Care2Report: Dialogue summarisation for geriatric performance assessments

Abstract

There is an administrative burden in healthcare. This thesis presents a solution to the administrative burden in geriatrics and nursing by creating a pipeline of linguistic techniques that allow for the automatic summarization of conversations for geriatric performance assessment. Ontology learning, Narrative information extraction and conversation interpretation techniques were used to extract the assessment-relevant information from interviews. We proposed using a sentence parser to extract the most essential syntactic components from sentences and using a sentence realiser to build new summarised sentences.