Matt Fletcher and Nicola Oliver

Every week, more is written on COVID-19 than any individual could possibly read. Collectively, the COVID19 Actuaries Response Group read more about the outbreak than most, so we’ve decided each Friday to provide you with a curated list of the key papers and articles that we’ve looked at recently.

Modelling – reports

Estimating infection numbers in European countries

A key variable in determining the timing and extent of easing lockdown and other restrictions is the proportion of the population that have already been exposed to the virus SARS-CoV-2 – if a large proportion have been exposed then re-opening with relatively light social distancing measures in place might be feasible without leading to a devastating second and subsequent wave of infection. If the proportion is small, then stricter measures may be required for a longer period.

There have been several recent studies aiming to determine the prevalence of infection – where some studies we have previously reported on have relied principally on estimates based on modelling of other variables (principally infections and deaths), data relating directly to infection is starting to become available in various countries:

Estimating the burden of SARS-CoV-2 in France (H. Salje et al, Science) 

This is a brief but wide-ranging report, looking at various aspects of the outbreak in France. The researchers estimate that, by 11 May 2020 (when interventions were scheduled to be eased), between 1.8 and 4.7 million people will have been infected, or between 2.8% and 7.2% of the population. This suggests that in France, population immunity is likely to be insufficient to avoid a second wave if all control measures are released.

This conclusion is based on modelling of hospitalisations and deaths, but notes that a prevalence of the virus of around 3% has been estimated among blood donors in Hauts-de-France, which is consistent with their findings.

Coronavirus: the first seroprevalence data estimates that 5% of the population has been infected (report (Spanish) & article (English))

The Carlos III public health institute collected blood samples from almost 70,000 participants, to determine the proportion of people in Spain that had developed antibodies following infection. Results suggest that around 5% of the Spanish population have been exposed to SARS-CoV-2, with prevalence highest in areas closest to Madrid, typically over 10%.

A phased approach to unlocking during the COVID-19 pandemic – lessons from trend analysis (Heald et al)

This is a report which has received some recent media attention, focusing on experience in England – the analysis set out information about COVID-19 cases in each of 149 Upper Tier Local Authorities and estimates that the reproduction rate R has fallen on average from 2.8 to 0.8, sufficient to suppress infection from the disease. They also estimate that 29% of the UK population has already been infected; however, they do note that this finding relies on a linear extrapolation of current trends, an assumption which may not hold. In addition, a significant proportion of deaths are not mapped to a Local Authority which may distort the trend within a given local authority.

It’s worth noting that there is a substantive seroprevalence study underway in the UK, with initial results expected in June – this should give a direct indication of population-level exposure without needing to extrapolate from alternative data sources.

The studies in France and Spain in particular show that, as expected, prevalence is higher than the official counts of positive tests – for example in Spain, over 200,000 cases have been recorded, where a 5% prevalence would indicate over 2,000,000 cases. But the figures are not indicative of a level of ‘herd immunity’ to SARS-CoV-2 being reached.

Clinical and Medical News

How Scientists Plan to Develop a Coronavirus Vaccine

We have covered the potential for vaccine development and examined the pharmaceutical pipelines previously. The process of exactly how a vaccine will eventually make it to market is outlined here: This includes a step-by-step overview from sequencing of the virus, through to testing, approval and manufacturing. This confirms that the process is far from simple and nothing is guaranteed to succeed.

Antibody Tests

An antibody test developed by Roche that has demonstrated excellent specificity and sensitivity following evaluation at the government’s Porton Down facility has gained approval from Public Health England . At this stage, there is no official announcement on whether these tests will be adopted by the NHS. The test already has approval from medical regulators in the EU and the United States.

The role of vitamin D?

The protective role of vitamin D in COVID-19 is unclear. In an insightful commentary piece, Dr JoAnn Manson, professor of medicine at Harvard Medical School and Brigham and Women’s Hospital discusses the potential benefits of vitamin D . It is known that vitamin D has an immune-modulating effect and can lower inflammation. Emerging evidence from COVID-19 patients suggests that vitamin D deficiency may be implicated in more severe disease.

A randomized clinical trial of vitamin D supplementation is planned in moderate to high doses, to see whether it has a role in the risk of developing COVID-19 infections and in reducing the severity of disease, hence improving clinical outcomes.

Heart Health

Interesting analysis of data derived from Fitbit users around the world suggests that, when comparing the baseline data from January to that of February, March, and April, the average resting heart rate declined . Three key trends were identified: step counts declined, but active minutes increased; sleep duration increased; and, bedtime variability decreased. This is all good news for cardiovascular health.

Air Pollution Reduction and Mortality Benefit

The stringent traffic restrictions and self-quarantine measures adopted by China have resulted in a reduction in transportation emissions. Emissions from residential heating and industry remained steady or slightly declined. . This analysis suggest that interventions to contain the COVID-19 outbreak led to improvements in air quality that brought health benefits in non-COVID-19 deaths. Whilst emissions are likely to increase again once restrictions are eased, perhaps there is now increasing appetite to maintain some transport restrictions considering the impact on health outcomes.

Kawasaki Disease

Kawasaki disease is a condition that mainly affects children and causes inflammation in blood vessels throughout the body. Early treatment results in rapid recovery for most; delayed treatment is associated with increased risk of prolonged disease with cardiac involvement. Clinicians are starting to report increasing cases of Kawasaki disease, probably associated with SARS-CoV-2. Analysis of paediatric patients in Bergamo reports a worrying 30-fold increased incidence of Kawasaki-like disease.

Research Research! Part Two

Last week we reported on a key repository for COVID-19 research in which studies published in pre-print or without peer review are assessed for their strengths and weaknesses .

This editorial in the BMJ highlights how poor quality research is damaging to an effective response . The authors note that many of the registered clinical trials are likely to be too small and too poorly designed to be of any use, that access to preprints has led to irresponsible dissemination as flawed studies are picked up by the media and that there is a plethora of duplication.

Data

ONS data

The Office for National Statistics continue to produce timely data on both COVID-19 and all-cause deaths. On 15 May, they published two new datasets:

Deaths involving COVID-19, England & Wales: deaths occurring in April 2020 (data & report)

This dataset sets out how many people have died from COVID-19, where deaths occurred up to 30 April 2020 and were registered up to 5 May. Information is provided about the characteristics of those who died (including pre-existing conditions), comparisons of COVID19 to other causes of death, along with other relevant information about the registration of deaths.

It shows that of the 33,841 deaths involving COVID-19 occurring between 1 March and 30 April, 95% had COVID-19 assigned as the underlying cause of death; for context, this is equivalent to the third highest cause of death for the whole of 2018. For April 2020, COVID-19 was the most frequent underlying cause of death for deaths occurring in April 2020.

Dementia and Alzheimer disease was the most common pre-existing condition found among deaths involving COVID-19, involved in around 20% of COVID-19 deaths.

Deaths involving COVID-19 in the care sector, England and Wales: deaths occurring up to 1 May 2020 and registered up to 9 May 2020 (provisional) (data & report)

This dataset sets out information on the number of deaths between 2 March and 1 May 2020 among care home residents, both from COVID-19 and other causes, and splitting out where these deaths occurred. In total, there were 45,899 deaths of care home residents; of these, 12,526 involved COVID-19 (27% of all deaths of care home residents). Of these 12,526 deaths, 72% (9,039 deaths) occurred within a care home and 27% (3,444 deaths) in a hospital. This means that, of all deaths in hospital from 2 March 2020 involving COVID-19, 15% were deaths of care home residents.

Health Data Research UK repository

Health Data Research UK is a group looking to bring together UK health data for research and innovation. Their COVID-19 response website can be found here. HDR UK have also set up a repository containing resources for COVID-19 research  – this pulls together a wide array of links to datasets and research relating to COVID-19 from the UK and elsewhere.

Et finalement …

Not sure if it’s a boy or a girl

One thorny issue that those on the continent are grappling with is whether the disease is masculine or feminine.

In France, many started off using the masculine ‘le covid’, but it seems that L’Académie Française, the esteemed guardians of the French language who are largely concerned with ensuring that the language is not infected by Anglicisms, have decided that it should in fact be ‘la covid’. On the academy’s website under the heading ‘Dire, ne pas dire’, (‘say, don’t say’), they announce:

‘On devrait donc dire la covid 19, puisque le noyau est un équivalent du nom français féminin maladie.’

‘We should therefore say (la) covid 19, since the nucleus is an equivalent of the French feminine name illness.’

In Italy, the situation is more complex – according to virologist Fabrizio Pregliasco, the disease Covid19 is feminine (la covid) , but the virus SARS-CoV-2 is masculine (il SARS) 

And in Spain, the debate continues …

We’ll keep you informed as data from other countries comes in – do let us know if you have the relevant information!

16 May 2020