Mental health services use and depressive symptom scores among gay and bisexual men in Canada

Increasing evidence indicates that gay and bisexual men (GBM) use mental health services (MHS) more frequently than heterosexual men [1‐3]. This is despite the historical lack of a sexual-minority affirmative stance of many MHS providers and the continued practice of conversion therapy by a minority of these providers [4]. This paradox can be explained by the greater MHS need of GBM, consistent with the higher burden of poor mental health among GBM compared to their heterosexual counterparts [5, 6]. Moreover, it has been observed that GBM use MHS at levels comparable to lesbian, queer, and bisexual women, effectively closing the gender gap that exists for MHS use between heterosexual women (high use) and men (low use) [1]. Most studies in the general population have not reported on the association between MHS use and mental health by sexual identity, impeding the extrapolation of study findings to GBM [7]. Conversely, studies comprised exclusively of GBM have mostly evaluated MHS in the context of factors associated with use or unmet MHS needs [8‐10]. At present, there is limited understanding of the association between MHS use and the mental health of GBM [11].

Depression is one of the most common mental health conditions reported among GBM [12]. Previous estimates from a meta-analysis using data from over 10,000 men showed that the prevalence of depression among GBM was more than twice that in heterosexual men [13]. Further, a recent modeling study using survey data from the US found that the prevalence and morbidity burden of depressive episodes in GBM have superseded that of HIV (a leading cause of morbidity among GBM) [14]. A public health concern in itself, depression among GBM is also associated with increased suicidality, substance use disorders, risky sexual behaviors and the acquisition of HIV and other sexually transmitted infections [6]. Thus, for GBM accessing MHS, these services may be instrumental not only in modifying depressive symptoms, but their use may also have associations with conditions/behaviors distally related to the symptoms.

In Canada, up to 80% of MHS have conventionally been provided by family physicians who use pharmacotherapy as a predominant treatment modality [15, 16]. Two provinces in the country (Ontario and Quebec) have recently committed to increase the delivery of publicly funded evidence-based psychotherapy for depression and anxiety, consistent with the United Kingdom’s Improving Access to Psychological Therapies (IAPT) model [17]. However, even though Canadian GBM have been identified as a group at risk for depression [12], neither of the two provinces have elucidated plans to adapt the proposed psychological services for GBM. Therefore, it is unclear if the proposed services will be successful in addressing depression among GBM beyond the existing MHS.

In this paper, we assess the association between MHS use in the Canadian health system and depressive symptoms and evaluate if the association is modified by sexual identity, that is, between GBM and heterosexual men. We frame these objectives keeping in mind the call by many countries for better assessment of MHS among GBM [7, 18].

The CCHS 2015–2016 cycles include data from 109,659 respondents representing 30,590,780 Canadians. After excluding women and those < 18 years of age, there were 46,191 (unweighted sample) respondents representing 13,941,459 (weighted sample) adult Canadian men. Our analytic sample was restricted to 21,383 respondents representing 6,326,364 adult men with available PHQ-9 scores, who self-reported their sexual identity from Ontario, Newfoundland and Labrador, Prince Edward Island, Nova Scotia, New Brunswick, Manitoba, Saskatchewan, and the Northwest Territories (excluding respondents from Quebec, Alberta, British Columbia, Yukon, and Nunavut, where the PHQ-9 was not administered). Of these, 509 respondents identified as GBM and 20,874 as heterosexual. These represented 173,394 GBM and 6,152,970 heterosexual men, respectively. The numbers of participants excluded by exclusion criteria are presented in Fig. 1.

Using data from a Canadian community-based survey of 21,383 men, we found that a higher number of mental health consultations were associated with higher average depressive symptom scores in both GBM and heterosexual men. Additionally, associations were not statistically different across sexual identity groups, despite GBM consistently having higher depressive symptom scores at all levels of MHS use. To our knowledge, this is the first analysis to use nationwide survey data from Canada to address a crucial research gap in understanding the association between MHS use and depressive symptom scores in GBM.

The higher average depressive symptom scores with more consultations could indicate that existing MHS are sub-optimal in mitigating depressive symptoms in all men with a greater need for MHS. As seen in our sensitivity analysis, this trend persists for both groups of men, albeit attenuated in magnitude, when adjustments for baseline PHQ-9 are made or men are assumed to start from similar PHQ-9 scores a year prior. Our findings suggest that for recurrent MHS users, changes that occur in depressive symptoms between consultations may better explain the severity of subsequent depressive symptoms than baseline depressive scores [23]. Previous work has suggested that recurrent treatment-seeking behavior is implicated in poorer treatment response [24]; a limitation of our sensitivity analysis is that we were unable to account for time-varying depressive symptom scores (occurring between consultations) which may confound the relationship between frequency of MHS use and subsequent depressive symptoms. This may be particularly important to address in future analyses to plan better MHS for GBM, who as our findings indicate, are more likely to exhibit recurrent MHS seeking behavior.

We observed that the association between MHS use and depressive symptom scores was non-differential by sexual identity. Our findings are consistent with findings from an evaluation of the aforementioned IAPT in the UK, in which gay but not bisexual men were found to have comparable average depressive symptom scores to heterosexual men [25]. For our data, given that GBM used proportionately more MHS than heterosexual men, we might have expected average depressive symptom scores among GBM to be lower. This assumes health professionals are familiar with the mental health needs of GBM and therefore more adept at providing MHS for them. However, as research over the past two decades indicates, explicit teaching on sexual minorities health has not achieved widespread curricular integration in health professional training [26‐28], which may have limited health professionals from providing context specific and sexual identity affirming MHS [29]. Furthermore, we also observe that GBM consistently had higher average depressive symptom scores for any number of mental health consultations. Therefore, we interpret the statistically non-significant difference in our associations between the sexual identity groups cautiously, as it is likely that we are not adequately powered to detect a significant difference. Based on our findings, we are unable to draw any conclusive inferences on whether the association between MHS use and depressive symptom scores are indeed similar between the two sexual identity groups. We believe that this is an important topic worthy of further study.

There are several limitations to our study findings. Given the cross-sectional nature of survey data, we are unable to explore the mechanisms for our observations. Nonetheless, by basing our hypotheses on existing literature, using a causal framework and appropriate sensitivity analyses, we are able to posit suppositions that can be further tested in future. Additionally, the use of survey data allowed us to explicitly set up comparisons between GBM and heterosexual men. We grouped different types of MHS providers together, reducing the heterogeneity of MHS that they provide. Different types of providers and treatment modalities they employ could be differentially associated with PHQ-9 scores. However, in this manuscript we position MHS from the larger perspective of the curricula in which mental health professionals receive their training, which consistently lacks in its focus on mental health issues of sexual minorities across disciplines (Medicine, Nursing, Social Work) [26‐28]. Similarly, we combined GBM into a single group cognizant that in doing so, we obfuscate the differences in mental health needs and MHS use patterns between them. However, we were not adequately powered to make meaningful distinctions in our findings when gay and bisexual men were separated i.e., interpretations of our findings remain the same as when gay and bisexual men are combined into a single group (as reported). In light of this limitation, we advocate for larger studies in future that are adequately powered to make an informative distinction in findings between gay and bisexual men. As the CCHS does not contain baseline PHQ-9 scores, we are unable to comment on the clinical significance of our findings using methods proposed by Jacobson and Traux [30]. We addressed the lack of baseline PHQ-9 scores using Monte Carlo simulations, but we refrain from making direct clinical interpretations of these findings, as simulation estimates are based on additional assumptions. Lastly, as already stated, our results may be most generalizable to Ontario.

We also make a few assumptions in our models. First, the associations observed are independent of the time when consultations occurred. This is less of a concern for a higher number of consultations, as it could be presumed that they would be distributed relatively equally over twelve months. Second, we ignore the cyclical nature of depressive symptoms [31] in our models. This should be addressed by better assessment methods (like studies using diagnostic interviews or ecological momentary assessments [32]) in future studies that can account for the cyclicity of depressive symptoms. Finally, our model fails to account for mental health disparities that exist across social hierarchies, by not being statistically powered to assess the heterogeneity of associations across racial identity, socioeconomic status or at the intersection of these social hierarchies.

Our findings highlight the importance of understanding better the inter-consultation changes in depressive symptoms, and the subsequent role that these may play in determining the effectiveness of MHS for men. We simultaneously advocate for the collection of more granular mental health data for sexual minority men accessing MHS and recommend larger comparative evaluations of MHS effectiveness between heterosexual men and GBM.