Examining Variations in the Prevalence of Diagnosed Mood or Anxiety Disorders Among Migrant Groups in Ontario, 1995–2015: A Population-Based, Repeated Cross-Sectional Study

Abstract

Background

International evidence on the frequency of mood or anxiety disorders among migrant groups is highly variable, as it is dependent on the time since migration and the socio-political context of the host country. Our objective was to estimate trends in the prevalence of diagnosed mood or anxiety disorders among recent (<5 years in Canada) and settled (5–10 years in Canada) migrant groups, relative to the general population of Ontario, Canada.

Methods

We used a repeated cross-sectional design consisting of four cross-sections spanning 5 years each, constructed using health administrative databases from 1995 to 2015. We included all Ontario residents between the ages of 16 and 64 years. We assessed differences in the prevalence of mood or anxiety disorders adjusting for age, sex, and neighbourhood-level income. We further evaluated the impact of migrant class and region of birth.

Results

The prevalence of mood or anxiety disorders was lower among recent (weighted mean  =  4.10%; 95% confidence interval [CI], 3.59% to 4.60%) and settled (weighted mean  =  4.77%; 95% CI, 3.94% to 5.61%) migrant groups, relative to the general population (weighted mean  =  7.39%; 95% CI,  6.83% to 7.94%). Prevalence estimates varied greatly by region of birth and migrant class. We found variation in prevalence estimates over time, with refugee groups having the largest increases between 1995 and 2015.

Conclusions

Our findings highlight the complexity of mood and anxiety disorders among migrant groups, and that not all groups share the same risk profile. These results can be used to help inform health service allocation and the development of supportive programs for specific migrant groups.

Introduction

Mood and anxiety disorders are leading contributors to the global burden of disease.^1,2 In Canada, these common mental disorders contribute to substantial morbidity and mortality.^3,4 Mood and anxiety disorders are responsible for three in every four contacts with the health care system for mental health reasons and are currently ranked as the sixth and seventh leading contributors to years lived with disability in Canada.^4,5

Immigrants and refugees often have unique experiences that can affect their mental health, which require culturally adapted mental health services and support.^6,7 Although migration is a global phenomenon, evidence points to geographical trends in the prevalence of mood or anxiety disorders among migrant groups compared to native-born populations. Evidence from outside of North America suggests that migrant status is associated with an increased risk of both mood and anxiety disorders.^8,9 Research from North America, including evidence from Canada, suggests the opposite trend, where migrant groups have a lower frequency of mood or anxiety disorders compared to the native-born population.^4,10–12

A recent systematic review identified that in Canadian settings, most research estimating the prevalence of mood and anxiety disorders among migrant groups has relied on population health survey data, and there is limited research estimating prevalence using health administrative databases. ¹⁰ This is an important gap, as recent work evaluating the concordance between survey and health administrative data sources suggests that the use of either survey or health administrative data sources alone may be insufficient in providing accurate estimates of mood and anxiety disorders at the population level. ¹³ A strength of health administrative data for estimating the prevalence of mental disorders among migrant groups is the availability of large samples to enable more granular analyses, which can help avoid misleading inferences drawn from higher-level aggregated analyses.^13,14 Current Canadian estimates have been limited in their evaluation of the characteristics of migrant groups, including differences in mood or anxiety disorders between migrant classes, as well as differences by region of birth.^14,15 Available evidence suggests there is variation in self-rated mental health and suicidal ideation by region of origin, and that ethnic migrant groups, regardless of immigrant status, reported a lower prevalence of mood or anxiety disorders compared to white people. ¹⁵ These factors have also shown to be associated with mood and anxiety disorders in other settings. ¹⁶ Another key consideration is the length of time spent in Canada. ¹⁷ Research in Canadian settings suggests that recent migrants have lower estimates of depression both within 5 years of migration and within 10 years of migration, relative to long-term migrants arriving 10–30 years prior.^10,18 Health administrative data also provide the opportunity to track estimates of mood or anxiety disorders over time. This data, when linked with immigration and refugee records, can be used to study the relationship between recent migrant status and mood or anxiety disorders at various time points, which can provide insight into the mental health and mental health service delivery for newly arrived migrants to Ontario over the past two decades. Ontario is the most populous and diverse province in Canada with ∼30% of its population being foreign-born, as such, it represents a strong setting to study migrant mental health. ¹⁹

The goal of this work was to estimate variation in the diagnosis of mood or anxiety disorders among migrant groups using linked population-based health administrative data. Specifically, our objectives were to: (1) compare age- and sex-adjusted estimates of the prevalence of mood or anxiety disorders among recent (<5 years in Canada) and settled (5–10 years in Canada) migrant groups, relative to the general population; and (2) estimate trends in diagnosis of mood or anxiety disorders among migrant groups between 1995 and 2015. We additionally examined the role of important risk factors, including migrant class and region of birth.

Methods

Sample

Using health administrative data from 1995 to 2015, we created four separate cross-sections of data, each five years in length. This allowed us to compare four consistent exposure groups over the 20-year period. We included all Ontarians who were between the ages of 16 and 64 years at the start of each repeated cross-section. Using these cross-sections, we were able to examine trends in 5-year prevalence estimates among migrant groups, relative to the general population, across the four time periods. Figure 1 presents the flowchart of our repeated cross-sectional design.

Figure 1.

Study flowchart.

Source of Data

We obtained data from ICES, which is an independent, non-profit research institute whose legal status under Ontario’s health information privacy law allows it to collect and analyze health care and demographic data, without consent, for health system evaluation and improvement. We accessed data on mood and anxiety disorders from the following databases: (i) Ontario Health Insurance Plan (OHIP; 1994–2015), which covers (>96%) of the population of Ontario with universal coverage of all medically necessary services. ²⁰ OHIP contains physician billing data, including billings from primary care and emergency department physicians; (ii) Ontario Mental Health Reporting System (OMHRS; 2005–2014), which includes data on adult inpatient psychiatric hospitalizations to designated psychiatric beds ²¹ ; (iii) Discharge Abstract Database (DAD; 1995–2014), which captures data on all inpatient hospitalizations, including psychiatric hospitalizations not captured in OMHRS. ²² We also linked data from the Immigration, Refugees, and Citizenship Canada (IRCC) Permanent Resident Database (1985-2014), which contains information on all permanent and temporary residents landing in Ontario, Canada, dating back to 1985. For demographic information, we used the Registered Persons Database (RPDB; 1995–2014), which contains information on people registered for OHIP, including date of birth, sex, and neighbourhood-level income quintile.

We followed the REporting of studies Conducted using Observational Routinely-collected health Data (RECORD) guidelines (Supplemental Material 1). ²³

Case Ascertainment

Diagnoses of mood or anxiety disorders were identified using the following standardized algorithm: (1) one hospitalization reported in OMHRS or DAD for a mood or anxiety disorder within a given 5-year cross-section or (2) one visit to a psychiatrist for a mood or anxiety disorder within a 5-year cross-section or (3) at least two outpatient OHIP physician billing claims (including billing from general practitioners and emergency department physicians) with a diagnosis code for a mood or anxiety disorder in any 12-month period during the 5-year cross-section, where one of the codes must be within the 5-year window. This algorithm is similar to a validated algorithm used to identify depressive disorders in other Canadian settings, which had a sensitivity of 61.4%, specificity of 94.3%, a positive predictive value of 69.7%, and a negative predictive value of 92.0%.^24,25 A list of diagnostic codes used to identify cases of mood or anxiety disorders is available in Supplemental Material 2.

Exposure Classification

Our exposure of interest was migrant status, stratified by time spent in Canada. We compared recent migrant groups (<5 years in Canada), with settled migrant groups (between 5 and 10 years in Canada), and the general population (including native-born Canadians, second-generation migrant groups, long-term residents of Canada for more than 10 years).^26,27 Exposure status was identified at the beginning of each of our four cross-sections and was identified using data from the IRCC Permanent Resident Database. We further explored trends in the prevalence of diagnosed mood or anxiety disorders by migrant class (economic, family reunification, and refugee), as well as by region of birth (European, Caribbean, South Asian, East Asian, Latin American, North African/Middle Eastern, and other African). We excluded data from the “other” migrant class (∼1% of all migrants), as this categorization is inconsistent between cross-sections. These categorizations are further defined in Supplemental Material 3.^28,29

Socio-demographic Factors

We obtained data on socio-demographic factors from RPDB and IRCC databases. For migrant groups, we obtained information on country of birth and migrant class. For both the migrant groups and the general population, we obtained data on age, sex, urban versus rural place of residence (greater or less than a core population of 10,000), and neighbourhood-level income quintiles.

Data Analysis

We computed descriptive statistics using means and standard deviations for continuous data and counts and proportions for categorical data. Standardized differences, which describe differences in means in units of standard deviation, were used to compare the distribution of baseline covariates between groups. ³⁰ A <10% standardized difference has been shown to be equivalent to having a phi coefficient of 0.05, indicative of a non-significant difference across groups. ³⁰ We calculated 5-year prevalence estimates of mood or anxiety disorders for each of our cross-sections stratified by our groups of interest. The stratification groups included: (1) migrant status (recent migrants, settled migrants, and general population); (2) migrant class (economic, family reunification, and refugee); and (3) region of birth.

We also calculated arithmetic weighted mean prevalence estimates and corresponding weighted variances to estimate 95% confidence intervals (CIs) combining estimates from our four cross-sections using the weights package in R. ³¹ We tested differences between our groups of interest using weighted Welch’s two-tailed t-tests, which were used to calculate weighted mean differences and standard errors. We used an alpha of 0.05, and the null hypothesis was equal means between groups.

We used modified Poisson regression models with robust variance estimators to estimate the prevalence ratios (PRs) and 95% CIs for mood and anxiety disorders in recent and settled migrant groups, relative to the general population. PRs were adjusted for age, sex, rural residence, and neighbourhood-level income. ³² Estimates from our four repeated cross-sections were assessed separately.

To explore differences between our four cross-sections, we performed a generalized estimating equation adjusting for the correlation between people included in multiple cross-sections. The model assumed an exchangeable correlation structure for the working correlation matrices, where it is assumed that the correlation between any pair of measurements on the same person is the same. ³³ We analysed differences between our groups of interest over the four cross-sections, adjusting for age, sex, rural residence, and neighbourhood-level income.

As a secondary analysis, we conducted a within-migrant analysis assessing differences between recent, settled, and long-term migrants living in Canada for more than 10 years, adjusting for migrant class and region of birth (modelled separately due to collinearity). Estimates were obtained from cross-sections 2–4, as we were unable to differentiate between the general population and long-term migrants in cross-section 1.

Results

Sample Characteristics

Our samples increased from cross-section 1 to cross-section 4, ranging from 7,357,862 to 9,502,713 people. The samples comprised ∼50% men and women in each cross-section, with the mean age ranging from 38 (standard deviation [SD]  =  13) to 40 (SD  =  14). Recent and settled migrant groups between the ages of 16 and 64 years made up between 8.1% and 9.3% of the samples. Trends throughout all four cross-sections suggest that compared to the general population, recent and settled migrant groups were significantly younger and more likely to be living in areas with the lowest income quintiles. Recent and settled migrant groups were also significantly less likely to live in rural settings. Table 1 provides our descriptive statistics for the socio-demographic characteristics of cross-section 2, which was similar to those from the other cross-sections (see Supplemental Material 4 for data from all cross-sections). There was minimal missing data (<1%) for our variables of interest.

Table 1.

Socio-demographic Characteristics of Migrant Groups and the General Population.

Cross-Section 2 – Followed: 2000–2005
Variable		General population n = 7,222,605 (n, SD/%)	Recent migrant n  =  336,098 (n, SD/%)	Standardized difference	Settled migrant n = 406,744 (n, SD/%)	Standardized difference
Age	mean (SD)	38.91(12.99)	34.88 (11.12)	0.33	36.64 (11.38)	0.19
Sex	Female Male	49.9% 50.1%	51.2% 48.8%	0.02	50.2% 49.8%	0.00
Income quintile	−5 (Highest) −4 −3 −2 −1 (Lowest)	20.0% 18.4%	23.6% 36.6%	0.08 0.09 0.42	17.7% 23.6% 34.7%	0.09 0.38
Rural	– Urban – Rural	86.3% 13.7%	99.1% 0.9%	0.51	98.9% 1.1%	0.50

Five-Year Prevalence

The 5-year prevalence of mood or anxiety disorders in the Ontario population ranged from 6.74% to 7.57%. Figure 2 presents 5-year prevalence estimates across our four cross-sections by the migrant group. The prevalence of mood or anxiety disorders was significantly lower in recent migrant groups (weighted mean  =  4.10%, 95% CI,  3.59% to 4.60%), and settled migrant groups (weighted mean  =  4.77%, 95% CI, 3.94% to 5.61%), relative to the general population (weighted mean  =  7.39%, 95% CI, 6.83% to 7.94%). We did not find a statistically significant difference in the weighted prevalence estimates for recent and settled migrants, though there is evidence there may be significant differences within various cross-sections (see Tables 3 and 4).

Figure 2.

Five-year prevalence estimates across our four cross-sections by migrant status, migrant class, and region of birth.

Table 3.

Modified Poisson Regression Analyses Assessing Prevalence of Diagnosed Mood or Anxiety Disorders Between Recent Migrant Groups, Settled Migrant Groups, and the General Population.

Outcomes	General population (PR; 95% CI)	Recent migrant PR (95% CI)	Settled migrant PR (95% CI)
Cross-section 1
Unadjusted	Ref	0.56 (0.55 to 0.57)	0.73 (0.72 to 0.74)
Adjusted	Ref	0.49 (0.49 to 0.50)	0.66 (0.65 to 0.67)
Cross-section 2
Unadjusted	Ref	0.58 (0.58 to 0.59)	0.65 (0.64 to 0.66)
Adjusted	Ref	0.52 (0.51 to 0.53)	0.58 (0.57 to 0.59)
Cross-section 3
Unadjusted	Ref	0.54 (0.53 to 0.55)	0.67 (0.66 to 0.68)
Adjusted	Ref	0.48 (0.47 to 0.49)	0.61 (0.60 to 0.62)
Cross-section 4
Unadjusted	Ref	0.54 (0.53 to 0.55)	0.59 (0.58 to 0.59)
Adjusted	Ref	0.47 (0.46 to 0.48)	0.53 (0.52 to 0.54)

Adjusted: adjusted for age, sex, income, rural status.

PR: Prevalence Ratio.

Table 4.

Modified Poisson Regression Analyses Assessing Prevalence of Diagnosed Mood or Anxiety Disorders Between Recent Migrant Groups, Settled Migrant Groups, and Migrants Living in Canada for >10 Years.

Outcomes	Migrants >10 years in Canada (PR; 95% CI)	Recent migrant PR (95% CI)	Settled migrant PR (95% CI)
Cross-section 2
Unadjusted	Ref	0.74 (0.73, 0.76)	0.83 (0.81, 0.84)
Adjusted 1	Ref	0.77 (0.75, 0.79)	0.85 (0.83, 0.87)
Adjusted 2	Ref	0.75 (0.73, 0.76)	0.81 (0.79, 0.83)
Adjusted 3	Ref	0.79 (0.77, 0.81)	0.85 (0.83, 0.87)
Cross-section 3
Unadjusted	Ref	0.70 (0.69, 0.71)	0.89 (0.87, 0.91)
Adjusted 1	Ref	0.72 (0.70, 0.73)	0.91 (0.89, 0.93)
Adjusted 2	Ref	0.72 (0.71, 0.74)	0.92 (0.90, 0.94)
Adjusted 3	Ref	0.76 (0.74, 0.77)	0.95 (0.93, 0.97)
Cross-section 4
Unadjusted	Ref	0.67 (0.66, 0.68)	0.75 (0.74, 0.76)
Adjusted 1	Ref	0.67 (0.66, 0.69)	0.76 (0.75, 0.78)
Adjusted 2	Ref	0.66 (0.64, 0.67)	0.76 (0.75, 0.77)
Adjusted 3	Ref	0.68 (0.67, 0.69)	0.78 (0.77, 0.79)

PR: Prevalence Ratio.

Within-migrant groups, there were differences among migrant classes. The prevalence of mood and anxiety disorders was lowest in economic migrant groups (weighted mean 4.20%, 95% CI, 3.85% to 4.55%), followed by family reunification migrant groups (weighted mean 4.77%, 95% CI, 4.32% to 5.21%) and was highest in refugee groups (weighted mean 7.02%, 95% CI, 6.01% to 8.03%). Refugee groups had a significantly higher weighted prevalence compared to economic and family reunification migrants. Family reunification migrants had a significantly higher weighted prevalence compared to economic migrants (see Table 2).

Table 2.

Weighted Welch’s t-tests Comparing Weighted Mean Prevalence Estimates of Mood or Anxiety Disorders Between Stratified Migrant Groups.

Comparison	Absolute weighted mean difference	Std. Error	p-value
Recent v. settled	0.67	0.29	0.066
Recent v. general	3.29	0.22	<0.001*
Settled v. general	2.61	0.29	<0.001*
Economic v. refugee	2.82	0.32	0.001*
Economic v. family	0.57	0.17	0.016*
Family v. refugee	2.25	0.33	0.002*
General v. European	1.38	0.23	0.0012*
General v. Caribbean	2.54	0.22	<0.001*
General v. South Asian	3.09	0.21	<0.001*
General v. East Asian	4.93	0.18	<0.001*
General v. Latin American	1.07	0.17	0.0014*
General v. North African & Middle Eastern	0.19	0.15	0.298
General vs other African	2.36	0.20	<0.001*

*p < 0.05.

Furthermore, the prevalence of mood or anxiety disorders varied by region of birth, with the lowest prevalence observed among migrants born in East Asian countries (weighted mean  =  2.57%, 95% CI, 2.2% to 2.93%), and the highest among migrants born in North African/Middle Eastern countries (weighted mean  =  7.68%, 95% CI, 7.49% to 7.88%). Migrants from all regions, apart from people born in North African/Middle Eastern countries, had significantly lower weighted prevalence of mood or anxiety disorder compared to the general population (weighted mean  =  7.49, 95% CI, 7.01% to 7.98%) (see Table 2).

Regression Analyses: Migrant Status

Results from our modified Poisson regression analyses with robust variance estimators suggest that the prevalence of mood or anxiety disorder is lower in the recent migrant group (range: PR  =  0.47, 95% CI , 0.46 to 0.48; PR  =  0.52, 95% CI, 0.51 to 0.53) and in the settled migrant groups (range: PR  =  0.53, 95% CI, 0.52 to 0.54; PR  =  0.66; 95% CI, 0.65 to 0.67) compared to the general population, when adjusting for age, sex, neighbourhood-level income quintile, and rural residence (see Table 3).

Generalized Estimating Equation

The findings from our 5-year prevalence estimates specified we should test whether there was a significant increase in the prevalence of mood or anxiety disorders among refugee groups over the four cross-sections. We tested the interaction between migrant class and our four cross-sections, adjusting for the correlation between people identified in multiple cross-sections, as well as age, sex, rural residence, and neighbourhood-level income. We found there was a significant increase in the prevalence of mood or anxiety disorders among refugee groups from the first cross-section to the second (PR  =  1.12, 95% CI , 1.09 to 1.17), third (PR  =  1.12, 95% CI, 1.09 to 1.16), and the fourth cross-section (PR  =  1.27, 95% CI, 1.23 to 1.31).

Secondary Analyses

For our migrant-specific analyses, adjusting for region of birth, we found that both recent migrants (range: PR  =  0.67, 95% CI, 0.66 to 0.69; PR  =  0.77, 95% CI, 0.75 to 0.79) and settled migrants (range: PR  =  0.76, 95% CI, 0.75 to 0.78; PR  =  0.91, 95% CI, 0.89 to 0.93) had significantly lower prevalence than migrants living in Canada for more than 10 years (see Table 4). Similar findings were observed when adjusting for migrant class (see Table 4).

Discussion

Our findings suggest that the majority of migrant groups in Ontario, Canada, have a lower prevalence of diagnosed mood or anxiety disorders, relative to the general population. Although this finding aligns with prior North American research, ^4,10 our work builds on this evidence by providing a detailed description of the relationship between migrant status, migrant class, and region of birth with mood or anxiety disorders using health administrative data. Our analyses highlight the complex relationship between migrant status and mood and anxiety disorders, and how estimates vary by time spent in Canada, region of birth, and migrant class. Taken together with other Canadian and international evidence, ^10,34 our findings confirm that not all migrant groups share the same burden of mood and anxiety disorders. As such, our findings highlight the importance of providing context when analysing the prevalence of mood or anxiety disorders among migrant groups.

Our findings suggest prevalence estimates were highest among refugee groups, compared to both family reunification migrants and economic migrants. Similar trends have been observed in prior Canadian research assessing primary care contacts for common mental disorders among Canada’s migrant population.^3,35 We also found variation in 5-year prevalence estimates of mood and anxiety disorders by region of birth, which align with prior work using self-reported population health survey data from Ontario. ¹⁵ Our analyses enabled an examination of more granular groups, and the findings suggest that people born in North African/Middle Eastern countries had a consistently higher prevalence of mood and anxiety disorders compared to migrants from other regions. These findings were consistent with those from the Longitudinal Survey of Immigrants to Canada. ³⁶ Variation in estimates of mood or anxiety disorders may reflect different post-migration mental health risks, and challenges related to mental health help-seeking, which may result from cultural differences, including expressions of illness, attitudes towards mental illness, in addition to language, and discriminatory practices.^6,37–39

Our findings also suggest that the prevalence of mood or anxiety disorders increases with longer time spent in Canada. These findings align with other Canadian research using survey methodology, which found lower estimates of depression both within 5 years of migrating and within 10 years of migrating to Canada, compared to migrants arriving in Canada between 10 and 30 years prior. ¹⁸ Although this work was not designed to follow trajectories, our findings align with the phenomenon of the healthy migrant effect, where migrant groups are generally healthier during the early years following migration, followed by a general deterioration in health status to more closely align with the health status of the host country. ⁴ This healthy migrant effect may be due to positive migrant selection, which arises when migrant groups are selected on a number of criteria related to skills and education, ⁴⁰ leading to migrant groups being healthier than the general population. ⁴¹ In Canada, economic migrants, who make up roughly 50% of all new Canadians, are subject to a human capital model of immigration.^42,43 This system favours people who are young, educated, and proficient in English or French. ⁴²

We identified temporal trends across the four cross-sections. We observed a steady and significant increase in 5-year prevalence estimates of mood or anxiety disorders among refugees across the four cross-sections. This trend may reflect a possible detection bias driven by an increase in the access and utilization of mental health services for various refugee groups in Canada, who may differ on pre-migratory risk factors and post-migration social and medical support. There is room to build on this work going forward by assessing trends among various migrant groups over longer periods of time post-migration, and by exploring the relationship between mental disorders and intersectionality with respect to gender, age, race/ethnicity, and other determinants of health.

Limitations

A limitation to this work is that our use of health administrative data is centred on billing codes and not standardized clinical or research diagnoses. As such, thresholds for diagnoses vary by clinician and setting. ⁴⁴ Furthermore, the accuracy of the algorithm used for case ascertainment may be different in Ontario compared to the provinces in which it has been validated, Alberta, and Newfoundland. There may also be racial bias in mental health diagnostic practices in Canada, which may lead to an under or overestimate of the prevalence of mood or anxiety disorders among various migrant groups. ⁴⁵ Furthermore, due to the lack of specificity of the diagnostic codes captured in health administrative data, we decided not to provide stratified estimates for mood and anxiety disorders separately. ⁴⁶

Our use of billing codes to estimate the prevalence of mood or anxiety disorders will underestimate the prevalence compared to estimates obtained from population-based surveys. ¹³ Evidence suggests that contact rates (identified in health administrative data) and population rates (identified in population surveys) may not align due to barriers that exist in accessing mental health care and receiving a diagnosis of a mood or anxiety disorder.^3,47,48 Research from Ontario suggests that 90% of people with depressive disorders have had contact with a primary care clinician for mental health reasons, however, only 50% have received a mental health diagnosis.^49,50 Due to the lack of specificity of the diagnostic codes captured in health administrative data, we decided not to provide stratified estimates for mood and anxiety disorders separately. ⁴⁶

Another limitation is that data from IRCC will allow us to identify people landing in Ontario but will misclassify migrants who land in other parts of Canada and moved to Ontario. Our general population comparison group will also include people who migrated prior to 1985 as well as second-generation migrants (∼22.5% of the population), who have been shown to have different frequency of mood and anxiety disorders compared to both migrant groups and the general population.^4,34,51,52 We will also lose migrants who were unable to be linked to the administrative data holdings, which represents ∼15% of migrants. ⁴² Evidence suggests that the linkage rates are not equal across ethnic groups, with East Asian immigrants having the lowest proportion of successful linkage. ⁵³ To maintain consistency between our cross-sections, we used billings to identify emergency department visits, rather than an emergency department database, which led to a loss of ∼17% of all emergency department visits. Furthermore, estimating the generalizability of our findings to other provinces in Canada is challenging due to provincial variation in mental health service provisions and the proportion of migrant groups living in various provinces.^54,55

Conclusion

Our findings highlight the complexity of the relationship between migration and mood or anxiety disorders and suggest that not all migrant groups share the same burden of mood or anxiety disorders. Prevalence of mood or anxiety disorders among migrant groups varies greatly by migrant class and region of birth. Furthermore, our findings suggest an increasing trend in the prevalence of mood or anxiety disorders among refugee groups over time, which may reflect changes in access and utilization of mental health services and requires further study. Our findings support the need to apply an equity lens in addressing risk and burden of mental health among migrant groups in Canada. These results can be used to help inform health service allocation and the development of supportive programs for specific migrant groups in Ontario.