We could be vastly overestimating the death rate for COVID-19. Here’s why