Open Access Thesis
Date of Award
Industrial and Manufacturing Engineering
Ratna B. Chinnam
Emergency Departments (EDs) in hospitals are experiencing severe crowding and prolonged patient waiting times. The reported crowding in hospitals shows patients in hospital hallways, long waiting times and full occupancy of ED beds. ED crowding has several potential unfavorable effects including patients and staff frustration, lower patient satisfaction and poor health outcomes. The primary motivations behind this study are shortening the patients’ waiting time and improving patient satisfaction and level of care.
The very initial interaction between clinicians and a patient is recorded on nurse triage notes which contain details of the reason for patient’s visit including specific symptoms and incidents. Triage notes and vital signs measured by triage nurse determine the complexity of the patient’s condition. If a minor illness or injury occurred, patient would be treated by nurse practitioners under ED physicians’ supervision. This process called fast track system which allows the main ED area to focus on more severe patient condition. The final decision should be made by physicians so patients have to wait to be seen in order to find out whether they need to be admitted in the hospital or be discharged.
In this study, we propose a decision support system based on nurse triage notes and vital signs that can automatically predict ICD9 code assigned to each patient prior to the visit time. We tested the model on 8000 patient records from VA Medical Center in Detroit for ICD9 classification and measured performance in terms of accuracy.
Tabaie, Azade, "Predictive Analytics For Disease Condition Of Patients In Emergency Department" (2015). Wayne State University Theses. 478.