https://orcid.org/0000-0001-8411-1653
Figures
Abstract
Background
Within the UK, more than a million people cannot speak English well or at all. The lack of data on English proficiency means that the link between English proficiency and health status and care utilisation is not comprehensively quantified.
Objective
Describe the association between English proficiency and patients’ health status and healthcare utilisation, and demonstrate that GP data can be useful in understanding the health burdens of those with poor language proficiency.
Methods
The Northwest London (NWL) Discover-NOW database contains linked, deidentified records from General Practices (GPs), hospitals, and social care in NWL. Using this data, we examined health outcomes and healthcare utilisation of people in Brent who are not proficient in English.
Results
Prevalence of age-sex-adjusted cardiometabolic conditions was higher in groups that were not proficient in English or spoke a main language other than English. Primary and secondary healthcare utilisation was also higher in groups that were not proficient in English.
Conclusion
This work is the first to quantify healthcare utilisation of those not proficient in English using a large, representative sample in a UK setting. It highlights poorer health outcomes in this group. There is a need to improve provision of language support, starting at registration, which would allow for this group to be better understood.
Citation: Yeoh S, Stafford M (2026) Exploring health inequalities by language proficiency; A routine health records study set in England. PLoS One 21(8): e0342624. https://doi.org/10.1371/journal.pone.0342624
Editor: Ari Samaranayaka, University of Otago, NEW ZEALAND
Received: January 26, 2026; Accepted: July 9, 2026; Published: August 19, 2026
Copyright: © 2026 Yeoh, Stafford. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: Data cannot be shared publicly because they were provided to researchers who met the criteria to access confidential data for a specified access period. Data can be applied for by application to the DiscoverNow North West London Data Access Committee. https://www.nwlondonicb.nhs.uk/professionals/digital-and-it/whole-systems-integrated-care-wsic/information-de-identified-users.
Funding: The author(s) received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
Abbreviations: COPD, Chronic obstructive pulmonary disease; GP, General practice; HRG, Healthcare Resource Group; IMD, Index of Multiple Deprivation; LSOA, Lower-level super output areas; LTC, Long-term condition; NHS, National Health Service; NP, Not proficient in English; NP-Prev, Not proficient in English, captured previously in GP record; NP-Eng, English as their first language plus a record indicative of poor English proficiency; NWL, Northwest London; ONS, Office for National Statistics; P-English, Proficient in English with English as their first language; P-Other, Proficient in English, but English is not their first language; P-Unknown, Proficient in English, unknown first language; UK, United Kingdom
1. Introduction
At the time of Census 2021, 16% of the population of England and Wales had been born outside the UK [1]. Out of this, more than a million could not speak English well or at all [2]. However, English proficiency is regarded as important for integration by both migrants and non-migrants. Migrants with better English abilities reside in less deprived neighbourhoods [3] and have more social support [4]. Migrants who speak English to their children at home are perceived by natives as being more integrated [5].
Language proficiency increases with time spent in the destination country [6, 7] – an effect largely associated with increased use of the language. Such opportunities arise during employment and conversing with others in the community [7]. There is therefore a reinforcing effect where migrants with limited or no English ability are less likely to be employed [6] and more likely to live in language enclaves [3], which in turn limit opportunities for English language use and learning. Children are more likely than adults to learn the language [7] older age is related to lower levels of English proficiency [6,8]. Women tend to have lower proficiency in the new language compared to men [8,9].
English proficiency is important for accessing services in the UK, such as healthcare. Research has demonstrated that even those comfortable with English can misunderstand medical terminology [10]. Despite the need for professional language services in the NHS [11,12], interpreting services are underused, with friends, family, and/or bilingual healthcare professionals filling the gap [13]. A study of 41 general practices in the UK found that healthcare professionals may attempt to act as translators even when their proficiency in the patients’ language is low [14]. In a recent UK study, only 63% of South Asian patients with limited or no English proficiency reported using professional interpreting services in primary care [15]. Interpreter services may also be hard to access; one study found that booking interpreter-assisted appointments required additional language support [16].
Extensive literature from the USA highlights how health outcomes are worse for non-proficient speakers. Low English proficiency has been associated with lower preventative healthcare uptake, such as screenings [17], and vaccinations [18]. While disease prevalence is not necessarily higher in those with low English proficiency, evidence suggests that symptom awareness is poorer [19,20], resulting in potential under- or misdiagnosis [20]. For non-proficient speakers with diagnosed conditions, disease awareness and management has been found to be poorer [21–26], and health interventions less effective, compared to English-proficient counterparts [27]. However, where patients receive language-concordant care, preventative healthcare uptake [17] and condition management [24, 25] are improved. As to secondary healthcare, a review concluded that non-proficient patients had fewer specialist physician visits, were more likely to forgo necessary medical care, less likely to receive preventative care, and more likely to be re-admitted into hospital within 30 days [28].
However, the healthcare system in the USA differs markedly from UK’s universal coverage system. Within the UK, 2011 Census data shows that only 65% of those who could not speak English well or at all had self-reported ‘good’ or ‘very good’ health - compared to 88% of those who could [29]. There is also quantitative evidence that non-English language preference is related to higher rates of chronic heart disease (CHD), but not stroke or hypertension [30], although the latter could be related to underdiagnosis [31]. Evidence on the relationship between English ability and diabetes is mixed [30,31], as is that between English ability and obesity [30,32]. However, these studies have focused on the South Asian population [31] and cardiometabolic disease [30–32]. Extant studies have not quantified the relationship between healthcare utilisation and language proficiency. Qualitative studies have focused on healthcare experiences. Those with limited proficiency in English report feeling discriminated due to their poor command of English [33–35]. In some cases, a limited command of English may affect migrants’ understanding of how the NHS works [36], which in turn leads to misuse of the healthcare system such as use of emergency services for non-emergency conditions [37,38]. How limited English proficiency differentiates migrants’ health and healthcare utilisation from natives in the UK is therefore neither well-described nor well-quantified.
Lack of data on English proficiency has contributed to this evidence gap. Routine healthcare records from primary and secondary care contain some data on English proficiency but have so far been under-used. This study uses healthcare records from patients in Brent, London, an area of high ethnic and language diversity, to quantify levels of health and use of care by language proficiency. A third of residents in Brent speak a main language other than English – the second highest rate of all boroughs in England and Wales – and 7% cannot speak English well or at all [39]. The contributions of this paper are to quantify non-proficient patients’ health status and healthcare utilisation and demonstrate that routine health data can be useful in understanding the health burdens of people not proficient in English.
2. Methods
2.1 Dataset
The Discover-NOW research platform contains linked, de-identified records from GPs, hospitals, and social care in Northwest London. Authors did not have access to information that could identify individual participants during or after the study. Apart from health data, the database also contains demographic information, such as birth month, sex, first language, and index of multiple deprivation (IMD) decile – which is a small-area measure of relative deprivation across England and Wales. The data were accessed for research purposes on 1st June 2025. As of June 2025, there were records for 2,872,782 currently registered, living patients in the database.
Since children generally attend appointments with parents, parents’ English proficiency is likely to matter more than children’s English proficiency. However, such parent-child relations are difficult to determine from the routine health records. Thus, children under the age of 18 were excluded from the sample. The final analytical sample was formed by patients over the age of 18, with a known, recorded gender, living in Brent, registered with a GP in the Brent health borough (previously Brent Clinical Commissioning Group), and alive from 1 January 2024 to 31 December 2024.
The NWL Discover-NOW research platform has ethics and HRA approval for all de-identified research REC Reference 23/WM/0196 IRAS ID 333128. The current study was approved by the Whole Systems Integrated Care Data Access Committee (ID-160).
2.2 Measures
2.2.1 Language proficiency.
Any patient with any GP record since 2010 of needing a translator, being unable to read or speak English, or speaking English poorly was categorised as not being proficient in English (henceforth, NP, Fig 1). It is noted, however, that language proficiency increases with time spent in the destination country [6,7]. A previous study found that the probability of self-reported English proficiency was significantly higher starting from the 4th year since migration, compared to the year of migration [7]. Thus, patients who had such a GP record previously may have since improved their English language proficiency. Patients who had such a record 4 or more years ago were therefore separately categorised as being not proficient in English previously (henceforth, NP-Prev).
P-Eng = Proficient in English with English as their first language; P-Unk = Proficient in English, unknown first language; P-Other = Proficient in English, but English is not their first language; NP-Prev = Not proficient in English, captured previously in GP record; NP = Not proficient in English; NP-Eng = English as their first language plus a record indicative of poor English proficiency.
It is also possible that there are patients who are not proficient in English, but who did not have such a record. Moreover, even if migrants have good levels of English ability, their health and care status may differ from locals. Patients with a first language other than English are likely to be such migrants. As such, patients who did not have any records indicating poor English ability (i.e., the assumed ‘Proficient’ group) were further subdivided on the basis of whether their first language was English (P-English), other than English (P-Other), or unknown (P-Unknown). See S1 Table for a summary of first languages.
There were 1,476 (0.5%) residents who had English as their first language but also had a record indicative of poor English proficiency. This group was designated ‘NP-English’. Of the NP-English group, 1,246 had an entry in their GP records indicating an interpreter was needed.
2.2.2 Health and healthcare utilisation.
Discover-NOW also includes de-identified records from GPs on whether an individual has long-term conditions (LTCs) and LTC diagnosis date. LTCs assessed in the primary care quality outcomes framework were included here.
Four measures were used as indicators for healthcare utilisation: number of GP encounters, number of emergency department attendances, number of avoidable emergency department attendances, and number of hospital admissions. Number of GP encounters was proxied by number of distinct days with an entry in the GP record. While this included administrative entries where there were no patient-GP interactions, this over-counting should nonetheless be similar across both proficient and non-proficient groups. Similarly, number of emergency department attendances and number of hospital admissions were proxied by number of distinct start dates. Emergency department attendances with a Healthcare Resource Group (HRG) code indicating no further investigation or significant medication were labelled as an “avoidable attendance”, based on the assumption that the issue could potentially have resolved in primary care. These measures were calculated for the period from 1 January 2024 to 31 December 2024.
2.3 Statistical analyses
To compare LTC prevalence, age- and gender-standardised ratios were calculated by indirect standardisation. Age- and gender-specific rates of the P-English group were applied to the age and gender composition of the five other language groups to derive expected counts. These expected counts were then compared against observed counts to derive ratios.
As there was a positive skew in the number of GP encounters, a log-transformation (plus one) was applied. A linear regression model was then built for the transformed model, adjusting also for age, gender, deprivation in the area of residence (based on the Index of Multiple Deprivation in the lower super output area), and number of LTCs.
For number of emergency attendances, avoidable emergency attendances, and hospital admissions, the measures were dichotomised, creating measures indicating whether each patient had had one attendance (admission) or not, and whether they had had multiple attendances (admissions) or not. A separate logistic regression model was created for each of these six binary measures, and also included age, gender, deprivation, and number of LTCs. Additionally, for the avoidable emergency department attendances regression models, patients who had not had any emergency department attendances were excluded from the model sample.
There may be ethnic differences in English language proficiency. Ethnic inequalities in health are also well-documented. As such, a sensitivity analysis was conducted by creating regression models that also adjusted for ethnicity to control for potential confounding, in addition to the above-mentioned variables. Statistical analysis was performed using RStudio and the R script is available by request to the authors.
3. Results
3.1 Demographics
The search yielded 314,754 residents in Brent over the age of 18, of whom 9,403 (3.0%) were NP, 12,836 (4.1%) were NP-Prev, 91,999 (29.2%) were P-Other, 97,611 (31.0%) were P-Unk. In the 2021 Census, 26.2% of the population did not have English as their main language but could speak English well, and 7.5% of the population could not speak English well or at all [40].
Demographics are presented in Table 1. All non-proficient groups had more females (NP: 58% females; NP-Prev: 57% females; NP-English: 54% females) than males. The P-English group had a roughly equal proportion of males and females, while the other proficient groups had more males than females. The P-Unknown group was the youngest (median age 37 years), while the NP-Prev group was the oldest (median age 47 years). A lower proportion of the P-Unknown and P-Other group lived in the most deprived deciles, while almost a quarter of the NP-English group (24.6%) lived in the two most deprived deciles.
3.2 Long-term conditions
Raw (unadjusted) prevalences of long-term conditions are summarised in Table 2.
After accounting for age and gender, the P-Other, P-Unknown, and NP-Prev groups had overall lower prevalence of LTC morbidity and multiple long-term conditions compared to the P-English group (Figs 2 and 3). Across several LTCs (anxiety, depression, serious mental illness, asthma, COPD, cancer, and CKD), the P-English group had higher prevalence than all other groups less the NP-English group. Outcomes were worst for the NP-English group, which had higher prevalences of various LTCs (including serious mental illness, dementia, epilepsy, and learning disability) compared to the P-English group.
P-Unk = Proficient in English, unknown first language; P-Other = Proficient in English, but English is not their first language; NP-Prev = Not proficient in English, captured previously in GP record; NP = Not proficient in English; NP-Eng = English as their first language plus a record indicative of poor English proficiency. 95% confidence intervals are shown.
P-Unk = Proficient in English, unknown first language; P-Other = Proficient in English, but English is not their first language; NP-Prev = Not proficient in English, captured previously in GP record; NP = Not proficient in English; NP-Eng = English as their first language plus a record indicative of poor English proficiency. 95% confidence intervals are shown.
All groups, less the P-Unknown group, had higher prevalence of CHD and diabetes, compared to the P-English group. The NP and NP-Prev groups additionally had higher prevalence of heart failure. The NP and NP-English groups also had higher prevalence of hypertension, and obesity. The P-Other group had higher prevalence of CHD and diabetes, while the P-Unknown group had lower prevalence across all LTCs.
3.3 Healthcare utilisation
The NP-English group had significantly higher use across all measures compared with the P-English group, after adjusting for age, sex, deprivation and number of LTCs (Table 3). They had 60% higher odds of multiple hospital admissions and 56% higher odds of multiple or avoidable ED attendances compared with the P-English group.
The NP group had significantly more GP encounters than the P-English group (β = 0.12; 95% CI: 0.11, 0.13; note that this estimate is for log + 1 number of GP encounters). This translates to an expected value of GP encounters + 1 per year that is 13% higher in the NP group compared with the P-English group. The NP group also had higher odds of hospital use, ranging from 30% higher odds of any admission to 42% higher odds of multiple ED attendances, compared with the P-English group.
The NP-Prev group had statistically significantly fewer GP encounters but the magnitude of differences was small. They were less likely to have multiple hospital admissions compared with the P-English group (OR: 0.79, 95% CI: 0.68–0.91) (Table 3).
The P-Unknown group had lower healthcare use across all measures, including 39% lower odds of multiple hospital admissions while the P-Other group also had significantly lower healthcare use across all measures less multiple avoidable ED attendances compared with the P-English group (Table 3).
In models that were additionally adjusted for ethnicity, trends were similar, less that the P-Other group did not have significantly higher odds of having had an avoidable emergency department attendances, the NP-Prev group had significantly higher odds of having had any (OR: 1.08, 95% CI: 1.04–1.13) or multiple emergency department attendances (OR: 1.10, 95% CI: 1.03–1.17), and the NP-English group did not have significantly higher odds of having had an avoidable ED attendance. Results are included in S2 Table.
4. Discussion
Based on diagnoses from GP records, the prevalence of LTCs and multiple LTCs was higher in the P-English group than all other comparison groups, less the NP-English group. Poorest health status was seen for the NP-English group, which had higher prevalence of serious mental illness, CHD, hypertension, epilepsy, dementia, diabetes, obesity, and learning disability, compared to the P-English group. After adjusting for age, sex, deprivation, and number of LTCs, the NP-English group also had higher GP and emergency department utilisation, and higher rates of hospital admission.
For the other two non-proficient groups (i.e., NP and NP-Prev), the prevalence of LTCs and multiple LTCs was generally lower. The prevalence for various LTCs (e.g., anxiety, depression, serious mental illness, asthma, and cancer) was lower in these non-proficient groups compared to the P-English group. However, prevalence of cardiometabolic conditions tended to be higher for these non-proficient groups.
Controlling for age, sex, deprivation, and LTCs, the use of healthcare services was higher across all measures for the NP group. The NP-Prev group tended to have fewer GP encounters and hospital admissions. Higher use of healthcare in non-proficient groups may be due to their higher levels of clinical need. However, differences in healthcare use remained after adjustment for number of long-term health conditions, which proxies clinical need to some extent. This suggests that need may not be the sole driver of higher activity although this statistical adjustment does not capture how advanced the condition is or how severely it impacts daily life, and acute conditions were not considered. Another potential explanation for the higher level of healthcare use is that difficulties in communication could result in a higher number of GP encounters being needed to reach a diagnosis or treatment plan. Low awareness or difficulties in accessing preventative and early intervention services could result in higher use of hospitals (either Emergency Department attendances or admissions) if patients present later in the course of the disease. This study of routine data could not explore possible explanations but qualitative reports set in the UK support the role of language [36–38] independently of level of need.
The P-Other group corresponds most closely to the non-English language preference group described in a study using routine patient data from general practices in South London, UK [30]. Similarly to the findings in that study, the P-Other group had higher rates of CHD compared to the P-English group (or the English language preference group), and no difference in rates of hypertension. However, where the South London study found higher rates of obesity and no difference in rates of stroke, rates of obesity and stroke are lower in the P-Other group here. The P-Other group also had fewer GP encounters, and lower odds of any or multiple emergency department attendances and any or multiple hospital admissions.
The NP-English group may have the lowest level of English proficiency of all groups. Members of this group were recorded as having English as their main language. However, more than 80% of this group had a record indicating an interpreter was needed – implying poor or no English ability. This contradiction could perhaps be explained by the fact that first language is typically recorded at registration and language barriers arising at the point of GP registration – which is often in English (e.g., [41,42]) – can lead to inaccurate recording of language preference. Individual language barriers can be ameliorated with the aid of friends and/or family [13]. Yet, the inaccurate recording of language preference at the point of registration suggests such aid was not available to members of the NP-English group, highlighting that the NP-English group may therefore reflect the group with the lowest English ability and less support with this.
4.1 Implications
There is a need to improve uptake of interpreting services by signposting to them, providing easy access, more proactively offering interpreting services, and improving their quality. A national survey of people with limited English language proficiency notes that interpreting services continue to be underused compared to the need for them [15]. The NP-English group additionally reflects that there is a clear and urgent need for interpreting services for some even when English is the main language, given their poor health status and high use of healthcare. Apart from this group, other non-proficient groups would also benefit from the use of interpreting services, since language barriers have been reported to negatively affect primary healthcare experience [33–35] and our study highlights higher use of secondary healthcare in the NP group.
This analysis has highlighted cardiometabolic conditions among those with low proficiency in the local setting. This shows where signposting, provision and training of interpreter services could be most beneficial locally. This analysis could be replicated using similar routine healthcare data in other localities to inform public health, general practice, other community and hospital teams about where to target interpreter provision and improvement.
While GP data can be a valuable resource for researching and understanding populations, accuracy in data recording should be improved. The scale of GP data means that even at smaller geographies such as local authorities, niche population groups can be studied at scale, which is invaluable to understanding health and planning interventions. However, the NP-English group suggests that, starting at the point of registration, there may be inaccuracies in recording. Indeed, in health records, inaccuracies and inconsistencies in recording tend to occur more frequently in minoritised groups [43]. Improving recording is key to improving our understanding of overlooked groups, and the first step in addressing and tackling health inequalities.
4.2 Limitations
A limitation of this study is the likelihood that there are under-estimates where people not proficient in English are concerned. This arises in two areas: under-count of non-proficiency, and under-diagnosis of some LTCs in non-proficient groups. Registered patients with an entry indicating non-proficiency are categorised as non-proficient (Section 2.2.1). Additionally categorising on the basis of recorded first language provides a further filter to proxy whether patients are proficient in English. However, as the NP-English group suggests, there may be patients for whom their first language is incorrectly-recorded as English. It is therefore also possible that a registered patient can be non-proficient in English, but have their first language incorrectly recorded as English, and no record indicating non-proficiency. The number of non-proficient English speakers is likely an under-estimate of the true figure. Secondly, some languages may not have equivalent terms to describe LTCs [44]. For instance, Urdu does not have an equivalent concept for ’anxiety’ [45], while Gujarati has no equivalent concept for ’depression’ [46]. Indeed, in languages (e.g., Somali, Bengali) that do not have equivalent conceptualisations of mental health conditions, these conditions may be stigmatised, which discourages help-seeking behaviour and diagnosis [45,47,48]. These languages are spoken by a large proportion of this study’s sample, and may therefore contribute to under-diagnosis of some LTCs.
4.3 Further research
An avenue for future work would be to examine English language proficiency for an even wider set of first languages and ethnic groups. Both this work and another study using routine patient data [30] investigated language proficiency and health within borough-level settings. When comparing the first languages spoken in these two studies, the differences in ethnic composition are apparent. European languages (e.g., Portuguese, Spanish, French, Polish) are spoken by more than half the non-English cohort in the South London study [30]. However, in this study, South Asian languages and Arabic comprise the most common first languages. Poor English language proficiency is not homogenous, and greater linguistic distance (e.g., between English and Arabic, vs. between English and French) increases the difficulty and cost of language acquisition [7]. The differences in the most common first languages – especially since neither of these studies has complete direct information on language proficiency – could contribute to some differences in results.
Finally, this study presents a cross-sectional perspective on the differences in health status and healthcare use. GP data provides an opportunity to examine and track anonymised individuals to understand how their health changes over time. Further study could therefore follow these groups – in particular, the P-English, P-Other, and NP groups – to understand how their health status and healthcare use changes over time.
5. Conclusion
This work is the first in the UK to quantify the health utilisation of those not proficient in English using linked GP and hospital data. It highlights poorer health outcomes in this group and reinforces the need for improved provision of language support. This work also demonstrates that GP data can be used to understand the impacts of language proficiency at scale for research and at small geographies for designing and improving language support locally. Better recording would improve this resource, which would facilitate understanding of overlooked groups.
Supporting information
S2 Table. Association between healthcare utilisation and language proficiency among adults in Brent, adjusted for ethnicity.
https://doi.org/10.1371/journal.pone.0342624.s002
(DOCX)
Acknowledgments
We thank the Northwest London Whole Systems Integrated Care De-identified Data team for their continued support with the data resource, metadata and data access. This work uses data provided by patients and collected by the NHS as part of their care and support.
References
- 1. Office for National Statistics. International migration, England and Wales: Census 2021; 2022 [cited 2025 May 13]. Available from: https://www.ons.gov.uk/peoplepopulationandcommunity/populationandmigration/internationalmigration/bulletins/internationalmigrationenglandandwales/census2021
- 2. Office for National Statistics. Language, England and Wales: Census 2021; 2022 [cited 2025 May 13]. Available from: https://www.ons.gov.uk/peoplepopulationandcommunity/culturalidentity/language/bulletins/languageenglandandwales/census2021
- 3. Aoki Y, Santiago L. Where to live? English proficiency and residential location of UK migrants. J Econ Behav Organ. 2024;221:73–93.
- 4. Kearns A, Whitley E. Getting there? The effects of functional factors, time and place on the social integration of migrants. J Ethn Migr Stud. 2015;41(13):2105–29. pmid:28473737
- 5. Sobolewska M, Galandini S, Lessard-Phillips L. The public view of immigrant integration: multidimensional and consensual. Evidence from survey experiments in the UK and the Netherlands. J Ethn Migr Stud. 2016;43(1):58–79.
- 6.
Fernández Reino M, Brindle B. English language use and proficiency of migrants in the UK. UK: COMPAS, University of Oxford; 2019.
- 7.
Isphording IE. What drives the language proficiency of immigrants? IZA World of Labor. 2015.
- 8. Morrice L, Tip LK, Collyer M, Brown R. ‘You can’t have a good integration when you don’t have a good communication’: English-language learning among resettled refugees in England. J Refug Stud. 2019;34(1):681–99.
- 9. Ng E, Pottie K, Spitzer D. Official language proficiency and self-reported health among immigrants to Canada. Health Rep. 2011;22(4):15–23. pmid:22352148
- 10. Gotlieb R, Praska C, Hendrickson MA, Marmet J, Charpentier V, Hause E, et al. Accuracy in patient understanding of common medical phrases. JAMA Netw Open. 2022;5(11):e2242972. pmid:36449293
- 11. Gill PS, Shankar A, Quirke T, Freemantle N. Access to interpreting services in England: secondary analysis of national data. BMC Public Health. 2009;9:12. pmid:19138392
- 12. Jones D, Gill P. Breaking down language barriers: the NHS needs to provide accessible interpreting services for all. Br Med J. 1998.
- 13. Whitaker KL, Krystallidou D, Williams ED, Black G, Vindrola-Padros C, Gill P, et al. Understanding uptake and experience of interpreting services in primary care in a south Asian population in the UK. JAMA Netw Open. 2022;5(11):e2244092. pmid:36445711
- 14. Gill PS, Beavan J, Calvert M, Freemantle N. The unmet need for interpreting provision in UK primary care. PLoS One. 2011;6(6):e20837. pmid:21695146
- 15. Hieke G, Williams ED, Gill P, Black GB, Islam L, Vindrola-Padros C, et al. Uptake and experience of professional interpreting services in primary care in a South Asian population: a national cross-sectional study. BMC Prim Care. 2024;25(1):405. pmid:39604912
- 16. Hieke G, Black GB, Yargawa J, Vindrola-Padros C, Gill P, Islam L, et al. South Asian patient experiences of professional interpreting service provision in general practice in England: a qualitative interview study. Int J Equity Health. 2025;24(1):104. pmid:40241154
- 17. Xie Z, Chen G, Suk R, Dixon B, Jo A, Hong Y-R. Limited English proficiency and screening for cervical, breast, and colorectal cancers among Asian American adults. J Racial Ethn Health Disparities. 2023;10(2):977–85. pmid:35297497
- 18. Quadri NS, Knowlton G, Benitez GV, Ehresmann KR, LaFrance AB, DeFor TA, et al. Evaluation of preferred language and timing of covid-19 vaccine uptake and disease outcomes. JAMA Netw Open. 2023;6(4):237877.
- 19. Latif Z, Inam M, Tummala A, Nelson B, Makuvire TT, Warraich HJ. Cardiovascular disease in patients with limited English proficiency: a narrative review. Curr Probl Cardiol. 2025;103:103107.
- 20. Holman H, Müller F, Bhangu N, Kottutt J, Alshaarawy O. Impact of limited English proficiency on the diagnosis and awareness of diabetes: the national health and nutrition examination survey, 2003-2018. Diabetes Care. 2022;45(8):e124–5. pmid:35763452
- 21. Müller F, Holman H, Bhangu N, Kottutt J, Azhary H, Alshaarawy O. Use of antihyperglycemic medications among US people with limited English proficiency. J Gen Intern Med. 2025;40(8):1803–10. pmid:39875767
- 22. Holman H, Müller F, Bhangu N, Kottutt J, Alshaarawy O. Impact of limited English proficiency on the control of diabetes and associated cardiovascular risk factors. The National Health and Nutrition Examination Survey, 2003–2018. Prev Med. 2023;167:107394.
- 23. Kim EJ, Kim T, Paasche-Orlow MK, Rose AJ, Hanchate AD. Disparities in hypertension associated with limited English proficiency. J Gen Intern Med. 2017;32(6):632–9. pmid:28160188
- 24. Parker MM, Fernández A, Moffet HH, Grant RW, Torreblanca A, Karter AJ. Association of patient-physician language concordance and glycemic control for limited-English proficiency latinos with type 2 diabetes. JAMA Intern Med. 2017;177(3):380–7. pmid:28114680
- 25. Clark JR, Shlobin NA, Batra A, Liotta EM. The relationship between limited English proficiency and outcomes in stroke prevention, management, and rehabilitation: a systematic review. Front Neurol. 2022;13:790553. pmid:35185760
- 26. Lehman R, Moriarty H. Limited English proficiency and outcomes in the intensive care unit: an integrated review. J Transcult Nurs. 2024;35(3):226–36. pmid:38351583
- 27. Njeru JW, Wieland ML, Kwete G, Tan EM, Breitkopf CR, Agunwamba AA, et al. Diabetes mellitus management among patients with limited English proficiency: a systematic review and meta-analysis. J Gen Intern Med. 2018;33(4):524–32. pmid:29256089
- 28. Twersky SE, Jefferson R, Garcia-Ortiz L, Williams E, Pina C. The impact of limited English proficiency on healthcare access and outcomes in the U.S.: a scoping review. Healthcare (Basel). 2024;12(3):364. pmid:38338249
- 29. Office for National Statistics. People who cannot speak English well are more likely to be in poor health; 2015 [cited 2025 May 12]. Available from: https://www.ons.gov.uk/peoplepopulationandcommunity/culturalidentity/language/articles/peoplewhocannotspeakenglishwellaremorelikelytobeinpoorhealth/2015-07-09
- 30. Mackay A, Ashworth M, White P. The role of spoken language in cardiovascular health inequalities: a cross-sectional study of people with non-English language preference. BJGP Open. 2017;1(4):bjgpopen17X101241. pmid:30564693
- 31. Mainous AG 3rd, Baker R, Majeed A, Koopman RJ, Everett CJ, Saxena S, et al. English language skills and diabetes and hypertension among foreign-born South Asian adults in England. Public Health Rep. 2006;121(3):331–6. pmid:16640158
- 32. Higgins V, Nazroo J, Brown M. Pathways to ethnic differences in obesity: The role of migration, culture and socio-economic position in the UK. SSM Popul Health. 2019;7:100394. pmid:31032393
- 33. Jackowska M, von Wagner C, Wardle J, Juszczyk D, Luszczynska A, Waller J. Cervical screening among migrant women: a qualitative study of Polish, Slovak and Romanian women in London, UK. J Fam Plann Reprod Health Care. 2012;38(4):229–38. pmid:22219504
- 34. Ekezie W, Cassambai S, Czyznikowska B, Curtis F, O’Mahoney LL, Willis A, et al. Health and social care experience and research perception of different ethnic minority populations in the East Midlands, United Kingdom (REPRESENT study). Health Expect. 2024;27(1):e13944. pmid:39102736
- 35. Phung V-H, Asghar Z, Matiti M, Siriwardena AN. Understanding how Eastern European migrants use and experience UK health services: a systematic scoping review. BMC Health Serv Res. 2020;20(1):173. pmid:32143703
- 36. Bell S, Edelstein M, Zatoński M, Ramsay M, Mounier-Jack S. “I don’t think anybody explained to me how it works”: qualitative study exploring vaccination and primary health service access and uptake amongst Polish and Romanian communities in England. BMJ Open. 2019;9(7):e028228. pmid:31289079
- 37. Richards J, Kliner M, Brierley S, Stroud L. Maternal and infant health of Eastern Europeans in Bradford, UK: a qualitative study. Community Pract. 2014;87(9):33–6. pmid:25286741
- 38. Leaman AM, Rysdale E, Webber R. Use of the emergency department by Polish migrant workers. Emerg Med J. 2006;23(12):918–9. pmid:17130598
- 39. Brent Council. Census 2021: Language in Brent; 2023 [cited 2025 Sept 11]. Available from: https://data.brent.gov.uk/dataset/language-in-brent-2021-census-topic-report-2nr88
- 40. Office for National Statistics. Proficiency in English Language - Census Maps. [cited 2025 Jun 26]. Available from: https://www.ons.gov.uk/census/maps/choropleth/identity/proficiency-in-english-language/english-proficiency-4a/main-language-is-not-english-english-or-welsh-in-wales-cannot-speak-english-or-cannot-speak-english-well?lad=E09000005
- 41. Bell S, Saliba V, Ramsay M, Mounier-Jack S. What have we learnt from measles outbreaks in 3 English cities? A qualitative exploration of factors influencing vaccination uptake in Romanian and Roma Romanian communities. BMC Public Health. 2020;20(1):381. pmid:32293379
- 42. Asif Z, Kienzler H. Structural barriers to refugee, asylum seeker and undocumented migrant healthcare access. Perceptions of Doctors of the World caseworkers in the UK. SSM Ment Health. 2022;2:100088.
- 43.
Scobie S, Spencer J, Raleigh VS. Ethnicity coding in English health service datasets. London, UK: Nuffield Trust; 2021.
- 44. Acharya B, Basnet M, Rimal P, Citrin D, Hirachan S, Swar S, et al. Translating mental health diagnostic and symptom terminology to train health workers and engage patients in cross-cultural, non-English speaking populations. Int J Ment Health Syst. 2017;11:62. pmid:29026440
- 45. Loewenthal D, Mohamed A, Mukhopadhyay S, Ganesh K, Thomas R. Reducing the barriers to accessing psychological therapies for Bengali, Urdu, Tamil and Somali communities in the UK: some implications for training, policy and practice. Br J Guid Couns. 2012;40(1):43–66.
- 46. Patel K, Shaw I. Mental health and the Gujarati community: accounting for the low incidence rates of mental illness. Ment Health Rev J. 2009;14(4):12–24.
- 47. Linney C, Ye S, Redwood S, Mohamed A, Farah A, Biddle L, et al. “Crazy person is crazy person. It doesn’t differentiate”: an exploration into Somali views of mental health and access to healthcare in an established UK Somali community. Int J Equity Health. 2020;19(1):190. pmid:33109227
- 48. Mohsin F, Aravala S, Rahman T, Ali SH, Taher MD, Mitra P, et al. Psychiatric healthcare experiences of South Asian patients with severe mental illness diagnoses and their families in New York City: a qualitative study. Community Ment Health J. 2025;61(1):39–49. pmid:39046623