**1. Introduction**

Strokes are a time-dependent medical emergency, in which treatment delay negatively influences patient prognosis [1]. For acute ischemic stroke patients with an evolution of less than 4.5 h of evolution, fibrinolysis therapy improves prognosis [2], ye<sup>t</sup> its benefits are limited for the subgroup of patients with large vessel occlusion (LVO) [3]. These patients benefit the most from endovascular thrombectomy, with a therapeutic window of up to 24 h from stroke onset according to multimodal neuroimaging criteria, at an adequate specialized tertiary hospital, which doubles their chances of clinical improvement [4]. For this reason, in recent years, stroke code (SC) systems have been developed to rapidly identify patients with acute stroke, allowing agile transfers to a specialized center [5,6]. This rapid assessment of acute stroke patients is paramount to obtaining the maximum benefits from reperfusion therapies. Thus, emergency medical services (EMS) are essential [7], not only for identifying stroke patients, but also for identifying a subgroup of patients with suspected large vessel occlusion (LVO), who would benefit the most from endovascular treatment [8]. The traditional prehospital assessment scales were developed to detect the typical symptoms of stroke patients [9]. Since then, several new scales for the specific detection of LVO patients have been designed, but few have been validated prospectively in prehospital care [10]. Implementing these new prehospital diagnostic tools as part of SCs is a priority to ensure familiarity with the protocol and to achieve current therapeutic standards [11]. Giving pre-notification of patients to the receiving center is also important to ensure allocation of in-hospital resources and to accelerate diagnostic and therapeutic decision-making through a minimum set of clinical data [12]. Additionally, international guidelines also emphasized the need to prioritize specific training in SCs, diagnostic tools and pre-notification systems for EMS professionals [13–15]. Following these recommendations, our group developed the Rapid Arterial oCclusion Evaluation (RACE) [16] scale for prehospital assessment of patients with a suspected LVO stroke (Figure S1). The RACE scale was validated in 2014, and international guidelines endorsed the RACE scale as a valid tool alongside others [17,18]. For implementation by EMS professionals, an online training intervention (OTI) was designed to update their knowledge on acute stroke recognition and the SC activation circuit, as well as to train them on the administration of the RACE scale.

The aim of this study was to evaluate the impact on an OTI focused on the RACE scoring for EMS professionals based on prehospital SCs in Catalonia from 2014 to 2018.

#### **2. Materials and Methods**

We performed a pre–post intervention study from January 2014 to December 2018 in the Catalonian EMS (prehospital care). This EMS provides care for 7.5 million people, employing more than 4000 professionals, and it activated approximately eight daily SCs before 2014. For this study, we included data from both EMS professionals and stroke patients. All EMS professionals (i.e., emergency technicians, nurses, and physicians) were invited to participate, and all of those who accepted were included, as no exclusion criteria were considered. A non-probabilistic sampling method was used. All clinical records of patients older than 18 years old and classified as primary acute stroke patients upon activation of SCs by the dispatch center were included. Records of patients who were being transferred between hospital settings were excluded.

#### *2.1. Online Training Intervention*

An online training intervention (OTI) was developed to provide 6 h of training through a learning managemen<sup>t</sup> system (i.e., Moodle). The programme comprised four modules: Three theoretical modules that addressed the (a) signs and symptoms of a stroke, (b) stroke treatment, and (c) prehospital managemen<sup>t</sup> of stroke, including the administration of prehospital scales and SC protocol; the final module was practical, and was introduced to address the application of the RACE scale using five clinical scenarios. The contents and evaluation methods considered the recommendations of the Cerebrovascular Disease Master Plan in Catalonia to attain content validity. Additionally, international recommendations from the European and American Stroke Organizations were incorporated into the curriculum. A pilot test was performed with an interprofessional group of 30 individuals that included neurologists and EMS professionals (i.e., physicians, nurses, and emergency technicians) between March and April 2014. The training was accredited by the regional council for the continuous education of healthcare professionals.

The OTI was administered progressively according to the Catalonian healthcare regions. All professionals from the same region participated in the course simultaneously for a 30-day period with on-demand access to the training platform. Fourteen replications of the training program were necessary to cover all regions and professionals. The training was completed by 2830 EMS professionals from May to September 2014. All EMS professionals were supported by forum interactions with the faculty, and additional resources were provided. A collaboration network was established by developing a Facebook group, a Twitter account (@escalaRACE), and a website (www.racescale.org).

#### *2.2. Assessment of the Online Training Programme*

The variables measuring the effectiveness of the OTI with regard to prehospital SCs were categorized using Kirkpatrick's [19,20] model of training evaluation as follows:


### *2.3. Data Collection*

This study was performed at five time-points: (a) Baseline (first quarter of 2014 (Q1)), (b) training intervention (second (Q2) and third quarter (Q3) of 2014), (c) immediate follow-up (fourth quarter of 2014 (Q4)), (d) follow-up after 1–2 years (2015–2016), and (e) follow-up after 3–4 years (2017–2018). The period between 2014 Q1 (baseline) and Q4 (immediate follow-up) was used to pilot the training intervention (March to April 2014); and to train the EMS professionals (May to September 2014).

Data from EMS professionals were obtained on the same learning managemen<sup>t</sup> system (i.e., Moodle) on which the course was provided. Socio-demographics were obtained at the beginning of the intervention (i.e., 2014 Q2 and Q3). Data on Kirkpatrick level 1 was recorded at the end of the intervention (i.e., 2014 Q2 and Q3). Kirkpatrick level 2 data were obtained at the beginning of the intervention (i.e., 2014 Q2 and Q3), and 3 months after the end of the training (2014 Q4). Data on all results from Kirkpatrick levels 3 and 4 were obtained through the Informatic System for Emergency Management (SITREM®) register between 2014 and 2018 (all periods except the intervention, that is, 2014 Q2 and Q3). This register prospectively records information about all Catalonian EMS activity, including details on SC activation, patients, time of call, first time of care, and arrival at the receiving hospital. From September 2014, RACE scores were also included in the register.

The data were processed in compliance with the European Data Protection Regulation 2016/679. The study was approved by the Clinical Research Ethics Committee of the University Hospital Germans Trias i Pujol (Badalona, Spain) with identification code PI-15-030.
