Linear programming applications to the evaluation of the pre-hospital care phase of an emergency medical services system

In the development of models for evaluating the quality of care rendered to ambulance transport patients, clustering of patients is a fundamental requirement. To determine the level of care, patients must be grouped into classes relative to severity, risk and quality. Sources of variation which affect care can be partially controlled by clustering patients into "homogeneous" groups. The most serious obstacle to the clustering of patients is the fact that most information available is not quantitative in nature. To resolve this problem, psychometric techniques are integrated into linear programming models to produce an evaluation methodology.