Cold working of holes generates compressive residual stresses resulting in a significant fatigue life improvement over a non-cold worked hole. Current fatigue life prediction methods for cold worked holes are based on two-dimensional (2-D) linear superposition of stress intensity factor, K, solutions of the non-cold worked hole and the residual stresses. Such predictions have shown various levels of agreement with the overall fatigue life and have generally underpredicted the crack growth over the majority of life. An inverse process was used to generate K solutions for the residual stresses of two experimental data sets using AFGROW and the crack growth data from the experiments. The inverse K solutions were inconsistent with the residual stress distribution indicating that it contained mechanisms or features not inherent to the 2-D weight function method. The predicted fatigue life was found to be very sensitive to a ± 1 % variance in the inversely generated K solution. This sensitivity of the K method is a very important issue that must be addressed in the future. A 2-D FEA model indicated that the crack remained completely closed over a range of crack lengths despite experimental crack growth indicating that the model was not an accurate physical representation of the real crack. The results of this study combined with the significantly faster crack growth observed on the side of the hole corresponding to the entry side of the mandrel and the through thickness residual hoop stress variation show that the current methodology based on a 2-D assumptions is inadequate in predicting the fatigue crack growth from cold worked holes for the range of specimen thicknesses in this study. It is suggested that further research focus on incorporating the through thickness stress variance in a solution that predicts crack growth both in the radial and through thickness directions to capture the peculiar crack growth associated with cold working.