Obsession with infection rates is hindering efforts to understand Covid data
Two thirds of the government’s original slogan for lockdown – “protect the NHS” and “save lives” – focused on the two key indicators of the impact of the virus on society: hospital admissions and death rates.
Now, as the government seems increasingly entranced by the infection rate metrics – despite even those panning out better than feared – they would do well to take a look at the those health figures once again.
We can start with the relationship between infections and hospital admissions. Across England for the four weeks until 5 October, the number of confirmed infections has been increasing at a rate of between 53% and 64%, week-on-week. Hospital admissions in the same period have increased by a slightly lower rate, by between 40% and 53% week-on-week.*
There are huge problems, of course, with comparing infection rate data with hospital admissions. “The more you test, the more cases you find – and of course testing has been prioritised towards areas with higher infection rates”, Professor Karol Sikora, medical director of Rutherford Health, told The Telegraph this morning. A widening gap between the two does not necessarily mean that fewer infected people are going to hospital now than before.
Other data suggest, however, that increasing testing may not be the whole story. We can turn to the NHS daily situation reports, published separately to the government and ONS data, for more context. These provide a breakdown of admissions for Covid-19 by hospital, region and age.
According to this data, those entering hospital for the virus are on average younger than before. A full 55% of those admitted to wards in England until the week ending 28th September were for those under the age of 65. This compares with just under 33% twenty weeks prior, in the seven days to 11 May.**
For the North West region and the North East and Yorkshire region, now the two regions most heavily affected by the virus, the trend holds. In the first, 50% of admissions in the latest recorded week were for those under 65, versus 30% twenty weeks earlier. In the second, the same timescale shows an increase to 54% from 34%.
Any conclusions drawn from this can only be tentative and would require more thorough investigation. For example, the data does not break down the age profile of those patients admitted between the ages of 18 and 64.
But does this data suggest that, in addition to rising testing, there has been an actual reduction in infections leading to hospital admissions?
“Without seeing precise information about individual patients… I have wondered for some time that the criteria for hospital admission might have changed in that they are now less demanding, with more admissions for patients who are not as ill as in March [and] April”, suggests Professor Hugh Pennington, Emeritus Professor of Bacteriology at the University of Aberdeen.
“A lot more testing has revealed a large number of cases (better called positives, because many only have trivial symptoms), the great majority of whom do not need hospitalisation.”
Indeed, the trend coincides not only with the reopening of universities and the rising number of cases attributed to students and young people, but with tangible improvements in medical treatment since at least June. Experimental treatments deployed by the UK’s Recovery programme, such as REGN-COV2, should allow further patients to be taken off ventilator beds.
Meanwhile, in the most infected areas now, the younger age profile of admissions may prove important for understanding why fewer of those positive cases entering hospital are dying from the virus. In the North West region hospitalisations increased by 58% between the week beginning 29th September and the previous week; deaths, however, increased by only 25%. In the North East and Yorkshire, hospitalisations increased by 51% while deaths in the North East and Yorkshire and Humber regions increased by only 24% in the same period.
Clearly, though, more questions must be asked of the data to establish underlying trends – what is the average time that patients spend with providers before being discharged? How many recorded patients come in for consultation, rather than to take up a hospital bed?
Some of the answers are buried in the available data. They provide crucial context to the increasingly frenzied debate over infection rates.
*Data obtained from the government’s coronavirus dashboard. For infections: 7-day average of daily new cases was obtained, by specimen date, for each of the four weeks beginning 08-09-2020. For hospitalisations: 7-day average of daily number of new patients admitted to hospital was obtained for same period. The writer takes full responsibility for calculations made.
**Data for English hospital providers as reported on 01-10-20 in the Covid-19 daily situation report, available here (“October 2020 COVID Publication”). 7-day average for new admissions obtained for period 22-09-20 through 28-09-20, compared with same average for period 05-05-20 through 11-05-20. The writer takes full responsibility for calculations made.