Abstract: In observational studies, different visit times may lead to missing data and disparate outcome measures. There are numerous ways to resolve these issues, and, in this dissertation, we apply several statistical methods to assess radiographic progression of joint damage in rheumatoid arthritis (RA) patients in a longitudinal study with irregular visit times. Traditionally, linear progression rates are used to assess radiographic outcomes. We instead examined radiographic profile patterns by using clustering algorithms to compute progression rates at set time intervals. Hands and feet radiographic scores were analyzed from a prospective RA cohort, and four progression rates were determined by interpolating between set time intervals past the first radiographic observation. Patients were grouped on their sets of progression rates by K-medians clustering algorithms based on Euclidean distances, and patterns were identified from the clusters. Cluster membership was then examined as a multinomial outcome measure and regressed on baseline covariates to identify baseline covariates that were associated with future radiographic patterns. To determine if these covariates affected radiographic outcomes longitudinally, radiographic scores were used as outcomes in repeated measures analyses. The scores, interpolated from the four segmented rates used in the cluster analyses, are regarded as count data. This allowed them to be analyzed using count models following the Poisson or negative binomial distributions. Many patients exhibited a “flat” radiographic pattern with a large proportion of zeros as scores at each time point, providing the rationale for using two-part models. These longitudinal analyses included generalized estimating equations models using an autoregressive working correlation matrix on the interpolated scores, and a two-part model using the zero-inflated Poisson distribution.