Smart Real-time Data vs. Models
Surveillance and Early Warning Systems > Models Using Bad Data and Speculative Assumptions
Low Cost Insurance > High Cost in Human Lives / Economic Damage
Manageable Healthcare Logistics > Healthcare Delivery Disaster
In war against deadly pathogens, planning is critical and correct interpretation of past (descriptive vs. real time) data helps. Yet everybody has a plan until you get punched in the mouth. Case precedent helps design better protocols yet the next pathogen is different. Beware of planning to fight the last war.
The solution is NOT better models with low quality / suspect data and unreasonable / speculative assumptions.
Pathogen phenomena at both individual and collective human levels is the bailiwick of complexity science. In high causal density, complex environments models are not appropriate for real world decision making considering unlimited freedom of model specifications and parameters.
All the King's data scientists and all the King's models will be unable to help decision makers make optimal policy decisions. Moreover, models provide wild best and worst case scenarios that:
1. Hurt policy decision making, and
2. Scare the public.
The solution is better real-time smart data.
An early warning system will allow fast collective action to minimize pathogen and economic damage. Once a deadly pathogen hits with exponential transmission rates the best policy response is immediate transmission suppression to avoid massive downside risk. The challenge is to: collect smart real-time data to turn a healthcare delivery disaster into a manageable logistics challenge; optimally balance the competing interests of pathogen suppression/mitigation and economic/social sustainability; and allow near real-time policy calibration.
For example, the delayed global response to the 2020 pathogen caused higher transmission, infection and mortality rates resulting in a healthcare delivery problem in hot spots that overwhelmed some healthcare systems. Early detection and actions with applied data science turns this problem into a manageable logistics solution. Yet we need better real-time data to measure and manage.
Today nation/states have varied criteria and standards for collecting and classifying data. The result is low quality and veracity data and flawed data interpretation resulting in suboptimal policy choices.
Therefore, we need to design a worldwide and nation/state/local pathogen surveillance system to collect and share the right smart data in near real-time to size up and optimally manage pathogen with minimum loss of life, social disruption and economic damage. Each sovereign also needs a surveillance system to detect, isolate and cure infected citizens to protect public safety.
Today we have the technology to engineer a reasonably good early warning and surveillance system at global, nation/state and local levels. Applied data science is developing into a profession and the study of complexity science holds great promise.
The trick will be obtaining global agreement on standardizing data collection (compare apples to apples) and designing legal architecture and procedures to protect privacy, civil liberties and human rights against a compelling public safety/health interest.
To correctly interpret, data science requires precise uniform definitions and uniform data collection standards. In addition, policy makers need to update decision making protocols to use real-time data instead of models.
I suggest the cost to build this system is in reality a cheap insurance policy compared to the huge cost in human lives and economic damage using the current model based architecture. It is our public and moral duty to better prepare for future deadly pathogens with potential existential implications.