Propensity Score-Matched Aspirin Effects

Categorical data, such as sex, comorbidities, lifestyle risk factors, AJCC stages of cancer, etc., were tested by either a two-sided Fisher’s exact test or a Pearson’s chi-squared test. The normally and non-normally distributed continuous data were analyzed using Student’s t-tests and Mann–Whitney U tests, respectively. In order to minimize the confounding effect of groups that are comparable due to non-randomized allocations, a 1:4 propensity score-matched study group (aspirin user vs. non-user) was created using the Greedy method with a 0.25 caliper-width using NCSS 10 software (NCSS Statistical Software, Kaysville, UT, USA) (Table S1) [41 (link),42 (link),43 (link),44 (link),45 ,46 (link),47 (link)]. The propensity scores were calculated using a logistic regression model with the following covariates: sex, age and AJCC stages of cancer. After adjusting for these confounding factors, the Kaplan–Meier method was used to evaluate the effects of aspirin use in the primary outcome (DSS). A univariate analysis and Cox proportional-hazards model were used to evaluate any parameters that could affect survival. All statistical analyses were performed using SPSS Statistics V22.0 software for Windows (IBM Corp., Armonk, NY, USA). Statistical significance was set for each analysis at p-values of <0.05.