Thread by @mgmgomes1

#epitwitter Just heard from JTB that our #COVID19 paper was the most downloaded in last 90 days. I plan to spend summer developing projects that build on same concepts and methods. Happy to accommodate collaborations with those interested. Just send note.
journals.elsevier.com/journal-of-the…

Broader scope viewpoints article outlining effects and inferences of unobserved variation in disease dynamics just updated on arXiv: arxiv.org/abs/2009.01354

Selection on individual variation in susceptibility or exposure to causative factors of disease (contagious or not) introduces biases when not considered in analysis.

Common errors include the following:

1- Overestimated R0 for endemic diseases, and overpredicted intervention impacts;
2- Overpredicted herd immunity thresholds (and epidemic final sizes for epidemic diseases, and overestimated intervention effects;
3- Overestimation of individual reinfection rates;

4- Overestimation of waning immunity.

Given that individual variation cannot be observed in its entirety we propose a method for its inference from patterns of disease. The method relies on "unfolding selection" along gradients.

It's been applied to malaria, TB and covid to infer human variation, to dengue and viruses of flies to infer insect variation and Wolbachia effects, to viruses of fish to infer fish variation and vaccine effects, and is being applied to human pathogens to infer vaccine efficacy.

The paper above provides references, but hopefully many more on these and other effects (beyond disease dynamics) will emerge over years to come.