We propose to develop methods of statistical inference for the analysis of longitudinal data and other challenging data structures. Much of the work will derive from specific areas of medical research but, generally, will have broader applicability. Methodological work on methods of analysis in the presence of informative observation will be considered, motivated by various longitudinal studies. Other methodological investigations will address the analysis of two part models for data which are “zero-heavy” and extensions of the analysis of plaid-square designs to generalised linear models. Our major collaborations in rheumatology will continue. Based on a cohort of patients with psoriatic arthritis, we will consider the modelling of individual joint damage and quality of life, the definition of outcome variables for clinical trials and genetic determinants of disease and disease course. For systemic lupus erythematosus patients, we will continue studies of neuropsychiatric events and the development of permanent damage and continue our investigations of measures of disease activity. We will also collaborate in a clinical trial to evaluate pulmonary metastectomy for colorectal cancer patients.