Veena Raleigh unpicks the recent ONS findings that ethnic minority groups in England and Wales live longer than White groups
The Office for National Statistics (ONS) findings that ethnic minorities in England and Wales had a higher life expectancy at birth than the White population during 2011-14 may come as a surprise to many, as ethnic minorities are widely reported to have worse health outcomes. [1,2] In fact, the picture is more complicated.
These headline life expectancy figures belie a complex picture of different ethnic groups being disproportionately affected by different causes of death. For example, White groups had the highest death rates from most cancers, dementia, and Alzheimer’s disease, which accounted for almost 40% of total deaths nationally. In contrast, South Asian and Black groups had higher mortality from diabetes and some cardiovascular diseases, such as heart disease, stroke, and hypertension. Mortality from specific cancers was also higher in some minority groups—for example, lung cancer among Bangladeshi males and prostate cancer among Black males. Other evidence shows South Asian and Black groups also have higher rates of infant and maternal mortality, although these cause relatively fewer deaths. [3,4]
Although overall mortality has declined in all ethnic groups, ethnic differences in overall and cause-specific mortality have remained similar through to recent years: in 2017-19 the White group had higher overall mortality than any other ethnic group and higher rates than most other ethnic groups for many leading causes of death. 
Many factors could be driving these ethnic differences in life expectancy. For example, research shows that migrants globally tend to be healthier than the host population. [6-13] However, this “healthy migrant effect” wanes with longer duration of residence, possibly as lifestyles change to mirror the majority population. [14-19] Black African and other Asian groups in the UK are more recent migrants, which may in part explain their higher life expectancy.  Moreover, some risk factors, e.g. smoking and alcohol consumption, are lower among ethnic minorities. [21,22] Deprivation is a significant driver of poor health in all groups and, although ethnic minorities experience greater deprivation on average than the White population, it’s not clear if the relationship applies equally across all ethnic groups. [23,24] We welcome ONS’ intention to examine the effects of these and other underlying drivers.
ONS notes caveats to its methodology (which it aims to refine further).  As death records don’t include ethnicity, they were linked to 2011 census records to derive the ethnicity of deceased individuals and adjustments were needed to allow for census under-enumeration, non-linkage, post-censal migration and other factors. The methodology is the most pragmatic given the limitations of currently available data, and similar to how ONS analyses ethnic differences in mortality from covid-19—findings which have greatly informed public understanding of and policies for reducing the pandemic’s impact on population sub-groups. Moreover, in Scotland, following poor data quality after ethnicity recording in death certification was introduced in 2012, an Expert Group’s recommendation that linkage of self-reported ethnicity in the census to health records is the best option for accurate ethnic group analysis has been accepted. 
Strengths of the ONS analysis are that it provides national data and is based on self-reported ethnicity which is not subject to the coding biases in health records. [27-29] The findings are also broadly consistent with evidence on ethnic differences in disease and mortality, and therefore have face validity. [30-38]
Covid-19’s disproportionate impact on ethnic minority groups reversed the previous pattern in some (notably Pakistani and Bangladeshi) ethnic minority groups who now experience higher overall mortality than the White population; other groups show no difference or retain their mortality advantage.  It’s premature to speculate whether pre-pandemic life expectancy patterns among ethnic groups will be restored when covid-19 is under control.
The covid-19 pandemic has highlighted the power of data nationally and globally in understanding the prevalence and impact of disease, and informing preventive interventions e.g. data on infections, hospital admissions, and deaths has been used internationally to manage the pandemic, despite significant data limitations including variations in coding this new disease.  The alternative to imperfect data would have been a blind spot. That’s why it’s imperative that available data are used to inform policy decisions, while simultaneously data quality is improved.
The ONS report adds significantly to our understanding of ethnic differences in mortality, brought sharply into focus by the covid-19 pandemic. The findings provide some clear pointers to where interventions are needed by government, health, and other agencies to reduce the burden of disease and health inequalities, including between ethnic groups. For example, cardiovascular disease and diabetes are potentially preventable diseases, yet they continue to cause significant morbidity and mortality among ethnic minority groups and the poor in all ethnic groups (including White). UK’s cancer survival rates compare poorly with many European countries. 
Health inequalities were wide and widening before the pandemic, and covid-19 has greatly amplified these inequalities. The need for a cross-government strategy that addresses the diverse health needs of all groups at risk of poor health and high mortality—including both ethnic minorities and the White poor—has never been more urgent.
Veena S Raleigh, Senior Fellow, The King’s Fund.
Competing interests: none declared.