Acquiring temporal knowledge for making decisions in medical processes : doctor of philosophy thesis

english

doktorska dela

2014

Univerza v Mariboru, Fakulteta za elektrotehniko računalništvo in informatiko

A medical process is a set of medical actions performed by healthcare professionals while making observations about signs and symptoms, ordering interventions, prescriptions, tests, and any other actions in order to solve a health problem that is affecting a particular patient. The objective is to offer a curative, chronic, palliative and/or symptomatic treatments. The capability of a physician to propose an appropriate treatment depends on his/her knowledge of similar clinical cases and by following advances in the treatment of particular diseases. Time is an important concept of the real world that has to be considered in regard to medical processes. Clinical Practice Guidelines (CPGs) are a narrative set of recommendations for treating patients suffering from a particular disease. By constantly referring to CPGs physicians can stay up to date with the best evidence-based medical care and with the recommendations of experts. The time dimension is, however, often omitted or only partially covered in CPGs. Some CPGs do contain certain recommendations considering time but often there are huge gaps that are supposed to be overcome by physicians own knowledge and experience. At the same time, healthcare centres hold healthcare records and information systems that register medical processes and patients data, including information about the times of any encounters, prescriptions, and other medical actions. Therefore healthcare records and information systems data can be a source for detecting temporal medical knowledge and sound evidence regarding healthcare.This thesis is centred on temporal knowledge acquisition and representation for the purposes of decision-making during medical processes. CPGs were analysed to obtain procedural knowledge models and Extended Timed Transition Diagrams defined for representations of the obtained knowledge. The data about the treatments of patients were analysed in order to detect temporal medical knowledge models that represented those medical procedures that were carried out while the data was being generated. These models thus provided an explicit representation of the time dimensions of past medical procedures. They could be used for complementing the knowledge provided by CPGs, for studying adherences to the CPGs and for representing a basic framework for medical procedural decision support systems development.

07.08.2023