If it is given that a program has a worst case running time of $O(n)$, then is it still okay to define the running time as being $O(n^2)$. By definition, this seems corrects since Big-Oh is essentially an upper bound.
But, since we already established a worst case running time, can this still be correct.
Yes, if a program runs in time $O(n)$, then it also runs in time $O(n^2)$. However, usually we try to give the best possible bound, to prevent readers from being misled. You can signify that the bound is tight using big theta: the worst-case running time of the program is $\Theta(n)$.
