Enhancing Community-Based Nursing Decision Support: Machine Learning Models for Diabetes Risk Prediction Using Home Health Nursing Notes.
Doyeon Lim, Aeri Kim, Hana Lee and 1 others
PMID 41957825WHAT IT FOUND
A machine learning model analyzing home health nursing notes predicted diabetes risk with high accuracy.
Hyperlipidemia, dysuria, and dressing care were strong indicators, offering nurses a potential tool for early detection in home settings.
Key findings
01The Random Forest model achieved an area under the curve (AUC) of 0.985 for predicting diabetes, outperforming other tested algorithms.
02Hyperlipidemia was the strongest predictor, with an adjusted odds ratio of 5.09, followed by depression and hypertension.
03Dysuria was significantly associated with diabetes (OR 1.51), whereas weakness, despite being common, was not significant after adjustment.
STILL TO COME
How it was doneWhat they foundWhat it means for RNs
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
The study used data from a single university hospital in Seoul, South Korea, limiting generalizability to other populations or healthcare systems. The analysis was retrospective and based on secondary data, so causality cannot be established. Key predictors such as family history, diet, physical activity, and specific laboratory values (e.g., HbA1c) were not included in the model. The use of Synthetic Minority Over-sampling Technique (SMOTE) to handle class imbalance carries a risk of overfitting, requiring cautious interpretation.
Declared interests
The authors declared no conflicts of interest. The study was funded by the Ministry of Education and the National Research Foundation of Korea.
The easy way to misread this
Do not assume this model is currently available for clinical use or that it replaces standard diagnostic testing. It is a research prototype that has not been validated prospectively or in diverse settings, and its high accuracy relies on specific Korean home health data structures.