Dynamic medical data fusion and data mining in the wound-healing process

Abstract

Wound healing assessment is dependent on the quality of data collected about a patient and his or her wound and the clinician’s ability to evaluate that data and produce a diagnosis. Wound data has very little structure, with most patient-based data being stored in hard copy patient notes. The process of wound healing would benefit from a structured database to store all the different data types required and provide the clinician with a quick and efficient tool to display any data required. To achieve this, data visualization and data fusion will be used to bring together different data modalities in real time. The application will allow users to decide which data, or combination of data, they wish to view. Data mining will be applied to the wound data set, in an attempt to extract any hidden knowledge that is not apparent. The database structure itself will use a novel approach: object-oriented data streaming. This will allow new data types to be incorporated into the database, dynamically and transparently. By structuring the wound data a first step will be taken towards laying down a classification standard for wound-healing assessment.