Price, Daniel (2021) Natural Language Processing to support Nurse-to-Patient allocation in acute care. [USQ Project]
|
Text (Project)
PRICE Daniel dissertation_redacted.pdf Download (1MB) | Preview |
Abstract
The modern Australian health care system demands a fast and high interpersonal level of care for all patients, with a stronger focus than ever on patient outcomes. A significant level of responsibility for these improved outcomes and reduced hospital stays is placed heavily on Nurses and the nursing profession. While nursing staff are supported by improved technology and workflow procedures it can be seen that some of these systems require heavy admin work and/or non-clinical skills to be effective, reducing time spent with patients and therefore potentially prolonging hospital stays.
Research into optimal staff management has provided evidence that effective distribution of staff skills can improve patient outcomes, staff satisfaction, and reduce costs. Healthcare workers appreciate a consistent and fair schedule that they can rely on. While systems and computer-based programs exist for staff-topatient allocation the process of assigning a staff member to a patient based on skills and patient needs is a time-consuming process.
This project aims to provide a solution to the implementation of the optimal model of care through the use of Natural Language Processing (NLP) and Machine Learning (ML). By automating the staff-to-patient allocation process the optimal model of care can be implemented with little to no admin overheads resulting in nursing staff spending more time with patients and therefore improved patient outcomes.
NLP algorithms and techniques including NER and TF-IDF for Topic modelling have been explored and analysed to determine if accurate extraction of critical patient information from nurse progress notes can be achieved. A Machine Learning neural network based on python’s TensorFlow and Keras ML libraries were also explored. The SpaCy NER model designed in the project return an accuracy of 89% for competency extraction and has proven to be the most reliable. This showed that an automation of staff-to-patient allocation is plausible provided a ward specific allocation is adopted.
|
Statistics for this ePrint Item |
| Item Type: | USQ Project |
|---|---|
| Item Status: | Live Archive |
| Faculty/School / Institute/Centre: | Historic - Faculty of Health, Engineering and Sciences - School of Mechanical and Electrical Engineering (1 Jul 2013 - 31 Dec 2021) |
| Supervisors: | Pythian, Mark |
| Qualification: | Bachelor of Engineering (Honours) (Computer Systems) |
| Date Deposited: | 03 Jan 2023 03:42 |
| Last Modified: | 26 Jun 2023 01:36 |
| Uncontrolled Keywords: | natural language processing, healthcare, staff skills, machine learning, TensorFlow, SpaCy, NER model |
| URI: | https://sear.unisq.edu.au/id/eprint/51826 |
Actions (login required)
 |
Archive Repository Staff Only |
