Development and Validation of a Nocturnal Hypoglycaemia Risk Model for Patients With Type 2 Diabetes Mellitus.
Chen Gong, Tingting Cai, Ying Wang and 5 others
PMID 39363560WHAT IT FOUND
A model using previous day CGM and clinical data flagged patients with type 2 diabetes likely to have nocturnal hypoglycaemia.
In this study it worked better than other models; previous day time below target range, diabetes duration under 5 years and bedtime insulin were key.
Key findings
01The LightGBM model had an AUC of 0.869 in the testing set, compared with 0.825 for logistic regression and 0.835 for random forest.
02Previous day time below target range was the most robust predictor, and diabetes duration under 5 years and bedtime insulin therapy were also linked to higher risk.
03Nocturnal hypoglycaemia was present in 573 of 4015 continuous glucose monitoring data points, or 14.30%.
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
All patients were from the same hospital in Shanghai, China, so the model may not work in other settings. External validation in other populations was not done; the paper says further external validation is needed. The study included 440 patients, although the paper states the sample size calculation suggested 1226. Behavioural, psychological and social variables were not included. Variables were chosen from prior literature rather than data-driven selection, which may leave important predictors out.
Declared interests
The authors declared no conflicts of interest. No funding source is reported in the supplied text.
The easy way to misread this
Do not assume the model is ready for bedside use or that it prevents hypoglycaemia. It was built from the same hospital's records, had no external validation, and the paper does not report that acting on the model improved patient safety.