Dear Mr Wragg,
As Director General for Regulation at the UK Statistics Authority and Head of the Office for Statistics Regulation (OSR), I write in response to your letter of 14 April to the National Statistician, covering the regulatory perspective on COVID-19 statistics and data. This letter summarises OSR’s review of data and statistics on COVID-19 cases and mortality, including trustworthiness, quality and value. It also considers broader regulatory points about the use of management information, and statistics on adult social care.
This is a very difficult time for everyone as the UK adjusts to rapid changes in society and the economy. OSR commends the flexibility and level of responsiveness shown by organisations that produce official statistics in adapting to this new environment and will continue to support further improvements to statistics and data on COVID-19.
Reviewing statistics and data
OSR has undertaken a review of all the data releases on COVID-19 cases and deaths – at a UK level and for each country within the UK – to help understanding of the available sources and to highlight strengths and our view on areas for improvement. The relative strengths and limitations were considered within the context of the three pillars of the Code of Practice for Statistics that you
mentioned, referred to as TQV:
• Trustworthiness: governance, including people, systems and processes
• Quality: robust data, method and statistics
• Value: statistics that answer people’s key questions
The document we published yesterday outlines the findings from our review. It acknowledges that there is value in having timely data, such as the daily surveillance data covering the UK that is published by DHSC less than 24 hours after the data reporting period. This output provides an important leading indicator of the trend in COVID-19 testing, cases and deaths. However, with this timeliness there is a trade-off with completeness, for example, publishing the setting where the death occurs. Because the data from England only captures deaths in hospitals and not deaths in the wider community, these UK daily outputs struggle to meet the needs of all users and require continuous innovation to include information about where the death took place. We understand that the ONS is working with the Care Quality Commission to publish further data on deaths in care home residences. Secondly, although we have seen notable improvements in the metadata that accompany the daily data for each nation, the nature and extent of the uncertainty around the UK estimates of deaths associated with COVID-19 could be clearer. Finally, we are concerned about the accessibility of the data from all four nations and have asked the Government Statistical Service to consider enabling users to navigate all COVID-19 related outputs from a central hub. None of the
daily data releases are designated as National Statistics.
In contrast, the weekly mortality statistics published by the ONS for England and Wales, and by the National Records of Scotland and the Northern Ireland Statistics and Research Agency, provide a more complete measure of the number of people whose deaths are associated with COVID-19, but these statistics are released with a greater time lag and are not designed to measure the spread of a pandemic in close to real-time. Weekly death registration reports focusing on COVID-19 use a variety of data recorded on the death certificate, such as setting of death, sex, age band and underlying diseases that may have contributed to a death. Overall, the weekly mortality statistics largely fulfil the Code’s expectations on trustworthiness, quality and value. Unlike the daily data, the weekly mortality statistics published for England and Wales, Scotland and Northern Ireland are all designated as National Statistics.
Broader regulatory points
There are two further regulatory points I wanted to draw to your attention. First, there is a wide range of information being used by Government to inform its understanding of the impact of COVID-19 on the economy and society. Where that management information is used as part of daily public briefings, it should be published and accessible to the wider public. Second, you note that concerns have been expressed about the limited information on deaths in care homes compared to deaths in hospitals. These concerns echo the findings of OSR’s review of statistics on social care for England, published in January 2020: we concluded that, when comparing social care to the data-rich health system, there are inadequate statistics. We called for greater parity of measurement between the two.
Our document published yesterday includes more detail on our work on COVID-19 statistics and data, and I hope the Committee finds this letter to be helpful.