Why are rates based on small numbers are unreliable in epidemiological research studies?

[sww]

Why are rates based on small numbers are unreliable in epidemiological research studies?

Thank you very much!

[Reeza]

Rates based on small numbers are unreliable because the confidence intervals around those rates are way too big and/or you're more likely to see extreme events.

For example, a hospital has 3 babies born in one day and another hospital has 50 babies born in one day. Which one is more likely to have all girls or all boys? The smaller hospital. Smaller N's allow for seeing more extreme events that may not accurately represent the population.