Dear Mr Wragg,
Thank you for your letter of 14 April regarding both Department for Health and Social Care (DHSC) data on cases of, and deaths from, COVID-19, and the Office for National Statistics (ONS) mortality data. I wanted to note, too, that as the UK’s National Statistician and Head of the Government Statistical Service, I have a responsibility to ensure the presentation of all data and analysis across Government is useful, transparent about its caveats, and of a high quality.
When making international comparisons it is important to compare figures that have been collected in the same way, using the same definitions. At this stage of the pandemic, international comparisons are difficult due to the differences in data collection. Most countries will be collecting data from hospital settings, some will require a positive test for COVID-19, others will include a diagnosis, and some will include any deaths with any suspicion of COVID-19. While they are not as complete as the ONS figures, the DHSC data are the nearest to international comparators.
The ONS weekly deaths publication is the most comprehensive source of deaths that have occurred in the pandemic, as it is complete, includes all the categories mentioned previously and all settings (hospital, care homes and in the community). However, although the UK produces these figures more quickly than most, these data are not as timely as the DHSC daily published numbers, which
can be used as a broad comparator for deaths that have occurred since the ONS publication. Because the ONS figures rely on the registration of deaths, more timely data would mean further burden on medical practitioners to register deaths on the day of death.
You also asked about policy decisions. Evidence is vital for good policy making, and the evaluation of that policy. When providing that evidence base, we must ensure that the accuracy and timeliness is communicated well. The ONS data provides a comprehensive and accurate measure of deaths according to the registration system. The DHSC daily deaths publication is less comprehensive, but more timely, and as such it is an important lead indicator for the spread and severity of the pandemic. The ONS is currently working with Public Health England, the Care Quality Commission, the Devolved Administrations and leading academics (through the Scientific Pandemic Influenza Group on Modelling, SPI-M) to reconcile the timely data from DHSC with the accuracy and coverage of the ONS data to provide an even better evidence base for policy making. This work will deliver over the next few weeks to improve the quality of the forecasts given to the Scientific Advisory Group for Emergencies (SAGE) to help manage the crisis.
It is a strength of the UK statistical system that we can provide equality of access to data on which important decisions will rely, while highlighting that they serve different purposes, which we did in a statement on our website published 31 March. The ONS publication is more comprehensive, but the daily DHSC numbers are the most timely available. They also serve the public good by providing insight into the pandemic in almost real time, even though we must note the limitations of this data.
Ed Humpherson, Director General for Regulation and Head of the Office for Statistics Regulation (OSR) will write separately on the approach of OSR to published information on COVID-19.
Please do let me know if I can be of any further assistance to yourself and the Committee.
Yours sincerely,
Professor Sir Ian Diamond
Related Links:
Ed Humpherson to William Wragg MP, Chair PACAC (April 2020)
ONS and OSR Oral evidence to PACAC (May 2020)