Cystic fibrosis (CF) is a genetic disease that dramatically decreases life expectancy and quality. Currently, the average life span for people with CF who live to adulthood is approximately 37 years. The disease is characterized by polymicrobial infections which lead to lung remodeling and airway mucus plugging. Data from a CF Clinic and a subset of data from the CF Registry will be analyzed. Both datasets consist of longitudinal lung function data, along with many other covariates for a large number of patients. This thesis presents a new approach to quantify disease severity by developing methodology for linear quantile mixed models using M-estimation to model longitudinal lung function data. The model development includes the implementation of an iterative estimation algorithm and the distributional theory for the inference of model parameters. Disease severity of CF patients is quantified by the quantile regression rankscores and the corresponding normalized ranks to describe both the within and between subject normalized ranks. Disease severity is based on both fixed regression coefficients as well as random coefficients. These ranks make use of the conditional distribution of the response across the quantiles of the entire distribution. In a simulation study, the proposed linear quantile mixed effects model is shown to model the dependence among the longitudinal data correctly and more efficiently the linear mixed effects models. The application of this methodology is presented using two datasets; CF clinical data and CF Registry data. The final model successfully predicts the severity of CF clinic patients based on a model built using a subset of the CF registry data compared to expert opinion.