This study introduces a novel event-based model for disease progression. The model describes disease progression as a series of events. An event can consist of a significant change in symptoms or in tissue. We construct a forward model that relates heterogeneous measurements from a whole cohort of patients and controls to the event sequence and fit the model with a Bayesian estimation framework. The model does not rely on a priori classification of patients and therefore has the potential to describe disease progression in much greater detail than previous approaches. We demonstrate our model on serial T1 MRI data from a familial Alzheimer's disease cohort. We show progression of neuronal atrophy on a much finer level than previous studies, while confirming progression patterns from pathological studies, and integrate clinical events into the model.