The onset of sub-critical crack growth during slow strain rate tensile testing (SSRT) is assessed through a combined experimental and modeling approach. A systematic comparison of the extent of intergranular fracture and expected hydrogen ingress suggests that hydrogen diffusion alone is insufficient to explain the intergranular fracture depths observed after SSRT experiments in a Ni-Cu superalloy. Simulations of these experiments using a new phase field formulation indicate that crack initiation occurs as low as 40% of the time to failure. The implications of such sub-critical crack growth on the validity and interpretation of SSRT metrics are then explored.