Appropriateness and determinants of outpatient antibiotic use among migrants in British Columbia, Canada: a population-based study Authors: Chia-Yuan Chang, Erica Chuang, Deva Thiruchelvam, David M Patrick, Ariana Saatchi, Hannah Lishman, Kevin Brown, Bradley J Langford, Valerie Leung, Nick Daneman, Michael Silverman, Kalisha Ramlackhan, Kevin L Schwartz, Fawziah Marra Published: 2026 ABSTRACT AbstractBackgroundMigrants represent a diverse population whose unique experiences and healthcare barriers may influence antibiotic prescribing. This study examined the appropriateness of outpatient antibiotic prescriptions and determinants of inappropriate prescriptions among migrants in British Columbia, Canada.MethodsWe conducted a population-based retrospective cohort study using health administrative databases. Antibiotic prescriptions were classified as appropriate, potentially appropriate, never appropriate, or unknown for migrants and non-migrants based on physician billing diagnostic codes. We used multivariable ordinal regression models to identify the determinants of antibiotic prescription appropriateness in 2022.ResultsFrom 2000 to 2022, the overall outpatient antibiotic prescribing rates declined by 45.3% in migrants and by 42.7% in non-migrants. Migrants had a higher proportion of never-appropriate prescriptions and a lower proportion of appropriate prescriptions than non-migrants. After adjustment for patient and prescriber factors, migrants were associated with less appropriate prescribing than non-migrants (odds ratio [OR] 1.12, 95% CI 1.11–1.13) in 2022. More than half of migrant prescription episodes were in the 35–64 years age group and were more likely to receive care from higher-volume physicians who were, on average, slightly older. Inappropriate prescribing was more likely among younger patients (0–11 versus 65+ years: OR 1.10, 95% CI 1.08–1.12), lower-income groups (lowest versus highest quintile: OR 1.04, 95% CI 1.03–1.05), and prescriptions from surgical (versus GP’s: OR 1.38, 95% CI 1.31–1.45) and urological specialists (versus GP: OR 3.06, 95% CI 2.67–3.51).ConclusionsWhile antibiotic prescribing appropriateness in BC has improved, inappropriate use remains common, particularly among the migrant population. Improving prescribing quality and equity requires strengthening stewardship in outpatient settings. Introduction The widespread use and misuse of antibiotics have driven the evolution of antimicrobial resistance.1 According to the World Health Organization, between 2018 and 2023, antibiotic resistance increased in more than 40% of monitored pathogen–antibiotic combinations across over 100 countries, with average annual rises of 5%–15%.2 Despite clear guideline recommendations, substantial antibiotic overprescribing continues to occur in primary care.3 Globally, based on clinical audits, patient interviews, and lack of indication, an estimated 30%–37% of antibiotic use is inappropriate.4 Understanding the patterns of antibiotic use is therefore essential to guide stewardship initiatives that optimize prescribing practices. In Canada, immigrants, defined by Statistics Canada as ‘persons who are, or who have ever been, landed immigrants or permanent residents,’ make up approximately one-quarter of the national population.5 The number of immigrants settling in British Columbia (BC) almost doubled between 2016 and 2021, and according to the 2021 Census, immigrants accounted for approximately 29% of the province’s population.6 This highlights the need to understand how this group engages with the healthcare system. BC offers a valuable context for such research due to its comprehensive, province-wide drug dispensing data and established stewardship programs, including Do Bugs Need Drugs? that was launched in BC in 2005, and the BC Centre for Disease Control’s (BCCDC) surveillance initiatives. The term ‘migrants’ is used broadly to encompass immigrants and non-immigrant residents born outside of Canada. Migrants represent a heterogeneous population that may face multiple barriers to healthcare access. They often encounter challenges such as differing expectations around antibiotics, language barriers, and lower health literacy, all of which can influence antibiotic-seeking behaviours.7–9 When patients believe antibiotics are necessary for symptom relief, general practitioners (GPs) frequently report prescribing in response to perceived patient expectations.10,11 A qualitative study from Europe found that healthcare providers perceived Turkish migrants to more often expect antibiotic prescriptions compared with non-migrant patients and noted communication challenges related to language and cultural context that can influence antibiotic use and prescribing behaviours in migrant groups.12 Previous Canadian research has identified inappropriate antibiotic prescribing in terms of drug choice, dose, and duration, but none have examined how these patterns differ within the migrant population.13–16 Given BC’s large and diverse migrant population, with nearly one-third being foreign-born, and its longstanding stewardship infrastructure, evaluating the quality of outpatient antibiotic prescribing in this setting can yield insights relevant both locally and nationally. We evaluated the appropriateness of outpatient antibiotic prescribing among migrants in BC. We compared prescribing trends between migrants and non-migrants from 2000 to 2022 and identified determinants of appropriateness for 2022. Methods Data sources We conducted a population-based retrospective cohort study in BC, Canada, using linked administrative health databases maintained by the BC Ministry of Health. Migrant status and demographics were determined using the Immigration, Refugees, and Citizenship Canada (IRCC) dataset.17 Demographic data on participants were obtained from the Central Demographic File (Consolidation File; CF).18 PharmaNet, a province-wide network, provided data on all antibiotic prescriptions dispensed by community pharmacies.19 This includes all outpatient prescriptions, from both primary care physicians and specialists. Data on diagnosis and prescriber characteristics were derived from the Medical Services Plan (MSP) billing system, which records reimbursement claims submitted by physicians for services provided to BC residents.20 Comorbidities were identified using MSP records, PharmaNet dispensing data, and the BC Cancer Registry, as well as hospital records from the Canadian Institute for Health Information (CIHI) Discharge Abstract Database (DAD)21 and National Ambulatory Care Reporting System (NACRS),22 which captures all inpatient and emergency department encounters in BC, respectively.19,23 Prescriber characteristics were obtained from the College of Physicians and Surgeons of BC.24 These provincial databases are well validated and have been used extensively to evaluate antimicrobial use at the population level.13,14,25,26 The datasets were linked using anonymized patient identifiers and extracted, anonymized, and made available to researchers by Population Data BC. Study population Our study population included all non-migrants and migrants residing in BC who received an outpatient antibiotic prescription between 2000 and 2022. Migrants were defined as foreign-born individuals who arrived in Canada and established residency in the province, including landed immigrants, permanent residents, refugees, and asylum seekers. Non-migrants were persons who were Canadian citizens by birth. International students and temporary workers were not captured in the IRCC database but have valid MSP coverage and could therefore not be distinguished from the non-migrant population. Residency was identified through registration in the MSP, which is mandatory for all eligible residents and their dependents. Therefore, registration was considered a reliable proxy for provincial residency. Prescriptions were excluded if, at cohort entry, the individual was not a BC resident or had missing or invalid age or sex data. Antibiotic classification and appropriateness determination Oral antibiotics dispensed in the outpatient setting were identified and classified using the World Health Organization’s Anatomical Therapeutic Chemical (ATC) system code J01 (Antibacterials for systemic use). Prescription episodes were defined as antibiotics dispensed, and if the same antibiotic was issued by the same physician to the same patient within 4 days of the prescription end date, it was considered a single episode. Each antibiotic prescription episode was linked to a diagnostic code based on the modified International Classification of Diseases, 9th Revision (ICD-9-CM) that matched the dispensing date to a practitioner service date within the preceding 5 days; then, it was assigned a hierarchy tier for antibiotic appropriateness.27 ICD-9-CM diagnostic codes were categorized into three hierarchical prescribing tiers, including ‘always,’ ‘sometimes,’ or ‘never,’ based on clinical treatment recommendations, using the classification framework described by Leslie et al.26 For example, bacterial pneumonia would be in the ‘always’ tier, acute pharyngitis would be ‘sometimes,’ and influenza would be ‘never.’ These tiers have been validated in several studies across the United States (U.S.) and Canada and prioritize diagnostic codes to reflect the most likely indication for antibiotic therapy, as well as the estimated frequency of appropriate use.16,28,29 When more than one diagnostic code was associated with a service date, or when multiple service dates occurred within a single 5-day window, precedence was given to diagnoses classified as ‘always’ requiring antibiotics, followed by ‘sometimes,’ then ‘never.’ If multiple diagnoses were present within the same tier, the primary physician billing code was used. Following the classification developed by Chua et al. and used in multiple studies,26,28,29 each antibiotic prescription episode was then grouped into four mutually exclusive outcomes: appropriate (linked to at least one ‘always’ code during the 5-day look-back period), potentially appropriate (linked to at least one ‘sometimes’ code but no ‘always’ code), never appropriate (linked only to ‘never’ codes), and unknown (no diagnostic codes identified or unmatched to the classification scheme). Statistical analyses Antibiotic prescription episode rates were derived using population estimates from the consolidation file, which includes all valid MSP patients. The consolidation file was linked with IRCC data to identify the denominator for the migrant population. The distribution of the appropriateness categories was examined by year and by migration status. A multivariable ordinal regression model with a cumulative logit link was used to identify factors associated with antibiotic prescribing appropriateness. The model included physician-level random intercepts to account for the clustering of prescriptions within prescribers. Covariates incorporated into the model included patient characteristics (migrant status, age, sex, income quintile, rural versus urban residence, comorbidities, prior hospitalization, and MSP visit frequency), physician characteristics (sex, specialty, years since graduation, origin of medical degree, practice location, and median daily patient volume), and continuity of care (measured using a score reflecting the proportion of visits a patient had with the same provider, with higher scores indicating more continuous care) and prescribing quarter (to account for seasonal variation in antibiotic use, particularly for respiratory infections, which are known to peak during winter months). We performed a sensitivity analysis to model factors associated with the unknown appropriateness of antibiotic prescriptions. We also conducted a subgroup analysis whereby we stratified by duration of residence in Canada (<5 years versus≥15 years) to explore whether differences in antibiotic prescribing varied by time since immigration. All models used non-migrants as the reference population. All analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC) and R version 4.5.1 (R Foundation for Statistical Computing, Vienna, Austria) (UBC Clinical Research Ethics Board #H23-02738). Results Characteristics of prescription episodes in migrants and non-migrants Table 1 compares the characteristics of the antibiotic prescriptions among migrants and non-migrants. In 2022, there were a total of 1 682 554 antibiotic prescription episodes, of which 283 214 (16.8%) were issued to migrants and 1 399 340 (83.2%) to non-migrants. The mean patient age was similar between groups (48.6 ± 18.0 years in migrants versus 48.7 ± 23.8 years in non-migrants; SMD 0.007), although a greater share of prescriptions among migrants were in the 35–64 years group (56.6% versus 36.0%). Antibiotics were prescribed to patients with at least one major comorbidity (Table 1, footnote a) in 45.1% of migrant prescription episodes compared with 54.5% in non-migrants (SMD 0.19). Specific chronic conditions, such as heart failure, chronic respiratory disease, and asthma, were all less frequently associated with antibiotic prescriptions among migrants than among non-migrants. Table 1.Patient and Physician Demographics among Migrant and Non-Migrant Patients in British Columbia, 2022. Numbers are antibiotic prescription episodesMigrantsn (%)Non-migrantsn (%)Standardized Mean DifferenceaTotal antibiotic prescription episodes283 214 (16.8)1 399 340 (83.2) Patient characteristics Sex0.024 Male107 601 (38.0)548 297 (39.2) Female175 613 (62.0)851 043 (61.0)Age group0.515 0–114604 (1.6%)126 310 (9.0%) 12–3460 685 (21.4%)319 652 (22.8%) 35–64160 339 (56.6%)503 567 (36.0%) 65+57 586 (20.3%)449 811 (32.1%)Income quintile0.165 Highest42 165 (14.9%)276 654 (19.8%) Next to highest49 433 (17.5%)268 538 (19.2%) Middle59 817 (21.1%)269 418 (19.3%) Next to lowest67 608 (23.9%)274 948 (19.6%) Lowest63 624 (22.5%)303 933 (21.7%) Missing567 (0.2%)5849 (0.4%)Location0.425 Urban271 490 (95.9%)1 163 135 (83.1%) Rural11 724 (4.1%)236 205 (16.9%)Comorbidities Any major comorbidityb127 653 (45.1%)763 038 (54.5%)0.190 Hypertension59 397 (21.0%)320 587 (22.9%)0.047 Ischaemic heart diseasec5603 (2.0%)41 256 (2.9%)0.063 Heart failure13 240 (4.7%)116 518 (8.3%)0.149 Chronic heart disease14 796 (5.2%)135 273 (9.7%)0.17 Atrial fibrillationc3643 (1.3%)45 106 (3.2%)0.131 Chronic respiratory disease53 353 (18.8%)425 565 (30.4%)0.271 Asthma39 124 (13.8%)278 953 (19.9%)0.164 Chronic obstructive pulmonary disease26 720 (9.4%)258 496 (18.5%)0.263 Diabetes45 165 (15.9%)185 494 (13.3%)0.076 Immunocompromising disorders14 980 (5.3%)96 754 (6.9%)0.068 Chronic kidney disease or dialysisc17 422 (6.2%)125 609 (9.0%)0.107 Advanced liver disease2129 (0.8%)21 921 (1.6%)0.076 Solid organ transplant or bone marrow transplant927 (0.3%)5677 (0.4%)0.013 Human immunodeficiency virus841 (0.3%)7412 (0.5%)0.036 Other immune system disorders14 135 (5.0%)89 444 (6.4%)0.060 Autoimmune disorders17 307 (6.1%)130 911 (9.4%)0.122 Rheumatoid arthritis7317 (2.6%)44 596 (3.2%)0.036 Inflammatory bowel disease9243 (3.3%)77 946 (5.6%)0.112 Dementia2776 (1.0%)32 211 (2.3%)0.104Prior physician visits (per year) [Mean (± SD)]21.9 ± 25.125.6 ± 29.50.135Prior hospitalizations (days) [Mean (± SD)]0.82 ± 6.61.73 ± 10.00.108 Physician characteristics Sex0.079 Male194 771 (68.8%)911 962 (65.2%) Female83 193 (29.4%)454 671 (32.5%) Missing5250 (1.9%)32 707 (2.3%) Practice location0.387 Urban271 240 (95.8%)1 183 797 (84.6%) Rural10 827 (3.8%)207 420 (14.8%) Missing1147 (0.4%)8123 (0.6%) Medical graduate0.123 Canada and the United States139 111 (49.1%)767 535 (54.8%) International138 853 (49.0%)599 098 (42.8%) Missing5250 (1.9%)32 707 (2.3%)Daily patient visits per days worked [Mean (± SD)]33.2 (59.6)28.5 (25.5)0.104Years of experience [Mean (± SD)]23.8 (13.0)22.2 (12.5)0.119Continuity of cared [Mean (± SD)]0.35 (0.27)0.30 (0.26)0.195 aStandardized mean difference (SMD) effect size: small ≈ 0.2, medium ≈ 0.5, large ≈ 0.8, as described by Cohen.30 bComorbidities include chronic respiratory diseases, chronic heart disease, hypertension, diabetes, immunocompromising disorder, autoimmune diseases, chronic kidney disease, advanced liver disease, dementia, and stroke or transient ischaemic attack. c5-year lookback; 3 consecutive months for dialysis. dThe proportion of visits a patient had with the provider who prescribed the antibiotic. Patient and Physician Demographics among Migrant and Non-Migrant Patients in British Columbia, 2022. Numbers are antibiotic prescription episodes aStandardized mean difference (SMD) effect size: small ≈ 0.2, medium ≈ 0.5, large ≈ 0.8, as described by Cohen.30 bComorbidities include chronic respiratory diseases, chronic heart disease, hypertension, diabetes, immunocompromising disorder, autoimmune diseases, chronic kidney disease, advanced liver disease, dementia, and stroke or transient ischaemic attack. c5-year lookback; 3 consecutive months for dialysis. dThe proportion of visits a patient had with the provider who prescribed the antibiotic. Physician characteristics associated with these prescriptions indicated that migrants received more prescriptions from urban prescribers than their counterparts (95.8% versus 84.6%, SMD 0.387) and were less likely to see a medical graduate from Canada or the U.S. (49.1% versus 54.8%, SMD 0.123). While years of experience were similar between the two groups (23.8 for migrants versus 22.2 for non-migrants, SMD 0.119), physicians prescribing to migrants saw more patients per day (33.2 versus 28.5, SMD 0.104). Table S1 (available as Supplementary data at JAC-AMR Online) shows that over half of these prescription episodes (56.6%) were for migrants who had lived in Canada for 15 years or longer, while approximately equal proportions occurred among those residing in the country for <5 years (14.9%), 5–9 years (14.2%), or 10–14 years (14.3%). Most prescription episodes involved migrants born in Asia (70.5%), followed by Europe (12.9%), North America (6.5%), and Africa (4.8%), with the leading countries of birth being India (23.4%), China (11.2%), and the Philippines (8.9%). Trends in antibiotic prescribing We saw a growth in total antibiotic claims, although overall outpatient antibiotic prescribing rates declined from 2000 to 2022, falling 45.3% from 492.1 to 269.4 per 1000 population among migrants and 42.8% from 590.8 to 338.2 per 1000 population among non-migrants (Figure 1). Antibiotic prescription episodes for migrants and non-migrants in British Columbia per 1000 population, 2000–2022. Figure 2 shows temporal trends in prescribing appropriateness between migrants and non-migrants. The proportion of appropriate prescriptions rose from 1.6% to 9.7% in migrants and from 2.6% to 11.4% in non-migrants—a 506% and 338% increase, respectively. In contrast, never appropriate prescriptions declined from 46.0% to 40.0% in migrants (a 13% decrease) and 44.3% to 36.9% in non-migrants (a 16.7% decrease). Potentially appropriate prescriptions remained stable at approximately 40% of all prescriptions in both groups, with a slightly lower rate among migrants. Unknown prescriptions were also relatively stable over time, accounting for 12%–15% of prescriptions in both migrants and non-migrant populations. Antibiotic prescription appropriateness for migrants and non-migrants in British Columbia, 2000–2022. Determinants of antibiotic prescribing appropriateness Using the 2022 cohort, we identified 1 461 044 prescriptions with known antibiotic appropriateness and 221 510 with unknown appropriateness. The known appropriateness prescription episodes were included in the regression models to identify the determinants of antibiotic prescribing appropriateness. In univariable analyses, lower prescribing appropriateness was associated with migrant status (versus non-migrant: OR 1.03, 95% CI 1.02–1.04), younger age (0–11 versus 65+ years: OR 1.14, 95% CI 1.12–1.15), increased continuity of care scores (OR 1.51, 95% CI 1.49–1.51), and prescriber specialty (internal medicine versus GP: OR 1.42, 95% CI 1.33–1.51; surgery versus GP: OR 1.34, 95% CI 1.26–1.43; urology versus GP: OR 3.91, 95% CI 3.26–4.69). Relative to their respective reference groups, female physicians (versus male sex: OR 0.86, 95% CI 0.83–0.90), those practicing in rural settings (versus urban: OR 0.94, 95% CI 0.90–0.98), and physicians with fewer years since graduation (<11 versus 25+ years: OR 0.61, 95% CI 0.58–0.64) were associated with higher prescribing appropriateness (Figure S1). In multivariable analysis adjusting for patient and prescriber factors, migrants had 12% higher odds of receiving an inappropriate antibiotic prescription than non-migrants (OR 1.12, 95% CI 1.11–1.13) (Figure 3). Prescribing was more appropriate among females (female versus male sex: OR 0.82, 95% CI 0.81–0.82), but inappropriate among lower-income groups (lowest versus highest quintile: OR 1.04, 95% CI 1.03–1.05), and those with chronic conditions such as diabetes (OR 1.12, 95% CI 1.11–1.13), hypertension (OR 1.07, 95% CI 1.06–1.08), immunocompromising disorders (OR 1.22, 95% CI 1.20–1.24), and chronic respiratory disease (OR 1.21, 95% CI 1.20–1.22). Physician-related factors also showed strong associations with prescribing appropriateness. Prescriptions issued by urologists (OR 3.06, 95% CI 2.67–3.51) and other surgeons (OR 1.38, 95% CI 1.31–1.45) were more likely to be inappropriate compared with those from general practitioners (GP), as well as those associated with higher continuity of care scores (OR 1.74, 95% CI 1.72–1.77). More appropriate prescribing occurred among physicians with fewer years since graduation (<11 versus 25+ years: OR 0.72, 95% CI 0.69–0.75). Multivariable analysis of factors associated with appropriateness of antibiotic prescriptions for migrants versus non-migrants in British Columbia, 2022. In the sensitivity analysis where we modelled factors associated with the unknown appropriateness of antibiotic prescriptions, the odds of inappropriate prescribing were similarly elevated across migrant duration groups compared with non-migrants (OR 1.09–1.13). (Figure S2). By migrant category, economic, family, and refugee classes all had higher odds of more inappropriate prescribing than non-migrants, with the highest risk among refugees (OR 1.19, 95% CI 1.16–1.22) (Figure S3). In the sub-group analyses stratified by duration of residence, compared with the non-migrant population, newcomers (<5 years) had 9% the odds of receiving inappropriate antibiotics, whereas individuals with ≥15 years of residence had approximately 13% the odds. The ≥15-year group was older and had more physician visits (greater healthcare utilization), and experienced longer durations of hospitalization than the newcomer group. Determinants of unknown antibiotic appropriateness In both the migrant and non-migrant groups, approximately 12%–15% of prescriptions could not be linked to MSP billing records and were therefore classified as unknown (Figure 2). Because physician characteristics are obtained from MSP records, these variables were unavailable for unlinked prescriptions. As such, physician characteristics were not included in the multivariable analyses. After adjusting for patient characteristics, higher odds of receiving a prescription with unknown appropriateness were observed for migrants (versus non-migrants: OR 1.10, 95% CI 1.08–1.11), females (versus male sex: OR 1.04, 95% CI 1.03–1.05), and across all income quintiles compared to the highest (lowest versus highest quintile: OR 1.11, 95% CI 1.10–1.13) (Figure S4). Discussion In this population-level study of BC residents, overall antibiotic prescribing rates declined in both migrants and non-migrants over two decades, reflecting improved provincial stewardship efforts. Although we found an approximately 45% decrease in outpatient antibiotic prescriptions, part of the marked decline likely reflects pandemic-related reduced transmission of respiratory infections.31–33 However, migrants consistently received a higher proportion of never appropriate prescriptions and a lower proportion of appropriate prescriptions than non-migrants. Importantly, more than half of migrant prescription episodes were in the 35–64 years age group, an age group with relatively low antibiotic use overall. This pattern likely reflects immigration policies that preferentially attract working-age adults, which, in turn, shapes both the age distribution of migrant patients and their patterns of healthcare utilization. Migrants were more likely to receive care from higher-volume physicians who were, on average, slightly older, both characteristics that have been associated with a higher likelihood of inappropriate antibiotic prescribing.34 Finally, urologists had over three times the odds of prescribing inappropriately, suggesting that antibiotics may be used more aggressively or prophylactically in their patient population, possibly due to complex urological conditions. Administrative data may not fully capture indication severity or procedural context, which could partly explain this finding. Our analysis stratifying by duration of residence suggests that the lower likelihood of inappropriate antibiotic prescribing is most pronounced among recent immigrants and decreases with longer duration of residence. This may reflect barriers to healthcare access among newcomers, leading to fewer encounters, the healthy migrant phenomenon, or a more selective group seeking care. Differences in case-mix or illness severity may also play a role. Clinicians may also prescribe more conservatively when there are language or cultural differences or greater diagnostic uncertainty. The attenuation of this effect among long-term residents may reflect increased healthcare access and utilization, as well as acculturation over time. The BCCDC Antibiotic Utilization Dashboard provides additional insight into inappropriate antibiotic use using the same antibiotic appropriateness classification. In 2024, the three most common diagnoses associated with never-appropriate antibiotic prescribing were acute upper respiratory infections of multiple or unspecified sites (ICD-9 465; 5.2%), acute bronchitis and bronchiolitis (ICD-9 466; 4.6%), and cough (ICD-9 786.2; 3.7%). These indications largely reflect upper respiratory tract conditions that are predominantly viral in aetiology and therefore unlikely to benefit from antibiotic therapy. However, assessing inappropriate antibiotic prescribing for migrants versus non-migrants was beyond the scope of this study and would be an important area for future research. A recent national analysis using data from the 2018 Canadian Community Health Survey reported higher odds of self-reported antibiotic use among immigrants than non-immigrants (adjusted odds ratio [aOR] 1.21, 95% CI 1.01–1.45).35 Although that study focused on immigrants only and did not evaluate appropriateness, its findings differ from ours. An Ontario study found that recent immigrants were more likely to receive antibiotics than long-term residents (sub-hazard ratio [sHR] 1.12, 95% CI 1.09–1.16), whereas regional analyses in Ontario found no difference in prescribing volume by immigrant density.36,37 Together, these results suggest that disparities in healthcare access and utilization may manifest as differences in prescribing quality rather than in total antibiotic exposure. International studies have also reported higher rates of inappropriate antibiotic prescribing among migrant and refugee populations. In Italy, children with parents born abroad had higher odds of receiving antibiotics for respiratory infections (versus native-born parents: aOR 1.71, 95% CI 1.36–2.15).38 Similar findings were observed in Germany, where 75% of antibiotic prescriptions for acute bronchitis among newly arrived refugees were inappropriate,39 and children of migrant parents were more likely to receive antibiotics for viral illnesses.40 Studies in low-resource settings, such as Uganda, have found comparable patterns, with 42% of prescriptions non-compliant with national treatment guidelines.41 Barriers such as differing expectations of treatment, language, and cultural factors likely contribute to these global trends.42,43 Provider-level factors also play an important role in appropriate prescribing and are shaped by complex interactions between physician perceptions, patient expectations, and barriers to communication. Family physicians frequently cite diagnostic uncertainty, parental beliefs, and a desire to maintain a positive doctor–patient relationship as reasons for prescribing antibiotics.38,44 These dynamics may be particularly relevant for migrant populations, who are more likely to receive care from international medical graduates, whose practice populations included an above-average number of patients from ethnic minority backgrounds.45 In a Quebec cohort of 852 primary-care physicians, international medical graduates (relative risk [RR] 1.78, 95% CI 1.30–2.44), those with a longer time in practice (RR 1.04, 95% CI 1.02–1.05), and higher patient volume practices (RR 1.27, 95% CI 1.09–1.48) were more likely to prescribe inappropriately for viral infections.34 Improving antibiotic appropriateness requires addressing both provider behaviours and system-level factors, including workload, clinical uncertainty, and targeted support for international medical graduates. This study has several limitations. First, we relied on administrative health data, which may be subject to misclassification or incomplete recording. In particular, some long-duration macrolide or tetracycline prescriptions may have been for non-infectious indications, which could lead to misclassification of antibiotic exposure and potentially bias our results. Having said that, these were relatively uncommon compared to shorter courses for acute infections, and therefore unlikely to have a major impact. Furthermore, the coding for the initial prescription episode would be captured and assigned the appropriate tier. These datasets have been validated and extensively used for epidemiological research and capture almost all publicly funded healthcare encounters and prescriptions in BC.46 Second, the IRCC registry began in 1985; therefore, migrants who arrived in Canada before this period were not identified (i.e. misclassified), potentially leading to under-ascertainment of long-term residents. However, because long-term migrants accounted for the largest proportion of migrants in our study, the impact of this is likely minimal. Third, because our analyses focused on antibiotic prescribing events, we were unable to assess the appropriateness of non-prescribing decisions, including potential undertreatment when antibiotics may have been clinically indicated. Finally, residual confounding from unmeasured factors such as physician attitudes, practice environments, or patient preferences remains possible despite adjustment. However, the results were similar across multiple sensitivity analyses using different modelling approaches. Further research should build on these findings by integrating additional data sources. Linking administrative health data with clinical information, microbiology results, and qualitative data on patient and provider perspectives could help elucidate whether differences in appropriate prescribing stem from clinical need, communication barriers, or access to care. Patient-centred interventions, such as culturally tailored educational campaigns and improved language accessibility, may help align expectations around antibiotic use. At the institutional-level, provider-focused strategies could include audit and feedback programs, targeted education for surgical specialists and international medical graduates, and integration of cultural competence training within provincial stewardship initiatives. Future policy should evaluate the effectiveness of migrant-specific interventions, such as language support and care navigation pathways, to promote equitable stewardship outcomes. Given the heterogeneity of the migrant population in Canada, to best tailor interventions on antibiotic use, further research is necessary to identify which ethnocultural and migrant groups are most at risk. In conclusion, this study provides the first population-based evaluation of overall outpatient antibiotic prescribing appropriateness among migrants in Canada. While the overall prescribing rates in BC have improved over time, inequities in appropriate antibiotic use persist. Strengthening culturally competent communication, prescriber training, and shared decision-making will be key to ensuring appropriate and equitable antibiotic use. Data statement Access to data provided by the Data Stewards is subject to approval but can be requested for research projects through the Data Stewards or their designated service providers. The following data sets were used in this study: Consolidation File (MSP Registration & Premium Billing), Medical Services Plan Practitioner File, National Ambulatory Care Reporting System, PharmaNet, Permanent Resident Database (IRCC), BC Cancer Registry Data, Discharge Abstract Database (Hospital Separations), and Medical Services Plan (MSP) Payment Information File. You can find further information regarding these data sets by visiting the PopData project webpage at: https://my.popdata.bc.ca/project_listings/19-080/. All inferences, opinions, and conclusions drawn in this publication are those of the author(s), and do not reflect the opinions or policies of the Data Steward(s). Supplementary Material