Official Statistics

Changes to OHID’s reporting of excess mortality in England

Updated 19 December 2024

Applies to England

Introduction

Excess deaths are normally defined as the difference between the actual number of deaths registered in a particular period and the estimated number of deaths expected in that period. Excess deaths are presented as negative values when the registered deaths are lower than expected.

Expected deaths can be derived in a variety of ways, which means that there are different estimates of excess deaths published by different organisations. Guidance on these can be found in Measuring excess mortality: a guide to the main reports.

The Office for Health Improvement and Disparities (OHID) has produced estimates of excess mortality for England and for population breakdowns since the start of the COVID-19 pandemic.

What has changed

OHID has released an updated version of its approach to excess mortality reporting: Excess mortality within England: post-pandemic method.

The previous version, Excess mortality in England and English regions, is still available but no longer being updated. 

The principal changes in the new version of the report are:

• estimates of excess deaths are provided by calendar month of death registration rather than by week

• the baseline period for estimating expected deaths will be updated every month, rather than being fixed on the 2015 to 2019 pre-pandemic period

• estimates by cause are based on underlying cause of death rather than on any mention of a disease on the death certificate

• reporting starts from January 2023 rather than the start of the pandemic

• populations have been updated to incorporate estimates and projections based on the 2021 Census

• the adjustment for bank holidays has been removed and there is now an adjustment for the number of working days in a month

All changes are summarised in tables 1, 2 and 3 below.

Although the majority of population breakdowns continue to be reported in the new report, estimates by ethnic group and place of death are no longer included, and the range of causes of death has been reduced.  

Estimates for ethnic groups will not continue as they rely on using personal identifiers to link deaths to hospital records, in order to obtain the recorded ethnicity of the deceased. OHID was permitted to do this during the pandemic under regulation 3 of the Control of Patient Information Regulations. Now that we are no longer monitoring the impact of the pandemic, the use of personal data for this purpose cannot continue.   

The range of causes reported has been reduced because numbers of deaths when based on underlying cause of death, rather than on any mention of a cause, will be too small to provide robust results for some causes.

Further work is needed on the updated methodology for place of death and so estimates for these are not included in the new report.

Updated baseline

Throughout the COVID-19 pandemic, OHID has based the number of deaths it would expect each week on the trend in mortality rates in the 5 years immediately preceding the pandemic (2015 to 2019). That was an appropriate method to use when comparing the number of deaths with the number expected, had the pandemic not occurred: the purpose of the reports was to measure the impact of the pandemic specifically.

However, it is not appropriate to continue to use that method for post-pandemic reporting. Rather than measuring the impact of a specific event (the COVID-19 pandemic) on mortality, the new approach provides a method for ongoing monitoring. OHID is therefore now going to update its baseline each month, using mortality rates in the most recent 5 years. These will be rolling 5-year periods, so estimates of expected deaths for January 2024 are based on mortality rates from February 2018 to January 2023.

This period includes the first waves of the pandemic, when the number of deaths was extremely high in some months. As it would not be appropriate to compare recent numbers of deaths with numbers in periods when they were far higher than average, those months which were most adversely affected by COVID-19 have been excluded from the model used to estimate expected deaths. (Those months are March, April, May, November, December 2020, and January and February 2021.)

Future reporting

In January 2024, OHID updated the Excess mortality in England and English regions report to ensure it provided data for December 2023, so estimates of excess deaths for the whole of 2023 will be available based on both the previous and updated models. Data for 2024 will only be published in the new version of the report. The new report will be updated monthly.

Excess mortality during COVID-19 pandemic

OHID will not update estimates of excess mortality for 2020, 2021 and 2022 reported using its previous model. That data will remain as a record of what was reported during the pandemic, showing the estimates which were used to inform responses to the pandemic and its aftermath.

Impact of the changes

A summary of results from both sets of analysis is published in Excess mortality within England: 2023 data - statistical commentary.

Methodology

See Methodology for excess mortality within England: post-pandemic method.

National estimates for England

OHID’s new report will provide estimates of excess deaths for English regions, upper tier local authorities and deprivation deciles, as well as breakdowns within England by age group, sex and cause of death.

Official estimates of excess deaths for England as a whole will remain those published by the Office for National Statistics (ONS). From 2024, ONS is implementing a new methodology that is closely aligned with OHID’s, as explained in Measuring excess mortality: a guide to the main reports.

Details of changes to new report

Tables 1 to 3, below, set out:

• a summary of changes between the previous version of OHID’s excess mortality report and the new (post-pandemic) version of the report
• changes to the breakdowns
• changes to the reporting of the causes of death

Table 1: Summary of changes

Measure	Previous report (Excess mortality in England and English regions)	New report (Excess mortality within England: post-pandemic method)
Time periods​	Weeks​	Calendar months​
Time coverage​	March 2020 to Dec 2023​	Jan 2023 onwards
Baseline​	Jan 2015 to Dec 2019​	Rolling 5-year periods​, for example baseline for Jan 2024 will be Feb 2018 to Jan 2023​, excluding peak COVID-19 months​
Working days	Model adjusts for bank holidays	Model now has an adjustment for number of working days in a month
Populations	Mortality rates calculated with population estimates based on 2011 Census	Mortality rates calculated with population estimates based on 2021 Census

Table 2: Changes to breakdowns

Previous report (Excess mortality in England and English regions)	New report (Excess mortality within England: post-pandemic method)
Age​	Age​
Sex​	Sex​
Deprivation​	Deprivation
Region​	Region​
Upper tier local authority	Upper tier local authority
Cause of death (mentions of a disease)​	Cause of death (underlying cause)​
Ethnic group	Not included
Place of death	Not included

Table 3: Changes in reporting of causes of death

In the previous report, causes of death were reported by ‘mentions’ whereas in the new version of the report they are reported by ‘underlying cause’. The breakdowns by cause of death for each report are listed below.

Previous report (Excess mortality in England and English regions)	New report (Excess mortality within England: post-pandemic method)
Mentions​	Underlying cause​​
All circulatory diseases	All circulatory diseases
Ischaemic heart disease	Ischaemic heart disease
Cerebrovascular disease	Cerebrovascular disease ​
Other circulatory diseases	Not included
Heart failure	Not included
Cancer	Cancer ​
Acute respiratory infections	Flu and pneumonia
Chronic lower respiratory diseases​	Chronic lower respiratory diseases ​
Other respiratory diseases​	Not included
Dementia and Alzheimer’s	Dementia and Alzheimer’s ​
Diseases of the urinary system	Not included
Cirrhosis and other liver diseases ​	Cirrhosis and other liver diseases
Diabetes ​	Not included
Parkinson’s disease ​	Not included ​

Feedback

Feedback from users on the new version of the excess mortality report would be welcomed. Responses and questions can be emailed to [email protected].