Figure. Cumulative Incidence of Loss-of-License Actions by Recertification Status.

The on-time recertification group was the reference group for significance testing in the Cox proportional hazards model. Proportionality assumption was tested using R software and was met for individual covariates and the model. HR indicates hazard ratio. The tick marks on the curves represent each 1-year discrete interval.

^aThe not recertified group includes 2452 surgeons at time 0, including 1045 surgeons with lapsed recertification who eventually recertified and 1407 who did not recertify. As those with lapsed certificates recertified, they moved to the lapsed recertification group (treating lapsed recertification as a time-dependent variable).