Following the merger, the performance of DC A showed a decrease in under-triage (3.0% vs 4.6%) without any increase in over-triage (70.8% vs 78%), a better sensitivity (86.8% vs 86.0%), and specificity (67.4% vs 48.0%) as a more restrictive use of L&S (39.9% vs 56.5%). DC A was already more efficient than DC B prior the merger but came even better after merging. The only explanation is the impact of the intensive continuous training taking place in this dispatch for many years regarding over and under-triage.

DC B’s performance compared with post-merger performance (A + B) showed a reduction in over-triage (70.8% vs 83.9%) and under-triage (3.0% vs 6.5) and improved specificity (67.4% vs 20.3%). As a result of the over-triage reduction, sensitivity decreased from 91.6% to 86.8%. The use of L&S strongly decreased from 81.4% of all missions to 39.9%.

Under-triage in dispatch may have a negative impact on patients’ safety [13]. Therefore, we can consider the reduced under-triage rate after the merger as an indicator of improved quality of services for the whole population. Over-triage is not harmful for the patients who benefit from it. It may however lead to an excessive use of L&S ambulances running hot for no or only little benefit for the patient [18, 19] and potentially fatal complications for the general population and EMS personnel [17, 20]. Over-triage may also lead to a scarcity of ambulances and endanger patients, as P1 dispatch will not be diverted to another suspected severe patient while P2 and P3 are very often in our system. We can therefore consider the reduction of the over-triage rate post-merger as an indicator not only of improved quality but also safety.

After the merger, the most frequent over-triage cases remained NACA 3 missions and under-triage cases NACA 4 missions. An explanation could be the subjectivity of the NACA score established by paramedics, and secondly, the difficulty for dispatchers, as they triage without visual cues, to differentiate future NACA 3 or 4 missions with their engagement criteria (P1, P2).

There is a lack of consensus on under- and over-triage rates in dispatch science, despite expert recommendations [21]. We should not forget that over-triage may reduce EMS capacities to respond to other patients, while eradication of under-triage is impossible without an increase of over-triage and reduced specificity.

This is due mainly to the heterogeneity of the dispatch system (CBD, MPDS, physician dispatch) and EMS system (two or three tiers), as well as to the absence of consensus on the definition of high- versus low-acuity cases [13]. The study design also plays an important role in benchmarking. Some compare dispatch priority and emergency department (ED) evaluation [22], others dispatch priority, EMS, and ED evaluations [23], and some dispatch priority and EMS evaluation [13], like in this study. As in Dami et al. [13], we decided to use the same methodology and compare only dispatch priorities and EMS field findings using the NACA score. We are deeply convinced that in a three-tier system the accuracy of priority dispatch should be evaluated by the first professional on scene for two reasons. First, it would reduce the impact of elapsed time from dispatch to clinical evaluation to its minimal, and second, an ED evaluation does not take into consideration possible improvement of the patient’s clinical condition due to EMS treatment. As stated in 2015 [13], it is still of prime importance for DCs to publish their results, as this may allow benchmarking and therefore ‘permit reaching of an international consensus on dispatch accuracy’.

This study has some limitations. It is an observational, retrospective study in a specific local setting and, as stated above, not applicable to other dispatch systems or two-tier EMS. The NACA score is known to be described as subjective [24] and not always reproducible [25], and the patient’s condition may change for better or worse while the EMS is on its way, inducing incoherence between the dispatch and EMS findings on scene. Therefore it may not always describe the patient’s clinical state on EMS arrival on site, as a minority of patient worsen during EMS care or transport, wich may overestimate under-traige. A clinical evaluation and score on EMS arrival would be the best to evaluate dispatch’s priority decision. The high rate of missing data post-merger is mainly due to omitted transmission of NACA scores. Those missing NACA scores were probably equally distributed between study groups, therefore mitigating any bias regarding over and undertriage.