Massive changes could be coming to Life Sciences due to the COVID-19 Pandemic. This series of posts will review technology strategies that should be leveraged to address the anticipated changes.

Business Drivers:

• Major increase in the amount of connected data within Life Sciences
• Automated external exchange of data with private and government agencies
• New regulations like the proposed Protecting the Pharmaceutical Supply Chain
• Predictive monitoring of events, production, vendors, and commitments
• Increase demand for productivity and speed to support new regulations

Technology to be reviewed in this series

• Graphs
  • Graphs are built and named after Graph Theory in Discrete Mathematics
  • Graphs are like NO SQL databases
  • Graphs have Algorithms for predictive analysis, etc.
  • Graphs are the base for the next advancement in AI technologies
  • Neo4j is the Graph that will be used in this series
• Artificial Intelligence (AI)
  • Natural Language Processing (NLP)
  • Machine Learning (ML)
  • Neural Networks (NN)
  • Robotics Process Automation (RPA)

COVID-19 Use Case

This series will use the COVID-19 use case to illustrate Graph technology benefits because it is timely, widely understood and covers a variety of issues within the Life Sciences industry:

• New virus and drug discovery
• Lab Testing
• Off Label Use / Vaccines
• Medical Devices / Supplies
• Supply Chain / Crisis Management

Data Exchange Hub

Hypothesis

An international Knowledge Graph data exchange hub will improve Pandemic responses and save lives.

An international data exchange will allow 3rd parties to share, validate, certify, and regulate data across a complex ecosystem. For example, the proposed Protecting Pharmaceutical Supply Chain legislation can be implemented by proposed standard.

Top 5 High-level Requirements

Time is of the essence

As the speed at which data is collected, shared, and actioned increases, the death / infection rate decreases.

For example, data shows that taking the following actions sooner rather than later — closing borders, locking down cities, enacting social distancing, providing non-pharmaceutical based interventions, encouraging the wearing of masks, and approving off label use of approved drugs — has improved outcomes during pandemics that have occurred over the past 5,0000 years:

Accurate Modeling / Simulations

The IHME model in the US appears to have been drastically wrong (~400%) although universally used by most government agencies. Accurate, predictive models and simulations are critical in decision making. Thus, these models need to be improved so that everyone can trust the data. If the airline industry provided a similar rate of success as that demonstrated by the IHME model, only 1 in 4 planes would be counted on to arrive safely at its destination. This would not be considered a viable mode of transportation by most of the population, if the risk profile was universally understood to be this high.

Bottom up approach

History shows that the general population rises to meet adversity. Thus, the data and process should be based on trusting the public.

Face Masks

It appears that the initial communication to the populace about the average person not needing to wear a mask were made based on the government’s fear of citizens hording masks, thereby rendering the masks unavailable to healthcare workers.

When the CDC finally admitted that masks help reduce the spread of the virus by reducing the droplets sprayed by carriers into common area and posted this update on the CDC website, the sudden profusion of masks was remarkable. Literally, the next day, I saw all kinds of homemade masks in the local markets. Of course, people realized that these homemade masks were not going to absolutely prevent people from getting the virus, but wearing a mask could reduce the spread, lower the viral load. It provided anxious people with another manner of protecting themselves at some level against an invisible enemy.

Retailers

If retailers had been alerted to the signs of a Pandemic, protocols could have been put in place early enough to mitigate issues with the supply chain. If retailers had been provided sufficient warning about Covid19, they could have monitored their inventories in real-time to determine what items were at risk. That would have provided the data needed to allow them to change policies to adapt to the situation, such as putting limits on purchases of items in high demand. If retailers had limited consumers to only 1 package of toilet paper before the store supplies were decimated, the panic buying and tendency toward hoarding would have been mitigated.

History Rhymes

Review of historical data related to pandemic trends could produce a model against which new and future pandemics can be analyzed. The model should be able to identify disasters with similar patterns in order to provide direction to the leaders and citizens. Early communications around COVID-19 compared the virus to the seasonal influenza virus. People wasted time explaining the differences between influenza and COVID-19. A more useful comparison via the model would produce information that shows how this virus is like another disruptive world event of any type (e.g. Pandemic Flu, Hurricane, Stock Crash, …)

Minimize Politics

The models and new standards should provide binary policy decision making indicators. A disaster either reaches a certain threshold to warrant action or it does not. There should not be political considerations in play to influence how decisions are made.

Better definitions: For example, what is a Pandemic?

Binary definitions are needed for clear actions by governments and citizens. For example, the definition and debate on when the World Health Organization (WHO) should have declared a pandemic should be eliminated. A clear definition and standard should be set.

WHO Definition:
A pandemic is the worldwide spread of a new disease.

WHO Pandemic Alert on 11-Mar-20

Why not declared on 29-Jan-20?

The point here is that the definition and thresholds should be based on a model. Real time conditions should be analyzed and compared to the model on an ongoing basis. Declaration of a pandemic should not rely on committee meetings, discussions, and the drive to reach consensus agreement amongst individuals with different agendas and potentially different data. When a hurricane hits and floods a major city, committees don’t first debate whether a hurricane has in fact struck and a city has in fact flooded. Just as in those cases of weather-based disaster, for pandemic response to be effective, time is of the essence.

29-Jan-2020	14-Feb-2020	9-Mar-2020

A group of medical professionals should come up with a better definition that is based on numbers that can be evaluated in real-time with no discussion needed.

Geneva Convention of Pandemics?

The Geneva Convention of 1949 set many of the rules by which future Wars would be measured. Specifically, on the topic of bayonets used prior to 1949, the Convention prohibits “bayonets with a serrated edge”.

This is an easy rule to implement in practical reality:

1. If there is a bayonet, check to see if it has a serrated edge.
2. If it has a serrated edge, then alert that an outlawed weapon has been found.

Using simple, science based, factual rules will increase the rate at which the world responds to a positively identified Pandemic.

There are currently allegations that China’s deception at the start or the Pandemic cost more lives globally. There have been calls in the US to make China make amends, including seizing China’s assets in the banking system by the US and allies. Currently, these calls are being made without a structure against which to assess the real cost of the delays.

Universal accurate definitions and models can help in determining damages and avoiding any excuses or ignorance. An accurate model would allow for the identification of how many lives could have been saved and the day to day cost of delays.

For example, a statistical model based on real world data could indicate how many lives would have been saved if the world knew about human to human transfer on 6-Dec-19 instead of 20-Jan-20? How much of the world’s GDP could have been protected? Could community spread have been avoided in the US if building test kits started in Dec-19?

Ideally, even if we are in a kinetic struggle with another government on other fronts, we would share Pandemic data to prevent civilian causalities.

Business Drivers

In the US, there have seen the following events in the COVID-19 Pandemic to date (April 2020)

• Lock-down: 1/2 of all humanity and 80% of Americans under stay at home orders
• ~$10 Trillion in US Aid / debt between the aid packages and US Federal Reserve actions
• Estimated 32 percent unemployment rates in the US by the Federal Reserve Bank of St. Louis
• Breakdown in Globalization: 78 countries have banned exports of goods needed in the Pandemic
• China threatens the US in denying critical supplies & APIs
• Defense production act invoked to address supply chain issues with tests, ventilators, masks, etc.

Right now, the focus is on saving lives. But when the dust settles, the world and all the IT systems will probably be changed. A quick look at similar regulations in the aftermath of major events in terms of debt to GDP.

The point is that major events generate the ‘government is here to help’ regulations.

Advisory Board

Lab Data Concepts is currently forming an advisory board to analyze and implement solutions to address these future changes and improving the efficiency of Life Sciences industry. If you would like to participate, please email [email protected].

Future Posts

Future post will focus on the approach and technology to help improve the Life Sciences industry. Please leave any comments, feedback, or suggestions.