lower participation is likely the case for Tromsø 6 as well. However, this applies particularly to serious psychiatric morbidity [27, 49]. Additionally, our data might underestimate psychiatric morbidity and treatment seeking due to perceived stigma [10], since the numbers failing to inform about psychiatric visits and/or morbidity were relatively high. The validity of symptoms scored by the EQ-5D may be disputable, although EQ-5D has been shown to achieve adequate levels of performance in depression and to some extent in anxiety [50]. It is thus possible that these symptoms are underestimated in our population. Our data might also underestimate treatment seeking since questions about psychiatric conditions and use of services were spread throughout the questionnaire, potentially increasing inaccuracies [51]. However, there is hardly any reason why people should report anxiety/depression but not use of psychiatric services, thus the relative validity between these variables should be quite robust. The validity of self-reported data as such may be questioned, although agreement between self-reported and registered health care utilisation is generally high [52]. It might also be easier to report anxiety/depression in a self-administered questionnaire than reporting to health care providers. Therefore, self-reported anxiety/depression might be the best available measure for our study purpose, since research based on doctor made diagnoses would make it difficult to include those who had not visited health care services. In our analyses we focused on anxiety/depression, but we cannot rule out the possibility that participants may have had other psychiatric ailments or diseases in addition. Furthermore, it might be a problem that we asked about anxiety/depression at the time of the survey, whereas health care utilisation was reported for the previous 12 months. However, the onset of these diseases is often ahead of 30 years of age [19, 20], making it unlikely that this have affected our study. Finally, we cannot exclude the possibility of unmeasured confounders of the reported associations.