Objectives: This study aimed to assess whether trigger tools were useful identifying triage errors among patients referred to non-emergency care by emergency medical dispatch nurses, and to describe the characteristics of these patients.

Design: An observational study of patients referred by dispatch nurses to non-emergency care.

Setting: Dispatch centres in two Swedish regions.

Participants: A total of 1089 adult patients directed to non-emergency care by dispatch nurses between October 2016 and February 2017. 53% were female and the median age was 61 years.

Primary and secondary outcome measures: The primary outcome was a visit to an emergency department within 7 days of contact with the dispatch centre. Secondary outcomes were (1) visits related to the primary contact with the dispatch centre, (2) provision of care above the primary level (ie, interventions not available at a typical local primary care centre) and (3) admission to hospital in-patient care.

Results: Of 1089 included patients, 260 (24%) visited an emergency department within 7 days. Of these, 209 (80%) were related to the dispatch centre contact, 143 (55%) received interventions above the primary care level and 99 (38%) were admitted to in-patient care. Elderly (65+) patients (OR 1.45, 95% CI 1.05 to 1.98) and patients referred onwards to other healthcare providers (OR 1.58, 95% CI 1.15 to 2.19) had higher likelihoods of visiting an emergency department. Six avoidable patient harms were identified, none of which were captured by existing incident reporting systems, and all of which would have received an ambulance if the decision support system had been strictly adhered to.

Conclusion: The use of these patient outcomes in the framework of a Global Trigger Tool-based review can identify patient harms missed by incident reporting systems in the context of emergency medical dispatching. Increased compliance with the decision support system has the potential to improve patient safety.

Background: Clinical feedback is often lacking in prehospital care, and while performance data is increasingly available to clinical and operational leadership, it is seldom made available to care providers themselves. In this study, we investigate the impact of a simple intervention consisting of the provision of monthly feedback reports via email to emergency medical dispatch nurses in three Swedish regions.

Method: Individualised reports consisting of 14 measures divided into descriptive (eg, priority-setting and call times), process (eg, dispatch times and documentation completeness) and outcome (eg, over/under triage rate) categories were developed with staff and management input. Report delivery was implemented using a stepped-wedge design, and effects were evaluated using a hierarchical regression-based interrupted time series analysis.

Results: 40 dispatchers were included in the study between March 2020 and October 2023, who handled a total of 246 353 incidents. Some impacts on documentation-related process measures were identified, with the odds of complete documentation increasing by 7.5% (95% CI 5.1 to 9.9) and the odds of having a documented contact reason increasing by 3.8% (1.5-5.9). These effects remained robust over the post-intervention period. Weaker impacts on outcome measures were identified which could be explained by a higher priority given to emergency medical dispatches overall.

Conclusion: Providing performance data can influence care providers to adjust their behaviour to improve process-related quality metrics under their direct control. The intervention may also have induced nurses to more often upgrade the priority of their patients. Improving outcome metrics may however require more intensive, multifaceted interventions.

BACKGROUND: The triage of patients in prehospital care is a difficult task, and improved risk assessment tools are needed both at the dispatch center and on the ambulance to differentiate between low- and high-risk patients. This study validates a machine learning-based approach to generating risk scores based on hospital outcomes using routinely collected prehospital data.

METHODS: Dispatch, ambulance, and hospital data were collected in one Swedish region from 2016-2017. Dispatch center and ambulance records were used to develop gradient boosting models predicting hospital admission, critical care (defined as admission to an intensive care unit or in-hospital mortality), and two-day mortality. Composite risk scores were generated based on the models and compared to National Early Warning Scores (NEWS) and actual dispatched priorities in a prospectively gathered dataset from 2018.

RESULTS: A total of 38203 patients were included from 2016-2018. Concordance indexes (or areas under the receiver operating characteristics curve) for dispatched priorities ranged from 0.51-0.66, while those for NEWS ranged from 0.66-0.85. Concordance ranged from 0.70-0.79 for risk scores based only on dispatch data, and 0.79-0.89 for risk scores including ambulance data. Dispatch data-based risk scores consistently outperformed dispatched priorities in predicting hospital outcomes, while models including ambulance data also consistently outperformed NEWS. Model performance in the prospective test dataset was similar to that found using cross-validation, and calibration was comparable to that of NEWS.

CONCLUSIONS: Machine learning-based risk scores outperformed a widely-used rule-based triage algorithm and human prioritization decisions in predicting hospital outcomes. Performance was robust in a prospectively gathered dataset, and scores demonstrated adequate calibration. Future research should explore the robustness of these methods when applied to other settings, establish appropriate outcome measures for use in determining the need for prehospital care, and investigate the clinical impact of interventions based on these methods.