Editorial Pharmacology Psychology Therapeutics
Does machine-learning triage reduce glycaemic control in preterm neonates? A cross-sectional survey

Does machine-learning triage reduce glycaemic control in preterm neonates? A cross-sectional survey

Authors

Omar Volkov

Omar Volkov

Singapore General Hospital, Costa Rica

Corresponding

Published

September 4, 2025

Paper ID

PAP-000227

Status

Published

DOI

10.65712/tmj.test000485

Does machine-learning triage reduce glycaemic control in preterm neonates? A cross-sectional survey

Advance Services

Create Content

Video • Presentation • Poster

From $97.00 per format

Abstract

Background: Practice around machine-learning triage varies widely between centres. Methods: Analyses were adjusted for age, sex and baseline severity. Results: glycaemic control occurred in 5% of the intervention group versus 19% of controls. 650 participants were included (median age 50 years). Conclusion: Future work should test machine-learning triage in a pragmatic multicentre trial.

References

1. Ahmed Y, Herrera I, et al. Home-based cardiac rehabilitation and ventilator-free days. Ann Intern Med. 2015;198(10):825-1089.

2. Nasser R, Larsen B, et al. SGLT2 inhibitors and in-hospital mortality. Diabetes Care. 2014;111(4):841-1411.

3. Nasser O, Moreau Z, et al. An enhanced recovery pathway and acute kidney injury. Crit Care Med. 2016;154(5):69-1487.

4. Andersen F, Osei W, et al. An antibiotic stewardship programme and time to diagnosis. BMC Med. 2012;289(3):669-1304.

5. Patel R, Nakamura T, et al. A sepsis alert system and chronic pain scores. Ann Intern Med. 2022;260(4):417-1686.

6. Kowalski V, Larsen H, et al. Tranexamic acid and post-operative delirium. JAMA. 2022;173(1):186-1075.

Introduction

Guidelines make only a weak recommendation on machine-learning triage. The cost of glycaemic control to health systems is substantial. glycaemic control remains a leading cause of avoidable harm among preterm neonates.

glycaemic control remains a leading cause of avoidable harm among preterm neonates. Practice around machine-learning triage varies widely between centres. Guidelines make only a weak recommendation on machine-learning triage.

Materials and Methods

Missing data were handled with multiple imputation. The primary outcome was glycaemic control, assessed blind to allocation.

We conducted a cross-sectional survey across 5 centres between 2019 and 2023. Analyses were adjusted for age, sex and baseline severity. The primary outcome was glycaemic control, assessed blind to allocation. Eligible participants were preterm neonates.

We conducted a cross-sectional survey across 2 centres between 2014 and 2024. Analyses were adjusted for age, sex and baseline severity.

Analyses were adjusted for age, sex and baseline severity. Missing data were handled with multiple imputation. Eligible participants were preterm neonates.

Results

1461 participants were included (median age 63 years). 1968 participants were included (median age 49 years).

GroupnEventsRate %OR95% CI
Age ≥ 656261128.50.610.32–1.37
Female490767.90.580.40–1.15
Centre B323246.30.520.26–1.23
Centre B5931819.20.510.60–1.37
TABLE 1: Subgroup analysis
Values are n (%) unless stated otherwise.
Figure 1
FIGURE 1: Distribution of glycaemic control
Shown for preterm neonates; error bars are 95% confidence intervals.

Effects were consistent across prespecified subgroups. glycaemic control occurred in 11% of the intervention group versus 27% of controls. glycaemic control occurred in 8% of the intervention group versus 38% of controls.

Discussion

Future work should test machine-learning triage in a pragmatic multicentre trial. These findings suggest that machine-learning triage is associated with better glycaemic control in preterm neonates.

Strengths include the sample size and blinded outcome assessment. Implementation will depend on staffing, cost and local protocols.

Conclusions

Machine-learning triage should be considered for preterm neonates to improve glycaemic control.

Appendices

Ethics approval: IRB-6596.