Which comes first: developing accurate diagnostic testing or accurate measures of death rates? Severity of COVID-19 fatality rates has been much debated in recent weeks with many calling for relaxing lockdown measures to re-start the economy. Antibody surveys from institutions in Germany (University of Bonn) and the Netherland (Sanquin blood bank) were among the first to suggest that more people had developed antibodies to SARS-CoV-19, leading to an earlier relief of lockdown in each country.[1] In late March, Stanford University suggested that a greater population had been infected and already developed antibodies in Santa Clara city (recruited via Facebook ads) and that COVID-19 was less fatal than originally thought.[2] Since the methodology was critiqued for not accounting for factors such as bias in population sample, an adapted study was conducted later by the University of Southern California (USC) and Stanford on a Los Angeles county. Subsequently, on 20 April 2020, the Los Angeles County Department of Public Health worked with USC to understand COVID-19 in the city and released preliminary results pointing to a much larger infection pool that developed antibodies, in other words, a lower fatality rate.[4]
The USC/LA County media release largely pointed to fatality rates always being dependent on how much accurate testing is being conducted. More accurate diagnostic testing is being approved or allowed for emergency use by the FDA, with 32 commercial kits as of 23 April 2020 having some sort of FDA oversight with qualification, such as access only to research institutions. However, for a month or so, there have been concerns over quality and false positives.[5] Reports of faulty commercial kits from Asia abounded in late April, such as Spain buying 640,000 rapid testing kits from un-authorised companies in China and South Korea.[6] The UK’s National Scientific Advisory Panel with lab work by the University of Oxford conducted a study on nine commercial kits and found them faulty on sensitivity.[7] The Association of Public Health Laboratories (APHL) in the United States has begun to release weekly surveys assessing the capacity of labs in the country to test accurately for SARS-CoV-2.[8] In a comment to Financial Times, chief executive Severin Schwan of Swiss pharma house Roche stated: ‘It is very easy to develop [antibody tests] but you have to validate them with enough patients….There are 150 companies producing them already, mostly from Asia. I just get mad, it’s so unethical…They are not only useless but can be damaging.’[9] Financial Times reports that now established pharma such as Roche and US Abbott are bringing serology tests to market. Accurate Diagnostic Laboratories and RUCDR Infinite Biologics are also working on saliva-based tests.[10]
In light of many concerned members of the public measuring the effectiveness of their government’s public health response in terms of meeting testing targets, it is worth reflecting on how the diagnostic industry has evolved since AIDS. Some of these companies I have never heard of, APHL Lab excel sheet here[11], and I need to investigate further why testing kits were not developed by pharmaceutical houses. Some event, maybe financing and a more concentrated focus on R&D by established pharmaceutical houses, might have split off where this commercial activity emerges from. Or perhaps, the timescale of COVID-19 responses are much more rapid than for HIV. I talked about Burroughs Wellcome’s Retrovir in an earlier post. As you can see below, it is a branch of the company, Wellcome Diagnostics, that had developed a testing kit for HIV-1 and HIV-2:
Advertisement for Wellcozyme Assay produced by Murex Diagnostics detecting HIV-1 and HIV-2 antibodies[12]
The rapidity of COVID-19 response in contrast to HIV might explain why accurate testing and commercial kits availability were not sorted earlier than epidemiological forecasting. Earlier in the COVID-19 crisis, there was much more focus on accuracy of death counts. The University of California, Berkeley used official Italian mortality data to suggest that the fatality rate was higher than those predicted by those two universities for COVID-19 and for general statistics of seasonal flu.[13] The Berkeley study, not yet peer-reviewed, aimed to understand true mortality using Italy’s case between Case Fatality Rates and Infection Fatality Rates.[14]
The global public health response was relatively more procedural in dealing with HIV testing. WHO in 1985 before it even had a Global Programme on AIDS established a network of AIDS Collaborating Centres with universities or labs in countries around the world with more advanced AIDS epidemics.[15] These WHO Collaborating Centres would send blood sera samples to WHO and WHO would give small grants for them to conduct sensitivity tests on local commercial testing kits. When it came to scaling up globally, WHO contacted diagnostic product units of major pharma houses to ask if they would make the tests available at a low price or free of charge to low-income countries. Accuracy was largely sorted by commercial competition of established pharma. When the Gallo-Montagnier 1983 debate on the HIV virus was settled, diagnostic testing was able to progress much faster and by the second generation, there were much less false positives.
How has data science and epidemiological research changed between AIDS and COVID-19? How have the global flows of commercial activity and product development become much more rapid and chaotic? I have criticized WHO in my upcoming article for not considering financing of AIDS programmes in the midst of the global economic recession of the 1980s. But the agency’s quiet achievements in biomedical and epidemiological aspects of disease response must be acknowledged.
[1] Gretchen Vogel, ‘Antibody surveys suggesting vast undercount of coronavirus infections may be unreliable’, Science, 21 April 2020. <https://www.sciencemag.org/news/2020/04/antibody-surveys-suggesting-vast-undercount-coronavirus-infections-may-be-unreliable#>. [2] ‘Is the Coronavirus as Deadly as They Say?’, Stanford Freeman Spogli Institute for International Studies, 25 March 2020. <https://fsi.stanford.edu/news/coronavirus-deadly-they-say>. [3] [4] While the results are preliminary, LA Public Health/USC joint research will be fascinating to follow in terms of connecting university-data science with local-level public health governance. Leigh Hopper, ‘Early antibody testing suggests COVID-19 infections in L.A. County greatly exceed documented cases’, <https://news.usc.edu/168987/antibody-testing-results-covid-19-infections-los-angeles-county/>. Interview with Dr Neeraj Sood, ‘Media Availability for Seroprevalence Study’, 21 April 2020, Los Angeles County Public Health YouTube, <https://youtu.be/C_jXKcp4Zyg>. [5] Kayla Matthews, ‘In-depth: 32 FDA-Approved COVID-19 Testing Kits’, 23 April 2020. <https://hitconsultant.net/2020/04/23/in-depth-32-fda-approved-covid-19-testing-kits/#.Xrp-QJNKg6g>. [6] Sam Jones, ‘Coronavirus test kits withdrawn in Spain over poor accuracy rate’, The Guardian, 27 March 2020. <https://www.theguardian.com/world/2020/mar/27/coronavirus-test-kits-withdrawn-spain-poor-accuracy-rate>. [7] Clive Cookson, Hannah Kuchler and Andrew Jack, ‘Quest for accurate antibody tests in fight against Covid-19’, 23 April 2020, Financial Times. <https://www.ft.com/content/a93e6b28-3778-4089-8d80-4e6775996aea>; Emily Adams, Mark Ainsworth, Rekha Anand, et. al., ‘Antibody testing for COVID-19: A report from the National COVID Scientific Advisory Panel’, medRxiv, not yet peer-reviewed, 15 April 2020. <https://www.medrxiv.org/content/10.1101/2020.04.15.20066407v2>. [8] APHL Week 4 Data Report on Lab Capacity for (4 May 2020). <https://www.aphl.org/programs/preparedness/Crisis-Management/Documents/COVID-19-Survey-Dashboard-Week4-5-04-2020.pdf>. [9] Severin Schwan (Roche executive) qtd. in interview with Clive Cookson, Hannah Kuchler and Andrew Jack, ‘Quest for accurate antibody tests in fight against Covid-19’, 23 April 2020, Financial Times. <https://www.ft.com/content/a93e6b28-3778-4089-8d80-4e6775996aea>. [10] ‘Accurate Diagnostic to launch antibody test for Covid-19 diagnostics’, Medical Device Network, 28 April 2020. <https://www.medicaldevice-network.com/news/accurate-diagnostic-antibody-test-covid-19/>. [11] <https://docs.google.com/spreadsheets/d/14rq10q0VkAIBh0Z_J0-PqLE-y8g4okwrhzt7_JwNhX4/edit#gid=983663582>. [12] This image is taken from Wellcome Collections, which has a digital collection on medical ephemera throughout history available with no copyright. <https://wellcomecollection.org/works/yw8arp2s/items?canvas=1&langCode=eng&sierraId=b28673554>. [13] Robert Sanders, ‘Study challenges reports of low fatality rate for COVID-19’, 24 April 2020, Berkeley News. <https://news.berkeley.edu/2020/04/24/study-challenges-reports-of-low-fatality-rate-for-covid-19/>. [14] Chirag Modi, Vanessa Böhm, Simone Ferraro, et. al., ‘Total COVID-19 Mortality in Italy: Excess Mortality and Age Dependence through Time-Series Analysis’, medRxiv, not yet peer reviewed, 15 April 2020. <https://www.medrxiv.org/content/10.1101/2020.04.15.20067074v2.full.pdf> [15] This then evolved into the WHO AIDS Reagent Project, which contributed to understanding HIV viral strains, essential for vaccine development. See Michael Merson and Stephen Inrig’s The AIDS Pandemic: Searching for a Global Response (Springer, 2018) for an insider’s view on WHO’s early global AIDS response.