Continuous and discrete global models of disease

Two models designed to identify and characterize homogeneous subgroups of patients, based upon analysis of large scale chronic disease data banks are discussed, suggesting that the evaluative tasks broadly labeled diagnosis and prognosis involve the rigorous and precise definition of the patient's disease.

Abstract

Most clinical activities (e.g. diagnosis and prognosis) implicitly assume some definition or interpretation of disease in order to provide a means for generating explanatory and predictive hypotheses. These serve as a basis for describing and treating patients. This paper discusses two models designed to identify and characterize homogeneous subgroups of patients, based upon analysis of large scale chronic disease data banks. From consideration of the formulation of these models, we suggest that the evaluative tasks broadly labeled diagnosis and prognosis involve the rigorous and precise definition of the patient's disease. The models also reveal a duality in the quantification of clinical judgment involved in the simultaneous description of diseases and patients. An illustration from the study of Systemic Lupus Erythematosus isprovided. It is concluded that both models have significant implications for future research, including the detailed study and modeling of chronic disease spectra, clinical decision making and the analysis of applied judgmental processes.