Multi-sample nonparametric treatments comparison in medical follow-up study with unequal observation processes through simulation and bladder tumour case study

Abstract

In medical follow-up study, the diseases recurrent processes evolved in continuous time and the patients are usually monitor at distinct and different intervals. Therefore, most of the existing methods that assumed identical observation processes might provide misleading results in this case. To address this, a nonparametric test based on integrated weighted different between the mean cumulative functions which characterized both the recurrent processes and observation processes with condition on treatment is proposed to allow unequal observation processes. The empirical power of the proposed test has been investigated via Monte Carlo simulation study and bladder tumour case study. The results arein line with earlier research; the proposed test procedure works well for practical situations and had a good power in detecting treatment difference.

Keywords: nonparametric; unequal observation; multi-sample; treatments comparison.