Nonparametric cure rate estimation with covariates

We propose a nonparametric method to assess the effects of one or more covariates on the cure rate for survival data with a cure fraction. The method extends the existing work on single sample and multiple sample cases and also allows for a continuous covariate. The proposed estimator is shown to be consistent and asymptotically normal. A simulation study shows that the proposed method estimates the covariate effect on cure rate with smaller biases than the existing semi‐parametric cure models, particularly when the semi‐parametric cure models mis‐specify the effect. The proposed method is applied to a study for leukaemia patients to assess the effect of age on the chance of being cured after bone marrow transplantation. The Canadian Journal of Statistics 42: 1–17; 2014 © 2013 Statistical Society of Canada