Human agency and infection rates: implications for social distancing during epidemics

Christopher Bronk Ramsey

The parameter t for each individual is likely to be very variable within a population and indeed the exact numbers will vary between countries and regions. There is some good data on contact numbers and intensity [4] , but these are not in exactly the form needed for this model. However, some reasonable upper and lower limits can be estimated. At the upper limit will be those who spend most of their day in close proximity to four other individuals in crowded spaces (ca. 50 person hrs/day), and at the other end of the scale, people who only have close interaction very rarely when doing essential errands, or those living alone who are rarely visited (ca. 0.5 person hrs/day). Whether contact within the household is considered is a difficult question, because households are particularly integrated. It might be better to consider each household as a superindividual -so with extra interactions and greater chances of picking up the infection together; this issue will not be addressed here, but obviously is within the simulation studies. The exact numbers are not important but the variability is, and here it is assumed that there is likely to be a 95% probability range of about two orders of magnitude in t. The distribution f (t) will be estimated to be log normal (using the estimates above, the median would be ca. 5 person hrs/day).