I found that inputting this complex expression gives quite a messy output. But you easily get a numerical result by setting k equal to some integer. Perhaps then make use of the results with k = 1, 2, 3 . . .
Thanks SM Blinder, in the end of the day I run simulations,so to run them I thought about having a formula as general as possible.
But yes, its quite a messy problem, my paper refers to extensions of binomial distributions, and its not the most elegant model:)