Written evidence submitted by Simon Briscoe (CLL0097)

I am an independent consultant and a director of a data science company. I have decades of experience in statistics relating to public policy, economics and business. I have worked at the FT (as statistics editor), in investment banking, and as a civil servant in the Treasury and ONS. Full details can be found at simonbriscoe.com. I have had a long-standing interest in data and have published several books on the topic.

Submission:

The care home sector is widely seen as being left behind in the Covid-19 pandemic. The latest ONS data[1] suggests that there have been around 60,000 excess deaths in E&W since Covid-19 descended upon us, of which over 40% were in care homes. Given under 1% of the population lives in care homes, something terrible has happened to the sector – even though the residents were predicted to be among the most vulnerable to such a virus. Had care homes been protected, the country would not be near the top of the death league tables.

It is often said that if you don’t measure it you can’t improve it. Indeed, if you don’t measure it, it is easy to ignore it – and that is what seems to have happened to care homes. It is remarkable how little information we have about care homes. If there is no data, it is easy to make bad decisions and without data the story of the thousands of early deaths in care home will never be fully told. Hopefully lessons will be learned but the apparent lack of progress to get new data during the summer is striking. The lack of data reflects the lack of interest that our politicians, civil service, NHS, local authorities, academics and statisticians have, as a rule, in nurturing and caring for the sector.

The sense of gloom that has accompanied the seemingly endless tales of death in part reflect the fact that the only statistic frequently published about care homes is the number of deaths.

So, what data would have been useful and what questions remain?

1. What is a care home? There is no agreed definition, no exhaustive list. Different bodies use different definitions. This would not happen with other areas of society and National Statistics. Data about the population of the country, hospitals, schools, businesses are all collected, noted, coded and categorised. Lists exist from which samples for surveys can be drawn.

2. Demographics. We should know the numbers of residents and fluctuations over time. We need to have access to all the usual variables such as age, sex and nationality. Data about disability, health and comorbidities would be valuable.

3. Testing. The failure to produce sector data on testing for Covid-19 remains a mystery. There have been attempts by DHSC to get test kits into care homes but no one knows how many tests were taken. Testing, for staff and residents, is widely seen as vital. When the government claims that tests have been carried out when they haven’t – or have no evidence that they have – it does not boost confidence. Regulators and local health teams should have collected data.

4. Cause of death uncertain. It is not clear how many people died due Covid-19. There has been the long-standing “died of” versus “died with” argument. An easy case can be made for the number of deaths in care homes due to Covid-19 being higher or lower than the various estimates published. I wonder what margin of error the National Statistician would put on a point estimate if he wanted to be certain that the real number fell in the range? Due to the regulations for certification of death being relaxed, we will never know the answer.

5. Movements out of hospitals. It remains extraordinary that the NHS has not seen fit to publish a detailed data set of the patients moved from the NHS into care homes to free-up hospital beds. Which hospitals, what type of care home, regionally, Covid-19 tested, etc? Did the NHS not know or not care? Or will it just not let it be known?

6. Devolution. Some of the statistical issues that have limited the creation of good data reflect old chestnuts. Devolution, for one, has had a devastating impact on attempts to create UK-wide data sets. Understandably the devolved bodies not only were keen to guard and publish their own data but often took their own definitions. Sometimes there might be good reasons for this but often there wasn’t. The ONS or another body needs to acknowledge the reality and produce England data. UKSA must work harder to force harmonisation where possible. 

7. PPE. Data on the care sector’s stocks and usage, pre-pandemic and during would be useful to have. What happened to supply chains? Did the NHS take from the care sector and in what quantity?

8. Staffing and absences. We know little about staffing. In other sectors of the economy there might be data on pay, permanent v agency, hours worked, sickness, zero hours contracts, and so forth.

9. Mental health. This applies to both staff and residents but seems not to be measured. Why is there no data on even the simple things like relative visits.

10. Resident health. Basic markers of health would be covered in the demographic data (see item 2 above) but what about interactions with health professionals? How many residents have visited hospital or had visits from a GP or nurse? Have there been visits from health protection teams?

11. Local and regional data. There need to be breakdowns of data where possible by region, county and constituency.

12. Co-ordination. There is no coordinated data collection and publishing regime between the various bodies involved in the sector. This probably needs to be regulated.

13. Financial audit. There have clearly been additional costs for care homes during the pandemic. But how large were they and how secure were the sector’s finances beforehand. What have been the reimbursements from central government? 

Worst of all, the Coronavirus Act suspended oversight just when it should have been increased. Any incentive or pressure for the parties to work together or innovate was gone.

The background

The Office for Statistics Regulation (OSR), the regulatory arm of UKSA, appreciated the shortage of data in a report produced in January 2020[2]. It called for data gaps to be plugged, more co-ordination between the various bodies in the business and an improvement to the data already published. OSR has its own very understated way of saying things but, as the list above suggests, there are huge gaps and no co-ordination between bodies. Sadly the weaknesses revealed by Covid-19 seem to be worse than imagined. It’s hard to imagine that the OSR intervention will change anything without some serious raising of expectations in UKSA and wider government.

There was a promising response from NHS Digital[3] but a search of their web pages revealed no outputs yet of interest.

The OSR produced a short position paper in June 2020[4]. It started:

“The impact of Covid-19 on those in care settings has received significant media attention. There is high demand for trustworthy, quality data and statistics to understand the large number of deaths in the care sector during the period of the pandemic.”

It said that OSR: “….. recognises producers have been making improvements to reporting in this area” and that while “these statistics start to provide a picture of the impacts on those receiving care and help decision-makers to understand and manage Covid-19 within care settings ……. further analyses are needed to provide context and facilitate a better understanding of key areas of concern.”

That said, the Office for Statistics Regulation (OSR) made few demands on the producers saying that they should “consider” attempts to:

• Explain the wider context of COVID-19 and the large number of deaths for those in care settings.
• Understand and assess the impact of any changes in the circumstances and context of data sources, and any implications for use should be clearly explained. 
• Provide or enable regional comparisons where possible, providing guidance and contextual information to support the interpretation of the statistics.

These might be worthwhile aims but even if they happened (and that’s unlikely) there’d not be much change in the data that users would be getting. Few if any of the gaps above would be filled. OSR did not seem to engage with other relevant bodies that could help to produce new data.

The NAO also produced a report in June. Only a small part of it was about data but it did say (para 3.7[5]): 

“….. the Department does not know, for example, how many people overall receive care in each area, including self-funders; and local authorities only have data on those whom they pay for. At the start of the outbreak, therefore, there was no systematic national process to collect a wide range of daily data from care providers. We have commented previously on gaps in adult social care data, particularly on self-funding recipients of care.”

It continued:

“For care homes, a tool commissioned by NHS England and NHS Improvement (NHSE&I) in 2019 to capture and share information between local authorities and the NHS on care capacity, such as vacant beds, was adapted. Following the Covid-19 outbreak, this tracker was expanded to include further data such as workforce absences, Personal Protective Equipment (PPE) levels, and overall risks in care homes. The Department told us that these data from nursing and residential homes registered with the CQC (the care home tracker) had been collated since early April. …… The government has asked care providers to provide data daily through this tracker exclusively. On average, around one-third of care homes, and half of domiciliary care providers had entered information during the preceding 24 hours between mid-April and mid-May.”

This looked promising but begs the question as to why data collection was not promoted – or made compulsory? Why is the data not published? Why has UKSA not argued for it?

The gaps in knowledge are huge and need to be plugged.

Nov 2020