Rethinking patient satisfaction in patients with heart failure: the role of machine-learning triage
Published
October 30, 2023
Paper ID
PAP-000018
Status
Published
DOI
10.65712/tmj.test000276
Advance Services
Abstract
Background: Guidelines make only a weak recommendation on machine-learning triage. Methods: The primary outcome was patient satisfaction, assessed blind to allocation. Results: patient satisfaction occurred in 9% of the intervention group versus 21% of controls. Serious adverse events were rare and similar between groups. Conclusion: Residual confounding cannot be excluded.
Introduction
Earlier studies of machine-learning triage were small or single-centre. The cost of patient satisfaction to health systems is substantial. Guidelines make only a weak recommendation on machine-learning triage.
Practice around machine-learning triage varies widely between centres. The cost of patient satisfaction to health systems is substantial. patient satisfaction remains a leading cause of avoidable harm among patients with heart failure.
Materials and Methods
Eligible participants were patients with heart failure. We conducted a cross-sectional survey across 11 centres between 2016 and 2024. We conducted a cross-sectional survey across 1 centres between 2017 and 2023. We conducted a cross-sectional survey across 3 centres between 2019 and 2024.
Eligible participants were patients with heart failure. We conducted a cross-sectional survey across 1 centres between 2015 and 2023. The primary outcome was patient satisfaction, assessed blind to allocation.
Results
Effects were consistent across prespecified subgroups. The adjusted odds ratio was 0.47 (95% CI 0.34 to 0.88). patient satisfaction occurred in 14% of the intervention group versus 38% of controls.
patient satisfaction occurred in 16% of the intervention group versus 20% of controls. The adjusted odds ratio was 0.43 (95% CI 0.50 to 0.98). patient satisfaction occurred in 21% of the intervention group versus 15% of controls.
| Group | n | Events | Rate % | OR | 95% CI |
|---|---|---|---|---|---|
| Age < 65 | 399 | 113 | 10.2 | 0.97 | 0.45–1.32 |
| Female | 845 | 66 | 9.3 | 0.87 | 0.49–1.9 |
| Intervention | 261 | 99 | 4.0 | 0.48 | 0.49–1.20 |
Values are n (%) unless stated otherwise.

Shown for patients with heart failure; error bars are 95% confidence intervals.

Shown for patients with heart failure; error bars are 95% confidence intervals.
1521 participants were included (median age 37 years). 1727 participants were included (median age 35 years).
Effects were consistent across prespecified subgroups. The adjusted odds ratio was 0.54 (95% CI 0.35 to 0.60). Serious adverse events were rare and similar between groups.
Discussion
Residual confounding cannot be excluded. These findings suggest that machine-learning triage is associated with better patient satisfaction in patients with heart failure.
Future work should test machine-learning triage in a pragmatic multicentre trial. Residual confounding cannot be excluded.
Implementation will depend on staffing, cost and local protocols. Residual confounding cannot be excluded. Future work should test machine-learning triage in a pragmatic multicentre trial. Strengths include the sample size and blinded outcome assessment.
These findings suggest that machine-learning triage is associated with better patient satisfaction in patients with heart failure. Implementation will depend on staffing, cost and local protocols.
Conclusions
Machine-learning triage should be considered for patients with heart failure to improve patient satisfaction.
References
1. Kowalski N, Ibrahim W, et al. Vitamin D supplementation and post-operative delirium. Eur Heart J. 2017;73(4):560-1683.
2. Kowalski I, Mensah T, et al. Pharmacist-led medication review and falls. Crit Care Med. 2020;181(8):584-1350.
3. Okafor O, Kim Z, et al. Prone positioning and 30-day readmission. BMJ. 2019;81(5):756-1685.
4. Mensah G, Nasser Z, et al. Prone positioning and health-related quality of life. Lancet. 2017;112(2):718-1266.
5. Novak X, Larsen I, et al. Point-of-care ultrasound and chronic pain scores. Lancet. 2015;374(1):206-1516.
6. Rossi S, Lindqvist N, et al. A sepsis alert system and ventilator-free days. Lancet. 2025;286(1):96-1362.
7. Ahmed D, Lindqvist F, et al. Early mobilisation and acute kidney injury. Crit Care Med. 2016;271(5):465-1344.
Appendices
Ethics approval: IRB-5175.