I am running up against some problems in my stats analysis. Please can anyone give me a hand on this...
I have data on Emergency Department Visits to an Urban Hospital in 2012 .
I am trying to compare the impact of the type of primary care provider (2 types) on the ED use. However, in our city the population from those 2 different types hasn't the same size. For instance:

2000 ED Visits
1500 Type A Primary Care Provider
500 Type B PCP

Population size Type A - 5000
Population size Type B - 1000

If I run a logit analysis (A,B - dependent variable) (independent variables like age, sex, distance to the ED, diagnosis,...), I would get that Type A is more likely to have people going to the ED. However, if we consider the population sizes, it actually isn't.
How can I incorporate these different population sizes in the regression analysis?