**4. Discussion**

Time is brain when dealing with acute stroke patients. International Stroke Organizations advocate for the development of specific training programs for healthcare professionals, as those who have not been trained specifically on strokes contribute to delay patients' access to adequate therapy [21]. On the other hand, the Stroke Alliance for Europe advocates for a systematic approach to continuous education in EMS as one of their 12 quality care indicators on strokes [22].

In this study, the OTI was well-received, scoring high in satisfaction (Kirkpatrick level 1). It was associated with a knowledge gain for all categories (Kirkpatrick level 2), as observed in similar studies with EMS professionals [23–25]. Participants accurately identified the signs and symptoms of a stroke and became aware of the differences between hemorrhagic and ischemic strokes. On the other hand, very few improvements were observed in recognizing transient ischemic attacks. Despite most EMS professionals (65%) having had previous education on strokes, a lot of heterogeneity was found at baseline for the recognition of very specific signs and symptoms. This training improved their competency, which is consistent with the improvements in knowledge observed by Hsieh et al. [23] in Taiwan, where 48% had previously received training. The window for thrombolytic therapy was only identified by a third of the participants, and half signaled the indications for endovascular treatment. However, these items might be very specific, which could explain the low number of correct answers

in our case, or why a similar study in Dubai [24], received no correct responses for these questions. It should be noted that while diagnostic and comorbidity scales determine the activation of the SCs, only 50.3% and 40.1%, respectively, were familiarized with them, which is similar to the finding of DiBiasio in Rhode Island [25]. An integrative review [26] on the impact of training programs on strokes found that only 1 of 21 courses was taught online. In that single UK-based study (RESPONSE [27]), there was a greater improvement in knowledge compared to our findings (95.6% vs. 82%), while the response rates varied (39% vs. 54%). These di fferences could be explained by the fact that our study managed to include a greater percentage of EMS professionals and that the context of the education was di fferent. Most professionals in RESPONSE were paramedics (55%), who had received 2–5 years of university training (Paramedic Sciences), in contrast to our 90.3% participation of emergency technicians, who had received a 2-year non-university qualification. This could also explain some of the very low scores at baseline when recognizing specific stroke signs and symptoms.

Improving pre-notification systems in SCs is essential for ensuring the allocation of in-hospital resources and for accelerating communication between EMS teams and receiving hospitals [28]. We observed a progressive improvement in the compliance with the pre-notification register in 80% of the items, increasing from an overall 53% compliance score in 2014 to a 73% in 2015–2016, and 83% in 2017–2018 (Kirkpatrick level 3). This increment was greater than in another study performed in the USA [29] that achieved an increase from 60.9% to 77.3%; the last peak of 10% increase in our case could have been due to the start of RACECAT, a clinical trial focused on di fferent transfer approaches for stroke patients. Pre-notifications systems have also been found e ffective in improving in-hospital times for therapy access [30,31]. We observed fewer notifications of TSO, which could be explained by the presence of more cases of awakening strokes in the last period (2017–2018).

New specific scales for LVO patients have been created, but most are still uncommon in EMS [32–34]. Our group developed and validated the RACE scale [12] in 2014, which has received endorsement by international guidelines [35–37]. We documented a grea<sup>t</sup> compliance with the prehospital assessment of LVO patients with the RACE scale, starting at 61% immediately after training (Kirkpatrick level 3), up to 71% at the 1- to 2-year follow-up, and 91% in the 3- to 4-year follow-up. During the last year (i.e., 2018) compliance reached 96.3%, which is similar to that found in a study in Ohio (USA) [38] that reached 100% compliance in recoding RACE scores. The results of studies using other scales for LVO recognition are varied; for example, an Australian study [39] reached 88% notification, while another study involving multiple EMS agencies involved only provided data in 53% of the cases. In our study, support from the EMS directorate and continuous education department, as well as the inclusion of the RACE scores in the EMS clinical register (i.e., SITREM ®), were paramount to achieving these positive long-term results.

Acute strokes are a time-dependent medical emergency where time between the onset of symptoms and treatment is essential. We recorded changes in prehospital care time as overall prehospital care time, which increased by 4.7 min (Kirkpatrick level 4). Additionally, on-site clinical care time increased by 2.8 min. A UK study, PASTA [40], showed an increase in time from assessment to thrombolysis by 8.5 min after a specific training programme for paramedics. Another UK project [41] focused on training at the dispatch center revealed a non-significant 2.8 min reduction in the time between alert activation to the arrival of the ambulance. The benefits of patient assessment using the RACE and the obtained pre-registration data (i.e., vital signs, assessment of stroke severity, and RACE scores) could outweigh the slight increase in the overall prehospital care time.

However, this study has some limitations. First, it was limited to the prehospital setting. Second, the correlation between the RACE score at prehospital assessment and endovascular therapy effectiveness remains unknown. Finally, we have no data on the prognosis and evolution of the stroke patients (i.e., final diagnosis, false positives or negatives, stroke mimics, and reperfusion therapy rates). Future studies should seek to include further in-hospital clinical variables.
