Naik HB, et al. - Deidentified data from the medical records of 1,203 patients with dermatologist-confirmed cases of HS from three North American hidradenitis suppurativa (HS) practices were collected and involved in a retrospective study by the experts in order to examine age at HS onset and its distribution. Collected data involved age at visit, gender, race, self-identified ethnicity, self-reported age at HS onset, and disease severity by Hurley stage. Because gender was significantly related to age at onset, mixture of regression with gender as a covariate was utilized, allowing recognition of subpopulations considering covariates. To sequentially test the number of components in multiple models, parametric bootstrapping was used, and two subpopulations were recognized. Estimation of distribution of subpopulations was done using the expectation maximization algorithm for the mixture of regressions with gender as a covariate.