A New Model to Assist Evaluation of COVID-19 Impact

Written by Benjamin T Solomon BSc (Eng), MA (Operations Research), MBS (Finance) on June 3, 2020. Posted in Current News

Benjamin T Solomon submits for wider consideration his paper, ‘A Generic Population Spread (GPS) Model Using US COVID-19 Data And Understanding Its Impact.’ We have pleasure in publishing the study in its entirety below. Feedback in the comments section is most welcome.

Summary

1. COVID-19 did not cause the coming recession. The economy will recover within 2 Quarters of suspending Shelter-In-Place policies. And about -3.0{154653b9ea5f83bbbf00f55de12e21cba2da5b4b158a426ee0e27ae0c1b44117} of the 1Q-2020 GDP contraction can be attributed to COVID-19.
2. The modeling show that there are 3 probability distributions that one has to determine to effectively manage public health, infectability, mortality and recovery. From this the 5-stages of an infectious disease and 6-triggers and 6-treatment strategies are determined.
3. Shelter-In-Place was misguided and should have been lifted in late April 2020 as herd immunity was arrived at in late April 2020. This has implications for the need for ultra rapid vaccine development.
4. The authoritative COVID-19 disease spread modeling is shown to have serious errors. This has implications in all fields of study as these models have been peer reviewed.

(A) Introduction

This paper is written to clear the misinformation and conflicting reporting about COVID-19 disease spread, not from a medical perspective but from a statistical perspective. The purpose of this article is to provide (i) A means of formulating an informed opinion about future disease spread. (ii) To understand at which stage in the disease lifecycle is a treatment effective. (iii) From this be able to evaluate its impact on a pharmaceutical company’s revenue and profitability. (iv) Would this fit, likely lead to an alpha investment? (v) Could an economic shutdown be a viable approach to disease spread management? And (vi) if so, should one reallocate assets for an imminent shut-down, per your personal risk profiles?

The primary objective of this paper is to characterize an infectious disease using statistical methods to rapidly develop health & treatment policies to counter the disease spread early. Maybe as early as 15 days into a pandemic.

COVID-19 has shown us that it is important to rapidly characterize an infectious disease spread using statistical methods, to develop health & treatment policies early in the disease spread. This paper explains why the authoritative models (reproduction models such as that of the Washington University’s Institute of Health Metrics & Evaluation model) used to manage health policy are dead wrong. Though the mathematics of these models are accepted on a peer-reviewed consensus basis to be valid, due to implicit assumptions the statistical implementation is wrong, and cannot be corrected. In the author’s opinion our (US) health policies were driven by misinformation (from relatively unfriendly countries) and therefore, the US is in urgent need of a substantially upgraded domestic health policy management system to avoid future economic shutdowns and ineffective quarantine policies.

The term context structure is used as in many cases many models are required to construct a context structure that fits the characteristics of the context structure. Models are best fits for the data and are only as good as the data available. The data is only as good as the test accuracy (whether the tests are accurate or provide substantial false positives or false negatives). It is interaction between the models that gives the context structure its power to provide deep insights into disease spread, i.e. context structure tells much more than the data alone could.

Note, that the Generic Population Spread (GPS) model presented here does not belong to any of the three know¹ model types, metapopulation model, cellular automaton model or gravity model.

(B) Structure of Context

The term context structure is used as in many cases many models are required to construct a context structure that fits the characteristics of the narrative. Models are best fits for the data and are only as good as the data available. The data is only as good as the test accuracy (whether the tests are accurate or provide substantial false positives or false negatives). It is the interaction between the models that gives the context structure its power to provide deep insights into disease spread. It will be shown that this context structure tells substantially more than the data alone could.

Note, that the Generic Population Spread (GPS) model presented here does not belong to any of the three know¹ model types, metapopulation model, cellular automaton model or gravity model.

Fig. 1: COVID-19 Generic Population Spread Context Structure

Fig. 1 shows several characteristics of this context structure,

1. Population Incubation Period (PIP), the period the pathogen goes unnoticed in the population until it reaches a critical level, the Ground Zero Population.
2. Ground Zero Population (GFP), is the number of known infected before CDC issues pandemic advisories.
3. Mortality Lag (ML), the time between onset of infection and death. The average ML in the early stages of the US COVID-19 pandemic is 14 days and concurs with Chinese findings^9,10. This modeling assumes an ML of 14 days. This is determined using minimized error sum of squares between new infected and new deaths when shifting the new infected model to the timeline of the new deaths model.
4. Increasing ML suggests that the pathogen has culled the weaker members of the population, resulting in a healthier (genetically modified?) population from the perspective of the population’s immunity, however, this is no consolation for those who lost their loved ones.
5. Known Infected, persons who test positive for the pathogen, provided the tests do not provide substantial false positives or negatives.
6. Known Dead, persons who have died from COVID-19 or comorbidities.

(C) Model Induced Heteroscedasticity of IHME and Similar Models

This methodology discussion included sources (1999-2020) from IHME’s website, National Review Institute, Epidemics, Nature, Nature Communications, International Encyclopedia of Statistical Science, SeekingAlpha.com and Real Estate Finance: The Quarterly Review of Commercial Finance Techniques.

The IHME model is essentially a variation of current epidemiological theory² (1),

(D) Implicit Axioms in Regression with Transformations

Ordinary Least Squares Regression with or with transformations is a type of mathematical programming technique defined by its objective function (least squares) for the data domain space of the original data. Transforming the data transforms this data domain space. The theoretical foundations for regression, therefore, has at least three implicit axioms that introduce optimization technique induced heteroscedasticity,

• There is no useful information in the model errors. Clearly this is incorrect. (7) shows that canceling the errors using the running cumulative on a statistic of interest introduces indeterminate biases in the model.
• It does not matter whether the mean of the transformed statistic of interest, when converted back to the original statistic, has a mean that is substantially different from the data. This is clearly incorrect, shown by (6), as minimizing errors is not based on the mean of the data but on the mean of the transformed data.
• Transformed model errors though Normally distributed, does not affect the how the original statistic of interest is modeled. This is clearly incorrect, as shown by (4) that converting back to the original statistic introduces an undesirable relationship with the transformed model errors.

One could counter that standard regression techniques are not used to build the disease specific IHME model, as IHME’s CurveFit³ uses alternative optimization search techniques. That may be true but the three implicit axioms above, must still hold in the negative for these alternative optimization search techniques. That is, there are restrictions as to how much one can transform the original data domain space. Therefore, getting extremely high correlations are indicative of either true model fit or breakdown in the statistical techniques used.

(E) Generic Population Spread (GPS) Basics

The GPS model presented in this paper is different. It is based on the axiom that infectious diseases can be characterize by 3 probability distributions, Individual Infectiousness Profile, Individual Mortality Profile and Individual Recovery Profile from a minimum set of 2 data types consisting of new infected, and new deaths. To determine the 3 individual probability distributions (profiles), the starting distributions was obtained from papers published in the New England Journal of Medicine, The Lancet, Journal of Virology and American Journal of Epidemiology between January and March 2020.

This GPS model uses a new class of solutions, Collated Distributions⁶ (these require specific rules to be obeyed), which was discovered and published by the author in 2020 even though a rudimentary version was first pioneered⁷ in the financial services industry for mortgage backed securities in 1999 (about 20-years ago).

(F) Generic Population Spread (GPS) Model

This GPS model is based on the first 42 days (2/23/2020 to 4/5/2020) of the COVID-19 pandemic data¹¹ (see Fig. 3). The primary new infections and new deaths data was taken from the Canadian source 1Point3Acres.com (also used by the CDC and John Hopkins) as they have a team verifying each data point which is currently an ongoing data collection process. This would facilitate a future study comparing US versus Canadian COVID-19 spreads and would not introduce additional data collection biases. The GPS model is defined by four equations.

To develop the context structure, the data and models were implemented in MS Excel, Word & PowerPoint 365. To determine big picture context structure of infectious disease spread, the disease lifecycle, the final version required more than 110 iterations of the models and write-ups to arrive at a model that works effectively.  In the search for a global minimum error sum of squares, 5-different objective functions (cums, growth ratios, weighted or some combination) were used with MS Excel 365’s Solver.

Early versions of the model used a geometric model which was eventually replaced with collated distributions of the individual profiles. With collated distributions it turned out that the testing with the original data set, a straight-forward error sum of squares provided the best results. To arrive at the individual probability distributions, published probability distributions were used as starter distributions. Using these starter distributions, MS Excel 365 Solver was used to optimize each individual distribution against the original data set of new cases and new deaths.

The Relative Mortality Rate R_RM was decreasing (see Fig. 4) until large companies, counties and states implemented Shelter-In-Place policies. Since then Relative Mortality Rate R_RM continues to increase. This concurs with the Infection Acquisition Rate R_IA observations. Together, these imply that the disease, having established a strong foot hold in the US population during its Population Incubation Period, is running its course despite Shelter-In-Place policies.

(H) US Shelter-In-Place

In the US public health management is a State & Local government prerogative and therefore, the news present is of state and local governments. To document US Shelter-In-Place, established news sources (March-April 2020) used included The Wall Street Journal, National Public Radio, The Washington Post, The New York Times, Fox News and The San Francisco Chronicle. The timeline of public health policies is not good. See Fig. 4.

1. 1^st month (February 24 – March 23) of COVID-19 pandemic:
   1. March 13, 2020: Large companies¹⁶ institute remote work prior to this date.
   2. March 16, 2020: California’s Six Bay Area Counties¹⁷ announced Shelter-In-Place.
   3. March 19, 2020: California Governor Gavin Newsom announces¹⁹ Stay-At-Home orders.
2. 2^nd month (March 24 – April 23) of COVID-19 pandemic:
   1. March 24, 2020: New York’s Governor¹⁸, Andrew Cuomo, announced Stay-At-Home orders. Other states followed soon after.
   2. April 9, 2020: IHME is publicly criticized for providing erroneous forecasts.
   3. April 13, 2020: CDC announces²⁰ Stay-At-Home guidelines, but this lateness may not be entirely the CDC’s fault^{21, 22}.

The implicit assumption with Shelter-In-Place policies is that the infectious disease is still in its early stages of spreading. Therefore, shutting down the economy is a feasible and viable approach to infectious disease management.

However, Shelter-In-Place is only effective if the non-infectious asymptomatic phase is substantially shorter than the duration of the Shelter-In-Place policies. Fig.5 shows that this non-infectious asymptomatic phase is 11 days. This is based on the new infected data and does not account for infection spread during the Population Incubation Period and therefore a lower bound. Therefore, if the true asymptomatic phase is substantially longer, say 180 days, then Shelter-In-Place is a misplaced strategy.

Fig. 4 shows that Shelter-In-Place was too little, too late, as the disease had gained a solid foothold in the country. Therefore, shutting down the economy in response to COVID-19 is not an effective public health policy. In fact, the R_IA and R_RM rate reversal and right shift of Fig. 3, or increased infections and mortalities, match the model under reporting of Fig. 4. This strongly suggests that Shelter-In-Place was the incorrect public health policy for COVID-19.

It is evident that during the Population Incubation Period (PIP), the spreaders¹² (high contact persons) of COVID-19 had spread the disease silently²² before any US agency was prepared to confront the disease. This is because of the duration of the stealthiness, the 7-day asymptomatic stage, which was followed by a much longer duration infectious period of 17 days. See Fig. 5.

Therefore, current health policy early warning systems need to be substantially upgraded to detect an infectious disease while the disease is still in its Population Incubation Period, and the disease-critical spreader job types.

Since Shelter-In-Place policies are not effective, it is proposed, all other factors being equal, that a robust exponential Mortality Rate is a good estimate of when COVID-19 disease exited the Population Incubation Period (PIP) by specific country. This explains the reported severity of the disease within a country. Note, this exit date is not the same as the date of arrival of first infected into a country. Table 2 shows the estimated PIP exit dates for sample countries¹⁵.

(I) The 5-Phases of the COVID-19 Disease Lifecycle and Respective Triggers

The modeling results in Fig. 5 (includes the data for the starter distributions used) shows that one can construct the US COVID-19 disease lifecycle. It is 76-days long. A physical process can be subdivided into concurrent or overlapping phases that are turned on or off by event triggers. Based on the statistical modeling analyses, the proposed 5-phases of the COVID-19 disease and triggers are as follows:

1. Phase 1, Non-Infectious Incubation Period (day 1 to 7): This duration starts with Trigger 1 (Person Infected) and ends with Trigger 2 (Pathogen Recognition).
2. Phase 2, Infectious Incubation Period (day 8 to 10): This duration starts with Trigger 2 (Pathogen Recognition) and ends with Trigger 3 (Immune System Activated).
3. Phase 3, Infectious Period (day 8 to 25): This duration starts with Trigger 2 (Pathogen Recognition) and ends with Trigger 5 (Infectiousness Process Terminated).
4. Phase 4, Recovery Period (day 11 to 33): This duration starts with Trigger 3 (Immune System Activated) and ends with Trigger 6 (Immune System Overwhelmed).
5. Phase 5, Mortality Period (day 22 to 76): This duration starts with Trigger 6 (Immune System Overwhelmed) until end of host life anytime between day 22 and 76.

The question that immediately comes to mind is, why is the infectious period early in the disease lifecycle, and not much later?

From a pathogen’s perspective, this infectious period should be late in the disease lifecycle as (i) the time is approaching for the pathogen to find its next host (before the host dies), (ii) whether or not the pathogen is successful in its current host, there would be a lot more pathogens available to infect a future host before day 22 when death is a real probability.

The literature does not discuss infectiousness as an evolutionary feature. Examples point definitively to pathogen evolution²³ as a major factor involved in disease emergence. However, much of the literature on infectious disease evolution²⁴ is focused on the pathogen side of this relationship.

This necessitates an alternative evolutionary perspective, that infectiousness is a host population’s strategy to inform other population members of the new pathogen. If infectiousness were late in the disease lifecycle, it would not be a productive signaling method as the host may have overcome the pathogen and there would no longer be any remaining pathogens to transmit to other members.

At the same time, transmitting live pathogens to other members is Nature’s method of culling the immunologically weaker members and thereby leaving a surviving population that is immunologically stronger. Thus, Trigger 2 (Pathogen Recognition) that starts the infectiousness period is earlier than Trigger 3 (Immune System Activated) when the immune systems starts the pathogen management and recovery process.

The implicit assumption here, is that naturally occurring pathogens cannot entirely eliminate their natural hosts otherwise both populations could not exist.

From a statistical perspective, the 4-stages of US COVID-19 Known Infected (see Fig. 6), the stratified quarantine strategy, and the respective triggers marking the transition to the next stage are:

1. Asymptomatic/Mildly Ill (day 1 & 11): Persons who acquires the pathogen on day 1 (trigger 1), but the host’s body has not recognized the pathogen. This time-period concurs with the published data²⁵. In effect carriers are asymptomatics who somehow never trigger immune system activation (trigger 3). Therefore, asymptomatics (test positive for the virus and negative for antibodies) are infected persons who have recently started the disease lifecycle journey and should be quarantined for at most 25 days until the infectious period is over.
2. Seriously Ill (between day 11 & 21): Immune system activated (trigger 3) to fight off the pathogen but there is no probability of death. The probability of recovering while in this phase is 40.94{154653b9ea5f83bbbf00f55de12e21cba2da5b4b158a426ee0e27ae0c1b44117}. These persons should be quarantined for at most 14 days until the infectious period is over.
3. Critically ill (between day 22 & 32): Immune system is burdened (trigger 4) in its ability to deal with the pathogen and secondary infections. There now exists an 35.94{154653b9ea5f83bbbf00f55de12e21cba2da5b4b158a426ee0e27ae0c1b44117} probability of death. These persons need to be quarantined for at most 3 days.
4. Terminally ill (between day 33 & 76): The immune system is overwhelmed (trigger 6) with no possibility of recovery and an 64.06{154653b9ea5f83bbbf00f55de12e21cba2da5b4b158a426ee0e27ae0c1b44117} probability of death. These persons need not be quarantined as they are no longer infectious. This is the realm of experimental drugs.

(K) The 6-Treatment Strategies for the US COVID-19 Disease

From a statistical perspective, the 3 distributions that comprise the US COVID-19 disease lifecycle (see Fig. 6) provide 6 treatment strategies:

1. Type 1 Treatment: Vaccination (prior to day 1). To prevent onset of infectious period and subsequent disease by introducing antigens (safe vaccinations) to prepare the host’s immune system.
2. Type 2 Treatment: Non-Infectious Disease (day 12 & 25). To turn-off the infectious period. Assuming that infectiousness is a host response, by intervening in the host’s infectiousness process by some yet to be determined treatment.
3. Type 3 Treatment: Pathogen Recognition (day 11 & 32). To shift the Individual Recovery Profile’s mode to the left. The two strategies here are,
   1. Cause the host’s immune system to recognize the new pathogen sooner rather than later. There are two reasons:
      1. The incubation period is the time it takes for enough pathogens to reach a critical level such that the host recognizes the pathogen’s presence and later activates its immune system.
      2. Handicap the pathogen’s ability to inflict damage as pathogen population has not reached the critical level it needs to overwhelm the host.
   2. Build a robust immune system. This is about immune fitness as opposed to physical fitness.
4. Type 4 Treatment: Accelerate Recovery (day 23 & 32). To skew the Individual Recovery Profile’s right tail to the left by a treatment that directly attacks the pathogen.
5. Type 5: Buying Time (day 23 & 32). To skew the Individual Recovery Profile’s right tail further to the right to give time for the host’s immune system and/or other treatments to overcome the pathogen and other secondary conditions that may be taxing the immune system. This is the lowering of the patient’s risk from terminally ill to critically ill. The use of respirators/ventilators are good example of Type 5 treatments.
6. Type 6: Non-Fatal Disease (day 22 & 76). Shift the Individual’s Mortality Profile to the right, thereby precluding the possibility of death. This type of treatment reduces the pathogen’s ability to inflict substantial damage on the patient.

(L) Evaluating Drug Strategies

Note that I am not a medical professional and the classification presented here is a best guess. The use of these treatments is not in any way an affirmation or denial of treatment effectiveness. The reader should consult a medical professional to confirm or modify these categories.  A sample list of treatments being considered have been categorized by treatment type:

1. Chloroquine²⁶: Is a Type 5 Treatment as it is a disease-modifying anti-rheumatic drug, and it regulates the activity of the immune system.
2. Chloroquine/Hydrooxychloroquine²⁶: Is a Type 4 Treatment as it prevents virus particles from using their activity for fusion and entry into the cell.
3. Convalescent Plasma (CP) Therapy²⁶: Is a Type 3 Treatment as it causes the white blood cells to recognize the pathogen.
4. Remdesivir²⁷: Is a Type 4 Treatment as it stops the replication mechanism of the coronavirus.
5. Lopinavir and Ritonavir²⁸: Is a Type 5 Treatment, as it may decrease your chance of developing acquired immunodeficiency syndrome (AIDS) and HIV-related illnesses such as serious infections or cancer.
6. Type 1 interferons²⁹: Is a step towards a Type 3 Treatment as it part of the body’s mechanism to recognize viral components.
7. Anti-Binding³⁰: Is a Type 1 Treatment to develop vaccines, essentially by preventing the coronavirus from binding to the human cell and thus prevent its replication.

One notes that the majority of the COVID-19 treatment types are 3 & 4. Obviously Type 1 Treatments have the largest market share of all the treatment types at 100{154653b9ea5f83bbbf00f55de12e21cba2da5b4b158a426ee0e27ae0c1b44117} of the population. Type 3 & 4 Treatments have much smaller market shares as the estimated infected is about 1.3{154653b9ea5f83bbbf00f55de12e21cba2da5b4b158a426ee0e27ae0c1b44117}³¹ to 6{154653b9ea5f83bbbf00f55de12e21cba2da5b4b158a426ee0e27ae0c1b44117}³² of the population or a median of 3.65{154653b9ea5f83bbbf00f55de12e21cba2da5b4b158a426ee0e27ae0c1b44117}.

However, since (per this article) COVID-19 has reached a natural herd immunity (not by the use of vaccines) it is doubtful that a COVID-19 vaccine would benefit the population, and thus the case for voluntary as opposed to mandatory vaccinations. Voluntary vaccinations would reduce its market share.

Finally, with a mortality rate below 0.1{154653b9ea5f83bbbf00f55de12e21cba2da5b4b158a426ee0e27ae0c1b44117} when community or herd immunity has occurred, it is doubtful that Type 6 Treatments would be financially viable without public funding as the only 0.00365{154653b9ea5f83bbbf00f55de12e21cba2da5b4b158a426ee0e27ae0c1b44117} (=0.01{154653b9ea5f83bbbf00f55de12e21cba2da5b4b158a426ee0e27ae0c1b44117} x 3.65{154653b9ea5f83bbbf00f55de12e21cba2da5b4b158a426ee0e27ae0c1b44117}) of the population, but this would require further analysis.

The medical industry uses Cost Effectiveness Analysis (CEA)³⁴, which can get very sophisticated³⁵, and is probably beyond the scope of most investors who do not specialize in the pharmaceutical industry. Therefore, the methodology presented here is a shortcut to estimating market share of a drug treatment, and its profitability by using generic ratios³⁶ provided by the Department of Veterans Affairs. This methodology also shows how drug treatments are segmented by their drug strategies and driven by competition between treatment types and within treatment types.

(M) A Case for Evolution

As this GPS model is about population spread further research is required to determine how suitable this model would be to modeling species evolution and is provided here as an example of how this GPS model can be used in other contexts. As a possible start the GPS model adaptations would include:

1. Timeline: Instead of days or year, the timeline would be expressed in generations.
2. Species Incubation Period: Equivalent to the Population Incubation Period. This is the number of generations it would take for a new species to evolve from its original species, given that the new set of mutations hold.
3. Generational Mutation Profile: Equivalent to the Individual Infectious Profile but shifted left into the Species Incubation Period. It describes the probability of the new mutation set holding.
4. Ground Zero Population: This is the minimum number of new species required before a species can be recognized as a new species.
5. Generational Reproductive Profile: Equivalent to the Individual Recovery Profile and determines the probability of each generation’s producing one or more young in total.
6. Generational Mortality Profile: Equivalent to the Individual Mortality Profile and determines the probability of each generation’s life expectancy.
7. Reversion: Equivalent to Vaccination. If the new species mutation set does not take hold the following generations revert to the original species type.

(N) Impact to the US Economy

The Fig. 7 legend for the left y-axis is as follows, 1 if GDP growth when negative, 1.05 represents 5-year treasury bond (YTB), 1.1 represents 10-year treasury bonds, 1.2 represents 20-year treasury bonds, and 1.3 represents 30-year treasury bond, when any of these treasury bond yields are below the 3-month treasury bonds (MTB) yields.

Continuing from an earlier article³³ Fig. 7 show the yield curve inversion, that a recession is guaranteed when all 4 treasury bond yields invert (drops below the 3-Month Treasury Bond yields) with three qualifications:

1. Testing shows that in the case of the 5-year bonds, this inversion only implies that GDP growth will dip below 0.25{154653b9ea5f83bbbf00f55de12e21cba2da5b4b158a426ee0e27ae0c1b44117} at some point in the future, but by itself does not guarantee a recession.
2. A full yield curve inversion implies a GDP contraction of at least -1.10{154653b9ea5f83bbbf00f55de12e21cba2da5b4b158a426ee0e27ae0c1b44117}.
3. The absence of the full yield curve inversion does not guarantee future positive GDP growth as occurred between (a) 3/7/2014 to 4/15/2014 and (b) 3/1/2011 to 10/1/2011.

The most important observation here is that the full yield curve inverted between (5-YTB) 10/25/2019 to 1/27/2019, (10-YTB) 5/23/2019 to 2/28/2020, (20-YTB) 8/12/2019 to 9/9/2019, and (30-YTB) on 8/28/2020. This is about 6 months earlier than COVID-19. That is, COVID-19 did not cause this coming recession (per NBER’s definition). However, public health policies in response to COVID-19 substantially increased the severity of this coming recession.

Fig. 8: Negative GDP Growth Duration versus Yield Curve Inversion Duration

The good news is that the period (from date of 1^st inversion to last inversion, however, there may be some days in between when the yield curve does not invert) of the 10-YTB and 30-YTB yield curve inversions are good indicators of the negative GDP growth period. The data suggests that up to -3.0{154653b9ea5f83bbbf00f55de12e21cba2da5b4b158a426ee0e27ae0c1b44117} of the -4.8{154653b9ea5f83bbbf00f55de12e21cba2da5b4b158a426ee0e27ae0c1b44117} 1Q-2020 GDP contraction could be attributed to Shelter-In-Place policies. That is, the economy will recover quickly (about 2 Quarters) once all Shelter-In-Place policies are lifted, provided that the yield curve does not invert again in the near future.

(O) Conclusion

The GPS model has shown that it is possible to statistically determine the disease lifecycle and health policies, treatment strategies and evaluate financial success of new treatment types, from only new cases and new deaths. The model provided should lead to a better understanding and parametrization of generic population spreads and therefore, make informed opinions (i) how public health management should be conducted and (ii) which drug maker is likely to succeed from a new treatment strategy.

References:

1. Caroline E.Waltersa1, Margaux M.I.Meslébc, Ian M.Halla, Modelling the global spread of diseases: A review of current practice and capability, Epidemics, Volume 25, December 2018 https://www.sciencedirect.com/science/article/pii/S1755436517301135
2. Mark E.J. Woolhouse, Daniel T. Haydon, Rustom Antia, Emerging pathogens: the epidemiology and evolution of species jumps, TRENDS in Ecology and Evolution Vol.20 No.5 May 2005
   https://www.cell.com/action/showPdf?pii=S0169-5347{154653b9ea5f83bbbf00f55de12e21cba2da5b4b158a426ee0e27ae0c1b44117}2805{154653b9ea5f83bbbf00f55de12e21cba2da5b4b158a426ee0e27ae0c1b44117}2900038-8
3. IHME Staff, CurveFit, https://ihmeuw-msca.github.io/CurveFit/methods/#statistical-model, accessed 04/23/2020.
4. Andrew C. McCarthy, COVID-19 Projection Models Are Proving to Be Unreliable, National Review Institute, 9 April 2020. https://www.nationalreview.com/corner/coronavirus-pandemic-projection-models-proving-unreliable/
5. Bucevska V. (2011) Heteroscedasticity. In: Lovric M. (eds) International Encyclopedia of Statistical Science. Springer, Berlin, Heidelberg https://link.springer.com/referenceworkentry/10.1007{154653b9ea5f83bbbf00f55de12e21cba2da5b4b158a426ee0e27ae0c1b44117}2F978-3-642-04898-2_628#toc
6. Benjamin Solomon, Solomon’s Method For Collated Distributions Used In Mortgage-Backed Securities, SeekingAlpha.com, April 20, 2020. https://seekingalpha.com/article/4338509-solomons-method-for-collated-distributions-used-in-mortgage-backed-securities
7. Esaki, L’Heureux & Snyderman, Commercial Mortgage Update, Real Estate Finance, The Quarterly Review of Commercial Finance Techniques, Spring 1999, Vol. 16, No. 1.
8. Bill Gates, The scientific advances we need to stop COVID-19, Gates Notes, April 23, 2020
   https://www.linkedin.com/pulse/scientific-advances-we-need-stop-covid-19-bill-gates/
9. Robert Verity, Lucy C Okell, Ilaria Dorigatti, Peter Winskill, Charles Whittaker, Natsuko Imai, Gina Cuomo-Dannenburg, Hayley Thompson, Patrick G T Walker, Han Fu, Amy Dighe, Jamie T Griffin, Marc Baguelin, Sangeeta Bhatia, Adhiratha Boonyasiri, Anne Cori, Zulma Cucunubá, Rich FitzJohn, Katy Gaythorpe, Will Green, Arran Hamlet, Wes Hinsley, Daniel Laydon, Gemma Nedjati-Gilani, Prof Steven Riley, Sabine van Elsland, Erik Volz, Haowei Wang, Yuanrong Wang, Xiaoyue Xi, Prof Christl A Donnelly, Prof Azra C Ghani, Prof Neil M Ferguson, Estimates of the severity of coronavirus disease 2019: a model-based analysis, The Lancet, Infectious Disease, March 30, 2020. https://www.thelancet.com/pdfs/journals/laninf/PIIS1473-3099(20)30243-7.pdf
10. Qun Li, M.Med., Xuhua Guan, Ph.D., Peng Wu, Ph.D., Xiaoye Wang, M.P.H., Lei Zhou, M.Med., Yeqing Tong, Ph.D., Ruiqi Ren, M.Med., Kathy S.M. Leung, Ph.D., Eric H.Y. Lau, Ph.D., Jessica Y. Wong, Ph.D., Xuesen Xing, Ph.D., Nijuan Xiang, M.Med., Yang Wu, M.Sc., Chao Li, M.P.H., Qi Chen, M.Sc., Dan Li, M.P.H., Tian Liu, B.Med., Jing Zhao, M.Sc., Man Liu, M.Sc., Wenxiao Tu, M.Med., Chuding Chen, M.Sc., Lianmei Jin, M.Med., Rui Yang, M.Med., Qi Wang, M.P.H., Suhua Zhou, M.Med., Rui Wang, M.D., Hui Liu, M.Med., Yinbo Luo, M.Sc., Yuan Liu, M.Med., Ge Shao, B.Med., Huan Li, M.P.H., Zhongfa Tao, M.P.H., Yang Yang, M.Med., Zhiqiang Deng, M.Med., Boxi Liu, M.P.H., Zhitao Ma, M.Med., Yanping Zhang, M.Med., Guoqing Shi, M.P.H., Tommy T.Y. Lam, Ph.D., Joseph T. Wu, Ph.D., George F. Gao, D.Phil., Benjamin J. Cowling, Ph.D., Bo Yang, M.Sc., Gabriel M. Leung, M.D., and Zijian Feng, M.Med., Early Transmission Dynamics in Wuhan, China, of Novel Coronavirus–Infected Pneumonia, The New England Journal of Medicine, January 29, 2020. https://www.nejm.org/doi/full/10.1056/NEJMoa2001316
11. com, https://coronavirus.1point3acres.com/en
12. Gabriel E. Leventhal, Alison L. Hill, Martin A. Nowak & Sebastian Bonhoeffer, Evolution and emergence of infectious diseases in theoretical and real-world networks, Nature Communications, January 16 2015. https://www.nature.com/articles/ncomms7101
13. Weier Wang, Jianming Tang,  Fangqiang Wei, Updated understanding of the outbreak of 2019 novel coronavirus (2019‐nCoV) in Wuhan, China, Journal of Medical Virology, 29 January 2020.
    https://onlinelibrary.wiley.com/doi/full/10.1002/jmv.25689?af=R
14. C. Ghani, C. A. Donnelly, D. R. Cox, J. T. Griffin, C. Fraser, T. H. Lam, L. M. Ho, W. S. Chan, R. M. Anderson, A. J. Hedley, G. M. Leung, Methods for Estimating the Case Fatality Ratio for a Novel, Emerging Infectious Disease, American Journal of Epidemiology, Volume 162, Issue 5, 1 September 2005. https://academic.oup.com/aje/article/162/5/479/82647
15. Maps & Trends: Mortality Analysis, Corona Virus Resource Center, John Hopkins University & Medicine, https://coronavirus.jhu.edu/data/mortality, Accessed 05/06/2020.
16. Alex Hern, Covid-19 could cause permanent shift towards home working, The Guardian, 03/13/2020, https://www.theguardian.com/technology/2020/mar/13/covid-19-could-cause-permanent-shift-towards-home-working
17. Erin Allday, Bay Area orders ‘shelter in place,’ only essential businesses open in 6 counties, San Francisco Chronicle, 16 March 2020. https://www.sfchronicle.com/local-politics/article/Bay-Area-must-shelter-in-place-Only-15135014.php
18. Kwame Opam, It’s Not ‘Shelter in Place’: What the New Coronavirus Restrictions Mean, The New York Times, 24 March, 2020. https://www.nytimes.com/article/what-is-shelter-in-place-coronavirus.html
19. Louis Casiano, California Gov. Gavin Newsom announces statewide coronavirus ‘stay at home’ order, Fox News, 19 March, 2020. https://www.foxnews.com/us/california-gov-gavin-newsom-announces-statewide-stay-at-home-order
20. William Wan, Philip Bump, What U.S. leaders say affects whether Americans stay at home, CDC data suggests, The Washington Post, 13 April, 2020. https://www.washingtonpost.com/health/2020/04/13/coronavirus-stay-home-orders-data/
21. Rebecca Ballhaus, Stephanie Armour, Health Chief’s Early Missteps Set Back Coronavirus Response, The Wall Street Journal, 22 April, 2020. https://www.wsj.com/articles/health-chiefs-early-missteps-set-back-coronavirus-response-11587570514?mod=hp_lead_pos7
22. Lauren Sommer, Why The Warning That Coronavirus Was On The Move In U.S. Cities Came So Late, National Public Radio, 24 April, 2020. https://www.npr.org/sections/health-shots/2020/04/24/842025982/why-the-warning-that-coronavirus-was-on-the-move-in-u-s-cities-came-so-late
23. Sonia Altizer, Drew Harvell and Elizabeth Friedle, Rapid evolutionary dynamics and disease threats to biodiversity, TRENDS in Ecology and Evolution Vol.18 No.11 November 2003.
    http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.513.5502&rep=rep1&type=pdf
24. Rustom Antia, Roland R. Regoes, Jacob C. Koella & Carl T. Bergstrom, The role of evolution in the emergence of infectious diseases, Nature, 11 December 2003. https://www.nature.com/articles/nature02104
25. Stephen A. Lauer, Kyra H. Grantz, Qifang Bi, Forrest K. Jones, Qulu Zheng, Hannah R. Meredith, Andrew S. Azman, Nicholas G. Reich, PhD; Justin Lessler, The Incubation Period of Coronavirus Disease 2019 (COVID-19) From Publicly Reported Confirmed Cases: Estimation and Application, Annals of Internal Medicine, 10 March 2020. https://annals.org/aim/fullarticle/2762808/incubation-period-coronavirus-disease-2019-covid-19-from-publicly-reported
26. Mahalaxmi Iyer, Kaavya Jayaramayya, Mohana Dev Subramaniam, Soo Bin Lee, Ahmed Abda Dayem, Ssang-Goo Cho & Balachandar Vellingiri, COVID-19: an update on diagnostic and therapeutic approaches, BMB Reports, April 30, 2020. http://www.bmbreports.org/journal/view.html?volume=53&number=4&spage=191
27. Victoria Rees, Mechanism of action revealed for remdesivir, potential coronavirus drug, Drug Target Review, March 3, 2020. https://www.drugtargetreview.com/news/56798/mechanism-of-action-revealed-for-remdesivir-potential-coronavirus-drug/
28. Lopinavir and Ritonavir, Medline Plus, accessed May 14, 2020. https://medlineplus.gov/druginfo/meds/a602015.html
29. Erwan Sallard, François-Xavier Lescure, Yazdan Yazdanpanah, France Mentre, Nathan Peiffer-Smadjab, Type 1 interferons as a potential treatment against COVID-19, Antiviral Research, Jun 2020. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7138382/
30. COVID-19 (coronavirus) vaccine: Get the facts, Mayo Clinic, Accessed May 14, 2020. https://www.mayoclinic.org/diseases-conditions/coronavirus/in-depth/coronavirus-vaccine/art-20484859
31. Anirban Basu, Estimating The Infection Fatality Rate Among Symptomatic COVID-19 Cases In The United States, Health Affairs, May 7, 2020. https://www.healthaffairs.org/doi/full/10.1377/hlthaff.2020.00455
32. Jacob Sullum, Official COVID-19 Numbers Represent Just 6{154653b9ea5f83bbbf00f55de12e21cba2da5b4b158a426ee0e27ae0c1b44117} of Total Infections, a New Analysis Suggests, Reason, April 12, 2020. https://reason.com/2020/04/12/official-covid-19-numbers-represent-just-6-of-total-infections-a-new-analysis-suggests/
33. Benjamin Solomon, Is The Recession Here?, SeekingAlpha.com, December 27, 2018. https://seekingalpha.com/article/4230396-is-recession
34. Cost-Effectiveness Analysis (CEA), Center for Disease Control, Accessed 05/21/2020 https://www.cdc.gov/policy/polaris/economics/cost-effectiveness.html
35. Thomas Gift, Cathleen Walsh, Anne Haddix, Kathleeen Irwin, A Cost-Effectiveness Evaluation of Testing and Treatment of Chlamydia trachomatis Infection Among Asymptomatic Women Infected With Neisseria gonorrhoeae, Journal of the American Sexually Transmitted Diseases, September 2002 – Volume 29 – Issue 9 – p 542-551. https://journals.lww.com/stdjournal/Fulltext/2002/09000/A_Cost_Effectiveness_Evaluation_of_Testing_and.9.aspx
36. Determining the Cost of Pharmaceuticals for a Cost-Effectiveness Analysis, Department of Veteran Affairs, Accessed 05/21/2020. https://www.herc.research.va.gov/include/page.asp?id=pharmaceutical-costs

About the author:Benjamin T. Solomon, BSc, DipOR, MAOR, MBS. Benjamin is President & CEO at QuantumRisk LLC and Principal Investigator at iSETI LLC. He recently completed a 12-year study into the theoretical and technological feasibility of gravity modification. He is the author of the book An Introduction to Gravity Modification: A Guide to Using Laithwaite’s and Podkletnov’s Experiments and the Physics of Forces for Empirical Results.

PRINCIPIA SCIENTIFIC INTERNATIONAL, legally registered in the UK as a company incorporated for charitable purposes. Head Office: 27 Old Gloucester Street, London WC1N 3AX. 

Please DONATE TODAY To Help Our Non-Profit Mission To Defend The Scientific Method.

Trackback from your site.