Applications of zero-heavy finite mixture models to identification of exacerbation-prone patients in COPDGene
Identifying latent subgroups in diseases with heterogeneous presentations can be difficult. It is clinically hypothesized that there are multiple types of patients prone to severe exacerbation in Chronic Obstructive Pulmonary Disease. Few studies have considered the longitudinal trajectories of such patients. We applied longitudinal finite mixture models to data from the Longitudinal Follow-up Survey from the COPDGene study with several potential mixture structures to identify subgroups of exacerbators. Results from the hurdle model component indicated that baseline spirometric ratios, ...
(For more, see "View full record.")