When is a death an excess death? Getting behind the numbers

The number of excess (ie, more than expected) deaths is a measure that’s probably now familiar to many since the Covid-19 pandemic started in 2020. It's widely used nationally and internationally to measure the mortality impact of Covid-19, including indirect impacts such as deaths resulting from unmet or delayed care. Rising numbers of excess deaths reported for the UK in the latter part of 2022 have received much attention, raising concerns about what's driving them.

Excess deaths is a relatively new measure, which can be produced quickly and was vital when strategies for controlling the pandemic and assessing its impact depended on having timely data. The need for timely data is still pertinent, but how reliable is excess deaths as a measure?

The Office for Statistics Regulation (OSR) notes that even within the UK there are four agencies that calculate excess deaths using different methodologies of varying complexity: Office for National Statistics (ONS), Office for Health Improvement and Disparities (OHID), Institute and Faculty of Actuaries (IFoA), and UK Health Security Agency. For example, choices about the baselines used for calculating excess deaths differ. And then some organisations (eg, ONS) simply take the difference between the current and baseline numbers of deaths, while others (eg, IFoA) adjust for demographic variables such as changes in population size and age structure, or factor in a wider set of variables including also ethnicity and deprivation (OHID). With such significant differences in methodologies between organisations, unsurprisingly, their results differ even though the trends are broadly similar (see, for example, Figure 1).

Figure 1 Cumulative excess deaths, England and Wales, 2022¹

Moreover, different mortality measures even from the same organisation can give differing results. For example, ONS data shows that excess deaths in 2022 were 6.3 per cent above the baseline average, but the age-standardised mortality rate (ASMR) was 0.7 per cent below. This is because the UK’s population is both growing and ageing, which will result in more deaths even if nothing else changes; as ASMRs adjust for these demographic changes, they will show a smaller excess over the baseline than the number of excess deaths.

International agencies also produce statistics on excess deaths: for example, the Organisation for Economic Co-operation and Development (OECD) and EU countries use the per cent change in the number of deaths from a baseline (rather than the absolute difference, like ONS), but they use different baselines, ie, 2015–19 (OECD) and 2016–19 (Eurostat) respectively. Although their results also differ, both have shown excess mortality for most constituent countries in the latter part of 2022. However, their results are also subject to the caveats about excess deaths outlined above.

While frequent and timely recording of excess deaths during the early, deadliest phases of the pandemic proved invaluable, as countries emerge from the ravages of Covid-19 we continue to need rigorous assessments of current mortality and trends – nationally and across countries. That’s because the pandemic has had lasting impacts in most countries, for example, a backlog of unmet care needs, large numbers with long Covid, an increased risk of cardiovascular disease. This means vigilance and epidemiological surveillance remain priorities, not least in the event of another pandemic. For example, is there continuing excess mortality, if so, of what magnitude and what's causing it?

The King’s Fund therefore welcomes the ONS announcement that it has established a technical working group with relevant stakeholders to review methods of estimating excess mortality, not limited to those currently used and to also include those used internationally. This should put mortality monitoring on a more solid foundation and could also inform analytical practice in other countries.

'While frequent and timely recording of excess deaths during the early, deadliest phases of the pandemic proved invaluable, as countries emerge from the ravages of Covid-19 we continue to need rigorous assessments of current mortality and trends.'

Policy-makers need timely but hard, reliable data to assess health problems resulting from the pandemic, and restore and improve the health of their populations to above pre-pandemic levels. While such data was always needed to make informed decisions, as health systems globally struggle to cope with the health fall-outs of the pandemic, it is indispensable. And the public, as taxpayers, need to be able to assess how well governments are doing in this role.

The UK also needs the data produced by international agencies to be able to benchmark itself against peer group countries. The UK’s pre-pandemic life expectancy and mortality during the pandemic (to mid-2022) compare poorly with most high-income countries (other than the US). Continuous monitoring of the UK’s international standing is needed, and also of any early signs of untoward mortality patterns in European infectious disease and mortality monitoring schemes. Having a national health service with universal coverage means that the UK has excellent public health and health data for the entire population, with a high-quality national data collection, processing and dissemination infrastructure. This infrastructure played an invaluable role in supporting pandemic control strategies, and now must continue to improve, adapt and support the recovery process and address upcoming health care needs – including in the event of another pandemic.