Machine Learning Model to Extract Malnutrition Data from Nursing Notes

Abstract

Malnutrition is a severe health problem that is prevalent in older people residing in residential aged care facilities. Recent advancements in machine learning have made it possible to extract key insight from electronic health records. To date, few researchers applied these techniques to classify nursing notes automatically. Therefore, we propose a model based on ClinicalBioBert to identify malnutrition notes. We evaluated our approach with two mainstream approaches. Our approach had the highest F1-score of 0.90.