Factors affecting the incidence of patient falls in hospitals.

Abstract

Patient falls affect the cost and quality of patient care in hospitals. Published studies of risk factors for patient falls are inconsistent and often methodologically weak. Existing programs of patient risk assessment rely on large sets of patient characteristics, not all of which may be necessary. This article describes an information-theoretic technique that is able to identify a limited set of patient characteristics that provide most of the statistically significant information regarding risk of falls that is available in a hospital's adverse patient incident data set. Analysis of data from four hospitals with different institutional and data system characteristics revealed a consistent pattern of risk factors.