18th Oct 2020 data update

Summary

• 108 people have tested positive for SARS-COV-2 in the first two weeks of October (1st-14th), compared to 51 in the two weeks previously (17th – 30th September). That more’s than doubling. Unfortunately, we do not know the number of local tests conducted, but it is extremely unlikely they have doubled, and across the country we know the proportion of tests that are positive – not just the number – is rising.
• 723 people tested positive in Gloucestershire in the same two week period, compared to 289 in the previous two weeks. Across the county a proportion of these tests will be associated with schools, colleges and the University of Gloucestershire – but the increasing numbers suggests outbreaks are not being contained in these places (unsurprisingly, as students and pupils interact with others in their households and wider community).
• The data records 28 admissions to Gloucestershire hospitals associated with Covid-19 for October so far (to the 15th) – compared to 20 during all of September, and 4 in the whole of August.
• The good news is at – at least as of the 2nd October, there have been no new cases where a person has died with Covid-19 mentioned on their death certificate in the district or county. We send our condolences to the families and loved ones of the 583 people who have died with Covid-19 mentioned on their death certificate in the county, 92 of them from Stroud district. It takes a few weeks for registrations of deaths to come through this data system, so we cannot be sure people haven’t died with/from Covid-19 recently, with registrations yet to be made/processed.

• 106,696 people tested positive (to 14th Oct). 102,075 tested positive in the previous week (to 7th Oct). A total of 698,341 people in the UK have now tested positive at one point since March
• 658 people died in the UK (to 14th Oct), 49 more than 609 last week, by the PHE “death within 28 days of a positive test” definition (total: 43,410).
• 4,713 patients in hospital at midnight on the 11th Oct – up 1,514 (47%) since last week.
• 664 patients admitted in England on the 12th Oct, compared to 472 on the 5th Oct – higher than the 586 on the 19th March, though lower than the 1,128 admitted on the 23rd March (the day the UK lockdown was introduced – and rising at a slower rate).
• 426 patients on a mechanical ventilation bed on 15th Oct, compared to 368 on the 8th Oct (and a peak of 2,849 in April).
• ONS data shows “Of the deaths registered in Week 40 [to the 2nd October], 321 mentioned “novel coronavirus (COVID-19)”, accounting for 3.2% of all deaths in England and Wales; this is an increase compared with Week 39 (when there were 215 deaths involving COVID-19, accounting for 2.2% of all deaths).”
• “Looking at the year-to-date (using the most up-to-date data we have available), the number of deaths up to 2 October 2020 was 463,720, which is 54,282 more than the five-year average. Of the deaths registered by 2 October, 53,187 mentioned COVID-19 on the death certificate, 11.5% of all deaths in England and Wales.” (ONS)

Local

There have been 42 positive test results associated with someone living in Stroud district providing a specimen up to the 17th October. However, not only does this figure not include the last day of the week, it will not include recent tests – specimens will still be being process. In short, the 42 cannot be considered a ‘drop’ on the 52 positives from the week before – the previous week was 24 at time of writing but now most test results have been received is 52, and there is a chance this still doesn’t include all test results. While data is not available for Stroud district on the number of tests done, this rapid rise in positive results cannot be explained purely by an increase in testing – particularly given increased testing took place from May till July and didn’t identify such a trend. More people are testing positive because there are more people who are infected with the virus. It is clear from the charts below that this pattern is occuring across Gloucestershire, the South West, England and the UK – and where data is available it confirms the proportion of tests that return positive results is rising, again indicating the rise is not a function of higher numbers of tests being done.

In the most recent week in Gloucestershire there have been 255 positive test results returned so far. This fugure will rise as new results are reported. In the previous week, the number is now 388 (up from 194 when we reported last week). Numbers appear to be doubling approximately every two weeks.

723 people tested positive in Gloucestershire in the first two weeks of October (1st – 14th), compared to 289 in the previous two weeks, and 427 for the whole of September, and 136 in August. Comparisons with March and April are included in the chart below in a different colour – as all of those cases were hospitalised, while since May testing has been for “anyone with symptoms” (and people’s contacts) – in an effort to control the spread of the virus and limit hospitalisations and deaths (see below for data on hospitalisations and deaths).

There is a new version of the government’s interactive map which shows weekly cases at a more localised scale than districts (an alternative map shows this as dots, which get larger per number of cases rather than being restricted to “400+” as in the map below).

In the most recent week of data, this shows:

• 8 people testing positive in Stroud Town
• 7 in Nailsworth
• 6 in Dursley
• 5 in Upton-St-Leonars
• 5 in Berkeley and Sharpness
• 5 in Wotton-under-Edge
• 4 in Painswick, Bisley & Eastcombe
• 3 in Leonard Stanley & Uley
• 3 in Minchimhampton

Where there are fewer than 3 positive tests, numbers are not released (to protect individuals). Hence, we only lmpw there have been 0-2 cases in Frampton, Whitminster and Eastington; Stonehouse; Rodborough & Thrupp; Chalford and Bussage; and Cam.

As across the country, the number of people admitted to hospital or diagnosed with Covid-19 in hospital is increasing.

The data for NHS Trusts by week summarises a variety of ways people in hospital can test positive*. The numbers are still low – at most 5 admissions a day – but you can see that were were many days in September where no-one was being admitted, but that there are far more admissions in recent weeks.

The data records 28 admissions for October so far (to the 15th) – compared to 20 during all of September, and 4 in the whole of August.

(*the chart below “Shows the number of patients admitted in previous 24 hours for the first time with COVID-19 plus the number of patients diagnosed in hospital in previous 24 hours where the test was within 48 hours of admission plus the number of patients diagnosed in hospital in previous 24 hours where the test was 3-7 days after admission (lagged by 5 days)”.

Data for the South West region is available to the 15th October – and shows a worrying trend where hospitalisations are rising

While numbers are lower than for other regions (46 on the 14th Oct is the highest recent daily number for the South West, compared to 258 on the 16th in the North West, and 184 in the North East and Yorkshire) – the trend is following a similar trajectory (see national section below).

In October so far there have been 358 Covid-19 admissions to hospital or diagnoses in hospital. 15 of these have come from care homes. The number of admissions and diagnoses is already nearly triple the 135 for the whole of September (0 from care homes).

Estimated number of people with the virus

Alongside positive tests, we can also look to estimates that can cover people who – for whatever reason – do not get tested. The Covid Symptom Study currently estimates (18th October) that 361 people [aged 20-69] may actively have the virus in Stroud district (up from 3016 on the 11th, but nearly double the 183 I reported on the 3rd Oct).

The Study also estimates 852 active cases in South Gloucestershire (up from 532, and double the 416 on the 3rd), 455 active cases in nearby Gloucester (up from 276 last week), 231 in Cotswold (up from 127), 396 in Cheltenham (223), 272 Tewkesbury (204) and 177 in the Forest of Dean (147). By these measures infection levels are rising less slowly in the recent week after a period of doubling. They are still well above what they were a month or more ago.

While the Covid Symptom Study app is an authoritative study, with over 4 million people contributing information about their daily health, estimates for local areas will have to be based on small smaples and are estimates – they are made to help us understand the situation, but are not definitive. They are particularly useful as a complement to positive test numbers – which we know do know cover all people with symptoms or who are infectious (but unable to get a test or do not realise they should get one). You can read about how these estimates are made on the Covid Symptom Study website.

There are 3,792 people contributing to the Covid Symptom Study app in the district (fewer than last week). The more people contribute, the more accurate local (and national) estimates will be, and the app is valuable for other reasons – drawing attention to the latest research from the study, for example. It is a different app to the NHS Covid-19 app. Download the ZOE Covid-19 app to a smartphone via this link.

People who have died locally

• Good news – there have been no cases where a person has died with Covid-19 mentioned on their death certificate in the district or county in the most recent data for death certifications (to week 40, 2nd October).
• 583 people have died in Gloucestershire with Covid-19 mentioned on their death certificate (ONS filtered data download). The last person to die with Covid-19 mentioned on their death certificate in the county had their death registered in the week ending 18th September.
• 92 of the 583 people to have died in Gloucestershire with Covid-19 mentioned on their death certificate were from Stroud district. The figure for Stroud district has thankfully not increased since week 24 (12th June), when two people from the district’s deaths were registered.
• Nationally, as the chart below shows, the rate at which people are dying with deaths attributed at least in part to Covid-19 is increasing. The default chart uses a “logarithmic” scale to better show the rate of increase, you can switch to “linear” to see that the recent numbers are still relatively low compared to the previous peak (for the time being).

National

We’ve been made aware of a website that collates national (and local, and some international) data nicely. This shows, for instance, that the 7 new cases in Stroud in the latest day’s data puts the area among the lowest in the country – in a large number of areas over 30 cases are being identified every day: e.g. 229 in Wigan. Explore the interactive map yourself.

For a full summary of national data, please watch the Independent SAGE video released on Friday 16th October.

I’m continuing to see people suggesting that the only reason more people are testing positive is because more tests are being done, so include the chart again below.

Ths is a chart showing the proportion of tests in the UK that are positive. The chart shows how:

• at the beginning of the epidemic in the UK, the proportion of people testing positive was very high – 20% or 1 in every 5 people tested. That was because very few tests were done (only 23,000/day even by the end of April), and – for a while – only on people going into hospital (with at least fairly significant symptoms).
• As more testing was done, through the spring and summer you can see how a lower and lower proportion of people tested positive – because even though from mid-May anyone with symptoms could get a test, there wasn’t that much of the virus around. Alongside this, the government started to ease restrictions.
• At the lowest point – at the end of July – just 0.4% of people tested were testing positive. At the end of July, there were 170,000/day PCR tests being processed. And only 0.4% tested positive.
• Since then, the numbers of tests have increased, nearly doubling, to around 300,000/day. But the number of people testing positive has risen much faster – from ~600/day at the end of July to around over 16,000 a day. That’s what this chart summarises – the proportion testing positive has gone from 0.4% (less than one in every 200 people tested) to over 5% (1 in in every 20 people tested).
• Similar charts are available by regions of England, and they show the same broad phenomenon is happening accross the country, though the proportion of tests that return positives are much higher in other parts of the country than in the South West – at least for now.

While data on the numbers of tests or share of tests that are positive isn’t available for Stroud or Gloucestershire specifically (at least so far), there’s no reason to assume that the same thing isn’t happening here, and no basis for arguing that the only reason we have seen higher numbers of positive tests lately is because there is more testing. We had more testing happening in July than in April but it still identified fewer positives, because the virus wasn’t spreading then.

The charts below from PHE’s “Weekly national Influenza andCOVID-19 surveillance report” (pdf) show the proportion of tests that are positive is rising in all English regions, including the South West (where it has risen from under 1% of tests in week 27 to around 3% of tests in week 40 (to 2nd Oct). Please note that this is just tests for Covid-19 – though the report also includes data on influenza, data is not combined (despite false claims circulating in the past week).

Across the UK, the number of confirmed COVID-19 patients in hospitals at midnight 11th Oct was 4,713. A rise of 1,514 (47%) since midnight on the 4th Oct. For comparison, on the 30th March – the earliest data for comparable data, there were 10,732 patients, and at the peak on 12th April 19,849 patients.

In England, there were 4,379 COVID-19 patients in hospital as of midnight 11th October. This is higher than the number for 25th March – 4,262 (two days after the national lockdown was announced).

Looking in more detail at the new hospital admission numbers, Independent SAGE member Kit Yates (Author and Mathematical Biologist at the University of Bath), highlights that it appears hospital admissions for England are doubling at a rate somewhere around 18-14 days (ie, faster than every three weeks, possibly as fast as every two weeks).

Kit Yates showed a chart showing the 3-day average of daily number of new hospital admissions with Covid-19 per day per million people is rising in all regions across England – but particularly in the North West, North West and Yorkshire and Humber. See more on the South West below.

Yates highlights that when looking at “The same plot as above, but with NW and NE+Yorkshire removed shows how the other regions of England are on a similar trajectory a few weeks behind the Northern regions, but still worryingly high and growing.”

ONS data shows “Of the deaths registered in Week 40 [to the 2nd October], 321 mentioned “novel coronavirus (COVID-19)”, accounting for 3.2% of all deaths in England and Wales; this is an increase compared with Week 39 (when there were 215 deaths involving COVID-19, accounting for 2.2% of all deaths).”

“Looking at the year-to-date (using the most up-to-date data we have available), the number of deaths up to 2 October 2020 was 463,720, which is 54,282 more than the five-year average. Of the deaths registered by 2 October, 53,187 mentioned COVID-19 on the death certificate, 11.5% of all deaths in England and Wales.”

Kit Yates adds: “Deaths are doubling roughly every 10 days, but we would expect that to slow to the two week trend seen in hospital admissions. Remember, deaths lag cases by around three weeks and these latest data are already two weeks old, so increases in deaths are locked in for the next month

There is some confusing/misleading coverage regarding deaths caused by Covid-19 and flu. The ONS has a good piece summarising this:

• “Between 1 January and 31 August 2020, 52,327 deaths in England and Wales involved COVID-19. Out of these, 48,168 deaths were due to COVID-19: that is, COVID-19 was the underlying cause. This was 12.4% of all deaths for the period (389,835 deaths). In the same period 69,781 deaths involved pneumonia and 506 deaths involved influenza: out of these, 13,619 and 394 deaths were due to pneumonia and influenza respectively (3.5% and 0.1% of all deaths).”
• “Therefore, there were 1.3 times as many deaths where influenza or pneumonia was a contributory factor than COVID-19, but COVID-19 was the underlying cause in 3.4 times as many deaths.”
• “Of the deaths where both influenza and pneumonia, and COVID-19 were mentioned on the death certificate, the underlying cause of death was COVID-19 in 95.8% (18,642 deaths) of cases. This is compared with 0.04% (eight deaths) of deaths where influenza and pneumonia were the underlying cause of death.”

Each week, the ONS does a testing survey. This goes out like other survey work it isn’t based on symptoms or people seeking a test. This week (9th October they say:

• “An estimated 336,500 people (95% credible interval: 312,200 to 362,000) within the community population in England had the coronavirus (COVID-19) during the most recent week, from 2 to 8 October 2020, equating to around 1 in 160 people (95% credible interval: 1 in 170 to 1 in 150).” This is up from 224,400 people the previous week – in other words, slowing from nearly doubling, but still triple the estimate for 116,600 estimated to have the virus on the 2nd October.
• The COVID Symptom Study estimate 390,871 adults aged 20-69 in the UK with symptomatic COVID (a significant rise from the 310,003 last week – though no longer rising rapidly close to doubling. It is still a very dramatic rise from the ~20,000-25,000 reported through August).
• Read more about the ONS methdology, or the COVID Symptom Study page, “What do all the different COVID figures mean and how do they compare?“

There is data on testing available. The below chart shows capacity and processing of tests based on “UK government data for week of 17–23 September 2020 and correspondence with Department for Health and Social Care” (Nuffield Trust)

International

As of the 18th October,

• Over 38.5 million people have been confirmed to have SARS-COV-2.
• Over 1.1 million people have died (another 60,022 people since our last report)
• Over 27.3 million people have been deemed to have recovered (however, this figure is likely an underestimate, just as confirmed cases is an underestimate of the number of people to have had the virus – as not everyone has had a test. Also, not all countries keep a track of ‘recoveries’ or do so fully – the UK government for instance only indicates 2,578 people have recovered (153 more than last week), but this is obviously well below the true number of people to have had the virus but survived.
• Above data is from John Hopkins University.

The chart below compares cumulative confirmed COVID-19 deaths per million people for selected countries as of the 17th October. Peru has the highest rate at over 1,000 people per million (or more than 1 in every 1,000 people). The UK is among the countries with the highest rates per million people: 639.73, lower than the US (660.41), but higher than Itality (602.48). There are a number of countries with very low rates: Taiwan (0.29), Vietnam (0.36) and Thailand (0.85) have extremely low rates. Rates that are much lower than those in the UK can also be found in other countries in Europe, though even Germany (116.57), Finland (63.35) and Greece (47.01) have done badly compared to the countries with the lowest rates.

You can compare different countries, and see trends over time, using the OurWorldInData website.