Analysing recovery from pandemics by Learning Theory: the case of CoVid-19

Romney B. Duffey; Enrico Zio

In this paper, we have originally proposed to adapt Learning Theory for describing the reduction of pandemic infections like that of CoVid-19. A key point is to look at infection rate, as a measure of error outcome, and time, as a measure of experience/knowledge or risk exposure which allows learning. The analyses of the currently available data show that the CoVid-19 infection rate data follow, after peaking, almost exactly the Universal Learning Curve describing the decreasing trajectory of many other instances where humans learn to apply effective countermeasures. More specifically, the learning curve is nearly the same (with universal constant K~3) as for any learning experience reducing outcomes, accidents and events for other modern technological system operated by humans. We claim that this trend decline due to learning is direct evidence of learning about risk reduction, also in this case of the pandemic, and call it the Universal Recovery Curve. It can be used to predict the expected time at which the pandemic will be under control, in terms of minimum achievable infection rate, and to test and demonstrate the relative author/funder, who has granted medRxiv a license to display the preprint in perpetuity.