Extracting Symptoms of Agitation in Dementia from Free-Text Nursing Notes Using Advanced Natural Language Processing.

Area of Science:

• Gerontology
• Computer Science
• Clinical Informatics

Background:

• Nursing notes are rich sources of patient data, including care needs and symptoms.
• Extracting specific information like agitation symptoms in dementia from free-text notes is challenging.
• Automated analysis of nursing notes can improve patient care and clinical decision-making.

Purpose of the Study:

• To develop and evaluate a deep learning and transfer learning-based named entity recognition (NER) model.
• To extract symptoms of agitation in dementia from free-text nursing notes.
• To lay the groundwork for machine learning models that recommend optimal nursing actions.

Main Methods:

• Utilized a Clinical BioBERT model for word embedding.
• Applied bidirectional long-short-term memory (BiLSTM) and conditional random field (CRF) models for NER.
• Trained and evaluated the model on nursing notes from Australian residential aged care facilities.

Main Results:

• The proposed NER model demonstrated satisfactory performance in extracting agitation symptoms.
• Achieved a 75% F1 score and 78% accuracy in identifying symptoms.
• Successfully processed free-text nursing notes to identify key patient observations.

Conclusions:

• Deep learning models, specifically NER with BiLSTM-CRF, are effective for extracting dementia agitation symptoms from nursing notes.
• The developed model shows promise for improving data extraction in aged care settings.
• Future work will focus on developing machine learning models for recommending nursing actions.