This was performed using Statistics Package for Social Sciences (SPSS) with guidance from a statistician. Pre-training and post-training and after surgery differences in VC, FCV, FEV1, MIP and MEP were analysed for significance. Wilcoxon signed rank test (non-parametric method) was used for analysis of the differences and a P-value < 0.05 was taken as significant. Since this was a pilot study, formal calculation of sample size was not necessary.