An enormous amount of television and print coverage has been devoted to the coronavirus pandemic, often accompanied by misleading and alarmist graphics culled from the raw data provided by the UK Coronavirus Dashboard and other sources. Without in any way wishing to question the severity of the epidemic in the UK, I suggest that it is possible to present a more realistic view of the actual situation to date.
Chart 1 below summarises key metrics from the UK coronavirus Dashboard scaled in a single graphic by dividing the weekly totals for each data series by the maximum weekly value recorded for that series, each series being represented on a scale from 0.0 - 1.0 This chart will be updated regularly during the course of the pandemic.
Standardised UK Data reported in the weeks ending 29th February to 28th November 2020
Data Series Chart Symbols Weekly data series description Minimum [ Week ] Maximum [ Week ] PCR Tests Processed [dotted blue line] Number of PCR [polymerase chain reaction] tests processed. 52,095 [ 4 Apr ] 2,322,046 [ 21 Nov ] Confirmed Cases [dotted red line] Number of positive tests reported. 4 [ 22 Feb ] 172,915 [ 14 Nov ] % of Positive Tests [red solid line] Percentage of the tests reported as positive. 0.44% [ 18 Jul ] 32.12% [ 11 Apr ] Patients Hospitalised [yellow solid line] Number of patients admitted to hospital 680 [ 22 Aug ] 21,209 [ 4 Apr ] Deaths [black solid line] Number of deaths within 28 days of a positive test 0 [ 29 Feb ] 7,812 [ 2 May ]
n.b. Patient hospitalisations are generally reported a few days later than the rest of the data and the most recent value is an estimate.
The statistically preferable but rather less spectacular Chart 2 plots the data after Z-score normalisation (subtracting the mean of each data series from each value in the series and dividing the result by the standard deviation of that series). As can be seen most values fall within two standard deviations from the mean [0.000].
As can be seen, the number of people testing positive each week [dotted red line] has surged since early September. However, this does not take into account the number of PCR tests conducted [dotted blue line], which increased roughly 15-fold from around 110,000 per week in early April to more than 1,500,000 per week by mid-September. Consequently, the percentage of positive tests [solid red line] is a much better metric, assuming that those tested are a consistently representative sample of the overall population. While this assumption is not necessarily robust, the standardised data plotted above do follow a consistent pattern during the first wave of coronavirus infections with the weekly percentage of positive tests and the number of hospital admissions peaking in mid-April, followed by a peak in deaths four weeks later. A similar trend appears to be developing in October.
The proportion of deaths now appears to be lower than in the Spring, probably because more effective treatments are available. Similarly, the actual proportion of weekly infections in the UK population in October is probably significantly lower than it was in March and April, the higher numbers reported now being largely a consequence of the increase in the number of tests conducted each week.
The Imperial College REACT immunoassay study of 109,000 randomly selected participants in England conducted between 20th June and 13th July estimated that around 3.3 million people, approximately 6% of the population of England, had been infected with SARS-CoV-2 by the end of June 2020, the overall infection fatality ratio being 0.90%. The proportion of the population who have had the SARS-CoV-2 virus will now be significantly higher, perhaps by a factor of two, three or even more.
Return to the previous page
© Chris Newall 2020
Author : Chris Newall
Page created on : 12th October 2020
Last updated on : 30th November 2020