Chapter 8 Conclusion

These notes took a relatively high level perspective on modeling of data related to covid-19.

Overall, there is a sense that these models are not very satisfying. They operate at one level of analysis: counts on aggregate geographic areas such as countries and states. They don’t get into the details of the biology or specifics of human behavior or evaluation of policies, contextual factors, health disparities and other aspects that are relevant to understanding the mechansisms. Agent-based modeling and network-based modeling provide possible frameworks to develop these more sophisticated models that span multiple levels of analysis.

The SARS-CoV-2 virus itself has many levels of analysis such as molecular, cellular, and genetic as well as complex interactions that we are just beginning to understand in how the virus interacts with its environment (e.g., how it travels in air, how long it can live on a surface, how it interacts with the host). Here is a wonderful video on some relevant biology including what the SARS-CoV-2 virus does to the lungs and how in some people it causes Acute Respitory Distress Syndrome (ARDS). ARDS is a common correlate of death among individuals with covid-19, which occurs because of the massive inflammatory reaction that covid-19 patients experience.

Right before the covid-19 pandemic occurred, our lab was about to launch a study with two collaborators (critical care doctors Jack Iwashyna and Hallie Prescott) on survivors of ARDS. Patients can take up to a year to recover from ARDS, even getting up from a chair can make these survivors winded. I had been learning a lot about ARDS when the pandemic started. One key piece of information I picked up from one of our collaborators (Jack Iwashyna) is that the human lung is a poorly designed system. If we wanted to design a way to bring oxygen into our body, process it, and eliminate the waste product, CO2, we wouldn’t have a designed what is essentially a sac where good air comes in from the top and the bad air goes out the same way the good air came in. An inhalation brings O2 in, an exhalation forces CO2 out. If the “sac” fills with liquid or is affected by inflammation, weexperience major complications. There are other ways of designing such a system, such as the gills in fish: as water flows through, the gills extract O2. No sac, no bottleneck between the incoming and outgoing channels, no opportunity for the sac to fill with liquid. Unfortunately, when the pandemic is over and our planned study can resume, there will be a larger pool of ARDS survivors to recruit into our study than we originally anticipated.

Fish Gills (from [Kumar, Gandhi and Majumder](https://arxiv.org/pdf/1805.07744.pdf))

FIGURE 8.1: Fish Gills (from Kumar, Gandhi and Majumder)

I would love to see deeper models that integrate the biological, behavioral, social, economic/financial and epidemiological factors. No single level of analysis will help us figure out the best way to respond to all aspects of the current pandemic.