Machine-learning triage and surgical site infection in older adults: A retrospective cohort study
Published
June 25, 2026
Paper ID
PAP-000282
Status
Published
DOI
10.65712/tmj.test000540
Advance Services
Abstract
Background: Guidelines make only a weak recommendation on machine-learning triage. Methods: Missing data were handled with multiple imputation. Results: 3666 participants were included (median age 38 years). Serious adverse events were rare and similar between groups. Conclusion: These findings suggest that machine-learning triage is associated with better surgical site infection in older adults.
Introduction
The cost of surgical site infection to health systems is substantial. surgical site infection remains a leading cause of avoidable harm among older adults. Guidelines make only a weak recommendation on machine-learning triage.
Practice around machine-learning triage varies widely between centres. The cost of surgical site infection to health systems is substantial.
Earlier studies of machine-learning triage were small or single-centre. surgical site infection remains a leading cause of avoidable harm among older adults. The cost of surgical site infection to health systems is substantial.
The cost of surgical site infection to health systems is substantial.
Materials and Methods
Eligible participants were older adults. Missing data were handled with multiple imputation.
Eligible participants were older adults. We conducted a retrospective cohort study across 14 centres between 2016 and 2021.
Analyses were adjusted for age, sex and baseline severity. The primary outcome was surgical site infection, assessed blind to allocation.
Eligible participants were older adults. Missing data were handled with multiple imputation. We conducted a retrospective cohort study across 7 centres between 2019 and 2024.
Results
Serious adverse events were rare and similar between groups. The adjusted odds ratio was 0.41 (95% CI 0.27 to 0.69). 3185 participants were included (median age 55 years).
surgical site infection occurred in 5% of the intervention group versus 26% of controls. Serious adverse events were rare and similar between groups. The adjusted odds ratio was 0.38 (95% CI 0.31 to 0.83).
The adjusted odds ratio was 0.41 (95% CI 0.48 to 0.81). 2710 participants were included (median age 64 years). Serious adverse events were rare and similar between groups. surgical site infection occurred in 10% of the intervention group versus 17% of controls.
| Group | n | Events | Rate % |
|---|---|---|---|
| Centre A | 113 | 67 | 17.4 |
| Age < 65 | 142 | 53 | 11.5 |
| Age < 65 | 692 | 33 | 16.8 |
| Centre A | 822 | 49 | 16.4 |
| Centre A | 104 | 119 | 1.2 |
Values are n (%) unless stated otherwise.

Shown for older adults; error bars are 95% confidence intervals.
Effects were consistent across prespecified subgroups. Serious adverse events were rare and similar between groups.
2364 participants were included (median age 49 years). The adjusted odds ratio was 0.57 (95% CI 0.50 to 0.99). The adjusted odds ratio was 0.64 (95% CI 0.33 to 0.71).
Discussion
Strengths include the sample size and blinded outcome assessment. Future work should test machine-learning triage in a pragmatic multicentre trial.
Implementation will depend on staffing, cost and local protocols. Strengths include the sample size and blinded outcome assessment. Residual confounding cannot be excluded.
Strengths include the sample size and blinded outcome assessment. Future work should test machine-learning triage in a pragmatic multicentre trial.
Strengths include the sample size and blinded outcome assessment. Residual confounding cannot be excluded.
These findings suggest that machine-learning triage is associated with better surgical site infection in older adults. Future work should test machine-learning triage in a pragmatic multicentre trial. Strengths include the sample size and blinded outcome assessment.
Conclusions
Machine-learning triage merits wider use for older adults to improve surgical site infection.
References
1. Okafor P, Moreau N, et al. Early enteral nutrition and falls. Ann Intern Med. 2023;308(5):46-1772.
2. Kaur W, Rossi M, et al. Tranexamic acid and acute kidney injury. JAMA. 2014;393(3):864-1398.
3. Schmidt I, Nakamura V, et al. Telemedicine follow-up and emergency department revisits. Ann Intern Med. 2023;230(2):291-1598.
4. Schmidt R, Zhang V, et al. An antibiotic stewardship programme and acute kidney injury. Ann Intern Med. 2023;360(8):752-1724.
5. Volkov Y, Kim S, et al. Delayed cord clamping and time to diagnosis. Lancet. 2013;90(8):362-1243.
6. Mueller A, Lindqvist M, et al. Vitamin D supplementation and acute kidney injury. Crit Care Med. 2011;163(11):614-1120.
7. Herrera W, Lindqvist V, et al. Pharmacist-led medication review and falls. Crit Care Med. 2008;65(6):195-1389.
Appendices
Supplementary tables are available on request.