This is very interesting and has helped me understand something I hadn't realised before - if a disease is primarily spread by a small subset of the population(superspreaders), but you assume otherwise, it gets very hard to do statistics.<p>For example, if you're trying to measure the R number, then you're sampling from a very fat tailed distribution, so it's easy to underestimate.<p>It also means that if an intervention (such as vitamin D) prevents transmission but not illness, it could be hard to detect/measure the impact.<p>(This comment is not a summary of the paper)