Join Hafta-Ichi to Research the article “How does the PHE statistics error change the UK Covid picture? | Society”
The “technical issue” that led to Public Health England undercounting last week’s coronavirus cases by almost 16,000 is an embarrassing error for the already beleaguered body, and could have significant implications.
What do we know so far?
On Sunday PHE announced that 15,841 confirmed cases of coronavirus had not been included in its daily published figures. Three-quarters of these related to the period from 30 September to 2 October.
The problem was attributed to a technical issue in the process for adding data on positive Covid-19 test results to the reporting dashboard. It appears there was an issue with the upper limits of file sizes, with the Daily Mail reporting that the data was being loaded into Excel.
PHE said the issue did not affect anyone receiving test results, and all those who tested positive were told to self-isolate.
How much does this change the overall picture?
According to a daily breakdown of the missing cases, the government dashboard under-reported the true number of cases by 41% on 2 October, the day in which there was the biggest discrepancy. The previous day, cases were under-reported by 37%.
The government has now added the 15,841 cases with a reporting date of 4 October, which brings the seven-day average for new daily cases to an inflated 9,716 – the figure would stand at 9,367 had the figures been recorded correctly in the first place.
Either way, the case rate is higher than any previous seven-day average figure, although it should be noted that testing was not widely available during the first wave of Covid-19 and official figures from that period were probably a significant undercount.
Worryingly for the authorities, the new data also indicates that the number of cases recorded day by day is growing at a faster rate than they had anticipated.
The old figures showed that after a fall in cases over the previous weekend there was a 75% increase in cases between last Monday and Tuesday . But if the new figures hold true then there was actually a 111% rise.
This one-day-doubling in the number of cases appears to have been an anomaly. But the latest data does show that the number of cases reported on 3 October was double the figure published six days previously, and week-by-week comparisons show that the rate of doubling is accelerating.
The error means that for three days the PHE reported that case numbers looked relatively static when they were in fact increasing.
The most immediate knock-on effect is on contract tracing. While those with positive results were asked to self-isolate, their contacts were not traced, meaning many thousands of potentially infected people were not reached by test-and-trace teams.
Is this the first major testing data glitch?
No. In April the UK Statistics Authority criticised the government over its misleading use of testing figures, after it conflated the number of home testing kits sent out with the number of tests processed in order to hit its targets.
In July it was reported that the daily cases dashboard showed only Pillar 1 cases, which are positive tests completed in PHE labs and the NHS. Pillar 2 cases, identified by commercial labs, were not made public. This meant that in many local areas the true extent of outbreaks was not apparent, and local government leaders complained that a lack of accurate data was hampering their ability to combat the spread of Covid-19.
In August the government removed 1.3 million tests from the testing dataset, citing double counting issues with at-home testing kits.
Source: The Guardian
Keyword: How does the PHE statistics error change the UK Covid picture? | Society