Language ability and virtual mental healthcare utilisation among immigrant and refugee youth: a population-based cohort study
WHAT IS ALREADY KNOWN ON THIS TOPIC
-
Virtual care accounts for a large proportion of mental healthcare and may improve access, particularly for those in rural and underserved areas.
-
Immigrant and refugee youth access mental healthcare at lower rates than native-born populations.
WHAT THIS STUDY ADDS
-
Paediatric refugees are less likely than non-refugee immigrants to use virtual mental healthcare, particularly those who do not speak English or French at arrival.
-
Both language ability and immigration category may influence the preference for inperson over virtual modalities for mental healthcare.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
-
These intersecting relationships between language and immigration category highlight the need to weigh patient-provider preferences, accessibility, and equitable service delivery of virtual care for youth, taking into account both immigration category and language ability.
-
Recognising these intersecting factors is essential for healthcare providers, policymakers, and stakeholders to create inclusive and effective mental health services for immigrant and refugee youth.
Introduction
In recent years, predominantly since the pandemic onset, there has been a notable shift from inperson to virtual platforms, particularly in the domain of mental health services, where physical examinations are infrequently required. In Ontario, Canada, through 2020 and 2021, virtual care constituted approximately 70% of paediatric mental healthcare visits.1–3 This shift holds promise in enhancing care accessibility, especially for underserved communities.4 5 Moreover, it has the potential to improve patient and provider choice, reduce geographical barriers, while concurrently reducing health system and patient costs, and saving families time (eg, eliminating transportation).4 5 However, the shift to virtual care may present additional challenges for youth who have historically faced difficulties in accessing mental health services, including refugees and immigrants. Among such challenges are technological obstacles, such as limited access to devices and reliable internet connections, along with concerns related to health system navigation, health literacy, privacy, safety and language barriers.4–8 The extent to which these factors affect the utilisation of virtual care is unknown.
Canada stands as one of the world’s most multicultural and ethnically diverse countries, with immigrants constituting 23% of its population, originating from over 100 source countries.9 ,10 11 ,11 Ontario receives the largest proportion of immigrants in Canada.9 Immigrants in Canada are admitted through three broad categories: economic class (eg, selected for skills and abilities to contribute to Canada’s economy, ~58%), family class (eg, sponsored by family in Canada, ~30%) and refugees (person who has fled their country due to fear of persecution, ~12%).9 12 Dependent children are admitted based on their parent/caregiver’s immigration category.13 25%–30% of immigrants arrive in Ontario, Canada without self-declared knowledge of French or English.14 Immigrant and refugee youth in Canada use physician-based and hospital-based mental health services at lower rates than Canada-born youth.15–17 This may, in part, be a reflection of the ‘healthy immigrant effect’, where immigrants, especially economic class, arrive in better health than native-born individuals.15 18 However, it underscores challenges in accessing culturally competent, high-quality and language-concordant mental healthcare.15 19 20 Further, limited language proficiency has been associated with poor health outcomes and difficulties accessing and using the health system for immigrant and refugee youth.21–25 While immigrants and refugees with language barriers may have difficulty accessing virtual care, it is also possible that to mitigate non-verbal and verbal barriers during a healthcare encounter, patients and/or providers preferentially choose inperson over virtual visits.
Considering the widespread uptake and continued use of virtual care following the onset of the COVID-19 pandemic, identifying populations and communities that could benefit from tailored approaches to virtual care provision may enhance the equity and accessibility of mental health services. We aimed to test the association between virtual mental healthcare utilisation and immigration category and Canadian language ability (CLA) at arrival of immigrant and refugee youth in Ontario, Canada. We hypothesised that those with limited CLA at arrival, especially refugees, would have lower utilisation of virtual mental healthcare compared with economic immigrants with CLA.
Methods
Study design and population
This is a population-based cohort study using linked health administrative databases in Ontario, Canada. Data sets were linked using unique encoded identifiers and analysed at ICES, a prescribed entity under Ontario’s Personal Health Information Protection Act (PHIPA). Section 45 of PHIPA authorises ICES to collect personal health information, without consent, for health system monitoring and evaluation and are exempt from research ethics board review. Use of the data in this project is authorised under Section 45 and approved by the Privacy and Legal Office of ICES.
We included all paediatric immigrants and refugees with permanent resident status (the right to live and stay in Canada), 3–17 years old, living in Ontario, Canada, and eligible for provincial health insurance between 1 March 2020 and 31 December 2021. We excluded individuals with invalid birth and death dates and those with missing sex (online supplemental eTable 1).
Data sources
We identified children with any record in the Immigration, Refugees and Citizenship Canada Permanent Resident Database (data before 1 October 2020) and included several immigration characteristics including immigrant category, duration of residence in Canada and World Bank region of birth (based on country of birth).26 We used the physician billings database (Ontario Health Insurance Plan database) with the most currently available data to identify outpatient mental health visits to family physicians, paediatricians or psychiatrists as per our previous work with validated mental health codes (online supplemental eTable 2).2 15 27 28 We used the Registered Persons Database to capture sociodemographic variables including date of birth, sex and postal code. We used the Ontario Marginalization Index for neighbourhood material deprivation quintile which combines census information on income and education as a measure of socioeconomic status.29 Rurality was defined using Statistics Canada Postal Code Conversion File definition (community size ≤10 000 persons).
Exposures
The main exposure, determined from official immigration records for all immigrants, was the combination of CLA at arrival (self-reported English and/or French language ability (as reported by the principal applicant parent or caregiver) (online supplemental appendix 1)) and immigration category (economic class, family class and refugees), with six exposure groups. Self-reported language proficiency measures in adults are correlated, reliable and valid compared with standardised language tests.30 31 Further, children’s self-reported language experiences are closely aligned with parent reports of language exposure.32 Children’s language acquisition after arrival is not measurable in existing data sets and likely depends on a variety of factors, including exposure to Canadian language, previous language skills, quality of educational programmes, family language proficiency, gender, cognitive abilities, economic (dis)advantage and immigration category.33 34 As a secondary exposure to provide deeper contextual information, we ascertained the country of birth and classified each country based on whether or not it had an official Canadian language.35
Outcome measures
The primary outcome was a binary measure (virtual (telephone and/or video) vs inperson) of the visit modality for all physician-based mental healthcare visits (from physician billing codes, online supplemental eTable 2),28 measured as counts and calculated as a rate denominated on the total number of mental health visits between 1 March 2020 to 31 December 2021. Visits to non-physicians (eg, psychologists and social workers) were not available in existing billing codes and were not included.
Covariates
We included several covariates considered to be potentially associated with the utilisation of virtual mental healthcare.2 27 This included age group (3–12 years and 13–17 years), sex (male/female), rurality (urban/ rural), neighbourhood material deprivation quintile, recency of immigration and region of birth.
Statistical analysis
The baseline characteristics of the cohort with a mental health visit were described using frequencies and means with SD or medians with IQRs, as applicable. We determined the visit number and rate (proportion of total mental health visits) for each modality by immigration factor groups. A modified Poisson regression model estimated rate ratios (RRs) with 95% CI of a virtual mental health visit by immigration category and CLA with a referent of economic class with CLA. Models were adjusted for age, sex, recency of immigration and months since the onset of the COVID-19 pandemic. Sensitivity analyses excluded the first 4 months of the pandemic and secondary analyses tested the association of CLA within immigrant categories. In additional descriptive analyses for context, we determined the proportion of virtual mental health visits by country of birth, plotted by the number of immigrants with a mental health visit from that country. All analyses were performed using SAS V.9.4 (SAS Institute).
Results
Baseline characteristics
During the study period, 22 420 immigrant and refugee youth had at least one mental health visit whereby 10 285 (46%) had CLA (economic class immigrant: 6293 (28%), family class immigrant: 1463 (7%), refugee: 2529 (11%)) and 12 135 (54%) did not have CLA (6310 (28%) economic class, 2207 (10%) family class and 3618 (16%) refugees; table 1). Half (50.3%) of the cohort was female with the cohort’s mean (SD) age 12.0 years (4.2). The majority lived in urban regions (98.8%), were non-recent immigrants (59.5%) and the largest proportion were born in the Middle East region (28.7%). Refugees disproportionately resided in the most materially deprived neighbourhoods. Refugees without CLA had the largest proportion of recent immigrants (62.4%; table 1).
Provision of virtual mental healthcare
Over the study period, we identified 71 375 physician-based mental healthcare visits. After an initial rapid uptake of virtual care at the pandemic onset by all exposure groups, 67% (47 989) of mental healthcare visits were performed virtually and this proportion was sustained at approximately 50%–70% across all exposure groups (figures 1 and 2).
Main analyses
Compared with economic class immigrants with CLA among whom 14 249 of 20 212 (70%) mental health visits were virtual, refugee youth with and without CLA had statistically and clinically important lower risks of utilisation of virtual care (CLA refugees: 4858 virtual of 7743 visits (63%), adjusted RR (aRR) 0.89, 95% CI 0.86 to 0.93; non-CLA refugees: 5213 virtual of 9437 visits (55%), aRR 0.80, 95% CI 0.77 to 0.83). Family class immigrants with a CLA had a modestly lower risk of virtual care utilisation (3329 virtual of 4972 visits; (67%), aRR 0.96, 95% CI 0.92 to 0.99) (figure 2). No important differences were observed in virtual care utilisation between the referent economic class immigrants and the other immigrant groups. Sensitivity analyses excluding the first 4 months of the pandemic did not meaningfully change findings (online supplemental eFigure 1). Analyses stratified by immigration category showed that non-CLA refugees had a 9% lower risk of virtual care utilisation compared with refugees with CLA (aRR 0.91, 95% CI 0.87 to 0.95) but other within immigrant group language differences were not observed (online supplemental eFigure 2).
Additional contextual analyses
There was heterogeneity in the utilisation of virtual care by country of birth and official language of those countries (online supplemental eFigure 3). Immigrants from several non-Canadian language countries had both high (Iran, United Arab Emirates, China) and low (Syria, Iraq, Afghanistan) rates of virtual care utilisation whereas for most countries where English or French are official languages, virtual care utilisation was high.
Discussion
In this population-based study of paediatric immigrants, we observed generally high utilisation of virtual mental healthcare with two-thirds of visits performed using a virtual modality during the study period. There were clinically important relative and absolute differences in virtual care utilisation between immigrant groups with refugees, especially those who did not self-report CLA on arrival, least likely to use virtual modalities for mental healthcare. Differences in tested associations between economic class immigrants with CLA and the other immigrant groups were not clinically important. Given the measure of language ability was reported at arrival and many will acquire Canadian language skills as they stay in Canada,36 observed findings likely underestimate the magnitude of our observed estimates. Taken together, results suggest the need to support the provision of virtual care in a way that ensures equitable access and aligns with the needs and preferences of specific subgroups of immigrant and refugee youth based on both language (recognising that variability exists in its acquisition) and the immigration category through which they come to Canada.
Evidence from various settings has consistently demonstrated the adverse effects of language barriers on mental healthcare utilisation in refugee and immigrant youth.5 21 24 25 Language barriers can lead to miscommunication between patients, caregivers and healthcare providers, missed appointments, decreased quality of care and patient/caregiver satisfaction, and importantly, poor health outcomes.37 38 The specific impact of language barriers on the use of virtual care, especially among immigrants and refugees who may acquire language ability at different rates,33 34 is complex and multifaceted. Qualitative studies from the USA have suggested challenges faced by immigrant and refugee youth in accessing virtual care are mainly due to limited access to technology as well as sociocultural and structural barriers.5 6 Limited language proficiency, needed for tasks like creating passwords and logging in, might add to difficulties for refugee patients, who may already struggle with health system navigation and health literacy. Providing interpretation services requires reliable technology and internet connectivity, which can present additional challenges for patients, healthcare providers and interpreters.39 Our study’s findings support the notion that language ability and refugee status may be barriers to virtual care utilisation. Alternatively, given the clinical limitations of virtual care, findings may reflect patient preference and clinically appropriate use of the right care modality for this patient group.
In Ontario, interpretation services are widely available (eg, LanguageLine)40 through either visit modality to mitigate language barriers. However, limitations around their use can negatively impact patient care.41 42 Utilisation of interpreters for refugees may limit the relationship between the patient and provider (eg, difficult to build rapport), can be time-consuming, provide inaccurate translations, create technical problems, and there may be concerns about a breach of confidentiality.39 43 Virtual interpretation services may further exacerbate these challenges, possibly explaining the lower utilisation of virtual care in favour of inperson care among this population.
The observed finding that family class immigrants with CLA have modestly lower utilisation of virtual care compared with economic class immigrants and non-CLA family class immigrants deviates from our initial hypothesis. Family class immigrants mainly arrive in Canada from a select number of countries including the Philippines, India, Jamaica and Pakistan.44 Though English is an official language in these countries, other factors may contribute to observed findings. Sociocultural experiences and differences in attitudes towards mental healthcare could contribute to their lower utilisation of virtual mental health services in favour of inperson ones. Additionally, decreased familiarity with virtual mental health services in their country of birth might lead to hesitancy or lack of awareness about the benefits and accessibility of these services in their new Canadian context.45 46
In this large population cohort study, we quantified the extent to which virtual mental healthcare is used across immigrant groups and where differences in care delivery exist. Strengths of this study include the use of detailed immigration data and the inclusion of immigrants from over 100 countries.14 However, our study was not without limitations. We were unable to capture individual-level socioeconomic status or track the acquisition of CLA over time, likely leading to underestimates of effect sizes and variations by immigration category. We were limited by virtual care fee codes, which do not distinguish between telephone and video visits, use of interpretation services, or provider-patient language concordance, all of which may affect the quality of assessment and counselling. We were also not able to capture non-physician visits, as these are not available in our data sets. We were not able to adjust for pre-existing healthcare utilisation or diagnoses prior to immigration. There may be residual confounding from unmeasured covariates. Virtual care implementation and utilisation is dynamic and findings may change as the use of virtual care matures.
Conclusion
Our study revealed immigration factors, including language ability and immigration category, are associated with virtual mental healthcare utilisation. We showed lower utilisation in both refugee groups compared with economic class immigrants, the magnitude of which is greatest among refugees who do have CLA at arrival. Language ability did not contribute meaningfully to findings in other immigration categories including family class and economic class immigrants. Differences across immigrant groups highlight the need to understand patient and provider preferences and to ensure equitable delivery of virtual care for youth that consider both immigration category and their arrival and evolved language ability. Recognising these intersecting factors is essential to create inclusive, effective mental health services for immigrant and refugee youth.
Data availability statement
Data are available upon reasonable request. The data set from this study is held securely in coded form at ICES. While legal data sharing agreements between ICES and data providers prohibit ICES from making the data set publicly available, access may be granted to those who meet prespecified criteria for confidential access, available at www.ices.on.ca/DAS (email: [email protected]). The full data set creation plan and underlying analytical code are available from the authors upon request, understanding that the programs may rely upon coding templates or macros that are unique to ICES and are therefore either inaccessible or may require modification.
Ethics statements
Patient consent for publication
Acknowledgments
The authors thank the Toronto Community Health Profiles Partnership for providing access to the Ontario Marginalization Index.
link