COVID-19 Definitions and Anomalies

How does each UK nation define COVID-19 deaths?

What you measure counts. Definitions of a ‘COVID-19 death’ differs within countries, as well as between them.

This article looks at the definitions in the four United Kingdom nations.

Four nations and no united definition

Governments often report the number of deaths associated with COVID-19 each day. How each country defines a ‘COVID-19 death’ differs. Those definitions can change.

The Department of Health and Social Care have paused daily updates of new reported deaths. Public Health England are reviewing their definition.

Every nation requires a positive SARS-CoV-2 lab test result.

There is a clear weekly cycle in reported deaths. (Image: PHE COVID-19 Dashboard — Beta)

Before the review, here are the definitions for each of the four UK nations:

  • England (Public Health England): deaths in all settings. Public Health England uses three data sources for deaths. These are: NHS England, Health Protection teams, and the NHS Demographic Batch Service. PHE link the notifications to the list of people with lab-confirmed cases of SARS-CoV-2. The cut-off time for reporting is 5pm on the previous day.
  • Wales (Public Health Wales): deaths of patients in Welsh hospitals and care homes. Clinicians suspect COVID-19 was a causative factor in the death. The cut-off time is also 5pm on the previous day.
  • Scotland (Public Health Scotland): deaths in all settings registered with National Records Scotland. The death must occur within 28 days of the first positive test result for SARS-CoV-2. The cut-off time is 9am on the previous day.
  • Northern Ireland (Public Health Agency): confirmed deaths reported to PHA in all places. The death must occur within 28 days of the first positive test result. The cut-off time is 9:15am on the previous day.

There are consistent differences in cut-off times for reporting. Both Scotland and Northern Ireland report the DHSC count a day earlier. Suppose it was Tuesday. The Scottish Government will report deaths recorded in the 24 hours before 9am. The DHSC count uses the Scottish figure for Monday. That statistic is the count for the 24 hours before 9am on Monday.

Three nations give counts of deaths where the person had a confirmed case of SARS-CoV-2. This is the virus that causes the disease.

The Public Health Wales measure has a conceptual difference. It is counting deaths caused by the disease, rather than with the virus.

All four measures count deaths recorded within 24 hours, not occurred in the past day. There are reporting lags. This is important for understanding the pandemic. There is a weekly cycle for when agencies report these deaths.

Changing measures

As I highlighted in an earlier article, the definition of a ‘COVID-19 death’ in England has changed.

Before 29th April, the England count was only for deaths in hospitals. The data source was NHS England. After that date, reporting deaths are in all places.

On 1st June, PHE expanded the pool of positive lab results to include commercial tests. Before this date, death counts were from confirmed cases in NHS and Public Health labs.

On 17th July, the Department paused publication of the daily count. This pause is while PHE reviews their definition.

Statistical anomalies

Each definition has its anomalies.

The 28-day window is arbitrary. The time from test to death may be longer than four weeks. This threshold is likely to exclude some deaths from COVID-19.

In this analysis, about a quarter of deaths occur 15 or more days after the onset of symptoms. (Image: Public Health Agency of Sweden)

Clinical suspicions can be wrong too.

Prof Loke (UEA) and Prof Heneghan (CEBM) highlight the issues with the Public Heath England definition. There is no time limit between the positive test result and the subsequent death.

A person could have a confirmed case of SARS-CoV-2 in March. They recover, but die in July from an unrelated cause, such as a traffic incident.

In England, that death would be a ‘COVID-19 death’. It would not be a ‘COVID-19 death’ in the other three nations.

This graph shows deaths by date of reporting. (Image: CEBM)

That definition made sense when the England count was in only hospitals. The extension to all places sets up these conceptual problems. The issue is not the possibility of such anomalous scenarios, but their prevalence.

Non-hospital deaths appear to decline slower than deaths in hospitals. It is unclear how these counts would change under an amended definition. Depending on the change, revisions could be modest.

A standardised definition across the United Kingdom may also be useful.

Other sources and measures

There are other sources for understanding the COVID-19 pandemic in the UK.

NHS England provide daily updates on deaths in English hospitals. Their statistics include collation by date of death.

The three statistical agencies also look at COVID-19 deaths. These counts are of registered death certificates which mention COVID-19. These death certificates take time to register and process.

Doctors may suspect the presence of COVID-19 without a positive lab confirmation. Death certificates also record the believed underlying cause of death. A prior lab confirmation does not mean death certificates will mention COVID-19. As in all human processes, there can be errors.

Excess deaths are a constructed measure. Statisticians define excess deaths as deaths above a baseline. For the ONS, ‘excess deaths’ are death numbers above the prior five-year average. Other institutes may use a different average or modelled baseline.

There are difficulties in interpretation, but this measure has several strengths. We can calculate excess deaths in a consistent way. It is free from the differences in reporting practices. Countries may have different testing volumes, test accuracy, and definitions.

It is also complete. Some people may die from pressure on healthcare services. This is not a direct death from the epidemic disease.

COVID-19 death definitions differ within countries, as well as between countries. Comparisons are challenging, and need care.

This blog looks at the use of statistics in Britain and beyond. It is written by RSS Statistical Ambassador and Chartered Statistician @anthonybmasters.

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store