I haven't thought about the proof yet but are you sure it is a IFF statement ? If it was, primality testing would be trivial, which is not the case. Unless I'm misunderstanding the statement which is quite probable.
Assume is prime. Then it divides the numerator but not the denominator. Since it is prime, we conclude it divides the binomial coefficient. (justify that properly)
Assume is composite. Let be any prime factor of . Of course, . It divides both the largest terms of the numerator and of the denominator of , and it doesn't divide any other factor since there are of them, and only 1 number among consecutive ones is divisible by . Thus, after simplification by , the numerator is not divisible by , so that the binomial coefficient isn't either.