Development and Validation of Predictors of Inpatient Hypoglycaemia Using Data from More Than 25,000 Admissions
Applied/Translational Research
Singapore Health Biomedical Congress
National Healthcare Group
31 December 2022
To identify independent risk factors associated with inpatient hypoglycaemia and develop a predictive tool for patients with. Logistic regression models are effective for predicting hypoglycaemia risk, laying groundwork for advanced models.
Year Submitted: 2022
Published Date: 31 December 2022
Tags: Applied/Translational Research, Quantitative Research
About this Content
Aims
To identify independent risk factors associated with inpatient hypoglycaemia and develop a predictive tool for patients with DM.
Background
Hypoglycaemia is a common inpatient issue linked to increased length of stay and significant morbidity/mortality.
Methods
Data from over 25,000 admissions were analyzed using logistic regression models to identify risk factors for hypoglycaemia.
Results
Models showed an AUC of 0.795 in the training dataset and 0.78 in the test dataset. Significant predictors included age, eGFR, weight, prior hypoglycaemia, and medication use.
Conclusion
Logistic regression models are effective for predicting hypoglycaemia risk, laying groundwork for advanced models.
Lessons Learnt
Data quality and interpretability are key for clinical integration. Variable engineering improves model performance.
Additional Information
Singapore Health & Biomedical Congress (SHBC) 2022: SHBC Best Poster Award (Clinical Research) – (Gold Award)
Keywords
Hypoglycaemia, Prediction Models, Inpatient Diabetes, Diabetes Mellitus
Innovators' Details
Innovators' Details
Healthcare Cluster(s) | National Healthcare Group |
Organization(s) Involved | Khoo Teck Puat Hospital |
Platform(s) | Singapore Health Biomedical Congress |
Healthcare Professional Group(s) | Allied Health, Medical |
Applicable Specialty or Discipline | Endocrinology, General Medicine, Clinical Research |
Project Lead(s) | Ester Yeoh |
Project Member(s) | Lin Yi |
Connect with this contributor!
Lin Yi - lin.yi@ktph.com.sg
Project Attachment
C_463_KTPH_SHBC_2022_Development_and_Validation_of_Predictors_of_Inpatient.pdf