US EditionPublished • loading... • Updated
New AI Tool May Help ICU Patients Get the Nutrition They Need
NutriSighT flags underfeeding risk in ventilated ICU patients with 76% accuracy on external data, aiding clinicians to prevent nutritional deficits during critical days 3–7.
- Mount Sinai researchers introduced NutriSighT, which predicts whether ventilated ICU patients will receive less than 70% of daily calories between days three and seven, updating every four hours.
- By day three, many ventilated ICU patients are already underfed, with 41%–53% underfed by day three and 25%–35% still underfed by day seven as muscle breakdown begins.
- Using transformer architecture, NutriSighT analyzed 62 patient measurements and achieved AUROC 0.81 internally and 0.76 externally, outperforming XGBoost.
- Flagging gives care teams time to change feeding plans, as NutriSighT flags patients so clinicians can adjust care or consider intravenous nutrition before deficits accumulate, but prospective multi-site trials and electronic health records integration are needed first.
- Because the study is retrospective, it faces selection bias and confounding, training and validation occurred only in Amsterdam and Boston ICU databases, while authors with declared ties received National Institutes of Health grant K08DK131286.
Insights by Ground AI
25 Articles
25 Articles
Early identification of nutrition risk in ICU patients using artificial intelligence
A new study by researchers at the Icahn School of Medicine at Mount Sinai suggests that artificial intelligence (AI) could help predict which critically ill patients on ventilators are at risk of underfeeding,
·United States
Read Full Article
Coverage Details
Total News Sources25
Leaning Left4Leaning Right3Center5Last UpdatedBias Distribution42% Center
Bias Distribution
- 42% of the sources are Center
42% Center
L 33%
C 42%
R 25%
Factuality
To view factuality data please Upgrade to Premium
