Understanding changes to mortality during the pandemic Non-COVID-19 excess deaths in England and Wales

During the coronavirus (COVID-19) pandemic, over 51,000 COVID-19 deaths have been recorded in England and Wales since March 2020. These are deaths recorded by the ONS as having COVID-19 on the death certificate. Worldwide, this number is almost 750,000 and continues to grow at worrying rates in many countries. However, counting only ‘COVID-19 deaths’ does not reveal the full picture of what has happened to people, health and care during the pandemic.

The number of excess deaths is a better measure of the pandemic’s total mortality than deaths attributed to COVID-19. It measures the additional deaths in a given time period compared to the number usually expected, and does not depend on how COVID-19 deaths are recorded. So far this year, England has had the highest excess mortality in Europe.

In England and Wales there have been 57,725 excess deaths from 7 March – the first week that a death with COVID-19 was registered – to 24 July. This is around 6,700 more than the 51,017 deaths where COVID-19 was mentioned on the death certificate.

With the possibility of future waves of COVID-19, it is critical we understand what has happened to all mortality, not just those attributed to COVID-19. Here, we focus on England and Wales because the same system for recording deaths is used by ONS, with some detailed data on non-COVID-19 deaths. We calculate excess deaths as the difference between deaths in the current year compared to average deaths over the past 5 years for the same period. Non-COVID-19 excess deaths are calculated in the same way, but COVID-19 deaths are excluded from the calculated difference.

What could have caused the excess deaths not recorded as ‘COVID-19’?

Excess deaths not attributed to COVID-19 could be:

• Deaths from other causes as a result of people not accessing care for non-COVID-19 related causes. For example, A&E attendances and emergency admissions have fallen significantly since February, with particularly large falls in March and April.
• Deaths from other causes as a result of changes in risk factors and/or the prevalence of certain illness or injuries. For example, the changing social conditions resulting from measures to combat COVID-19, such as social distancing and increased self-isolation, could cause stress leading to an increase in deaths from stress-related causes. Alongside these increases, there may be decreases in deaths from other causes such as road deaths and injuries.
• Deaths that are actually COVID-19 related but not recorded as such.

The relative contribution of these factors has implications for health and care policy. For example, if a large proportion of non-COVID-19 excess deaths were from people

not accessing care when needed, in any possible future outbreaks greater consideration should be given to ensuring people who need to use emergency care services are able to – and understanding why they may not have. If a large proportion of non-COVID-19 excess deaths are due to increased prevalence of health conditions, such as increases in stress, in any future pandemic there should be an increased focus on those at risk. And if a large proportion of these deaths are truly COVID-19 deaths, the death toll from the virus is bigger than officially reported. This leaves questions about why people did not access care, along with issues regarding testing and reporting on COVID-19 deaths.

Non-COVID-19 excess deaths in England and Wales

In this section, we examine a few possible explanations for non-COVID-19 excess deaths in England and Wales. We explore changes in mortality over time, by place of occurrence and by age. Non-COVID-19 excess deaths refer to all deaths excluding COVID-19, compared to the average deaths in the same period over the past 5 years.

If non-COVID-19 deaths are unreported COVID-19 deaths, then we would expect to see the trends over time following those of COVID-19 deaths. However, if these are genuine non-COVID-19 deaths – such as deaths caused by missed care –  we may expect different trends to COVID-19 deaths. We would also expect to see changes in where deaths occur as people avoid hospitals. If there are non-COVID deaths caused by changes in risk factors or prevalence of other conditions, we might expect to see changes in cause of death, particularly for specific conditions.

Trends over time

The chart below shows that COVID-19 and non-COVID-19 excess deaths follow a similar trend over the pandemic period. Early in the pandemic (up to week 18) COVID-19 deaths represented 72% of the total excess deaths. These trends and patterns are consistent with a significant proportion of non-COVID-19 deaths being COVID-19 deaths. Early on in the pandemic there was insufficient testing (outside of hospitals) so people may have died from COVID-19 without having been tested and therefore having COVID-19 would not necessarily be mentioned on their death certificate.

Figure 1

However, it is also likely that some excess deaths are not caused by COVID-19. A&E attendances fell sharply during the pandemic, which could lead to non-COVID-19 deaths caused by missed care. For example, attendances for acute coronary syndromes reduced significantly, leading to concerns over out of hospital deaths. While they remain far below normal levels, A&E attendances have since risen. At the same time non-COVID-19 excess deaths have fallen below average, which could be partially driven by newly increased use of health care. To explore further we examine deaths by place of occurrence.

Deaths by place of occurrence

The chart below shows deaths by place of occurrence, with all deaths and COVID-19 deaths compared to average deaths over the past 5 years. Deaths in care homes and hospitals show similar and substantial increases during the peak of the pandemic. In both private homes and care homes, the increases in all deaths are greater than COVID-19 deaths, leaving significant numbers of non-COVID-19 excess deaths.

Figure 2

While total deaths in care homes and hospitals are now below the average of the previous years, this is not the case for private homes. Deaths in hospitals, even including COVID-19, have been below average since mid-May. This is consistent with a change in deaths from people not accessing or receiving care. However, from the data, we cannot identify whether these deaths would still have occurred if that care had been received.

In previous years, deaths in private homes accounted for about 45% of weekly deaths for this period. However, as COVID-19 deaths rose, they become a much larger proportion of deaths – 59% of weekly deaths during the worst weeks of the pandemic. NHS capacity was not breached at a national level. But some of the additional capacity was created by people (some of whom may have subsequently died) not attending hospital or being discharged early.

The next chart shows the same data, but excluding COVID-19 deaths. We calculate non-COVID-19 excess deaths as deaths over the period, excluding those attributed to COVID-19, compared to previous years.

Figure 3

Deaths in care homes rise and fall with similar trends to the COVID-19 deaths shown in Figure 2, which would be consistent with unreported COVID-19 deaths. From ONS data, deaths from dementia and Alzheimer’s rose substantially in April and May. Additionally, Public Health England found that for deaths listing dementia and Alzheimer’s as the primary cause of death, there were very few with COVID-19 mentioned on the death certificate. These are highly prevalent conditions among care home residents, which can additionally make it harder to identify the symptoms of COVID-19. However, these are also conditions that can cause people to deteriorate quickly without care. We also know in March and April there were significant reductions in hospital attendances among care home residents.

A recent study on care homes in Scotland shows that excess deaths in care homes occurred almost entirely in places that had a COVID-19 outbreak. While it is thought that these are undercounted COVID-19 deaths, it is also likely that a COVID-19 outbreak makes care more difficult to deliver. All these trends in care homes are consistent with many non-COVID-19 excess deaths in care homes being unreported deaths from COVID-19.

Excluding reported COVID-19 deaths, deaths in hospitals have fallen sharply and remained below average since March. At the same time, deaths in private homes have risen by similar amounts. This trend has continued for many weeks even as testing has significantly expanded. It seems probable that these trends may be driven by changes in behaviour accessing care, along with changes in access to health and care delivery. For example, polling the Health Foundation funded from Ipsos Mori showed 47% of respondents were uncomfortable with attending their local hospital if needing treatment, with catching the virus the most significant concern.

Recent research also shows significant falls in hospital admissions for acute coronary syndrome. The authors conclude this is likely to have resulted in increases to out-of-hospital deaths, which would explain some of the trend seen in Figures 2 and 3. This is also consistent with ONS data showing overall increases in deaths from ischaemic heart disease. While these deaths decreased in hospitals, in total the increase was driven by large increases in private and care homes.

Increases in deaths in private homes from diabetes and asthma provide further evidence of problems accessing health care. In late April, these were more than double their normal level. However, as the ONS note, they are also risk factors for COVID-19. Deaths from diabetes in private homes increased in April by more than the decrease in these deaths in hospitals, which could suggest an increase from not receiving hospital care. As a result, all deaths from diabetes have increased.

Recent Health Foundation research also examined deaths among those receiving domiciliary care, noting these have risen substantially since March and far above normal levels. This indicates that some of the excess deaths in private homes are among those receiving social care in their own homes. This indicates a trend of deaths in private homes consistent with an increase in deaths outside hospitals, which could be from not accessing care.

Another explanation explored by the ONS are stress-related causes. There were substantial increases in deaths due to hypertension. However, it is difficult to assign a cause of this cause, as there could be many contributing factors. Deaths due to suicide or drug use take far longer to register so may not be shown yet in the data.

Deaths by age

Deaths by age may also reveal more about whether these excess deaths are COVID-19 or non-COVID-19 deaths. If these are COVID-19 deaths, we would expect a similar age profile to COVID-19 deaths. While this may also be true for non-COVID-19 deaths, if the deaths were caused by a lack of access to treatment, these may not follow the same age profile.

Table 1: Excess and COVID-19 deaths by age group, 7 March to 24 July 2020.

Source: ONS weekly deaths registrations.

*Small number suppressed. Excess deaths number differs slightly to death numbers in the text as not all deaths in previous years had ages.

Deaths rise with age, which is consistent both with all-cause mortality and those with COVID-19. COVID-19 deaths as a proportion of overall excess deaths differ between age groups. It is very high in the 15–44 age group, indicating the low risk of death usually faced by this age group. In the age groups older than 45, it is relatively similar in all age groups, reflecting that deaths with COVID-19 increased somewhat proportionally to the average deaths for each group. However, the proportion of excess deaths made up of deaths with COVID-19 was not highest in the 85+ group. This is consistent with our commentary above, particularly as the care home deaths are likely to be in this population, with possibly underreported care homes deaths with COVID-19.

Discussion

We have investigated changes in all mortality during the year, focussing on deaths not attributed to COVID-19. The data highlight several factors that may have contributed to this: unreported COVID-19 deaths, deaths from a lack of access to care, and changes in disease patterns.

Evidence suggests a significant number of unreported deaths with COVID-19, particularly in care homes. However, it is difficult to know the proportion this represents . In particular, those with dementia and Alzheimer’s have had significantly higher mortality rates, indicating a population particularly vulnerable to COVID-19 under-diagnosis, but also conditions which may cause people to deteriorate quickly from reduced or lack of access to care.

There is also a concerning trend regarding deaths in private homes, which remain far above historical averages. The ONS has noted some of these may be among those requiring end-of-life care, but much more evidence is required. We know that deaths for those receiving domiciliary care have risen substantially since March, which are likely to form a significant component of the excess deaths in private homes. We also know that A&E attendances were 57% lower in April 2020 than April 2019, and are still far below normal levels in England.

Some of those dying at home may have survived with hospital care. And if people are not accessing care when they need it, it is also likely there are many living with conditions that would benefit from treatment. In recent research, we previously noted a need for a well-targeted communications campaign alongside monitoring, to ensure those who would normally access and need care continue to do so. In addition, many of the impacts of delayed and unmet need for care will be longer term. Urgent referrals for cancer were down 60% in April, for example, and the number of first treatments fell by 21%. It may be some time before we know the full impacts of these.

Taken together, it is concerning that there have been increases in deaths among populations (in private homes and care homes) where data on the care received by individuals is more limited than in hospitals. The health and mortality impact of COVID-19 can only partially be quantified from looking at COVID-19 deaths.

The UK has had nearly the highest – and England the highest – age standardised excess mortality rates in Europe. The impact not been equal across all communities, with a far greater toll on black and minority ethnic communities. The Academy of Medical Sciences recently outlined the planning required for COVID-19 for winter 2020/21, highlighting the potential for the resurgence of COVID-19 and disruption to health and social care. With risks of further outbreaks and waves, it is crucial to build an understanding of the impacts over time of COVID-19 on mortality and morbidity – both from the virus itself, and from changes to health and care services, and human behaviour.