Introduction:Traditional percutaneous nephrolithotomy (PCNL) training involved subjective award of cases to the trainee. We restructured this according to the Guy's stone score (GSS) such that each trainee stepwise completed 25 cases of each grade before progressing. This study compares the outcomes of training with traditional versus stepwise approach.Methods:Four hundred consecutive cases equally distributed for two trainees in each group were compared in terms of complications (Clavien-Dindo), stone free rate (SFR), operative and fluoroscopy time. External comparison was also done against a benchmark surgeon. Multivariable regression model was created to compare SFR and complications while adjusting for comorbidity, Amplatz size, access tract location, number of punctures, body mass index, stone complexity, and training approach.Results:The distribution of cases in terms of calculus complexity was similar. Overall, in comparison to traditional training, stepwise training had significantly shorter median operative time (100 vs. 120 min, P < 0.05), fluoroscopy time (136 vs. 150 min, P < 0.05) and fewer overall (29.5% vs. 43.5%, P < 0.005) as well as major complications (3% vs. 8.5%, P - 0.029), though initial SFR was higher but not statistically significant (77% vs. 71.5%). On multivariable analyses, stepwise training was independently associated with lower complications (odds ratio 0.46 [0.20–0.74], P - 0.0013) along with GSS grade, number of punctures, and Amplatz size. Stepwise training had similar fluoroscopy time, major complications and final clearance rate compared to expert surgeon.Conclusions:PCNL has a learning curve specific for each grade of calculus complexity and stepwise training protocol improves outcomes.