Raising ethnicity recording in NHS Lothian from 3% to 90% in 3 years: processes and analysis of data from Accidents and Emergencies.

BACKGROUND: The disproportionate burden of COVID-19 on ethnic minority populations has recently highlighted the necessity of maintaining accessible, routinely collected, ethnicity data within healthcare services. Despite 25 years of supportive legislation and policy in the UK, ethnicity data recording remains inconsistent, which has hindered needs assessment, evaluation and decision-making. We describe efforts to improve the completeness, quality and usage of ethnicity data within our regional health board, NHS Lothian. METHODS: The Ethnicity Coding Task Force was established with the aim of increasing ethnicity recording within NHS Lothian secondary care services from 3 to 90% over 3 years. We subsequently analysed these data specifically focusing on Accident and Emergency (A&E) use by ethnic group. RESULTS: We achieved 91%, 85% and 93% completeness of recording across inpatients, outpatients and A&E, respectively. Analysis of A&E data found a mixed pattern of attendance amongst ethnic minority populations and did not support the commonly perceived relationship between lower GP registration and higher A&E use within this population. CONCLUSIONS: We identified a successful approach to increase ethnicity recording within a regional health board, which could potentially be useful in other settings, and demonstrated the utility of these data in informing assessment of healthcare delivery and future planning.

Effects of dietary and physical activity interventions on the risk of type 2 diabetes in South Asians: meta-analysis of individual participant data from randomised controlled trials.

AIMS/HYPOTHESIS: Individuals of South Asian origin have a high risk of type 2 diabetes and of dying from a diabetes-attributable cause. Lifestyle modification intervention trials to prevent type 2 diabetes in high-risk South Asian adults have suggested more modest effects than in European-origin populations. The strength of the evidence of individual studies is limited, however. We performed an individual participant data meta-analysis of available RCTs to assess the effectiveness of lifestyle modification in South Asian populations worldwide. METHODS: We searched PubMed, EMBASE, Cochrane Library and Web of Science (to 24 September 2018) for RCTs on lifestyle modification interventions incorporating diet and/or physical activity in South Asian adults. Reviewers identified eligible studies and assessed the quality of the evidence. We obtained individual participant data on 1816 participants from all six eligible trials (four from Europe and two from India). We generated HR estimates for incident diabetes (primary outcome) and mean differences for fasting glucose, 2 h glucose, weight and waist circumference (secondary outcomes) using mixed-effect meta-analysis overall and by pre-specified subgroups. We used the Grading of Recommendations, Assessment, Development and Evaluation (GRADE) system to rate the quality of evidence of the estimates. The study is registered with the International Prospective Register of Systematic Reviews ([PROSPERO] CRD42017078003). RESULTS: Incident diabetes was observed in 12.6% of participants in the intervention groups and in 20.0% of participants in the control groups. The pooled HR for diabetes incidence was 0.65 (95% CI 0.51, 0.81; I2 = 0%) in intervention compared with control groups. The absolute risk reduction was 7.4% (95% CI 4.0, 10.2), with no interactions for the pre-specified subgroups (sex, BMI, age, study duration and region where studies were performed). The quality of evidence was rated as moderate. Mean difference for lifestyle modification vs control groups for 2 h glucose was -0.34 mmol/l (95% CI -0.62, -0.07; I2 = 50%); for weight -0.75 kg (95% CI -1.34, -0.17; I2 = 71%) and for waist -1.16 cm (95% CI -2.16, -0.16; I2 = 75%). No effect was found for fasting glucose. Findings were similar across subgroups, except for weight for European vs Indian studies (-1.10 kg vs -0.08 kg, p = 0.02 for interaction). CONCLUSIONS/INTERPRETATION: Despite modest changes for adiposity, lifestyle modification interventions in high-risk South Asian populations resulted in a clinically important 35% relative reduction in diabetes incidence, consistent across subgroups. If implemented on a large scale, lifestyle modification interventions in high-risk South Asian populations in Europe would reduce the incidence of diabetes in these populations.

Heterogeneity/granularity in ethnicity classifications project: the need for refining assessment of health status.

BACKGROUND: Identifying ethnic inequalities in health requires data with sufficiently 'granular' (fine detailed) classifications of ethnicity to capture sub-group variation in healthcare use, risk factors and health behaviors. The Robert Wood Johnson Foundation (RWJF), in the USA, commissioned us to explore granular approaches to ethnicity data collection outside of the USA, commencing with the European Union. METHODS: We examined official data sources (population censuses/registers) within the EU-28 to determine the granularity of their approach to ethnicity. When ethnic information was not available, related variables were sought (e.g. country of birth). RESULTS: Within the EU-28, we found 55% of countries collected data on ethnicity. However, only 26% of these countries (England, Wales, Northern Ireland, Scotland, Republic of Ireland, Hungary, Poland and Slovakia) had a granular approach, with half of these being within the UK. Estonia, Lithuania, Croatia, Bulgaria, Republic of Cyprus and Slovenia collected one to six categories. A 'write-in' option only was found in Latvia, Romania and the Czech Republic. Forty-five percent of countries did not collect ethnicity data but collected other related variables. CONCLUSIONS: (i) Although there is reasonable attention to the diversity of ethnic groups in data collection, a granular approach does not predominate within EU-28 classifications. (ii) Where ethnicity is collected, it is conceptualized in different ways and diverse terminology is used. (iii) A write-in option provides the most granular approach. (iv) Almost half of the countries did not collect data on ethnicity, but did collect related variables that could be used as a proxy.

Mortality, ethnicity, and country of birth on a national scale, 2001-2013: A retrospective cohort (Scottish Health and Ethnicity Linkage Study).

BACKGROUND: Migrant and ethnic minority groups are often assumed to have poor health relative to the majority population. Few countries have the capacity to study a key indicator, mortality, by ethnicity and country of birth. We hypothesized at least 10% differences in mortality by ethnic group in Scotland that would not be wholly attenuated by adjustment for socio-economic factors or country of birth. METHODS AND FINDINGS: We linked the Scottish 2001 Census to mortality data (2001-2013) in 4.62 million people (91% of estimated population), calculating age-adjusted mortality rate ratios (RRs; multiplied by 100 as percentages) with 95% confidence intervals (CIs) for 13 ethnic groups, with the White Scottish group as reference (ethnic group classification follows the Scottish 2001 Census). The Scottish Index of Multiple Deprivation, education status, and household tenure were socio-economic status (SES) confounding variables and born in the UK or Republic of Ireland (UK/RoI) an interacting and confounding variable. Smoking and diabetes data were from a primary care sub-sample (about 53,000 people). Males and females in most minority groups had lower age-adjusted mortality RRs than the White Scottish group. The 95% CIs provided good evidence that the RR was more than 10% lower in the following ethnic groups: Other White British (72.3 [95% CI 64.2, 81.3] in males and 75.2 [68.0, 83.2] in females); Other White (80.8 [72.8, 89.8] in males and 76.2 [68.6, 84.7] in females); Indian (62.6 [51.6, 76.0] in males and 60.7 [50.4, 73.1] in females); Pakistani (66.1 [57.4, 76.2] in males and 73.8 [63.7, 85.5] in females); Bangladeshi males (50.7 [32.5, 79.1]); Caribbean females (57.5 [38.5, 85.9]); and Chinese (52.2 [43.7, 62.5] in males and 65.8 [55.3, 78.2] in females). The differences were diminished but not eliminated after adjusting for UK/RoI birth and SES variables. A mortality advantage was evident in all 12 minority groups for those born abroad, but in only 6/12 male groups and 5/12 female groups of those born in the UK/RoI. In the primary care sub-sample, after adjustment for age, UK/RoI born, SES, smoking, and diabetes, the RR was not lower in Indian males (114.7 [95% CI 78.3, 167.9]) and Pakistani females (103.9 [73.9, 145.9]) than in White Scottish males and females, respectively. The main limitations were the inability to include deaths abroad and the small number of deaths in some ethnic minority groups, especially for people born in the UK/RoI. CONCLUSIONS: There was relatively low mortality for many ethnic minority groups compared to the White Scottish majority. The mortality advantage was less clear in UK/RoI-born minority group offspring than in immigrants. These differences need explaining, and health-related behaviours seem important. Similar analyses are required internationally to fulfil agreed goals for monitoring, understanding, and improving health in ethnically diverse societies and to apply to health policy, especially on health inequalities and inequities.

All-cause and cause-specific mortality of different migrant populations in Europe.

This study aimed to examine differences in all-cause mortality and main causes of death across different migrant and local-born populations living in six European countries. We used data from population and mortality registers from Denmark, England & Wales, France, Netherlands, Scotland, and Spain. We calculated age-standardized mortality rates for men and women aged 0-69 years. Country-specific data were pooled to assess weighted mortality rate ratios (MRRs) using Poisson regression. Analyses were stratified by age group, country of destination, and main cause of death. In six countries combined, all-cause mortality was lower for men and women from East Asia (MRRs 0.66; 95 % confidence interval 0.62-0.71 and 0.76; 0.69-0.82, respectively), and Other Latin America (0.44; 0.42-0.46 and 0.56; 0.54-0.59, respectively) than local-born populations. Mortality rates were similar for those from Turkey. All-cause mortality was higher in men and women from North Africa (1.09; 1.08-1.11 and 1.19; 1.17-1.22, respectively) and Eastern Europe (1.30; 1.27-1.33 and 1.05; 1.01-1.08, respectively), and women from Sub-Saharan Africa (1.34; 1.30-1.38). The pattern differed by age group and country of destination. Most migrants had higher mortality due to infectious diseases and homicide while cancer mortality and suicide were lower. CVD mortality differed by migrant population. To conclude, mortality patterns varied across migrant populations in European countries. Future research should focus both on migrant populations with favourable and less favourable mortality pattern, in order to understand this heterogeneity and to drive policy at the European level.

The HAPPY (Healthy and Active Parenting Programmme for early Years) feasibility randomised control trial: acceptability and feasibility of an intervention to reduce infant obesity.

BACKGROUND: The prevalence of infant obesity is increasing, but there is a lack of evidence-based approaches to prevent obesity at this age. This study tested the acceptability and feasibility of evaluating a theory-based intervention aimed at reducing risk of obesity in infants of overweight/obese women during and after pregnancy: the Healthy and Active Parenting Programme for Early Years (HAPPY). METHODS: A feasibility randomised controlled trial was conducted in Bradford, England. One hundred twenty overweight/obese pregnant women (Body Mass Index [BMI] ≥25 kg/m(2)) were recruited between 10-26 weeks gestation. Consenting women were randomly allocated to HAPPY (6 antenatal, 6 postnatal sessions: N = 59) or usual care (N = 61). Appropriate outcome measures for a full trial were explored, including: infant's length and weight, woman's BMI, physical activity and dietary intake of the women and infants. Health economic data were collected. Measurement occurred before randomisation and when the infant was aged 6 months and 12 months. Feasibility outcomes were: recruitment/attrition rates, and acceptability of: randomisation, measurement, and intervention. Intra-class correlations for infant weight were calculated. Fidelity was assessed through observations and facilitator feedback. Focus groups and semi-structured interviews explored acceptability of methods, implementation, and intervention content. RESULTS: Recruitment targets were met (~20 women/month) with a recruitment rate of 30 % of eligible women (120/396). There was 30 % attrition at 12 months; 66 % of recruited women failed to attend intervention sessions, but those who attended the first session were likely to continue to attend (mean 9.4/12 sessions, range 1-12). Reaction to intervention content was positive, and fidelity was high. Group clustering was minimal; an adjusted effect size of -0.25 standard deviation scores for infant weight at 12 months (95 % CI: -0.16-0.65) favouring the intervention was observed using intention to treat analyses. No adverse events were reported. CONCLUSIONS: The HAPPY intervention appeared feasible and acceptable to participants who attended and those delivering it, however attendance was low; adaptations to increase initial attendance are recommended. Whilst the study was not powered to detect a definitive effect, our results suggest a potential to reduce risk of infant obesity. The evidence reported provides valuable lessons to inform progression to a definitive trial. TRIAL REGISTRATION: Current Controlled Trials ISRCTN56735429.

Ethnic variations in upper gastrointestinal hospitalizations and deaths: the Scottish Health and Ethnicity Linkage Study.

BACKGROUND: Upper gastrointestinal (GI) diseases are common, but there is a paucity of data describing variations by ethnic group and so a lack of understanding of potential health inequalities. We studied the incidence of specific upper GI hospitalization and death by ethnicity in Scotland. METHODS: Using the Scottish Health and Ethnicity Linkage Study, linking NHS hospitalizations and mortality to the Scottish Census 2001, we explored ethnic differences in incidence (2001-10) of oesophagitis, peptic ulcer disease, gallstone disease and pancreatitis. Relative Risks (RRs) and 95% confidence intervals were calculated using Poisson regression, multiplied by 100, stratified by sex and adjusted for age, country of birth (COB) and socio-economic position. The White Scottish population (100) was the reference population. RESULTS: Ethnic variations varied by outcome and sex, e.g. adjusted RRs (95% confidence intervals) for oesophagitis were comparatively higher in Bangladeshi women (209; 124-352) and lower in Chinese men (65; 51-84) and women (69; 55-88). For peptic ulcer disease, RRs were higher in Chinese men (171; 131-223). Pakistani women had higher RRs for gallstone disease (129; 112-148) and pancreatitis (147; 109-199). The risks of upper GI diseases were lower in Other White British and Other White [e.g. for peptic ulcer disease in men, respectively (74; 64-85) and (81; 69-94)]. CONCLUSION: Risks of common upper GI diseases were comparatively lower in most White ethnic groups in Scotland. In non-White groups, however, risk varied by disease and ethnic group. These results require consideration in health policy, service planning and future research.

Pregnancy glycaemia and cord-blood levels of insulin and leptin in Pakistani and white British mother-offspring pairs: findings from a prospective pregnancy cohort.

AIMS/HYPOTHESIS: To determine the extent to which gestational fasting and postload levels of glucose explain differences in infant fat mass between UK-born Pakistani and white British infants. METHODS: Analyses were undertaken in a prospective pregnancy cohort study of 1,415 women and their singleton live-born infants (629 white British and 786 Pakistani). Infant fat mass was assessed by cord-blood leptin levels and fetal insulin secretion by cord-blood insulin levels. Maternal OGTTs were completed at 26-28 weeks of gestation. RESULTS: Pakistani women had higher fasting and postload glucose levels and greater incidence of gestational diabetes than white British women. Higher fasting and postload glucose levels were associated with higher cord-blood levels of insulin and leptin in all participants, irrespective of ethnicity. Cord-blood leptin levels were 16% (95% CI 6, 26) higher in Pakistani than in white British infants. After adjustment for fasting glucose levels, this difference attenuated to 7% (-3, 16), and with additional adjustment for cord-blood insulin levels it attenuated further to 5% (-4, 14). Path analyses supported the hypothesis that fasting glucose levels mediate the relationship of Pakistani ethnicity to greater fat mass at birth, as measured by cord-blood leptin levels; on average, 19% of this mediation involved fetal insulin secretion. Postload glucose levels did not act as an important mediator of ethnic differences in cord-blood leptin levels. Results were very similar when 130 women with gestational diabetes were removed. CONCLUSIONS/INTERPRETATION: These novel findings suggest a role of maternal pregnancy glycaemia in mediating differences in fat mass between Pakistani and white British infants.

Potential social, economic and general health benefits of consanguineous marriage: results from the Born in Bradford cohort study.

BACKGROUND: More than 1 billion people live in societies where consanguineous marriages are common. When children are born to consanguineous unions, there is an increased probability of the expression of single-gene disorders with a recessive mode of inheritance. There are presumptive social benefits of consanguineous marriages reported in the literature. METHODS: The UK's Born in Bradford birth cohort study recruited 12 453 women at 26-28 weeks' gestation between 2007 and 2010. In all, 11 396 completed a questionnaire, including questions about their relationship to their baby's father. We compared Pakistani and Other ethnic groups in consanguineous relationships and Pakistani, Other and White British groups not in consanguineous relationships, calculating percentages and age-adjusted prevalence ratios (95% confidence intervals). RESULTS: In the Pakistani group, 59.3% of women (n = 3038) were blood relatives of their baby's father. Consanguinity was uncommon in the Other ethnic group (7.3%, n = 127) and rare (n = 5) in the White British group. Compared with non-consanguineous counterparts, mothers in consanguineous relationships were socially and economically disadvantaged (e.g. never employed, less likely to have higher education). The Pakistani consanguineous group's social, economic and health lifestyle circumstances were equivalent to, in some cases better than, women in non-consanguineous relationships (e.g. up-to-date in paying bills, or in disagreeing that they wished for more warmth in their marital relationship). The consanguineous relationship group had less separation/divorce. Rates of cigarette smoking during pregnancy were lower in mothers in consanguineous relationships. CONCLUSION: Debate about consanguinity should balance the potential protective effect of consanguineous relationships with established genetic risk of congenital anomaly in children.

Ethnicity and first birth: age, smoking, delivery, gestation, weight and feeding: Scottish Health and Ethnicity Linkage Study.

BACKGROUND: We linked census and health service data sets to address the shortage of information comparing maternal characteristics and pregnancy outcomes by ethnic group in Scotland. METHODS: Retrospective cohort study linking the 2001 National Census for Scotland and hospital obstetric data (2001-08), comparing maternal age, smoking status, gestational age, caesarean section rates, birthweight, preterm birth and breastfeeding rates by ethnic group. RESULTS: In all, 144 344 women were identified as having had a first birth between 1 May 2001 and 30 April 2008. White Scottish mothers were younger [mean age 27.3 years; 95% confidence interval (CI): 27.3, 27.4] than other white groups and most non-white groups. They had the highest smoking rates (25.8%; CI: 25.5, 26.0) and the lowest rates of breastfeeding at 6-8 weeks (23.4%; CI: 23.1, 23.6), with most of the other groups being around 40%. Women from non-white minority ethnic groups in Scotland tended to have babies of lower birthweight (e.g. Pakistani mean birthweight-3105 g, white Scottish-3356 g), even after adjustment for gestational age, maternal age, education, smoking and housing tenure. This effect was more noticeable for women born in the UK. White English, Irish and other white babies tended to have higher birthweights. There was little variation between groups in caesarean section rates. CONCLUSIONS: Pregnant women from ethnic minority populations in Scotland have more favourable health behaviour than the white Scottish, although the non-white groups tend to have lower birthweight. Further exploration of the reasons for these differences has potential to benefit women from the majority population.

Consideration of ethnicity in guidelines and systematic reviews promoting lifestyle interventions: a thematic analysis.

BACKGROUND: There is a growing body of evidence supporting lifestyle interventions for the prevention of chronic disease. However, it is unclear to what extent these evidence-derived recommendations are applicable to ethnic minority populations. We sought to assess the degree of consideration of ethnicity in systematic reviews and guidelines for lifestyle interventions. METHODS: Two reviewers systematically searched seven databases to identify systematic reviews (n = 111) and UK evidence-based guidelines (n = 15) on smoking cessation, increasing physical activity and promoting healthy diet, which were then scrutinized for ethnicity-related considerations. Evidence statements were independently extracted and thematically analysed. RESULTS: Forty-one of 111 (37%) systematic reviews and 12 of 15 (80%) guidelines provided an evidence statement relating to ethnicity; however, these were often cursory and focused mainly on the need for better evidence. Five major themes emerged: (i) acknowledging the importance of diversity and how risk factors vary by ethnicity; (ii) noting evidence gaps in the effectiveness and cost-effectiveness of interventions for ethnic minorities; (iii) observing differential effects of interventions where these have been trialled with ethnic minority populations; (iv) suggesting adaptation of interventions for ethnic minority groups; (v) proposing improvements in research on interventions involving ethnic minority populations. CONCLUSIONS: Despite increasing recognition of the challenges posed by ethnic health inequalities, there remains a lack of guidance on the extent to which generic recommendations are applicable to, and how best to promote lifestyle changes in, ethnic minority populations. These important evidence gaps need to be bridged and tools developed to ensure that equity and population context is appropriately considered within evidence syntheses.

The potential for measuring ethnicity and health in a multicultural milieu--the case of type 2 diabetes in Australia.

OBJECTIVE: Ethnicity influences health in many ways. For example, type 2 diabetes (T2DM) is disproportionately prevalent among certain ethnic groups. Assessing ethnicity is difficult, and numerous proxy measures are used to capture its various components. Australian guidelines specify a set of variables for measuring ethnicity, and how such parameters should be categorised. Using T2DM data collections as an illustrative example, this study sought to examine how ethnicity is measured in Australian health databases and, by comparing current practice with Australia's existing benchmark recommendations, to identify potential areas for improvement of the health data landscape. DESIGN: We identified databases containing information from which ethnic group-specific estimates of T2DM burden may be gleaned. For each database, details regarding ethnicity variables were extracted, and compared with the Australian guidelines. RESULTS: Data collection instruments for 32 relevant databases were reviewed. Birthplace was recorded in 27 databases (84%), but mode of birthplace assessment varied. Indigenous status was commonly recorded (78%, n=25), but only nine databases recorded other aspects of self-perceived race/ethnicity. Of 28 survey/audit databases, 14 accommodated linguistic preferences other than English, and 11 either excluded non-English speakers or those for whom a translator was not available, or only offered questionnaires in English. CONCLUSIONS: Considerable variation exists in the measurement of ethnicity in Australian health data-sets. While various markers of ethnicity provide complementary information about the ethnic profile within a data-set, non-uniform measurement renders comparison between data-sets difficult. A standardised approach is necessary, and identifying the ethnicity variables that are particularly relevant to the health sector is warranted. Including self-identified ethnicity in Australia's set of recommended indicators and as a core component of the national census should be considered. Globalisation and increasing migration mean that these findings have implications internationally, including for multi-ethnic countries throughout North America and Europe.

Smoking cessation interventions for ethnic minority groups--a systematic review of adapted interventions.

OBJECTIVE: Existing smoking cessation interventions tend to be under utilized by ethnic minority groups. We sought to identify smoking cessation interventions that have been adapted to meet the needs of African-, Chinese- and South Asian-origin populations, to increase understanding of the approaches used to promote behavior change, to assess their acceptability to the target populations, and to evaluate their effectiveness. METHODS: Two reviewers independently searched for, identified, critically appraised and extracted data from studies identified from 11 databases (January 1950-April 2013). Study quality was assessed using validated instruments (EPHPP and STROBE). Adaptations were independently coded using an established typology, and findings descriptively summarized and thematically synthesized. RESULTS: 23 studies described interventions adapted for African-Americans, and five for Chinese-origin populations. No intervention adapted for South-Asian populations was identified. Six studies directly compared a culturally adapted versus a non-adapted intervention. Adapted interventions were more acceptable to ethnic minority groups, but this did not translate into improvements in smoking cessation outcomes. CONCLUSIONS: Given the evidence of greater acceptability of adapted interventions, it may be ethically preferable to use these. There is, however, no clear evidence of the effectiveness of adapted interventions in promoting smoking cessation in ethnic minority groups.

Sizable variations in circulatory disease mortality by region and country of birth in six European countries.

BACKGROUND: Circulatory disease mortality inequalities by country of birth (COB) have been demonstrated for some EU countries but pan-European analyses are lacking. We examine inequalities in circulatory mortality by geographical region/COB for six EU countries. METHODS: We obtained national death and population data from Denmark, England and Wales, France, the Netherlands, Scotland and Sweden. Mortality rate ratios (MRRs) were constructed to examine differences in circulatory, ischaemic heart disease (IHD) and cerebrovascular disease mortality by geographical region/COB in 35-74 years old men and women. RESULTS: South Asians in Denmark, England and Wales and France experienced excess circulatory disease mortality (MRRs 1.37-1.91). Similar results were seen for Eastern Europeans in these countries as well as in Sweden (MRRs 1.05-1.51), for those of Middle Eastern origin in Denmark (MRR = 1.49) and France (MRR = 1.15), and for East and West sub-Saharan Africans in England and Wales (MRRs 1.28 and 1.39) and France (MRRs 1.24 and 1.22). Low ratios were observed for East Asians in France, Scotland and Sweden (MRRs 0.64-0.50). Sex-specific analyses showed results of similar direction but different effect sizes. The pattern for IHD mortality was similar to that for circulatory disease mortality. Two- to three-fold excess cerebrovascular disease mortality was found for several foreign-born groups compared with the local-born populations in some countries. CONCLUSIONS: Circulatory disease mortality varies by geographical region/COB within six EU countries. Excess mortality was observed for some migrant populations, less for others. Reliable pan-European data are needed for monitoring and understanding mortality inequalities in Europe's multiethnic populations.

Research agenda for tackling inequalities related to migration and ethnicity in Europe.

Reducing ethnic inequalities and inequities in health needs to become a higher priority for public health research. Active involvement of migrant and ethnic minority populations in European population health research is necessary, for data show important inequalities but evidence, particularly on effectiveness, is sparse and strategic overviews rarer still. Ethnically disaggregated health surveillance systems are developing slowly, and pragmatically, often using country of birth. The principles to adopt, given the gaps between the ideal and the current reality, need wider discussion. Ethics may provide both principles and impetus. Doing no harm, doing good, respecting the research participants' autonomy are good starting points. More emphasis is needed on justice, fairness and equality, participation and communicating effectively. Ethnic minority groups are willing participants in trials once linguistic and trust-related barriers are overcome, though recruitment costs are higher. Guidelines and strategies by European bodies with research funding or enforcement responsibilities are needed. The research community needs an infrastructure within which to collect and utilize evidence. Research can help migrant and ethnic minority groups to participate more fully in our multi-ethnic societies.