Infection rates fall by 30% in England: is this cause for celebration?

Yesterday news media and government highlighted the recent epidemiological evidence of a substantial fall in cases of infection since the second lockdown was introduced in England.  What did the data show and what are the implications for what is being proposed next to control the epidemic?

(Although this is very much an English story, for readers of this blog in the rest of the UK and beyond, I believe the lessons are important for all countries)

What are the data?

• As regular readers of this blog may recall, no reliance can be given on any of the official national data on the reports number of people who tested positive for Covid-19, given the variability in those who are encouraged or turn up for testing
• In England since May there has been a major epidemiological study, taking random samples of the population, who have provided a swab for analysis and other background information
• There have been 8 such rounds of sample collections, the first in May and the latest finished last week
• Sample sizes are substantial (120,000 to 170,000) for each round
• The estimates of the prevalence of infection (percent of samples that are positive for Covid-19) are adjusted so that the results reflect the age, gender, region and social characteristics of the country as a whole

How many people were infected in each of the surveys?

• I have plotted on this chart the percent of people who were positive during each period of sample collection
• The width of the bars varies to reflect difference in the periods over which the samples were collected and analysed 
• Given the concerns about the second wave, the researchers made their data collection more frequent from October 
• What do I take from this graph:
  • Currently just under 1% of the English population are infected
  • Yes, there has been a fall of 30% since the peak in early November 
  • Equally important, the current rate is 30% higher than when the tiered lockdown started in England on 14^thOctober

But shouldn’t we look at trends as well as the current level?

• The decline in cases is important of course
• Indeed, it would have been extremely worrying if the prevalence of cases had not fallen since the lockdown was introduced on 5 November
• It is interesting to consider the trajectory of the decline, and the researchers have estimated a halving time of 14 days.  Obviously this is only an estimate, but if the reduction continues at this rate then as ‘a worked example’ I have produced the table below:

• Thus, if we continued on the current level of lockdown, it would take until early February to achieve the same low level of the population infected as existed in late July

Would mass testing affect these trends? 

• Mass testing of high-risk groups for transmission eg school children, university students could, only if followed by rigid contact tracing and isolation,  drive the transmission down further 
• If this were to happen, then the rates above could potentially fall faster

Do the data tell us anything about who is most infected?

• The numbers in some subgroups are quite small but there are some interesting findings
• The prevalence by age confirms that the younger age groups (school children and students) have the highest rates
• The over 65s – who are presumably socially distancing more – have the lowest rates 

• There are also some interesting data by household size
• Perhaps not surprisingly, the larger the household size the greater the risk – my understanding of the data is that this takes account of age
• Households with 7 people have 7 times the risk of single person households

• One final statistic to point out is that restricting the analysis to those participants who had not had contact (as far as they knew) with anyone who had tested positive or had any symptoms, 0.5% of these also tested positive 

The example from Melbourne

Before giving my conclusions I thought it useful to look at another country’s experience

• Extrapolation of examples from other countries have always to be treated with some caution as local circumstances vary

BUT

• At its peak the rate of new cases in Melbourne was around two thirds of that the rate of new cases is in England today
• Melbourne entered a 112 day severe lockdown which restricted movement outside 5K from home and only for essential shopping etc
• The results are illuminating (see below) and emphasise it takes time to get this disease under control (without a vaccine, but it is possible!)

What do I take from the UK figures?

• A prevalence of 1% in a disease as infectious as Covid-19 suggests a very high number of infected individuals in the population
• The infection remains at an epidemic level in all age groups and all regions
• The tighter lockdown is now beginning to have an effect in reducing the numbers; whether the tiered lockdown introduced for just 3 weeks from 14 October would have had the same effect is not calculable, but certainly not proven
• If we continue with the same level of isolation the epidemic could (even without a vaccine) get down to acceptable levels by February 
• The government have announced mixing of households over Christmas, and are likely to respond to political pressure to reduce the burden from the tiers.
• This mixing will presumably increase inter-generational exposure to include those in age groups who are the largest transmitters
• I believe these actions will adversely influence the current downward trajectory
• I try to make these blogs non-political, but I do not believe the current plans are helpful

The opposing view

• I will finish by mentioning again the ‘Barrington accord’
• The argument is that (with my comments in bold):
  • More than half the cases of Covid-19 remain clinically well – true
  • The infection is really only a problem (in terms of significant numbers) in those aged over 60  – true
  • Death rates are coming down due to better treatment – half true: less than 0.5% in those aged under 60, (but 1 in 6 in those aged over 80)
  • There are major health consequences, both physical and psychological, of lockdown and similar restrictions – true, but remain to be calculated
  • There is a direct connection between a fall in GDP and health – true in non Covid-19 situations 
  • The financial hit also affects health expenditure – it could

My take home message

• The reduction reported is welcomed, but given the lockdown was expected
• The epidemic is still very much present in the whole of the UK
• We still need to reduce transmission much further and this cannot be achieved overnight

To readers!

I found this a difficult piece to write as I do aim to be positive and also keep to the facts!  Do you agree with my conclusions on the data presented.  Feel free to post your comments – it is an interesting and indeed crucial debate!

If you would like to receive email notifications of new posts on this blog just click this link and enter your email at the bottom of the home page: https://makingsenseofcovid19withs.com