Last week – following extensive inquiries from ourselves and others – it was revealed 400,000 people who received negative PCR results, mainly in the South West, may have received inaccurate results – around 43,000 potentially received false negative results between 8th Sept and 12 Oct.
There are wide-ranging impacts. We’ve been trying to advise people to trust their lateral flow/symptoms since 29th September (and had started to do so in individual cases since 16th September), but unfortunately many people will have been told to stop isolating and go into school/work/otherwise mixed with people while infectious.
Tonight, Inside Science on Radio 4 explore “The Possible Impact of false-negative PCR Tests“. They included estimates that if 43,000 people with false negative PCR results went out and mixed at current levels of spread, that might have caused 200,000 additional infections so far and therefore potentially 500-1,000 additional deaths. And those 200,000 people will – to some extent – continue to infect others and so on, so the consequences of the mistakes will continue after the lab problems are fixed.
I’m not capable of that level of analysis – but below I’ve done my best to present charts that present current data, but also contextualise it. Apologies that as a result this week’s update is a bit longer than normal – but hopefully it’s helpful to people to explore the current issues in depth. Two key points first:
- Numbers of infections are very high at the moment – locally particularly but also nationally, no two ways about it. If you want to avoid being infected, or reduce the chances of passing the virus on to someone at risk, please take care – wear masks in crowded/indoor settings, if you want to protect yourself consider upgrading to a more protective mask, meet people outside if you can, and if meeting people indoors keep a flow of fresh air. Regular testing with lateral flow devices will help catch infections early so people can isolate and avoid passing the virus on – and you might particularly want to do them or ask others to do them before mixing indoors with large numbers of people. For those most at risk in our community, you may want to consider reducing contacts to the extent this is possible until rates are lower.
- The reported data on case rates, and particularly reports about increases, are basically broken at the moment. Health Security Agency epidemiologist Meaghan Kall has a good twitter thread outlining six reasons why, but two key ones are that apparent ‘increases’ are from points that were artificially low due to people who were infected not showing in the data because they had false negative PCRs. It will be a another week or so before the data settles down and we see only true new cases, rather than some people who have recently re-tested being added to the totals for the last seven days (when they should have been included in the week before that). Cases are likely rising at the moment – but not at the rates some reporting is implying. Hopefully the charts below help explain this.
Gloucestershire – positive tests, hospital admissions, people in hospital
First, a chart to try and explain the problems with the reported data at the moment. Some people who got false negatives and didn’t show up on the yellow dashed line, which should have been higher (box marked FN), have now been re-tested and are showing up on the red line, pushing it higher than it would otherwise be (RT area).
The green dot trend line isn’t based on clever modelling of where we were headed (just a linear trend based on reported data from 1st June – 7th Sept. It’s there just to illustrate that the yellow dashed line is lower than it should have been because of false negatives. The true number of new cases should have been closer to the blue double line – maybe roughly on the green dot trajectory.
If the data wasn’t borked by a bunch of false negatives, followed by the effect of people showing up as testing positive later than they would have done because of re-testing, the number of people testing positive in Glos’ each day might have been somewhere around that green line. It could have been lower if people had been doing things to reduce spread, or maybe higher if not – we don’t know, and we won’t get an idea for a little while.
One way we can get a better idea of what’s been going on recently, less affected by the false negative testing issue, is to look at the number of people admitted to local hospitals with Covid-19. The chart below shows how the 7 day total has been bouncing up and down but broadly rising over the past few months, hitting a recent six-month high of 86 patients in the 7 days to the 15th of October, double the 43 in the week to the 14th September (doubling in a month).
Again, I’ve added an unsophisticated linear trend-line to give some indication of where things might be headed – but the trend could easily change, it is affected by many factors. These numbers are still much relatively compared to all admissions or to previous periods locally BUT the NHS locally and nationally is under serious pressure and any admissions that can be avoided will help reduce that pressure
We know that hospital admissions are rising nationally – so, is there any indication in the data that the Immensa false negative PCR testing fiasco has caused more spread and more hospital admissions? Well… maybe – see the next chart. This shows that for the last 7 days, admissions for Covid-19 to Gloucestershire hospitals are at 41% of their peak – in January. They are more erratic than for larger areas (because of the smoothing effects that average such ups and downs over larger numbers of people), but this is much higher than for the South West overall – where weekly admissions are 28% of their peak, and indeed twice as high as for England as a whole – where, though rising, weekly admissions are 20% of their peak. This is achieved through indexing, and peaks in the South West and England were of different scale to Gloucestershire – but you can see how, other than in Novemebr-January last year, the rates have largely tracked each other in indexed terms. However, please see the next chart after the one below too, which covers the number of people currently staying in hospital
People that are admitted to hospital may be discharged fairly quickly. However, some Covid-19 patients will need long hospital stays for treatment before recovery, particularly if they need Critical Care. The chart below shows how the number of beds occupied by Covid-19 patients has changed compared to the peak. For Cloucestershire the current number of patients is 18% of the peak in January, the same proportion as across England as a whole, and slightly lower than the 23% for the South West region overall. It’s worth re-stating that this is despite much higher case numbers than previously – fewer people end up in hospital per case because of the protection vaccination provides people, and because cases are more often in unvaccinated people, who tend to be younger – and less likely to be hospitalised than older unvaccinated people (who were the people being hospitalised in the previous wave, before we had access to Covid-19 vaccines).
I usually show a comparison of the key data for Gloucestershire. This week I’ve taken the tests out of this chart because of the issues with the testing data. The chart shows
- Admissions to hospital (purple line) show a slow rise – as in the chart above
- Numbers in hospital are also clearly rising. There were a total of 41 beds occupied by COVID-19 patients in Gloucester/Cheltenham hospital as of 19th October – up from 26 on the 5th October. There were last 41 people with Covid in these hospitals on the 2nd March.
- Eight more people who lived in Gloucestershire have died with COVID-19 mentioned on their death certificate in the most recent week of data (to 8th October). We send our condolences to their loved ones. The total number of people who lived in Gloucestershire to die with COVID-19 mentioned on their death certificate is now 1,233. Because we still hear people suggesting these deaths aren’t really caused by COVID-19, I continue to encourage people to read about how Covid-19 deaths are measured from the Office for Statistics Regulation. Again, the number of people dying suggests a rising trend, but as with hospital admissions is much lower compared to cases than in the time before people could be vaccinated.
People who have tested positive – districts, areas within districts, and age profile
Data on rates of people testing positive per 100,000 is still affected by the issues with the Immensa lab – as indicated on the chart. You can see how rates for Gloucestershire and the districts were in it went from being above/in-line with the England rate to diverging from it as people received false negative PCR results. The England rate will have been affected by this, but to a much lesser extent, and likely provides a guide to roughly where our local numbers ‘should’ be – though I expect our numbers will rise substantially above the England rate due to the impact of additional mixing where infectious people have not isolated.
The trend for England, shows a fairly rapid rate of increase – generally since mid-September – to 417 per 100,000 (or around one in every 240 people). However, this only covers people who have been tested. Estimates by ONS of true prevalence based on random-sampling of the population (rather than people seeking tests), find that “in the week ending 9 October 2021… 890,000 people in England had COVID-19… equating to around 1 in 60 people.”
Colin Angus has a great chart below showing how Stroud and nearby local authorities also affected by the Immensa PCR testing failures are outliers across the UK at the moment both in terms of ‘new’ cases in the past week, and the change in case rate compared to the preceding week (again, comparisons to the preceding week don’t really work because the previous week’s data was wrong).
You can view daily numbers for Gloucestershire, or districts within it, on the government’s dashboard, and there is more information below. You can also enter your postcode into the government’s dashboard to get more data on your local area. However, for the next week or so please take into account that the reported data in terms rates per population and of increase are a mess because of the Immensa false negative scandal.
One dataset that’s hard to view on the dashboard – although you can see a heatmap, is cases by age. The below chart is from Professor Colin Davis. While this suffers in part from the same testing related issues – with the numbers in September and early October artificially low, the chart reveals that on October 13th nearly 3 in every 100 children aged 10-14 in Stroud district had had a positive test in the past week. The government dashboard has more up to date data, showing that 4.5% of 10-14s in Stroud district tested positive in the week to 16th October (and that still includes days to the 12th October when false negatives were being reported). Rates are also high for 15-19 year olds (2.2%), 5-9 year olds (1.3%), and the age groups we might associate with parents/teachers: around 1.2% for ages 40-44, 45-49, and 50-54. While the actual rates of current infections will be lower, these proportions of the people in the age groups have still tested positive in a remarkably short space of time.
We can maybe get a better idea of true prevalence in the South West (and hence a guide to Stroud district/Gloucestershire) through the ONS infection survey. Rather than relying on people seeking tests (and the Immensa lab reporting false negatives) this is based on sampling the population and making estimates – for the South West there has been a clear recent rise to a point where 2% of people were expected to have the virus in the week to 9th October. This diverges from many but not all other regions so may in part show the impact of many people not being told to isolate even though they were likely infectious. The North East, East Midland, South East and London all have proportions lower than 2% and falling or flat trends. However, the East of England, West Midlands, North West and Yorkshire and Humber have similar increasing trends and rates (1.6%-2%, with confidence intervals overlapping 2%).
It’s quite interesting to look at the positivity rate for Gloucestershire – this is the proportion of people who submit PCR tests that return a positive result. You can clearly see when the problem with the Immensa lab started affecting results with the dramatic drop in positivity. The positivity rate has now returned to roughly where it was before September – a little higher in part because people who received potentially false negatives are getting ‘re-tested’ may be more likely an someone otherwise seeking a test to test positive.
In the week to 14th October (when re-testing was yet to begin officially – though people may have been seeking second/third tests themselves), 10.8% of the 27,486 people to get tested were positive. This compares to 7.7% of the 27,548 people to get tested in the week to 7th September, before the problem arose. Interestingly, the number of people getting tested was similar, which doesn’t suggest a huge rise in people with symptoms (though the number of people being tested as contacts would have been reduced as those with false negatives didn’t encourage contacts to get tested).
Vaccination
Data from the NHS on Covid-19 vaccinations is now collated in a monthly vaccination update (due to the Immensa issues we haven’t had time to update this for a while, but the numbers haven’t shifted much. We’ll hope to update it next week).
National context
New government guidance has been out since 19th July. While many restrictions have been lifted, please see the NHS webpage on “How to avoid catching and spreading coronavirus“. Please also read our SCCR statement on the current situation where cases are very high, which underlines that “we can still choose to act conscientiously and with compassion for others“. There are people in our community who have do not have the protection granted by vaccination (including those who cannot have jabs, or do not benefit as much from them because of a health condition/treatment, as well as those who for whatever reason have not taken up the offer of vaccination yet).
You can see a summary of the trends on the government dashboard at this link, and in the image below (Described in text below the image)
Across the UK, the number of people to have tested positive in the past 7 days (to 21st October) is 327,537. This is 17.9% higher than the number that tested positive in the previous week – 49,662 more people. Deaths within 28 days of a positive test have risen to 912 people in the past week (10.8% increase). Having fallen for a few weeks, sadly admissions to hospital are also rising – which means we can expect the numbers of people dying to continue to rise for some weeks too. 6,315 people were admitted to hospitals in the week to 17th October, 845 more than in the previous week (15% increase)
In terms of people aged 12+ eligible for vaccination, across the UK four-fifths (79%) have been fully vaccinated and 86.2% have received at least one dose.
Estimates by ONS of true prevalence based on random-sampling of the population (rather than people seeking tests), find that “in the week ending 9 October 2021; we estimate that 890,000 people in England had COVID-19 (95% credible interval: 839,700 to 941,300), equating to around 1 in 60 people.” Hence, the numbers below are a clear underestimate.
Similarly, across the UK as the whole, KCL/ZOE app team independently estimate around 1 million people had an active infection on the 22nd October (compared to 934,000 on the 7th October), based on symptom reporting by up to 4.6 million app users. This essentially confirms the ONS estimate above.
It’s important to note that the age breakdown at the moment is very distinct. The ONS esimtate that 8.1% of those in school years 7-11 would have tested positive in the week ending 9th October – but rates are much lower for other age groups: 3.1% for age 2 to school year 6, 1.1% for school year 12 to age 24, 0.6% for 25-34, 1.2% for 35-49, 0.7% for 50-69 and 0.6% for 70+.
The Covid-19 Actuaries Response Group has updated their useful chart on how modelled estimates of daily hospital admissions in England compare the reality. They describe the chart as showing: “Admissions stayed below central scenarios modelled ahead of “Freedom Day” but now appear to be rising to another peak, as modelled by Centre for Mathematical Modelling of Infectious Diseases at the London School of Hygiene and Tropical Medicine. Since the summer peak was smaller than modelled, will this one be larger?!”
International context
The number of people testing positive globally each day has been falling for some time: around 410,000 (7-day-rolling average to Oct 19th) compared to around 626,000 to Sept 4th. The situation is different in different continents (and countries within them). Confirmed cases have been falling consistently for some time in South America – to the lowest levels since spring last year, and have recently been falling in North America, and Asia too. Across Europe as a whole cases are rising quite dramatically – though this is limited to some countries, with others maintining very low rates. While there seems to have been a fall across Africa, data is much more limited for the continent.
The numbers of people dying with their deaths attributed to Covid-19 continue to rise, particularly in Asia. Over 4.8 million people have had their death attributed to Covid-19 since the start of the pandemic. This is believed to be a significant underestimate due to limited death registration data in many countries – including India for example. The Economist estimates that excess mortality at 16.4 million (which will include deaths from causes other than Covid-19 as well as deaths caused by Covid-19 but not recorded as such).
Notes
The core advice remains: If you have symptoms (or if you are asked to by contact tracers), self-isolate until you have a negative test – or for 10 days since your symptoms appeared if you test positive or are asked to by Test and Trace. If you are struggling with self-isolating, please get in touch with us or with one of the local support groups, or call 0808 196 3646 between 8am and 8pm, 7 days a week to self-refer or visit NHS Volunteer Responders. You may be able to receive financial support to self-isolate from Stroud District Council.
Book a test via this link. You can now do this whether or not you have symptoms – it’s really important you isolate and get tested if you have symptoms (fever, new cough, loss of smell/taste). The link will tell you which type of test to book if you have symptoms or not. Twice weekly rapid tests are available to everyone in England without symptoms. If you have symptoms, there is a permanent unit at Hempsted Meadow in Gloucester, and a walk-in unit in Stratford Park. See this link for details of testing locations in Gloucestershire.
Whether or not you have symptoms, please still follow the public health advice to meet outside when possible, keep indoor spaces well ventilated with fresh air, wear masks when appropriate (they will help prevent spread of the virus if you have it but don’t have symptoms yet, or are asymptomatic – meaning you have the virus but without ever getting any symptoms), keep distance from people, and wash your hands regularly.
If there is a piece of guidance you have a question about, again – please ask in our Facebook group.
These updates are designed to improve understanding of the pandemic and its impacts, with the hope this can help us to reduce those impacts locally. We appreciate they do not involve space to properly convey the full impact of the virus nor the restrictions that are making life difficult for many people. We’re also volunteers with no public health expertise – collating and signposting to other sources for guidance.
Please remember we have a list of resources to support your emotional and mental health during this time on our website (and welcome further recommendations). The following numbers may be useful:
- Samaritans: 116 123
- Domestic Violence Hotline: 0808 2000 247
- Mind: 0300 123 3393
- Age UK: 0800 169 6565
- Childline: 0800 1111.
Your suggestions for inclusion of data in these summaries are welcome. Please submit posts to our Facebook group.