A Hybrid-Structure Framework for Multivariable Health Assessment and Conditional Short-Horizon Projection

A mathematically transparent framework is developed for assessing a person’s current health status and conditionally projecting the status to a selected future time. The method uses a hybrid structure in which each biomedical parameter contributes both a set of supported monitoring states and a fuzzy confidence value. Eight specific physiological and metabolic parameters are converted by means of explicit and repeatable fuzzy calibrations, followed by combination of fuzzy and set-valued information to form an aggregate monitoring vector and Hybrid Health Severity Index. A local trend model is developed using recent observations, and the same hybrid approach is used for future evaluation. There are three synthetic illustrations given here to demonstrate the computations, including one with a deliberately severe profile to examine the upper end of the proposed scale. The work is purely methodological, and validation of the calibrations on clinical data is required before any health care application.