incidence

First Aid 2015 pg.49 tells us that both prevalence and incidence rate have the denominator of number of people at risk. I'm wondering if this is wrong for prevalence. Won't prevalence have the denominator of total population instead...