The Imperial College COVID-19 Model Predicted a 98% Reduction in COVID-19 Deaths WITHOUT Closing Non-Essential Businesses

When Oklahoma Governor Kevin Stitt announced on March 24 that he was issuing an order closing all non-essential businesses in 19 counties (which he subsequently extended to all 77 counties), my heart sank.  I knew of no evidence supporting such a drastic move, and I was skeptical that any such evidence actually existed.  In other words, I was skeptical that such a move would significantly reduce the death toll from COVID-19 and I was certain that the fallout (in terms of actual harm caused to individual citizens) would be dramatic.

As a very technical person who remains inherently skeptical of claims not supported by solid evidence and/or well-reasoned arguments, I took it upon myself to review a plethora of reports and data analyses regarding the spread of the deadly SARS-CoV2 virus (the virus that causes COVID-19).  In that respect, I strongly prefer going to the source documents written by acclaimed ‘experts’ rather than merely relying upon interpretations (or interpretations of interpretations) presented by journalists.

As a direct result of my aforementioned skepticism, I have closely reviewed the COVID-19 data analyses and reports produced by three distinguished teams of researchers – the Imperial College in London, Oxford University, and the University of Washington.

In a subsequent blog post, I will discuss in some detail the research and reports produced by each of the aforementioned teams of scholars.  However, for this post, I want to hone in on the Imperial College model and subsequent report, because it is the one that predicted the most dire potential outcomes (and thus started the ball rolling with respect to media hysteria and extreme political reactions).  The Imperial College model predicted 2.2 million deaths in the U.S. if we were to ‘do nothing’ to prevent the spread of SARS-CoV2.  The media, of course, picked up on that prediction of 2.2 million deaths and used it to garner the rapt attention of citizens and politicians alike (there is no better motivator than fear coupled with anxiety – forget FDR’s quip about having nothing to fear but fear itself).

However, what the media seem to have lost sight of (or perhaps never even looked into at all) is the fact that the Imperial College epidemiologists provided a prescriptive plan for reducing the number of COVID-19 deaths, by up to 98% (i.e. from 2.2 million U.S. deaths to less than 50,000).  And, equally important to the potential reduction in deaths, the Imperial College plan detailed a path for reducing COVID-19 deaths (from cataclysmic levels to ‘normal flu season’ levels) WITHOUT shuttering ANY businesses (other than schools).  Their comprehensive COVID-19 suppression plan (the one they predicted would reduce deaths by 98%) involved four specific ‘social isolation’ strategies: case isolation (CI), household quarantine (HQ), social distancing (SD), and school closures (PC).  More specifically, the above strategies can be grouped into three distinct responses:

• case isolation at home (CI) and household quarantine (HQ) (if you have symptoms, everyone in your household stays at home for 14 days, i.e. no direct contact with anyone outside the household, with the exception of medical professionals, as needed for testing and/or treatment),
• social distancing of the entire population (SD) (if no one in your household has symptoms, you can still go to work, but you are expected to reduce social contact by 25% at work and by 75% everywhere else), and
• closure of schools and universities (PC) (government closes all schools, 25% of universities remain open).

Another extremely important aspect of the Imperial College COVID-19 suppression plan (that has been completely ignored by the media and politicians and just about everyone else) is that the strategies were designed to be applied LOCALLY or REGIONALLY, based on triggers tied to the # of empty ICU beds available within a given community.  

“Given local epidemics are not perfectly synchronised, local policies are also more efficient and can achieve comparable levels of suppression to national policies while being in force for a slightly smaller proportion of the time.” (italics added, p. 15)

Also, the Imperial College epidemiologists analyzed the effect of stopping mass gatherings and concluded that stopping mass gatherings “is predicted to have relatively little impact … because the contact-time at such events is relatively small compared to the time spent at home, in schools or workplaces and in other community locations” (italics added, p. 8).  In other words, the experts who predicted up to 2.2 million deaths in the U.S. due to COVID-19 also presented plausible analyses that those deaths could be reduced by 98% WITHOUT closing ANY businesses and WITHOUT forcing Christians to violate the scriptural admonition against “giving up meeting together” (Hebrews 10:25, NIV).  It is a shame that, on their holiest day of the year, Christians in the U.S. were prohibited from gathering together for corporate worship, even though the experts who provided the most dire predictions also predicted that those dire outcomes could be avoided, and that they could be avoided without requiring Christians to abandon their scripturally-mandated commitment to regularly gather together for worship, prayer, and mutual encouragement.

Lastly, it bears noting that, although the Imperial College epidemiologists presented and analyzed strategies that are far less draconian than those that are currently in place (both in Oklahoma and across the nation), they did suggest that such measures (which should be triggered on and off based on ICU bed usage), may need to be in place for up to 12 or even 18 months (i.e. until a vaccine or effective treatment is developed).  Personally, I have no doubt that strictly following the Imperial College’s COVID-19 suppression plan for 12 to 18 months would be FAR less disruptive to individual lives and the national economy than even one month (much less 2 or more) of the economy-shattering, business-shuttering, job-killing measures that are currently in place.

I will be following up in a day or two with a post also examining the Oxford College model and the University of Washington model.  All three take a very different approach.  Each has something to offer.  None warrant the kind of heavy-handed government responses we have seen these past few weeks.

UPDATE #3 (4/14/20):  After preparing Update #2, I noticed two additional anomalies with Table 3 and Table A1.  THESE ANOMALIES SERIOUSLY CALL INTO QUESTION THE ENTIRETY OF TABLES 3 & A1. First, in all instances in Tables 3 & A1, increasing the trigger threshold results in an IMPROVEMENT in both the number of deaths and the number of ICU beds required.  However, increasing "cumulative ICU case count triggers" should mean a SLOWER response which should result in more need for ICU beds and more deaths, but the tables show a decrease.  Second, and more importantly, when comparing Table A1 (p. 20) to Table 4 (p. 13), the number of deaths associated with CI_HQ_SD in Table A1 (for R0=2.4) averages 499,000, but the number of deaths for CI_HQ_SD in Table 4 (for R0=2.4) averages only 91,000 (for the exact same combination of NPIs).  EITHER THERE IS A SERIOUS FLAW IN THE UNDERLYING MODEL, OR THE MODEL RESULTS WERE INCORRECTLY TABULATED IN TABLES 3 & A1 OR TABLE 4, OR BOTH.

UPDATE #2 (4/14/20):  After preparing Update #1, I noticed another 'anomaly' with the Imperial College model results.  In Tables 3 & A1, the model predicts (in the UK) 328,000 deaths under the implementation of CI_HQ.  However, when SD is added to the model (widespread social distancing), the predicted number of deaths climbs to 422,000 (an increase of 94,000).  By contrast, when SDOL70 is added (instead of SD), the predicted number of deaths drops to 247,000 (a reduction of 82,000).  This pretty much validates what I have been thinking from the beginning: WE NEED TO FOCUS THE MAJORITY OF OUR EFFORTS AND ENERGY ON ISOLATING AND PROTECTING THE MOST VULNERABLE MEMBERS OF OUR SOCIETY.  The models assume a 50% increase in at-home contact when SD is in place (compared to only 25% increase with SDOL70).  This increased interaction in the home with family members over 70 years of age dramatically increases the risks to our elderly population (who are by far the most vulnerable, and most susceptible to dying from COVID-19).

UPDATE (4/14/20):  I forgot to include in the original post the fact that the Imperial College model, in one of their simulations, actually predicted that widespread school closures would INCREASE the number of COVID-19 deaths.  They do not point this out explicitly (i.e. it is buried in the results shown in Table 3, p. 9, and Table A1, p. 20).  In particular, the 'Table 3' and 'Table A1' version of the model compares CI_HQ_SDOL70 (case isolation, household quarantine, social distancing of those over 70) to PC_CI_HQ_SDOL70 (same, but with the addition of school closures).  The addition of 'school closures' (to that specific combination of NPIs) results in the model predicting 110,000 MORE COVID-19 DEATHS, in the UK, compared to those same NPIs being implemented without closing schools.  The reason for this dramatic difference is unclear; however the Imperial College researchers acknowledge (on pp. 1-2) that closing schools can be expected to create unintended negative consequences.

Steve Trost is Associate Director of the Institute for the Study of Free Enterprise and can be contacted at trost@okstate.edu.  He has a bachelor’s degree in engineering from MIT, a master’s degree and PhD in engineering from Oklahoma State University and a PhD in entrepreneurship (also from OSU).

Follow Dr. Trost on twitter: @TrostParadox

Disclaimer: All comments, observations, and statements presented herein represent the opinions of the author and in no way reflect the views of Oklahoma State University or the Institute for the Study of Free Enterprise.