Wednesday 19th August 2020

Every week we share a summary of local, national, and international data on the coronavirus/Covid-19 pandemic. We really welcome questions, feedback, or suggestions for inclusion of data you have found useful, in our Facebook group, or email us at:

Key points

  • The big statistical news this week is a change to the total number of deaths attributed to Covid-19 in the Public Health England (PHE) dataset (now 41,466). Below we explain this change, and why it does not impact the ONS register of deaths where Covid-19 is mentioned on the death certificate (now 51,879 in England and Wales, to 7th August).
  • As of 18th August there have been 277 total confirmed cases of Covid 19 in Stroud district. This number is five higher than we reported last week. The data of specimen for the last confirmed case was 15th August.
  • 1,942 cases of Covid-19 have been confirmed in Gloucestershire. 30 cases have been added to this total since we reported on this data last week.
  • As of 7th August, no-one has died with Covid-19 mentioned on their death certificate in Stroud district since 12th June, and in Gloucestershire since 19th June – a month ago today.
  • Swindon is identified in PHE’s weekly surveillance report as having a rising trend of individuals tested per day and being an area of “Concern” – it is the only local authority area in the south of England with a rate of Covid-19 cases confirmed during the week of 45 or above.

Local statistics

Confirmed cases in Stroud district

  • As of 18th August there have been 277 total confirmed cases of Covid 19 in Stroud district. This number is five higher than we reported last week. The data of specimen for the last confirmed case was 15th August. This data and that below is based on analysis of spreadsheet data downloaded from the government’s coronavirus data dashboard. The dashboard provides the totals and current rates, but the local daily/weekly confirmed numbers are contained in a spreadsheet which you can download.
  • Stroud district has a confirmed cases rate of 230.9 per 100,000 people. There are 315 “Lower Tier Local Authorities” of which Stroud District Council is one. Stroud district continues to have among the lowest rates (27th lowest / 289th highest). In plain language: after adjusting for population, fewer people have been confirmed to have the virus in Stroud district than in most parts of the country. It is important to note that “confirmed” cases may not reflect the true number of people who have had the virus – but nonetheless, the difference between authorities is more likely to be because of different prevalence of the virus than differences in testing (given access to testing is largely similar around England). For comparison, the highest rate after Leicester (a Unitary Authority also covered below) is Blackburn with Darwen (1,139 confirmed cases per 100,000 people) – over four times the rate in Stroud district.
  • The chart below shows confirmed cases in Stroud district by week. Data for week 33 ending the 14th August may still be revised, but while recent confirmed case numbers are higher than since the end of May, there is little sign of a trend of rising cases. It may be that a higher proportion of cases are being picked up since 24th July and in August than was the case for June and most of July. However, Stroud district isn’t an island – please see the data below for Gloucestershire where there does appear to be some evidence of a rising trend, and for the UK as a whole.
chart described in text
Source: Government coronavirus dashboard data download

You can see where cases are being confirmed by neighbourhood (MSOA – an ONS geographic unit that covers areas of Stroud district like Stroud Town, Stonehouse, Nailsworth etc).

Confirmed cases in Gloucestershire

  • As of 18th August a total of 1,942 cases of Covid-19 have been confirmed in Gloucestershire. 30 cases have been added to this total since we reported on this data last week (very similar to the 29 we reported last week), with the last case confirmed from a specimen submitted on the 16th August.
  • Gloucestershire has a lab-confirmed cases rate of 304.8 cases per 100,000 people. Gloucestershire still has among the lowest rates (22nd lowest / 128th highest). For comparison, the lowest rate is in North East Lincolnshire at 144.8, and the highest rate is in Leicester at 1,550.4. Neighbouring South Gloucestershire has an even lower rate than Gloucestershire: 275.3.
  • The chart below shows confirmed cases in Gloucestershire by week. Data for week 33 ending the 14th August may still be revised, but it is beginning to look more like there is a trend of slowly rising cases since week 25 (ending 19th June), even if the latest week’s data is lower than for last week. While confirmed case numbers are still low week by week – and in part represent greater availability of testing, the prospect of getting to “Covid Zero” in Gloucestershire – as in the rest of the country – feels distant given this apparent trend. Furthermore, Gloucestershire isn’t an island either – and cases confirmed elsewhere can introduce further cases in the county.
chart describred in text
Source: SCCR analysis of data from the government Coronarvirus Dashboard
rates by Local Authoirty are much higher in the north in general, particularly in the north west
Source: PHE

Estimate of local cases

The Covid Symptom Study was estimating a rate of 1,412 per 1,000,000 people for Stroud district a few weeks ago – resulting in an estimate by Gloucestershire Live that 168 people might have the virus in the district. This week the figure is 137 (lower even than the 282 active cases per million people of last week). As of the ONS latest population estimate (2018) there were 110,019 people living in Stroud district, by our calculations that means an estimate of just 15 people with the virus in Stroud district. This estimate is low and consistent with the confirmed cases number to the extent is it made from a small sample and it would be extremely difficult to be statistically accurate at this level. Nonetheless, this is a reasssuring estimate.

You can read about how these estimates are made on the Covid Symptom Study website. We encourage anyone with symptoms (a new continuous cough, high temperature, or loss of smell/taste) to please book a test. There is a permanent unit at Hempsted Meadow in Gloucester, and mobile units tour Gloucestershire.

If you don’t have symptoms, please still follow the guidelines to 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

People who have died because of Covid-19 in Stroud and Gloucestershire

National level data

Public Health England (PHE) update to total number of deaths attributed to COVID-19.

On July the 16th the Centre for Evidence Based Medicine published a piece by Yoon K Loke and Carl Heneghan, titled “Why no-one can ever recover from COVID-19 in England – a statistical anomaly“. It highlighted the fact the way that PHE counted deaths – looking “for people on the NHS database who have ever tested positive, and simply checks to see if they are still alive or not… Anyone who has tested COVID positive but subsequently died at a later date of any cause will be included on the PHE COVID death figures.”

At the beginning of the pandemic, this made sense. PHE explain in a detailed “Technical Summary” that the methodology was based “a statement from the World Health Organization (WHO) Regional Office for Europe on 4 June 2020 that“WHO has no defined time-limitation for COVID-19 related deaths. This should be a clinical decision” (as indeed it is in death-certificate data).

PHE has now “undertaken an assessment ofthe impact ofdifferent timeperiodsfrom the first positive COVID-19 specimen and date of death(with and without mention of COVID-19 on the death certificate) on the mortality series which has undergone statistical peer review”. This finds that: 88% “of COVID-19 related deaths occurred within 28 days of the first positive specimen date, and that for those deaths 95% have a mention of COVID-19 on the death certificate. The proportion of deaths witha mention of COVID-19 on the certificate decreases for deaths witha longer interval between the first positive specimen and date of death;howeverthisis 73.4% for those with an interval of 29-42 days and 48.1%for those with an interval of 43-60 days between specimen date and date of death”. This is detailed in their Table 1 included below:

Table providing details. At 43-60 days, 48.1% of deaths identified through the PHE method have COVID-19 mentioned on the death certificate
Source: PHE

The PHE data series is now being revised to include two measures:

  • “deaths in laboratory-confirmed positive individuals where the death occurred within 28 days”, and
  • “deaths within 60 days or if the death occurred after 60 days, COVID-19 is listed on the death registration.”

Both measures will bepublished daily on the GOV.UK dashboard and weekly in the PHE surveillance report.

As a result, the total number of deaths attributed to Covid-19 by PHE has been reduced by 5,377 (11%). However, it is important to note that the ONS data below uses a different methodlogy – focusing only on mentions of Covid-19 on death certificates (“ONS death registrations which can be linked to laboratory confirmed COVID-19 tests are included in the PHE data series, but ONS death registrations without laboratory confirmation are not” this is a reasonable approach to take, but means that when people who died before tests were available for them will not be included even if clinicians are convinced Covid-19 played a part).

While the two methods produce different results, it remains the case that tens of thousands of people have died with Covid-19. As fuller analysis of data is possible, there may be further adjustments to estimates, but these shouldn’t be taken as an indication that Covid-19 is not a serious and regularly fatal disease, or that there is any deliberate attempt to either underestimate or overestimate the number of deaths related to the virus. The chart below shows the effect of excluding deaths where a first positive test specimen was collected over 28 days before death on the cumulative number of Covid-19 deaths PHE report.

increasing proportion of deaths in red as "28 days" limitation affects cumulative total
Source: PHE

PHE provide a list of Frequently Asked Questions about this change, including: “Does this mean that deaths have been over-reported up to now?“, to which they answer “The way deaths were countedwas based on the best available evidence at each stage of the epidemic. A time limit could only be recommended once enough data were available to analyse the time between positive test result and death, cross checking against death registration”.

The latest data from the Office for National Statistics on deaths:

  • “The number of deaths registered in England and Wales in the week ending 7 August 2020 (Week 32) was 8,945; this was 1 death fewer than in Week 31.”
  • “In Week 32, the number of deaths registered was 1.7% below the five-year average (157 deaths fewer); this is the eighth consecutive week that deaths have been below the five-year average.”
  • “Of the deaths registered in Week 32, 152 mentioned “novel coronavirus (COVID-19)”, the lowest number of deaths involving COVID-19 in the last 20 weeks and a 21.2% decrease compared with Week 31 (193 deaths), accounting for 1.7% of all deaths in England and Wales.”
  • “Looking at the year-to-date (using the most up-to-date data we have available), the number of deaths up to 7 August was 389,008, which is 52,737 more than the five-year average. Of the deaths registered by 7 August, 51,879 mentioned COVID-19 on the death certificate, 13.3% of all deaths in England and Wales.”
  • “the number of deaths for England was 365,354, which is 51,151 (16.3%) more than the five-year average. Of these, 49,267 (13.5%) mentioned COVID-19.”
deaths in March and April elevated, particularly by Covid-19 - and now back to average levels
Source: ONS

You can read the full article and access the full data on which the article is based on the ONS website.

Cases – confirmed and estimated

On the 16th August 1,040 cases were confirmed (9th August: 1,062), and the 7-day rolling average was 1,043 (last week 859.9). As you can see, this represents a slight rise since 5th July when 516 cases were confirmed and the rolling average reached a low-point of 546.1. While numbers of confirmed cases are affected by availability of testing, the Independent Sage noted this week that an increase in testing cannot alone account for the increase in total confirmed cases. The latest chart also suggests that measures to control the virus in the areas where it is most prevalent may be stemming the rising trend

daily confirmed cases rising slightly but with a dink down - and still much lower than during April and May

Prevalence and Incidence – ONS and COVID Symptom Study app

Prevalence is the total number of estimated cases at any given time, whereas incidence is the estimated number of daily new cases.

Incidence estimates are higher than the the government’s number of “lab confirmed cases” because they are designed to estimate cases that have not been confirmed through testing

The ONS also report that “Between 26 April and 26 July, 6.2% of people tested positive for antibodies against SARS-CoV-2 on a blood test, suggesting they had the infection in the past.”

The COVID app figures:

  • COVID Symptom Study app incidence: 1,214 daily new cases (lower than last week’s 1,626)
  • COVID Symptom Study app prevalence: 20,300 symptomatic COVID (lower than last week’s 24,720). The chart below shows the trend: numbers now appear to be falling, a good sign that “Zero Covid” in the UK could be achieved if people continue to adhere to precautions.
estimated number of active cases falling
Source: COVID Symptom Study App estimates

The two different methods produce different result in a similar ballpark: 20,000-28,000 symptomatic cases. Read about the difference in methodology that leads to these different estimates on the COVID Symptom Study page, “What do all the different COVID figures mean and how do they compare?” – or follow the links above for further details on each estimate.

Hospital admissions and Covid-19 patients in hospital

The governments dashboard shows 16 Covid-19 patients were admitted to hospital in England on the 16th August, together with 38 for Wales – making 54 for the UK as a whole. That is down from 94 a week earlier (10th August – 63 for England and 31 for Wales). By contrast, at the peak – on the 12th April – a total of 19,872 Covid-19 patients were admitted to hospital. The trend appears to be continuing downward after a brief small rise. We can perhaps again imagine days in the near future where noone will be admitted to hospital with Covid-19 across the UK (as now appears to have been the case in Scotland and Northern Ireland for a number of individual days).

number of patients admitted to hospital falling

As of the 17th August, there were 895 Covid-19 patients in hospitals in the UK. 73 of these patients are in mechanical ventilation beds – and we can send our hopes that these people will recover. The number is down from 999 a week previously on 10th August, though we do not know to what extent this is because people have died, or because they have recovered and been discharged.

number of patients in hospital falling

PHE surveillance report

Every week PHE publish a “surveillance report” on the pandemic. This week “The Secretary of State for Health and Social Care, drawing on epidemiological advice from the CMO, NHS Test and Trace, JBC and PHE, has determined the following Watchlist (Table 1), highlighting the local authorities of greatest con-cern.”
We have reproduced this below to note the inclusion of nearby Swindon – identified as having a rising trend of individuals tested per day and being an area of “Concern”, defined as being “for areas with the highest prevalence, where the local area is taking tar-geted actions to reduce prevalence e.g. additional testing in care homes and increased com-munity engagement with high risk groups”.

table identifying Swindon
Source: PHE

The comparative position of Swindon as an upper-tier local authority can be seen in the chart below: it is the only local authority area in the south of England with a rate of Covid-19 cases confirmed during the week of 45 or above.

other local authorities with high test numbers are in the North West of England - the South West has the lowest rates of any region in each local authority bar Swindon
Source: PHE

For the South West as a whole, PHE report 13,684 positive cases, and an incidence rate of 244/100,000 people. This is the lowest rate for an English region, the highest rate is 667/100,000 in the North West.

International data

It is important to say that different countries are testing and collecting data on deaths in different ways, making fair comparisons difficult. However, there are reputable sources doing their best to make data available, which we summarise below

People who have died with deaths attributed to Covid-19

Last week we highlighted the Financial Times’ interactive chart that can show the “Cumulative deaths (per million), by number of days since 3 total deaths (per million) first recorded”. This shows the UK with the third highest number (619.5) after Belgium (867.2), and Peru (820) – with deaths still rising fast in Chile, Brazil and the United States which will potentially achieve even worse rates. Other countries have far lower rates, despite the virus having been present for a similar time – such as Germany (111 deaths per million), Cuba (7.8), South Korea (5.9), and New Zealand (4.5).

Another option on the FT site is to compare the “Seven-day rolling average of new deaths, by number of days since 3 average daily deaths first recorded”. This makes it possible to see whether the worst impacts of the virus have peaked in a country, and identify any “second waves”. In the chart below, we have picked out the UK – where the 7-day average number of daily deaths has been falling consistently for a long time, but appears to have stalled. The pattern is similar to that for Italy. The chart also shows the pattern in Spain – where a sharper decline in daily deaths is accompanied by a more recent uptick that may represent the foothills of a “second wave” – though this also appears to have stalled. The chart also shows that daily deaths continue to be very high in other countries such as Peru, Brazil and the United States (note that the left-hand scale representing numbers of deaths is “logarithmic” – to make it easier to compare very large numbers with much smaller ones on the same chart).

chart described
Source: Financial Times

Last week, we also pointed to the Office for National Statistics useful “Comparisons of all-cause mortality between European countries and regions: January to June 2020“. We also recommend the Financial Times‘ visual narrative of the spread of Covid-19”, with charts showing comparisons between 20 countries on the best available testing measure – “excess deaths”, as well as global regional comparisons to the 13th July. Sadly this no longer seems to be being updated. We will include a chart again when there is an update.

Our analysis of data from the Johns Hopkins University tracker shows

The number of people who have died is now well over three-quarters of a million – 781,932

40,209 people died in the past week (very similar to the number who died in the previous week).

There are fourteen countries where over 10,000 people have died where Covid-19 was involved:

  • The USA – over 170,000 people haved died – 171,833 (7,296 people died in the past week, a 4.4% increase. This is a slightly lower rate of increase compared to last week, but still higher than three weeks ago)
  • Brazil – 109,888 (8,136, 8.0% – the highest weekly number since 17th June, after last week we found the lowest number of people to die in a week since 2nd June)
  • Mexico – 57,774 (3,845, 7.1% – the lowest rate of increase we have reported in a week for the country)
  • India – 52,888 (6,797 – 14.7% – a higher number of people than last week)
  • The UK – 41,466 (232 people who did have been added to the total in the past week, a 0.5% increase – the lowest weekly number we have reported so far. However, this number is affected by methodlogy changes so comparisons are difficult. As explained in detail above, the total official number for the UK is technically 5,145 lower than last week. However, this is because PHE have altered their methodology and – by limiting deaths they attribute to COVID-19 to those that took place within 28 days of a positive test, reduced the total number of deaths they report by 5,377). Please note that the ONS data covered above – which appears at a lag – uses a different methodology and records that “Of the deaths registered by 7 August, 51,879 mentioned COVID-19 on the death certificate, 13.3% of all deaths in England and Wales.”
  • Italy – 35,405 (190, 0.5% – the highest weekly number since the 24th June)
  • France – 30,434 (106, 0.3% – the highest weekly number since the 22nd July)
  • Spain – 28,670 (89 – 0.3% – the highest number of people to die in a week for seven weeks)
  • Peru – 26,658 (5,157, 24%, higher than last week)
  • Iran – 20,125 (1,325 – 7.0%, higher than last week but lower than the week before)
  • Russia – 15,951 (720, 4.7, lower than last week)
  • Colombia – 15,619 (2,144, 15.9%, lower than last week).
  • South Africa – 12,264 (1,513, 14.1%, lower than last week)
  • Chile – 10,546 (368, 3.6, lower than last week%)

Confirmed Cases

Confirmation of cases is one way to measure the spread of the virus. It is worth bearing in mind that not everyone who has the virus, particularly if they do not develop senses, will get a test – so confirmed numbers are an underestimate of total cases. However, changes in confirmed case numbers can give us an indication of whether the virus is still spreading or is coming under control in a country. Our analysis of Johns Hopkins University data shows:

  • Over 22 million cases of Covid-19 have been confirmed (22,173,973)
  • That’s an an increase of over 1.8 million cases during the week (1,867,117, the eighth week in a row with over 1 million cases confirmed. The total number of cases confirmed in a week is higher this week than last week.
  • Global confirmed cases increased by 9.2% in the past week (this percentage increase is falling slowly, a sign that global confirmation of cases may be peaking).
  • 20 countries have over 200,000 cases (the same number as last week), and three have over 2 million confirmed cases. Below we cover details for countries with over 300,000 cases – as this now includes 14 countries (two more than last week).
  • The USA has confirmed over 5 million cases – 5,486,232 cases, a weekly increase of 369,689 (6.7% – a lower increase than the previous week for the fourth week running, the lowest increase since the 1st July)
  • Brazil has confirmed over 3 million cases – 3,407,354 cases, a weekly increase of 349,884 (11.4% – the highest weekly increase in cases we have recorded)
  • India has confirmed over 2 million cases – 2,767,253 – a weekly increase of 437,615 (18.8% – a higher number but lower percentage increase than last week)
  • Russia – 935,066, a weekly increase of 34,321 (3.8% – lower than last week)
  • South Africa – 592,144, a weekly increase of 26,035 (4.6% – lower than last week and a new lowest increase since 8th July when we started compiling weekly data for the country)
  • Mexico – 531,239, increase of 38,717 (7.9% – the second consecutive fall in weekly number of confirmed cases)
  • Peru – 549,321, increase of 59,641 (12.2% – higher than last week)
  • Colombia – 489,122, increase of 78,669 (19.2% – higher confirmed cases than last week, and the highest rate of increase for any country. The percentage increase is falling, which hopefully signals an approaching peak)
  • Chile – 388,855, increase of 12,239 (3.2% – lower than last week, the lowest weekly total since 17th June when we started recording numbers for the country)
  • Spain – 364,196, increase of 37,574 (11.5%). This is a nearly twice as many confirmed cases as last week. Confirmed cases have been increasing for 10 weeks in a row – just 2,034 cases were confirmed in the week ending 9th June.
  • Iran – 350,279 cases – a weekly increase of 19,090 (5.8% –higher than last week)
  • The UK – 322,178 cases, increase of 8,776 (2.8%). This is a higher number and percentage than last week, and indeed any week since 17th June (9,025, 3.1%). Confirmations of cases have risen for four consecutive weeks (though it is worth bearing in mind that this partly reflects more widespread testing, this is not the only cause). There has not yet been a week with fewer than 4,000 cases since the peak.
  • Argentina – 305,966 cases, increase of 45,055 (17.3% – lower than last week).
  • Saudi Arabia – 302,686 cases, increase of 11,218 (3.8% – higher than last week)
  • The remaining countries with over 200,000 cases are Pakistan (290,445), Bangladesh (285,091), France (256,534), Italy (254,636), Turkey (251,805), and Germany (228,543). Weekly case numbers are higher than last week in all but Pakistan.

Our condolences

The aim of these updates is to improve understanding of the pandemic in the belief this can ultimately reduce the number of people affected. We are aware that presenting the information in this way can seem cold – we send our condolences and best wishes to the many individuals and their loved ones affected.


We welcome your suggestions for inclusion of data in these summaries. Please submit posts to our Facebook group.

Please remember we have a (growing) list of resources to support your emotional and mental health during this time on our website. 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.