Discussion
Between 2020 and 2023, we observed a substantial increase in the total number of EMCC dispatches for patients with suspected stroke, despite no change in the number of patients discharged with a stroke diagnosis. There was no evidence that this increased use of ambulance resources impacted on key quality indicators, in terms of the proportion of stroke patients admitted to hospital within four hours of symptom onset, patients treated with thrombolysis, and thrombolysis initiated within three hours of symptom onset.
Almost all EMCCs reported increases in EMCC dispatches for suspected stroke, though with considerable variation between centres. EMCC Oslo showed the largest increase, dispatching more than five ambulances per patient ultimately diagnosed with stroke, while EMCC Nord-Trøndelag only dispatched 1.5 ambulances per diagnosed patient. The EMCC Oslo findings are consistent with a previous study at this centre, which showed a PPV for stroke recognition of 16 % (14). EMCC Bergen experienced approximately a 30 % increase in dispatches for suspected stroke, while the number of patients discharged with a stroke diagnosis in this area decreased by 5 % over the same period, consistent with an earlier study at this centre (18).
High sensitivity has been a central focus in most studies on stroke recognition in EMCC, as it is assumed to reduce prehospital delays and thereby increase the proportion of patients treated with thrombolysis (13). It is therefore notable that the proportion of stroke patients admitted to hospital within four hours of symptom onset remained stable. Similarly, although there was a slight increase in the proportion of cerebral infarction patients treated with thrombolysis in 2023, there was no corresponding increase in the proportion receiving treatment within three hours of symptom onset, the window during which thrombolysis is most effective (19).
Given that both the overall number of strokes and the number of patients admitted to hospital within four hours remained stable during the study period, neither the observed increase in EMCC dispatches nor the variations between EMCCs can be explained by changes in stroke incidence. Multiple factors are likely to influence EMCC operators' assessments and contribute to overtriage. The fact that EMCC Oslo had both the largest increase in EMCC dispatches for suspected stroke and the highest number of dispatches relative to the number of patients discharged with a stroke diagnosis makes it particularly interesting to consider potential underlying causes.
EMCC Oslo is the largest EMCC in Norway, serving a population of approximately 1.7 million. In 2023, more than 260,000 calls were made to the medical emergency number 113, in addition to a large number of other enquiries. Previous studies have shown that workload at this centre has been high, leading to shorter assessment time intervals – the time interval from receipt of an emergency call to the decision on the appropriate emergency medical response. For instance, the assessment time interval for stroke patients not identified by the EMCC dispatchers was only 55 seconds. The high workload also meant that dispatchers rigidly followed the Norwegian Index for Medical Emergency Assistance without making independent clinical judgements, and that they rarely received feedback on whether their assessments of symptoms, severity and triage category were accurate (14, 20).
In the period 2020–2023, several research projects were undertaken to improve stroke recognition within both EMCC Oslo and the Ambulance Service in Oslo (14, 20, 21). Although it is difficult to demonstrate any definite association between these studies and the observed overtriage, it is likely that training and involvement in these research initiatives influenced operators' tendency to suspect stroke.
The Norwegian Stroke Registry is a quality register with high validity and reliability, but it does not include patients discharged with transient ischaemic attack (TIA). TIA is defined as a clinical event with sudden, ischaemia-related focal neurological deficits that fully resolve within 24 hours (22). EMCC dispatchers are rarely able to distinguish patients with TIA from those with other acute stroke symptoms, as EMCCs are typically contacted shortly after symptom onset. In most studies of stroke recognition at EMCCs, stroke patients are collectively defined to include both stroke and TIA, making it difficult to estimate the proportion of each (18, 23–25). Nevertheless, two Norwegian studies indicate that patients with TIA account for approximately 20 % of all patients with suspected stroke (14, 26).
Although the medical emergency number 113 is primarily used by the public to contact EMCCs, general practitioners (GPs) also use it to request ambulance services for patients they consider in need of urgent care. In a previous study verifying calls to EMCCs via 113, patients referred from GP offices accounted for approximately 8 % of all incidents meeting stroke criteria (14). This study does not assess the sensitivity of stroke recognition at EMCCs. Despite these limitations, it provides new and important insights into whether increased use of ambulance services leads to improvements in key quality indicators for the treatment of stroke patients. The identified limitations are likely consistent over time and therefore have minimal impact on the observed increase in EMCC dispatches. It is also possible to adjust for the limitations. This approach could likely be applied to other patient groups with diagnosis-specific quality registries that correspond to the Norwegian Index for Medical Emergency Assistance, providing a basis for future research.
Analysis of anonymised registry data revealed a substantial increase in EMCC dispatches in which EMCCs suspected stroke from 2020 to 2023, which cannot be explained by an increased incidence of acute stroke in the population. There was no evidence that this increase impacted on the proportion of stroke patients admitted to hospital within four hours of symptom onset, patients treated with thrombolysis, or thrombolysis initiated within three hours of symptom onset. The findings therefore indicate a marked increase in resource use, without a corresponding improvement in key quality indicators for the treatment of stroke patients.