Friday Report: Issue 35

By Matt Fletcher, Nicola Oliver and John Roberts

Matt

Nicola

John

COVID-19 Actuaries Response Group – Learn. Share. Educate. Influence.

COVID-19 is still one of the hottest topics for scientific papers and articles. The COVID‑19 Actuaries Response Group provides a regular Friday update with a summary of key papers and articles.

Vaccines

Demographic Characteristics of First Month Vaccinations (United States)

The CDC report on the characteristics of those receiving a COVID-19 vaccine during the first month of the programme in the US. The figure below displays the daily and cumulative total up until 14 January 2021. Of these, 37% were male, 60% were White, non-Hispanic, and 55% were aged 50 or more.

Latest monitoring data confirms safety of COVID-19 vaccines

The Medicines and Healthcare products Regulatory Agency (MHRA) has today published analysis that shows that the safety of the COVID-19 vaccines remains as high as expected from the clinical trial data that supported the approvals (link). The safety profile of the vaccines remains positive and the benefits continue to far outweigh any known side-effects.

As of 24 January 2021, Yellow Card (suspected side effect) reports for the UK were:

• 16,756 reported for the Pfizer/BioNTech vaccine;
• 6,014 reported for the Oxford University/AstraZeneca vaccine;
• 50 reported where the brand of the vaccine was not specified.

The majority of reports relate to injection-site reactions (e.g. sore arm) and generalised symptoms such as ‘flu-like’ illness, headache, chills, fatigue (tiredness), nausea (feeling sick), fever, dizziness, weakness, aching muscles, and rapid heartbeat. Generally, these happen shortly after the vaccination and are not associated with more serious or lasting illness.

Clinical and Medical News

Characteristics of those testing positive for COVID-19

The ONS have this bulletin on the characteristics of those testing positive for COVID-19 in England.

In contrast to earlier analysis, ONS report that the percentage testing positive for the coronavirus (COVID-19) has decreased in non-patient facing job roles but increased amongst those in patient-facing roles in England.

The figure displays the estimated percentage of the population testing positive for COVID-19 on nose and throat swabs by patient-facing role and age from 1 September 2020 to 9 January 2021.

In addition, symptom profiles were presented in cases compatible with the new UK variant. Symptoms were categorised as:

• any: all reported symptoms, including reporting symptoms compatible with COVID 19 whilst not naming specific symptoms
• classic: cough, fever, shortness of breath, loss of taste or loss of smell
• gastrointestinal (GI): abdominal pain, nausea, vomiting or diarrhoea
• loss of taste or smell only

People testing positive compatible with the new UK variant were more likely to report any symptoms and the classic symptoms, but less likely to report loss of taste and smell. The largest differences in reported symptoms between the new variant positive and non-new variant (triple positive[1]) were found in cough, sore throat, fatigue, and myalgia.

The figure above right displays the percentage of people with symptoms by variant, from 15 November 2020 to 16 January 2021.

UK Biobank SARS-CoV-2 Serology Study

In this study, UK Biobank collected blood samples from ~19,000 participants monthly for six months from the end of May to the beginning of December 2020 in order to ascertain the duration of SARS-CoV-2 sero-positivity.

The seroprevalence of SARS-CoV-2 infection was 6.6% at the start of the study period (end-May to June), increasing to 8.8% by the end of November/early December (Figure 1, above left). 88% of participants remained seropositive for at least six months after infection (Figure 12, above right).

Having at least one of the three classic COVID-19 symptoms (fever, persistent dry cough or loss of sense of taste or smell) was associated with a 12-fold higher risk of being seropositive.

Other key points:

• There was no evidence of a difference in seroprevalence by gender.
• The seroprevalence of SARS-CoV-2 infection was consistently lower among those over 50 years of age across all the study months.
• Seroprevalence was highest among those of Black ethnicity (16%) and lowest among those of White (9%) and Chinese ethnicities (8%). The increased risk in those from a black and South Asian background was attenuated following adjustment for socioeconomic status and other demographic factors, but still remained.
• Seroprevalence of SARS-CoV-2 infection varied by region being highest in London (12%) and lowest in Scotland (6%).
• Seroprevalence was lower in rural locations compared to urban locations, and higher in those with higher socioeconomic deprivation vs. those with lower.

Use of rapid testing technologies for COVID-19

In this review, Alex Crozier and colleagues from UCL look at how new technologies can be most appropriately used to support different testing strategies, and examine the benefits and risks. Sensitivity (proportion of infected people who test positive) and specificity (proportion of non-infected people who test negative) are central to meaningful interpretation of any test.

Real-time reverse transcription polymerase chain reaction (PCR) was the first, and still the most widely used, test, but the time lag between requesting a test and getting and acting on a result risk the opportunity for onward transmission in an asymptomatic individual.

Rapid antigen lateral flow tests provide a rapid result, but it is reported that they are less able to detect infections.

A comparison of the two types of testing is shown in the following table.

In the appendix to the main article, novel types of assay are presented in terms of how they work, sensitivity and specificity, advantages, and limitations. In addition, there is a summary of the Innova Rapid Antigen Lateral Flow Test performance data, challenges to large scale asymptomatic testing, and an overview of the principal testing strategies and examples of countries deploying them.

Key points:

Lateral flow tests

1. Relatively cheap do not require laboratories, and provide results rapidly
2. Most appropriate for widespread community testing
3. Generate more false negatives used as a test of infection rather than infectiousness vs. PCR
4. Susceptible to sampling quality
5. Narrow window to detect infectious cases
6. Less likely to detect post-infectious shedders vs. PCR

Mass testing

2. Tremendous logistical challenges
3. A focused approach to community testing such as that adopted in Liverpool, UK, may help (SMART – see figure to the right)
4. Identifies cases that would have gone undetected
5. Advantages of identifying infectious cases early to break transmission chains must be balanced against any potential false reassurance given to people with false negative results, and the risks that confirmed cases do not isolate

Testing to improve cluster identification, outbreak response, testing to protect, release and enable

3. In high-risk settings for outbreaks (workplaces, care homes, schools, universities, prisons, and hospitals), repeated and frequent lateral flow tests improve outbreak containment
4. Robust communication that a negative lateral flow result does not mean “not infectious” is essential, and symptomatic people must continue to isolate
5. Outbreak response and community testing can also be improved using mobile or pop-up PCR or LamPORE laboratories
6. Repeated testing in high infection risk settings can protect people who are either clinically vulnerable or vulnerable to infection (and transmission)
7. Any test-to-release policy must account for the incubation period, mitigate the risks of premature return or hazardous behaviours, and be shown to be cost effective
8. Test-to-enable policies seek to lift the current restrictions on social contact that are causing wider public health and economic harms in a way that minimises risk – disadvantaged areas harmed more by control measures could get most benefit from locally sensitive responses

Country-level determinants of the severity of the first global wave of the COVID-19 pandemic

This ecological study of publicly available data (from European Centre for Disease Control) aimed to identify the country-level determinants of the severity of the first wave of the COVID-19 pandemic. The outcome was the steepness of the ascending curve of country-specific daily reports of COVID-19 related deaths between 31 December 2019 and 8 June 2020.

The following is the list of determinants considered:

• Demographics
• Economic determinants
• Environmental determinants
• Prevalent comorbidities
• Health system determinants
• International arrivals
• Stringency index
• Exposure to UV radiation
• BCG vaccination coverage
• Testing capacity

The following showed a significant relationship with mortality in univariate analyses: population size, international arrivals, neoplasms, hypertension, GDP per capita and BCG vaccination coverage.

On multivariable adjustment, international arrivals in 2018 (a marker of global connection) was significantly associated with an increase in mortality rate. In addition, country-level BCG vaccination coverage was associated with decreases in COVID-19 mortality rate during the first wave of the pandemic.

Finally, in the secondary analyses including only countries with available testing capacity data, the total number of COVID-19 tests performed per 1000 population until 8 June 2020 was also associated with increases in COVID-19 mortality. (Caution, residual confounders may influence this finding, as may be expected in an ecological study.)

Modelling

Direct and Indirect Impacts of COVID-19 on Excess Deaths and Morbidity: November 2020 Update

Department of Health and Social Care, Office for National Statistics, Government Actuary’s Department, Home Office (executive summary, full paper)

This paper, presented to SAGE on 17 December 2020 and published on 29 January, updates a paper submitted to SAGE in July, discussing the direct and indirect mortality and morbidity impacts of COVID-19 on the UK population.

The paper uses an epidemiologically-derived “Winter Scenario” from 30 September 2020 to 30 March 2021, rather than attempting to model policy or interventions. This means that the near-term figures in particular are scenario-driven and can’t be compared directly to actual outcomes in the winter months. The definition used in the paper of “excess deaths” is simply “any deaths due to the COVID-19 pandemic which would not have occurred otherwise within one year” – this is somewhat different from other measures, which typically compare experience in 2020-21 to a baseline based on previous years’ experience.

The paper estimates 100,000 excess deaths over October 2020 to February 2021, with 70,000 as a direct result of contracting COVID-19, 22,000 due to changes in adult social care (for example, early hospital discharge or non-COVID-19 medical emergencies), and 10,000 due to changes in emergency care (either due to reluctance to attend or changes in protocols). The overall impact on other causes is expected to be a slight reduction in deaths (due, for example, to better air quality, lower levels of infectious disease).

The paper also looks at the impact of the pandemic on quality-adjusted life years (one lost QALY equal to losing one year of life in perfect health) and at years of life lost – QALYs in particular being a measure of morbidity as well as mortality.

Since the July paper, the estimated direct negative impact of the pandemic on mortality and morbidity over the medium term has increased significantly. The indirect impact on mortality is little changed, but on morbidity has increased, principally due to the longer periods of lockdown (the previous method accounted only for a 2-month lockdown period).

The paper also looks at long-term impacts, estimating an additional 61,000 excess deaths over the coming 50 years. Some of these are expected to be due to longer waiting times for non-urgent elective care and delays in cancer diagnoses; the majority are linked to reduced income and wealth due to the pandemic induced recession.

We will produce a separate short bulletin on this paper.

Data

Excess Mortality by Country (Link)

A key debating point over the pandemic has been to compare different countries to see who has done best or worst, often in part to assess the performance of a particular approach in comparison with others.  A new paper looks at excess mortality of 77 countries, ranging from the USA and Russia to Liechtenstein.

The extract below shows a small extract of the ‘all countries’ graphic.

There are many factors that may influence the impact of the pandemic on a country (notably the age spread of the population), and the pandemic is not yet over, so material changes may occur within countries in the future. Nevertheless, this paper is of interest in comparing similar countries’ experiences to date – the range of patterns is striking at even a cursory glance.

The Full Fact Report 2021: Fighting a pandemic needs good information (link)

This report, put together by Full Fact (fullfact.org) and supported by the Nuffield Foundation (link) uses the pandemic as a case study to show how collecting and communicating good information can be of benefit both to individuals and to society as a whole. It looks at information collection (for example, relating to the spread of the pandemic in care homes, but also on supply and demand of PPE), interpretation (early confusion over mortality data, misleading comparisons of COVID-19 with other diseases, lack of transparency over data sources and in particular exaggeration of performance against targets) and communication (a major challenge being regular changes to rules and guidance, as well as the issuance of conflicting advice).

ONS Infections and Antibody Surveys

This week, as well as the usual Friday infection survey from the ONS, we had the first of a new more frequent fortnightly update on antibody prevalence.

In the infection survey (link) there’s good news in England (1.55%) and Northern Ireland (1.56%) with noticeable falls in infection levels. In contrast, Wales (1.40%) and Scotland (0.92%) remain more stable, although note that Scotland is at a much lower level than the other three nations, and has remained very stable throughout the second wave.

The antibody survey (link) shows a significant jump in prevalence since the last survey – itself not  particularly surprising when one considers the high infection levels we’ve seen since November. The overall level has increased from 10.7% to 15.3%. London has the highest regional prevalence, as it has done for the majority of the surveys since they began last summer.

The most interesting aspect of the survey however is the age distribution, a new addition. This shows that the over-80 age group has the highest level of antibodies, at over 25%. The report notes that the surveillance will now be picking up those with antibodies from vaccination, as well as infection. This is one ONS measure we look forward to increasing dramatically in future reports!

R Estimate

SAGE’s latest estimate of R is put at 0.7 to 1.0 for the UK and 0.7 to 0.9 for England. This is very consistent with our latest view as measured by hospital admissions in England, which is 0.8. However, we expect that the vaccination roll out will soon reduce admissions disproportionately to any reduction in R, as they are heavily weighted towards the older population who are being vaccinated first. Accordingly, we will shortly discontinue this monitoring.

And finally …

Is it a bird? Is it a plane?

Each year, the Scots celebrate Burns Night on 25 January where the life and works of poet Robert Burns are remembered. This year, Scottish butcher Simon Howie took this celebration to new heights. Here you can see the launch of the very first haggis into space, which flew for two hours, 37 minutes, and covered a distance of 52 miles. It has now been safely transported back to company HQ, where it will be preserved for years to come as the first haggis in space.

[1] Swabs are tested for three genes present in the coronavirus: N protein, S protein and ORF1ab. Each swab can have any one, any two or all three genes detected. Positives are those where one or more of these genes is detected in the swab other than tests that are only positive on the S-gene, which is not considered a reliable indicator of the virus if found on its own.

The new UK variant of COVID-19 has genetic changes in the S-gene. This means the S-gene is no longer detected in the current test, and cases that would have previously been positive on all three genes are now positive only on the ORF1ab and the N-gene (not the S-gene).