UK ethnicity data collection for healthcare statistics: the South Asian perspective

Underrecording and underreporting of participant ethnicity in clinical trials is persistent and is a threat to inclusivity and generalizability

OBJECTIVES

People from ethnic minority groups are underserved by randomized trials, and poor representation of these groups reduces generalizability of results. There is no guidance on which ethnicity categories are appropriate for use in trials and thus inconsistency exists. The purpose of this study is to establish, in a large sample of trials, if participant ethnicity is recorded, how it is obtained (categories used), and if its reporting varies from its recording.

STUDY DESIGN AND SETTING

We reviewed trial documentation for 407 randomized controlled trials published in the UK National Institute of Health Research library from 2016 to 2021. We extracted data on the recording (if it was recorded and the categories used) and reporting (if the categories remained the same as those obtained, or not) of ethnicity for each trial along with demographics. In the analysis we categorized the manner of recording and reporting of ethnicity in the trials according to UK Census ethnicity categories.

RESULTS

Ethnicity was recorded in 67.3% (n = 274) of trials. The location in the trial report where ethnicity was recorded was available for 42% (n = 116) of trials. The details on how ethnicity was collected (predefined categories or self-defined) was available for 54/274 (20%) of trials and details on the specifics of the categories recorded was available for 44 (16%) trials. Of the 44, 6 of those did not go on to report on ethnicity in the trial report. Of the remaining 38, only 13 reported ethnicity exactly as it had been recorded. Taken as a whole from the 407 trial reports examined 9.3% (38/407) of trials demonstrated exactly how they both recorded, and reported, ethnicity. Authors made reference to whom results were relevant in terms of ethnicity in 80/407 (19.7%).

CONCLUSION

Ethnicity is underrecorded and underreported in clinical trials. This is a threat to the generalizability of the findings and needs to be improved.

JBI systematic review protocol of text/opinions on how to best collect race-based data in healthcare contexts

INTRODUCTION

Racialized population groups have worse health outcomes across the world compared with non-racialized populations. Evidence suggests that collecting race-based data should be done to mitigate racism as a barrier to health equity, and to amplify community voices, promote transparency, accountability, and shared governance of data. However, limited evidence exists on the best ways to collect race-based data in healthcare contexts. This systematic review aims to synthesize opinions and texts on the best practices for collecting race-based data in healthcare contexts.

METHODS AND ANALYSES

We will use the Joanna Briggs Institute (JBI) method for synthesizing text and opinions. JBI is a global leader in evidence-based healthcare and provides guidelines for systematic reviews. The search strategy will locate both published and unpublished papers in English in CINAHL, Medline, PsycINFO, Scopus and Web of Science from 1 January 2013 to 1 January 2023, as well as unpublished studies and grey literature of relevant government and research websites using Google and ProQuest Dissertations and Theses. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement methodology for systematic reviews of text and opinion will be applied, including screening and appraisal of the evidence by two independent reviewers and data extraction using JBI's Narrative, Opinion, Text, Assessment, Review Instrument. This JBI systematic review of opinion and text will address gaps in knowledge about the best ways to collect race-based data in healthcare. Improvements in race-based data collection, may be related to structural policies that address racism in healthcare. Community participation may also be used to increase knowledge about collecting race-based data.

ETHICS AND DISSEMINATION

The systematic review does not involve human subjects. Findings will be disseminated through a peer-reviewed publication in JBI evidence synthesis, conferences and media.

PROSPERO REGISTRATION NUMBER

CRD42022368270.

Ethnicity and the surgical management of early invasive breast cancer in over 164 000 women

Background

Limited information is available about patterns of surgical management of early breast cancer by ethnicity of women in England, and any potential inequalities in the treatment received for breast cancer.

Methods

National Cancer Registration and Analysis Service data for women diagnosed with early invasive breast cancer (ICD–10 C50) during 2012–2017 were analysed. Multivariable logistic regression was used to estimate odds ratios (ORs) and 95 per cent confidence intervals for the risk of mastectomy versus breast‐conserving surgery by ethnicity (black African, black Caribbean, Indian, Pakistani and white), adjusting for age, region, deprivation, year of diagnosis, co‐morbidity and stage at diagnosis.

Results

Data from 164 143 women were included in the analysis. The proportion of women undergoing mastectomy fell by approximately 5 per cent between 2012 and 2017 across all the ethnic groups examined. In unadjusted analyses, each ethnic minority group had a significantly higher odds of mastectomy than white women; however, in the fully adjusted model, there were no significantly increased odds of having mastectomy for women of any ethnic minority group examined. For example, compared with white women, the unadjusted and fully adjusted ORs for mastectomy were 1·14 (95 per cent c.i. 1·05 to 1·20) and 1·04 (0·96 to 1·14) respectively for Indian women, and 1·45 (1·30 to 1·62) and 1·00 (0·89 to 1·13) for black African women. This attenuation in OR by ethnicity was largely due to adjustment for age and stage.

Conclusion

Allowing for different patterns of age and stage at presentation, the surgical management of early breast cancer is similar in all women, regardless of ethnicity.

Engagement barriers and service inequities in the NHS Breast Screening Programme: Views from British-Pakistani women

Objectives

Previous research has largely attempted to explore breast screening experiences of South Asian women by combining opinions from Pakistani, Bangladeshi, and Indian women. This research often fails to reach the most underserved sub-groups of this population, with socioeconomic status not routinely reported, and English fluency being a participation requirement. With uptake low amongst British-Pakistani women, this study explores the experiences these women encounter when accessing the NHS Breast Screening Programme.

Methods

19 one-to-one semi-structured interviews were carried out with British-Pakistani women from East Lancashire, UK. 14 interviews were conducted via an interpreter.

Results

Data were analysed using thematic analysis. Three themes were identified: ‘Absence of autonomy in screening and healthcare access’ describes how currently the screening service does not facilitate confidentiality or independence. Access requires third-party intervention, with language barriers preventing self-expression. ‘Appraisal of information sources’ makes distinctions between community and NHS communication. Whereas community communication was invaluable, NHS materials were deemed inaccessible due to translation incongruences and incomprehensible terminology. ‘Personal suppositions of breast screening’ explores the subjective issues associated with disengagement, including, the cultural misalignment of the service, and perceiving screening as a symptomatic service.

Conclusions

British-Pakistani women face some unique challenges when accessing breast screening. To promote uptake, the service needs to address the translation of screening materials and optimize upon community networks to disseminate knowledge, including knowledge of the screening environment within the context of culture to promote informed choice about attendance.

Patient perspectives on routinely being asked about their race and ethnicity: Qualitative study in primary care

OBJECTIVE

To understand patients' perspectives on responding to a question about their race and ethnicity in a primary care setting.

DESIGN

Qualitative study using semistructured individual interviews conducted between May and July 2016.

SETTING

An academic family health team in Toronto, Ont, where collection of sociodemographic data has been routine since 2013.

PARTICIPANTS

Twenty-seven patients from 5 of the 6 clinic sites of the family health team, ranging in age, sex, educational background, and immigration status.

METHODS

Semistructured interviews were conducted with patients who completed a sociodemographic questionnaire after registration for their medical appointment. Patients were asked whether responding to the question was difficult or uncomfortable, how they interpreted the term race and ethnicity, and what response options they considered. Interviews were audiorecorded, transcribed, and coded iteratively.

MAIN FINDINGS

Patients did not report discomfort with responding to a question about race and ethnicity in their family doctor's office. Although many patients considered the question straightforward, some patients reported different interpretations of the question. For example, some thought the question about race and ethnicity related to parental origin or ancestry, whereas others considered the question to be about personal place of birth or upbringing. Many patients appreciated being able to select from a variety of specific response options, but this also posed a difficulty for patients who could not easily find an option that reflected their identity. Patients with mixed heritage experienced the most challenges selecting a response.

CONCLUSION

Patients attending a primary care clinic were not uncomfortable responding to a question about race and ethnicity. However, patients had different interpretations of what was being asked. Future research should explore perspectives of patients in other primary care settings and test different methods for collecting data about their race and ethnicity.

Challenges of the heterogeneous nutrition response: interpreting the group mean

Extensive research demonstrates unequivocally that nutrition plays a fundamental role in maintaining health and preventing disease. In parallel nutrition research provides evidence that the risks and benefits of diet and lifestyle choices do not affect people equally, as people are inherently variable in their responses to nutrition and associated interventions to maintain health and prevent disease. To simplify the inherent complexity of human subjects and their nutrition, with the aim of managing expectations for dietary guidance required to ensure healthy populations and individuals, nutrition researchers often seek to group individuals based on commonly used criteria. This strategy relies on demonstrating meaningful conclusions based on comparison of group mean responses of assigned groups. Such studies are often confounded by the heterogeneous nutrition response. Commonly used criteria applied in grouping study populations and individuals to identify mechanisms and determinants of responses to nutrition often contribute to the problem of interpreting the results of group comparisons. Challenges of interpreting the group mean using diverse populations will be discussed with respect to studies in human subjects, in vivo and in vitro model systems. Future advances in nutrition research to tackle inter-individual variation require a coordinated approach from funders, learned societies, nutrition scientists, publishers and reviewers of the scientific literature. This will be essential to develop and implement improved study design, data recording, analysis and reporting to facilitate more insightful interpretation of the group mean with respect to population diversity and the heterogeneous nutrition response.

Potential harms associated with routine collection of patient sociodemographic information: A rapid review

Background

Health systems are recommended to capture routine patient sociodemographic data as a key step in providing equitable person‐centred care. However, collection of this information has the potential to cause harm, especially for vulnerable or potentially disadvantaged patients.

Objective

To identify harms perceived or experienced by patients, their families, or health‐care providers from collection of sociodemographic information during routine health‐care visits and to identify best practices for when, by whom and how to collect this information.

Search Strategy

We searched OVID MEDLINE, PubMed “related articles” via NLM and healthevidence.org to the end of January 2018 and assessed reference lists and related citations of included studies.

Inclusion Criteria

We included studies reporting on harms of collecting patient sociodemographic information in health‐care settings.

Data Extraction and Synthesis

Data on study characteristics and types of harms were extracted and summarized narratively.

Main Results

Eighteen studies were included; 13 provided patient perceptions or experiences with the collection of these data and seven studies reported on provider perceptions. Five reported on patient recommendations for collecting sociodemographic information. Patients and providers reported similar potential harms which were grouped into the following themes: altered behaviour which may affect care‐seeking, data misuse or privacy concerns, discomfort, discrimination, offence or negative reactions, and quality of care. Patients suggested that sociodemographic information be collected face to face by a physician.

Discussion and Conclusions

Overall, patients support the collection of sociodemographic information. However, harms are possible, especially for some population subgroups. Harms may be mitigated by providing a rationale for the collection of this information.

Dietary intake and factors influencing eating behaviours in overweight and obese South Asian men living in the UK: mixed method study

OBJECTIVE

It is widely recognised that South Asian men living in the UK are more likely to develop type 2 diabetes mellitus (T2DM) than their white British counterparts. Despite this, limited data have been published quantifying current dietary intake patterns and qualitatively exploring eating behaviours in this population. The objectives of this study were to (1) assess diet, (2) explore perceptions of T2DM, (3) investigate factors influencing eating behaviours in overweight/obese South Asian men and (4) determine the suitability of the UK Diet and Diabetes Questionnaire (UKDDQ) for use in this population.

SETTING

Community-based setting in the Greater London, UK area.

PARTICIPANTS

South Asian men aged 18-64 years, with a body mass index of over 23.0 kg/m², not previously diagnosed with T2DM.

METHODS

A cross-sectional mixed-methods design, including assessment of dietary intake using UKDDQ (n=63), followed by semistructured interviews in a purposive sample (n=36).

RESULTS

UKDDQ scores indicated 54% of participants had a 'healthy' diet with a mean sample score of 3.44±0.43 out of a maximum of 5. Oily fish consumption was low (1.84±1.85). Body weight was positively associated with a high-added sugar subscore (r=0.253, p=0.047), with 69.8% of the men having 'unhealthy' intakes of sugar-sweetened beverages. Cultural commitments (eg, extended family and faith events), motivation and time were identified as key barriers to dietary change, with family support an important facilitator to making healthy dietary changes. Participants stated that UKDDQ was suitable for assessing diets of South Asians and made suggestions for tailoring questions related to rice consumption, providing examples of Indian sweets, and including ghee as a fat source.

CONCLUSION

Many of the areas of dietary improvement and factors affecting eating behaviours identified in this study are similar to those observed in the general UK population. Consumption of sugar-sweetened beverages in particular was high; given the association between their consumption and the risk of T2DM, this should be an area of primary focus for healthcare professionals. Nevertheless, there are sociocultural factors unique to this population that need to be considered when designing culturally specific programs to reduce the development of T2DM in this high-risk population.

'Foreigners', 'ethnic minorities', and 'non-Western allochtoons': an analysis of the development of 'ethnicity' in health policy in the Netherlands from 1970 to 2015

BACKGROUND

The Netherlands, because of the sustained and systematic attention it paid to migrant and minority health issues during the last quarter of the twentieth century, has been depicted as being progressive in its approach to healthcare for migrants and minorities. Recently, however, these progressive policies have changed, reflecting a trend towards problematising issues of integration in order to focus on the responsibilities that migrants and ethnic minorities bear in terms of their health. This article explores these shifts and specifically the development of particular categories of ethnicity, and examines the wider consequences that have arisen as a result.

METHODS

The analysis presented here entailed a qualitative content analysis of health policies for migrants and ethnic minorities from 1970 to 2015, and examined various documents and materials produced by the institutions and organisations responsible for implementing these healthcare policies during the period from 1970 to 2015.

RESULTS

Four distinct periods of political discourse related to health policy for migrants and ethnic minorities were identified. These periods of political discourse were found to shape the manner in which ethnicity and various categories and representation of foreigners, later ethnic minorities, and at present non-Western allochtoons are constructed in health policy and the implantation practices that follow. At present, in the Netherlands the term allochtoon is used to describe people who are considered of foreign heritage, and its antonym autochtoon is used for those who are considered native to the Netherlands. We discuss the scientific reproduction and even geneticisation of these politically produced categories of autochtoon, Western allochtoon, and non-Western allochtoon-a phenomenon that occurs when politically produced categories are prescribed or taken up by other health sectors.

CONCLUSIONS

The categories of autochtoon, Western allochtoon, and non-Western allochtoon in the health sciences and the field of ethnicity and health in the Netherlands today have been co-produced by society and science. Policy formulated on the basis of specific political discourse informs the conceptualisations about groups and categories, issues, and solutions, and when these are institutionalised in subsequent health policy, databases, research, and care practices, these ethnic categorisations are replicated in a manner that renders them 'real' and enables them to be applied both socially and scientifically, culminating in pronouncements as to who is the same and who is different in Dutch society and science.

Recent migrants' perspectives on antibiotic use and prescribing in primary care: a qualitative study

BACKGROUND

Currently there is great interest in antibiotic prescribing practices in the UK, but little is known about the experiences of the increasing numbers of recent migrants (those present in the UK for >1 year but <5 years) registered at GP practices. Qualitative research has suggested that reasons for not prescribing antibiotics may not be clearly communicated to migrants.

AIM

This study aimed to explore the factors that shape migrants' experiences of and attitudes to antibiotics, and to suggest ways to improve effective communication around their use.

DESIGN AND SETTING

A qualitative study on recent migrants' health beliefs, values, and experiences in a community setting in primary care.

METHOD

Twenty-three recent migrants were interviewed in their preferred language by trained community researchers. The research team conducted a thematic analysis, focusing on health beliefs, engaging with health services, transnational medicine, and concepts of fairness. Experiences around antibiotics were a strong emerging theme.

RESULTS

Three reasons were identified for antibiotics seeking: first, holding an 'infectious model' of illness implying that antibiotics are required quickly to avoid illness becoming worse or spreading to others; second, reasoning that other medications will be less effective for people 'used to' antibiotics'; and third, perceiving antibiotic prescription as a sign of being taken seriously. Some participants obtained antibiotics from their country of origin or migrant networks in the UK; others changed their mind and accepted alternatives.

CONCLUSION

Primary care professionals should aim to understand migrants' perspectives to improve communication with patients. Further research is needed to identify different strategies needed to respond to the varying understandings of antibiotics held by migrants.

Responsiveness of primary care services: development of a patient-report measure – qualitative study and initial quantitative pilot testing

Background

Primary care service providers do not always respond to the needs of diverse groups of patients, and so certain patients groups are disadvantaged. General practitioner (GP) practices are increasingly encouraged to be more responsive to patients’ needs in order to address inequalities.

Objectives

(1) Explore the meaning of responsiveness in primary care. (2) Develop a patient-report questionnaire for use as a measure of patient experience of responsiveness by a range of primary care organisations (PCOs). (3) Investigate methods of population mapping available to GP practices.

Design setting

PCOs, including GP practices, walk-in centres and community pharmacies.

Participants

Patients and staff from 12 PCOs in the East Midlands in the development stage, and 15 PCOs across three different regions of England in stage 3.

Interventions

To investigate what responsiveness means, we conducted a literature review and interviews with patients and staff in 12 PCOs. We developed, tested and piloted the use of a questionnaire. We explored approaches for GP practices to understand the diversity of their populations.

Main outcome measures

(1) Definition of primary care responsiveness. (2) Three patient-report questionnaires to provide an assessment of patient experience of GP, pharmacy and walk-in centre responsiveness. (3) Insight into challenges in collecting diversity data in primary care.

Results

The literature covers three overlapping themes of service quality, inequalities and patient involvement. We suggest that responsiveness is achieved through alignment between service delivery and patient needs, involving strategies to improve responsive service delivery, and efforts to manage patient expectations. We identified three components of responsive service delivery: proactive population orientation, reactive population orientation and individual patient orientation. PCOs tend to utilise reactive strategies rather than proactive approaches. Questionnaire development involved efforts to include patients who are ‘seldom heard’. The questionnaire was checked for validity and consistency and is available in three versions (GP, pharmacy, and walk-in centre), and in Easy Read format. We found the questionnaires to be acceptable to patients, and to have content validity. We produced some preliminary evidence of reliability and construct validity. Measuring and improving responsiveness requires PCOs to understand the characteristics of their patient population, but we identified significant barriers and challenges to this.

Conclusions

Responsiveness is a complex concept. It involves alignment between service delivery and the needs of diverse patient groups. Reactive and proactive strategies at individual and population level are required, but PCOs mainly rely on reactive approaches. Being responsive means giving good care equally to all, and some groups may require extra support. What this extra support is will differ in different patient populations, and so knowledge of the practice population is essential. Practices need to be motivated to collect and use diversity data. Future work needed includes further evaluation of the patient-report questionnaires, including Easy Read versions, to provide further evidence of their quality and acceptability; research into how to facilitative the use of patient experience data in primary care; and implementation of strategies to improve responsiveness, and evaluation of effectiveness.

Funding

The National Institute for Health Research Service Delivery and Organisation programme.

Accuracy of routinely recorded ethnic group information compared with self-reported ethnicity: evidence from the English Cancer Patient Experience survey

OBJECTIVE

To describe the accuracy of ethnicity coding in contemporary National Health Service (NHS) hospital records compared with the 'gold standard' of self-reported ethnicity.

DESIGN

Secondary analysis of data from a cross-sectional survey (2011).

SETTING

All NHS hospitals in England providing cancer treatment.

PARTICIPANTS

58 721 patients with cancer for whom ethnicity information (Office for National Statistics 2001 16-group classification) was available from self-reports (considered to represent the 'gold standard') and their hospital record.

METHODS

We calculated the sensitivity and positive predictive value (PPV) of hospital record ethnicity. Further, we used a logistic regression model to explore independent predictors of discordance between recorded and self-reported ethnicity.

RESULTS

Overall, 4.9% (4.7-5.1%) of people had their self-reported ethnic group incorrectly recorded in their hospital records. Recorded White British ethnicity had high sensitivity (97.8% (97.7-98.0%)) and PPV (98.1% (98.0-98.2%)) for self-reported White British ethnicity. Recorded ethnicity information for the 15 other ethnic groups was substantially less accurate with 41.2% (39.7-42.7%) incorrect. Recorded 'Mixed' ethnicity had low sensitivity (12-31%) and PPVs (12-42%). Recorded 'Indian', 'Chinese', 'Black-Caribbean' and 'Black African' ethnic groups had intermediate levels of sensitivity (65-80%) and PPV (80-89%, respectively). In multivariable analysis, belonging to an ethnic minority group was the only independent predictor of discordant ethnicity information. There was strong evidence that the degree of discordance of ethnicity information varied substantially between different hospitals (p<0.0001).

DISCUSSION

Current levels of accuracy of ethnicity information in NHS hospital records support valid profiling of White/non-White ethnic differences. However, profiling of ethnic differences in process or outcome measures for specific minority groups may contain a substantial and variable degree of misclassification error. These considerations should be taken into account when interpreting ethnic variation audits based on routine data and inform initiatives aimed at improving the accuracy of ethnicity information in hospital records.