A Novel Method of Identification of Delirium in Patients from Electronic Health Records Using Machine Learning
R. Kavitha, KDV Prasad, S Archana Shreee, Bhavesh Maheshwari, G Jeevitha Sai, Dankan Gowda
Abstract
In most cases, the mental impairment caused by delirium may be treated and eventually reversed. Lack of concentration, disorientation, incoherent thought, and fluctuating degrees of awareness (consciousness) are all symptoms. Delirium, an acute neuropsychiatric disorder characterised by inattention and generalised cognitive impairment, is common, hazardous, and generally linked with poor results. Patients with delirium are at increased risk for adverse outcomes throughout their time in the critical care unit. It requires time and medical competence to diagnose delirium. Those at risk of developing delirium should be identified as soon as possible. Once a diagnosis has been made, the treatment process may be lengthy and include several groups working together. This paper’s goal is to show how a model may be built to diagnose delirium using Electronic Health Record data employing a Machine Learning technique.