Written evidence submitted by Junade Ali (CLL0007)

 

This evidence is made in submission to the "Coronavirus: lessons learnt" inquiry which is jointly held by the Health and Social Care Committee and the Science and Technology Committee. This evidence specifically focuses on the use of death rate statistics, successes on vaccines and treatments, failures on non-pharmaceutical interventions and contact tracing apps.

 

The use of Statistics

Before the Science and Technology Committee, Professor Chris Whitty stated the metric for international comparisons "should be all-cause mortality, age adjusted, and possibly seasonally adjusted, over the course of the epidemic, and we are not there yet" (Science and Technology Committee 2020). I agree with this statement.

 

International comparisons presented by the Government so far have focussed on the use of “COVID-19 deaths”. Due to different counting techniques, “COVID-19 deaths” is a challenging metric for COVID-19 international comparisons. The University of Oxford’s Centre for Evidence Based Medicine has previously highlighted such challenges in pan-UK context (Loke and Heneghan 2020).

 

“Excess deaths” is another flawed metric that has been used to attempt international comparisons. Instead of understanding the real number of deaths (per million population), “excess deaths” compares a country to its own historical performance and then compares that excess to other countries. This neglects the fact that countries with more effective healthcare systems may have more vulnerable members of the population or that “excess deaths” can be less noticeable in lower mortality countries.

 

Whilst the pandemic certainly isn’t over yet, I would like to paint the picture of how all-cause mortality looks so far. Using 2020 death data to Week 35 (Eurostat 2020a) and population data (Eurostat 2020b), we are able to calculate the overall death rate per million people. For convenience, I’ve split these deaths into non COVID-19 and COVID-19 deaths using the data from Our World in Data (as reported by national governments) from the 30th August 2020 (end of Week 35) to help our understanding (Our World in Data 2020).

 

Eurostat data is not yet complete; countries like Italy and Ireland haven’t yet reported deaths to Week 35 to Eurostat, and this is done without the age-adjustment or seasonal-adjustment that Chris Whitty alluded to, but provides the best picture of the pandemic so far. This chart is shown in Figure 1.

Fig 1: 2020 All-Cause Deaths per Million divided by COVID and non-COVID deaths

 

This chart shows a startling image that is seldom seen; the UK all-cause mortality rate is about near the middle of the pack in Europe. The real risk of death in countries like the UK and Sweden in reality remains far lower than in countries like Portugal or Germany.

 

Some have previously attempted to claim that death data shows that the multi-payer health care system in Germany is superior to the universal “free at the point of use” NHS system in use in the UK. The reality is that to Week 35 2020, Germany had a death rate over 9% higher per head of population than the UK. Germany has both higher 5-year average death rates and 2020 death rates than the UK. This highlights just how much is missing when we look purely at “excess deaths” and “COVID-19 deaths”.

 

The lowest death rates in countries being in countries with high GDP per capita (like Iceland, Liechtenstein, Luxembourg and Norway) perhaps highlights how essential economic well-being is to lowering national death rates.

 

Whilst the doomsday picture of “COVID-19 deaths” or “excess deaths” metrics may well heighten the level of fear in the population, it may also fuel distrust if people feel such metrics have been hidden from them.

 

Sadly, this failure is not only confined to the Government itself. The Office for National Statistics published a report claiming to be a comparison of "all-cause mortality between European countries and regions" (Office for National Statistics 2020); unfortunately the report only contains “excess deaths” metrics, not “all-cause mortality” as the title of the report would imply. The raw data contained a note that they had superimposed their own data on the Eurostat data so it was "subject to caveats in relation to the temporal comparability". Whilst the ONS report claimed to compare 20 European countries, 5 countries in the ONS data set had missing data for the weeks they compared data up to. Countries with a higher all-cause mortality than the UK, like Germany and Italy, were amongst those with missing all-cause mortality data.

 

This provides the picture on pan-European mortality data as things stand. I make no claim as to the impact of how non-pharmaceutical interventions may have prevented the growth in such death rates, but I will contrast the evaluation of scientific evidence on pharmaceutical and non-pharmaceutical interventions next.

 

Success on Finding Vaccines and Treatments

It is unquestionable that the UK has been a world leader in the development of treatments and vaccines, using robust scientific evidence. At the time of writing, dexamethasone remains the first and only life-saving COVID-19 drug - found by the University of Oxford’s RECOVERY trial run in NHS hospitals. It is estimated that 650,000 lives will be saved globally by the discovery of the course of 6 months from discovery (Aguas et al. 2020), over 15 times the UK death toll.

 

Further such discoveries remain in the pipeline; the RECOVERY trial continues as the largest trial of repurposed drugs as COVID-19 treatments in the world, Oxford University’s ChAdOx1 nCoV-19 vaccine was first globally to Phase III randomised control trials and Synairgen Plc’s SNG001 treatment was shown in Phase II trials by the University of Southampton (in NHS hospitals) to reduce the severity of disease by 79% (Synairgen Plc 2020).

 

There have no doubt been plenty of ancillary measures to achieve such a breakthrough; a wide pool of expertise in the NHS to identify good candidate drugs, the power of the Government to secure drug supply early by banning hoarding and parallel export using The Human Medicines (Amendment etc.) (EU Exit) Regulations 2019 and central Government funding for the trials.

 

Peter Horby and Martin Landray serve as the Co-Chief Investigators of the RECOVERY trial and Horby chairs the Government’s NERVTAG committee. Whilst this may have played a role in the Government’s action, we should note the lack of diversity of thought on the Government’s SAGE advisory committee. On the public list of SAGE members there are 29 members affiliated with universities, but 22 are from "Golden Triangle" universities (Oxford, Cambridge and London-based universities).

 

As a slight digression; it is perhaps important to remember where the three Oxford professors driving forward advances in vaccines and treatments are trained. Martin Landray did both his undergraduate degree and PhD at the University of Birmingham. Sarah Gilbert, driving the Oxford/AstraZeneca vaccine candidate, did her undergraduate degree at East Anglia University and her PhD at the University of Hull. Peter Horby is the only one of these three who was educated in a London-based university and the only one holding a seat on SAGE; indeed he has previously tweeted that he was only educated at LSHTM after got B, C and D grades in his A Levels, not meeting his lowest university offer (Horby 2020). The enormous contribution made by these three Professors highlights the benefits in listening to a wide range of experience and I hope that in the future the Government or Parliament will request Her Majesty to confer some signal mark of Her Royal favour upon them.

 

Nevertheless; the driving factor behind these successes has been the confidence to pursue robust scientific proof instead of relying on poor observational data. In other countries, cocktails of “promising drugs” were given to patients early on in the pandemic - causing harm to patients in many cases and leading to retracted non-randomised observational studies like Mehra et al. 2020. In the UK, full confidence was given to the Randomised Control Trial approach, with all four CMOs wrote to NHS doctors to “strongly discourage the use of off-licence treatments outside of a trial, where participation in a trial is possible” (Whitty 2020).

 

This brings us to the fundamental difference in approaches to pharmaceutical and non-pharmaceutical research. Figure 2 shows the evidence-based medicine pyramid; quality of evidence represented by each level of the pyramid, with quality increasing from bottom to top. For example; Randomised Control Trials are of significantly higher quality than Expert Opinion, but there are far less Randomised Control Trials than Expert Opinions.

 

Fig. 2: Evidence-based medicine pyramid (Golden and Bass 2013)

Failure in Research on non-Pharmaceutical Interventions

 

This brings us to the fundamental failures in identifying and finding non-pharmaceutical interventions. When a scientist encounters uncertainty, they ordinarily work to do the research work to resolve such uncertainties.

 

Masks are one such area where we’ve failed to resolve uncertainty by performing trials, causing opinions to divide towards the extremes (Jefferson and Heneghan 2020). Another example is Contact Tracing apps; we have modelling claiming it fails to reduce the r rate below 1 (Kucharski et al. 2020) but observational historical data from the Isle of Man showing it may be effective (Kendall et al. 2020). This is especially tragic given the ease of performing Randomised Control Trials on mobile apps.

 

This also highlights a key difference between engineers and scientists. When scientists find uncertainty, they use research to find the gap in knowledge. When engineers identify risk, they seek to mitigate the impact of that risk and communicate clearly the balance of risk and benefit (Engineering Council UK, n.d.).

 

Contact Tracing Apps

 

To avoid making an already long evidence submission any longer, I shall be brief here. For the most part, engineers have made a quiet but essential contribution to the pandemic. For example; at the start of the pandemic, the Royal Academy of Engineering working with Professional Engineering Institutions provided volunteers to help establish the Nightingale hospitals. Nevertheless, it is not hard to feel that the role of engineers has been somewhat sidelined by scientists (even when operating directly in the area of knowledge of engineers). I understand that some of these frustrations have been already vocalised before the Science and Technology Committee (Hebard 2020).

 

Contact Tracing apps are one area that could have benefited from improved engineering oversight. Whilst the National Cyber Security Centre did seek to communicate on cybersecurity advice, industry engagement seemed more lacking on the side of NHSx alongside developers of the underlying protocols at Apple and Google. Low risk de-anonymization attacks have been identified within the Apple/Google contact tracing protocols (Ali and Dyo 2020) and their real-world effectiveness continues to be questioned; effective industry communication would have allowed such issues to either be fully resolved or there to be effective public communication on the risk/reward of such engineering challenges.

 

Like the UK has a Chief Medical Officer and Chief Scientific Advisor, it may help in future for there to be a Chief Scientist (perhaps even part of the proposed ARPA research agency). This would allow there to be someone accountable for working alongside the Government’s scientific infrastructure to help manage the risk of engineering decisions and be able to communicate risk/reward decisions. This would also allow for scientific research to be transferred to effectively implemented solutions.

 

About Junade Ali

 

Junade Ali is a British computer scientist whose work on anonymous communication protocols has changed industry practice on password security and has helped lay the foundations for the Contact Tracing protocols developed for the COVID-19 pandemic. He is a Chartered Engineer via the Institution of Engineering and Technology and his professional experience includes developing software for embedded systems used in mission critical road safety applications and cyber security. He is currently studying Contact Tracing protocols as part of a PhD in anonymous communication protocols. He holds a Masters degree in Computer Science, with his thesis focussing on applications of Operations Research and data science to road traffic data collection.

 

Bibliography

Aguas, Ricardo, Adam Mahdi, Rima Shretta, Peter Horby, Martin Landray, and Lisa White. 2020. “The potential health and economic impact of dexamethasone treatment for patients with COVID-19.” medRxiv, (July). https://doi.org/10.1101/2020.07.29.20164269.

Ali, Junade, and Vladimir Dyo. 2020. “Cross Hashing: Anonymizing encounters in Decentralised Contact Tracing Protocols.” arXiv, (October). https://arxiv.org/abs/2005.12884.

Engineering Council UK. n.d. “Guidance on Risk.” Engineering Council. Accessed October 26, 2020. https://www.engc.org.uk/standards-guidance/guidance/guidance-on-risk/.

Eurostat. 2020a. “Deaths by week and sex.” Eurostat [demo_r_mwk_ts]. https://appsso.eurostat.ec.europa.eu/nui/show.do?dataset=demo_r_mwk_ts&lang=en.

Eurostat. 2020b. “Population on 1 January.” Population on 1 January. https://ec.europa.eu/eurostat/databrowser/view/tps00001/default/table?lang=en.

Golden, Sherita H., and Eric Bass. 2013. “Validity of Meta-analysis in Diabetes: Meta-analysis Is an Indispensable Tool in Evidence Synthesis.” Diabetes care, (October). https://care.diabetesjournals.org/content/36/10/3368.

Hebard, Peter. 2020. “Written evidence submitted by Peter Hebard (C190042).” Science and Technology Committee. https://committees.parliament.uk/writtenevidence/6592/pdf/.

Horby, Peter. 2020. “Peter Horby on Twitter: "Amy, push on with confidence. I got BCD (yes D!) A level grades. Not what I needed from the lowest offer I got; but doctor I am, professor now @ Oxford."” Twitter. https://twitter.com/PeterHorby/status/1294689182632345601.

Jefferson, Tom, and Carl Heneghan. 2020. “Masking lack of evidence with politics.” University of Oxford's CEBM. https://www.cebm.net/covid-19/masking-lack-of-evidence-with-politics/.

Kendall, Michelle, Luke Milsom, Lucie Abeler-Dörner, Chris Wymant, Luca Ferretti, Mark Briers, Chris Holmes, David Bonsall, Johannes Abeler, and Christophe Fraser. 2020. “Epidemiological changes on the Isle of Wight after the launch of the NHS Test and Trace programme: a preliminary analysis.” The Lancet, (October). https://doi.org/10.1016/S2589-7500(20)30241-7.

Kucharski, Adam J., Petra Klepac, Andrew J. Conlan, Stephen M. Kissler, Maria L. Tang, Hannah Fry, Julia R. Gog, and W. J. Edmunds. 2020. “Effectiveness of isolation, testing, contact tracing, and physical distancing on reducing transmission of SARS-CoV-2 in different settings: a mathematical modelling study.” The Lancet, (June). https://doi.org/10.1016/S1473-3099(20)30457-6.

Loke, Yoon K., and Carl Heneghan. 2020. “Why no-one can ever recover from COVID-19 in England – a statistical anomaly.” Why no-one can ever recover from COVID-19 in England – a statistical anomaly - CEBM. https://www.cebm.net/covid-19/why-no-one-can-ever-recover-from-covid-19-in-england-a-statistical-anomaly/.

Mehra, Mandeep R., Sapan S. Desai, Frank Ruschitzka, and Amit N. Patel. 2020. “RETRACTED: Hydroxychloroquine or chloroquine with or without a macrolide for treatment of COVID-19: a multinational registry analysis.” The Lancet, (May). https://doi.org/10.1016/S0140-6736(20)31180-6.

Office for National Statistics. 2020. “Comparisons of all-cause mortality between European countries and regions: January to June 2020.” Comparisons of all-cause mortality between European countries and regions: January to June 2020. https://www.ons.gov.uk/peoplepopulationandcommunity/birthsdeathsandmarriages/deaths/articles/comparisonsofallcausemortalitybetweeneuropeancountriesandregions/januarytojune2020.

Our World in Data. 2020. “Data on COVID-19 (coronavirus) by Our World in Data.” Data on COVID-19 (coronavirus) by Our World in Data. https://github.com/owid/covid-19-data.

Science and Technology Committee. 2020. Oral evidence: UK Science, Research and Technology Capability and Influence in Global Disease Outbreaks, HC 136. N.p.: House of Commons. https://committees.parliament.uk/oralevidence/309/html/.

Synairgen Plc. 2020. “Synairgen announces positive results from trial of SNG001 in hospitalised COVID-19 patients.” Synairgen plc. https://www.synairgen.com/wp-content/uploads/2020/07/200720-Synairgen-announces-positive-results-from-trial-of-SNG001-in-hospitalised-COVID-19-patients.pdf.

Whitty, Chris. 2020. “Novel Coronavirus: Clinical Trials.” MHRA Central Alerting System. https://www.cas.mhra.gov.uk/ViewandAcknowledgment/ViewAlert.aspx?AlertID=103012.

 

Nov 2020