The impact of the COVID-19 pandemic may be worse among certain groups, for example ethnic minorities, people living in poorer areas, or with pre-existing illness. This could be due to the epidemic starting in cities where proportionately more Black, Asian and Minority Ethnic (BAME) or poorer people live, or there may be health-related reasons, such as BAME or poorer people having more existing illness. This has been highlighted in many provisional reports focusing on hospital deaths directly attributed to COVID-19. Simply counting these gives an incomplete and possibly misleading picture of the epidemic in the community. However, a wider perspective can be obtained by counting all deaths and comparing this to the number expected from the same season over previous years. This is known as excess mortality.
Our study will use anonymised general practice medical records to examine whether excess mortality during 2020 is greater among certain groups (BAME, more deprived, pre-existing illness) compared to patterns from 2010-9. It will seek to explain variations by risk factors such as smoking or obesity, and by a wide range of medical conditions.
Additionally, we will investigate patterns of illness before death for all patients who died over the last 10 years. We will assess if people dying recently are spending more or less of their final years in good health compared to patients who died at the beginning of the study, focusing on ethnic and socio-economic differences. We will also investigate the how COVID-19 has impacted any trends here.
Preliminary analyses during the COVID-19 pandemic have reported higher risks among people from Black, Asian and Minority Ethnic (BAME) backgrounds, those living in more deprived areas, and those with existing co-morbidities including diabetes. Regional variations in BAME groups and deprivation, combined with the progress of the UK epidemic may partially explain these differences, as could higher prevalences of obesity, cardiovascular disease, diabetes and associated treatments. Initial reporting largely focused on COVID-19 hospital deaths, where coding may not have been consistent throughout. Focusing on these deaths assumes that there has not been an indirect impact on mortality, with people avoiding secondary care. Therefore, a fuller assessment of the impact of COVID-19 needs to consider excess all-cause mortality as a measure of impact, while still accounting for these individual characteristics.
We will combine data from Aurum and GOLD to estimate excess mortality rates among registered adults during the pandemic period (mid-March to mid-June 2020) compared to corresponding periods in the previous ten years. Poisson regression will be used to compare excess mortality rates between pre-defined groups of interest (BAME, deprivation, co-morbidities), estimating how much the excess rate widened in 2020 for these groups. The models will seek to explain ethnic and socio-economic variations by including other individual level modifiers of COVID-19 risk such as age, smoking, obesity and pre-existing medical conditions.
Finally, we will investigate how COVID-19 has impacted on compression of morbidity. Focusing on the last 5 years of life, we will compare morbidity in all patients who died between 2010-2020 to see whether those dying recently have more morbidity before death than those dying 10 years previously, and whether there are variations by ethnicity and deprivation. Analyses will control for age at death, and a matched comparison group will be used to account for any temporal trends in recording.
Health Outcomes to be Measured:
Mortality (All and by Cause of Death); Emergency Hospitalisations (All and by reason for admission)
HES Admitted;ONS;Patient IMD