Statistical Analyses

Data were analyzed June 9, 2021, to May 31, 2022. Differences between groups were measured using t test or analysis of variance and χ² tests for continuous and categorical variables. We constructed cumulative incidence plots using the survminer package in R software (R Foundation for Statistical Computing). A Fine-Gray competing risk model was used to estimate the relative implications of the covariates for the subdistribution hazard of skin cancer development using the cmprsk package in R software.⁹ The models included all cancers with non-0 follow-up times, which removed 7 synchronous cancers from the first to second skin cancer model and 3 synchronous cancers from the second to third skin cancer model. Age at transplant was converted into a categorical variable with 10-year groupings for modeling purposes. For all analyses, the baseline was either time of transplant or time of the previous skin cancer. A 2-sided P < .05 was considered to be statistically significant. R software, version 4.0.2 was used to conduct all analyses.