Cardiopulmonary resuscitation (CPR) training strategies in the times of COVID-19: a systematic literature review comparing different training methodologies

Background Traditional, instructor led, in-person training of CPR skills has become more challenging due to COVID-19 pandemic. We compared the learning outcomes of standard in-person CPR training (ST) with alternative methods of training such as hybrid or online-only training (AT) on CPR performance, quality, and knowledge among laypersons with no previous CPR training. Methods We searched PubMed and Google Scholar for relevant articles from January 1995 to May 2020. Covidence was used to review articles by two independent researchers. Effective Public Health Practice Project (EPHPP) Quality Assessment Tool was used to assess quality of the manuscripts. Results Of the 978 articles screened, twenty met the final inclusion criteria. All included studies had an experimental design and moderate to strong global quality rating. The trainees in ST group performed better on calling 911, time to initiate chest compressions, hand placement and chest compression depth. Trainees in AT group performed better in assessing scene safety, calling for help, response time including initiating first rescue breathing, adequate ventilation volume, compression rates, shorter hands-off time, confidence, willingness to perform CPR, ability to follow CPR algorithm, and equivalent or better knowledge retention than standard teaching methodology. Conclusion AT methods of CPR training provide an effective alternative to the standard in-person CPR for large scale public training.


Background
Sudden Cardiac Death (SCD) refers to an unexpected death from cardiac arrest [1]. Worldwide, SCD is the most common cause of death accounting for 17 million deaths every year or 25% of all global mortality [1]. Outof-hospital cardiac arrest (OHCA) is a global health issue with incidence reported as 40.6 per 100,000 person-years in Europe, 47.3 in North America, 45.9 in Asia, and 51.1 in Australia [2][3][4][5].
Decreasing the time to initiation of CPR is crucial for improving outcomes in cases of cardiac arrest [6,7]. This is where the role of the bystanderany layrescuer (non-medical professional) who witnesses a medical emergencycomes into play [8]. In fact, bystander CPR before arrival of EMS is independently associated with up to a threefold increase in survival [9]. Various attempts have been made to increase the number of people trained in CPR and therefore improve bystander CPR rates, including organization of mass CPR training events. These attempts, particularly when backed by effective legislation mandating CPR training, result in significantly more laypersons trained in CPR as demonstrated by efforts led in Norway [10,11], Singapore [12], and Denmark [13].
CPR has traditionally been taught face to face using a mannikin as a proxy for a patient. In 2015, the American Heart Association introduced the concept of blended learning that involved the use of online videos and simulated Voice Assisted Mannikins to replace instructors. CPR self-instruction through video-and/or computerbased modules paired with hands-on practice may be an effective alternative to instructor-led courses and such technologies can be utilized more easily to facilitate safe and effective learning [14,15]. This has become particularly relevant now that the COVID-19 pandemic, where wide spread restrictions on in-class training and potential risk of virus spread during face-to-face sessions, has caused organizations to reconsider how trainings are allowed to be conducted [16,17].
The aim of this systematic review is to compare the learning outcomes between standard instructor-led classroom-based CPR training with the alternative training methods among laypersons.

Study design
A research question was identified using the PICO strategy (Population (P): laypersons not trained in CPR, Intervention (I): alternative CPR training methodologies, Comparison (C): standard CPR training methodology, Outcome (O): CPR knowledge, quality, and skill performance). After establishing the research domain, inclusion and exclusion criteria were established to identify and select relevant articles. After assessing the quality of the studies included, data was extracted, organized, summarized, and charted accordingly. The results were analyzed and reported. The primary research question guiding this review is: "What are the differences in CPR knowledge, skill performance, and quality in laypersons receiving alternative CPR training when compared to standard training methodology?"

Search strategies
We searched PubMed or Medical Literature Analysis and Retrieval System Online (Medline), and Google Scholar for relevant articles from January 1995 to May 2020. Medical subject headings (MeSH) were searched using Boolean operators "OR/AND". The search terms were: ("hands-only CPR" OR "cardiopulmonary resuscitation" OR "CPR") AND ("teaching methodologies" OR "training methods") AND ("medical students" OR "bystanders" OR "laypersons" OR "health-care workers" OR "school children" OR "physicians" OR "nurses" OR "paramedical staff" OR "technicians").

Inclusion and exclusion criteria
We included studies which compared two or more CPR training methodologies targeting laypersons with no previous CPR training. Studies describing a single methodology with no comparison group were excluded as were the case reports, case series, and non-English articles.

Identification and selection of studies
The studies were selected after two stages of screening. Two researchers (DMA and BH) independently, extracted data. In stage 1, we screened the article titles and abstracts and those which matched the inclusion criteria were selected for full text review. In the final stage, we reviewed full texts of the articles and determined their inclusion in this review. Any conflicts between researchers during the article screening process was resolved by the senior researcher (JR). Data was organized using a simple database on Microsoft Excel. Figure 1 presents a Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram showing the process of searching and selecting the research articles.

Data extraction from included studies
After article selection, we extracted and recorded data in a data extraction form in an excel spreadsheet. The domains in the data extraction form were: year and country of publication, intervention tested, study design, sample size, study population, presence of prior training, outcome measures, and key findings.

Quality assessment of studies
The quality was assessed using the Effective Public Health Practice Project (EPHPP) Quality Assessment Tool [18]. Two researchers (DMA and BH) reviewed each study using EPHPP. The results of the quality assessment are summarized in Table 2. No studies were excluded on the basis of quality assessment, as this quantitative evidence synthesis aimed to include all articles relevant to our review question.

Summarizing the findings
We summarized our findings into the following research domains: standard instructor-led classroombased CPR training, non-standard face to face CPR training, hybrid CPR training, and online CPR training.

Definition of terms
The definitions of commonly used terminologies in this article are detailed in Table 1.

Studies' characteristics
A total of 978 articles were retrieved from PubMed and Google Scholar. Four hundred and twenty duplicate articles were excluded. Out of the remaining 558 articles, 537 articles were either not comparing different teaching methodologies, were case reports or case series, or were written in a language other than English, and no English translation was available and therefore were excluded. Among the remaining 21 articles, 1 articles did not have available full texts. Twenty full-text articles were reviewed and included in this study. Out of these twenty articles, ten had a moderate global rating, while ten had a strong global rating based on Effective Public Health Practice Project (EPHPP) Quality Assessment Tool (Table 2).

Research domains
Among the twenty studies included in this review, eleven compared online CPR training with the standard training, six studies compared non-standard face to face CPR training with the standard training, and three studies compared the standard CPR training with hybrid training methodologies. Among the included studies, fourteen studies were randomized controlled trials, two had an interventional study design, two were cluster randomized controlled trials, and two studies had a casecontrol study design. The study population comprised of school children, laypersons, medical students, and nursing students. The details of individual studies are summarized in Table 3.

Characteristics of different CPR training methodologies
The CPR training methodologies were divided into two broad categories including standard instructorled classroom-based CPR training and alternative CPR training. The alternative CPR teaching methodology was further classified as non-standard face to face CPR training, hybrid CPR training, and online CPR training. The comparison of content, duration, mode of delivery, standard of content, and measured outcomes between different training methodologies are detailed in Fig. 2 and Table 4. Significant difference was noted between the duration of the teaching methods. The studies reported a longer duration of standard CPR training (20 min to 6 h) when compared to non-standard face to face (45 min to 3 h), hybrid (4 min to 1.5 h), and online CPR training methods (1 min to 1.5 h). Moreover, variability was also noted in the standard of content taught between different training methods and within each training method as well. Although "Einlebenretten" ("save one life") educational framework [20] and European Resuscitation Council (ERC) 2010 guidelines [21,34] were the two contents similar between standard and non-standard face to face CPR training, the standard training group also used contents from ERC 2005 guidelines [34], American Heart Association (AHA) Heartsaver Citizen CPR course [27,28,31,38], AHA 2010 guidelines [25], National Safety Council Adult CPR training program [29], HeartCode BLS course [33], Dutch Resuscitation Council course [37], and Danish Red Cross course [36]. Although the computer-based HeartCode BLS course [33] and National Center for Early Defibrillation course [27] were similar between hybrid and online CPR training methodology, the standard of content was also adopted from other sources in these instructional methods. The hybrid teaching methodology had contents from Japanese Red Cross Society [25] and AHA 2010 guidelines [25], while online training method adopted content from National Safety Council Adult CPR training program [29] and Tryg-Fonden foundation [36]. The content (CPR, ventilation, and breathing) and outcomes measured (CPR performance, quality, knowledge, attitude, self-   Three months after initial training, the video-reminded group showed more accurate airway opening (P < 0.001), breathing check (P < 0.001), first rescue breathing (P = 0.004), and hand positioning (P = 0.004) than controls. They also showed significantly higher selfassessed CPR confidence scores and increased willingness to perform bystander CPR in cardiac arrest than the controls at 3 months (P < 0.001 and P = 0.024, respectively). confidence, and willingness to perform CPR) were similar between the training methodologies.

Comparison of outcomes between different training methodologies
The studies assessed three main outcomes after CPR training which included CPR skill performance, CPR quality, and CPR knowledge. The difference in each outcome was compared between the standard instructor-led classroom-based CPR training and alternative (nonstandard face to face, hybrid, and online) CPR training methodologies. The detailed description of these differences is illustrated in Table 5.

Standard versus non-standard face to face CPR training
The non-standard face to face CPR training included simplified (hands-only) CPR, peer-based CPR training, Jigsaw model CPR training, flowchart-supplemented CPR training, and a multi-staged approach to CPR training. Out of the twenty studies, five randomized controlled trials and one prospective case-control study fell under this domain. Two studies compared CPR performance and one study compared CPR quality. More than one outcome was measured by three studies in which one study compared CPR performance and quality, one study compared CPR quality and knowledge, and one study compared CPR performance and knowledge between the instructional methods.
In CPR quality, the simplified CPR group performed better on CPR algorithm (p < 0.01), had higher number and adequate chest compressions (p < 0.01), and shorter hands-off time (p < 0.001) when compared with the standard training group [19]. Although the flowchart-supplemented group showed shorter handsoff time (147 ± 30s vs. 169 ± 55 s, p = 0.024), the time to chest compression was longer (60 ± 24 s vs. 23 ± 18 s, p < 0.0001) as compared to the instructor-led group [22]. The staged CPR group had better "shout for help" (p = 0.02 to p < 0.01) and more adequate compressions (p = 0.05 to p = 0.04) when compared to standard training [23].
Although better CPR knowledge retention was seen in the multi-staged approach CPR training when compared to the standard group [23], no difference in retention was seen between peer-assisted (61.76 ± 17.80) and professional instructor groups (60.78 ± 39.77, p = 0.848) [24].

Standard versus hybrid CPR training
The hybrid CPR training included a kiosk group, an interactive computer training group plus an instructorled training group, and a video learning group followed by hands-on CPR training. Three studies fell under this domain. One study compared CPR quality, while one study compared CPR performance and quality and one study compared CPR performance and knowledge between the instructional methods. In CPR performance, although the kiosk group outperformed the instructor-led group on hand placement (+ 4.9), they scored lower on compression depth (− 5.6) [26]. Moreover, for all outcome measures, mean scores were higher in the interactive-computer training group plus instructor-led practice group when compared to the instructor-led group [27].
In terms of CPR quality, no significant difference was noted in time to first chest compression (33 s vs. 31 s, U = 1171, p = 0.73) and number of total chest compressions (101.5 vs. 104, U = 1083, p = 0.75) between the instructor-led group and flipped learning group [25].
Furthermore, the kiosk group and the instructor-led group had similar total scores after training [26].
Lastly, use of a computer program resulted in higher knowledge retention (80%) as compared to the instructor-led group (75%) two days after training [27].

Standard versus online CPR training
The online CPR training methodology included video self-instruction, interactive computerized module with video, mobile phone video clips, a computer-based course with Voice Advisory Mannequin (VAM), and virtual reality CPR training. Eleven studies fell under this  3. Flowchart group was more confident than non-flowchart group (7 ± 2 vs. 5 ± 2, p = 0.0009) [22].
2. For all outcome measures, mean scores were higher in the interactive-computer training plus instructor-led practice as compared to the instructor-led group [27].  domain. Five studies compared CPR performance, two studies compared CPR quality, and one study compared CPR knowledge between the instructional methods. More than one outcome was compared by three studies in which, two studies compared CPR performance and knowledge while one study compared CPR quality and knowledge between instructional methods. In CPR performance, video self-instruction group had superior overall performance scores with only 19% non-competent trainees as compared to 43% non-competent trainees in the instructor-led group [28]. Moreover, another study also reported similar findings in which, 40% of the video self-instruction group were competent when compared to only 16% competency in the instructor-led group [31]. The group which received video-based training also had more accurate airway opening (p < 0.001), breathing check (p < 0.001), first rescue breathing (p = 0.004), hand positioning (p = 0.004), and higher confidence and willingness to perform CPR at 3 months when compared to the instructor-led group [32]. Furthermore, another study showed that the video-based group performed better scene safety (95.2% vs. 76.1%, p < 0.05), call for help (97.6% vs. 76.1%, p < 0.05), and had shorter response to compression time (35 ± 9 s vs. 54 ± 14 s, p < 0.001) as compared to the standard instructor-based group [35]. A study in United States showed higher overall performance (60% vs. 42%), appropriate responsiveness assessment (90% vs. 72%), adequate ventilation volume (61% vs. 40%), and correct hand placement (80% vs. 68%) in the video group as compared to instructor-led training [38]. However, one study reported lower compression depth scores (− 9.9) [26] while another study had lower scores in calling 911 (71% vs. 82%) [38] in the video group as compared to the instructor-led group. Voice Advisory Mannequin (VAM) feedback was another methodology adopted for online training in one of the studies and those participants trained using this method had more correct hand position (73% vs. 37%, p = 0.014) and better compression rate (124 vs 135, p = 0.089) than the instructor-led group [34]. A study in Netherlands compared standard instructorled training with Virtual Reality (VR) CPR teaching methodology. Although the VR group had better chest compression rates (114/min vs. 109/min) and proportion of compressions with full release (98% vs. 88%, p = 0.002), the instructor-led group had higher overall scores (12 vs. 10, p < 0.001), better chest compression depth (57 mm vs. 49 mm), adequate chest compression fraction (67% vs. 61%, p < 0.001), higher proportion of participants fulfilling depth (75% vs. 51%, p < 0.001), and rate requirements (63% vs. 50%, p = 0.01) [37].

CPR
In CPR quality, the instructor-led training group had better quality of CPR compressions (location, rate, depth, and release) as compared to the computer-based training group [29]. Moreover, the chest compression depth was also better in the instructor-led group when compared to the group trained using brief videos [30]. Although the VAM feedback group showed similar compression rates, they had more compressions with adequate depth and hand placement, and had more ventilations with adequate volume than the instructorled group [33].
Although some studies showed no significant difference in the CPR-related knowledge scores between the instructional methods [29,31], other studies highlighted significant differences. A study in Denmark highlighted that after 3 months, although the DVDbased group had better average inflation volume (844 ml vs. 524 ml, p = 0.006) and chest compression depth (45 mm vs. 39 mm, p = 0.005), the instructor-led group was superior in assessment of breathing (91% vs. 72%) [36]. At 2 months post-training, another study illustrated that although the video group had higher scores in overall  [38].

Discussion
This is a comprehensive systematic review that compares CPR performance, quality, and knowledge between different teaching methodologies including standard instructor-led, non-standard face to face, hybrid, and online CPR trainings. This review includes 20 studies and 5961 participants and illustrates significant differences in both the characteristics and the outcomes between the instructional methodologies. All the included articles had an experimental study design and had a moderate or strong global rating based on our quality assessment tool. Our results suggested that the standard instructor-led CPR training had a longer duration (20 min to 6 h) as compared to alternative CPR trainings (1 min to 3 h). Moreover, the standard of content also varied significantly between the instructional methods. Interestingly, our review also showed variability in the content within the standard instructorled CPR training methodology in which the teaching material was adopted from multiple sources including "Einlebenretten" ("save one life") educational framework [20], European Resuscitation Council (ERC) 2005 and 2010 guidelines [21,34], American Heart Association (AHA) Heartsaver Citizen CPR course [27,28,31,38], AHA 2010 guidelines [25], National Safety Council Adult CPR training program [29], HeartCode BLS course [33], Dutch Resuscitation Council course [37], and Danish Red Cross course [36].
The instructional methods were compared on the basis of CPR performance, quality, and knowledge which were the three primary outcomes of the studies. In CPR performance, when compared to the standard instructor-led CPR training, the non-standard face to face CPR trained group were although more confident in performing CPR [22], similar performance was seen in the peer-led [20,24] and the jigsaw model groups [21]. Although the hybrid CPR training methodology led to higher overall performance scores including better hand placement, the instructor-led methodology outperformed on the chest compression depth scores [26,27]. When compared to standard CPR training, online instructional methodology not only resulted in a higher percentage of competent trainees [28,31], but it also resulted in more performance of scene safety, assessing responsiveness, calling for help, accurate airway opening, breathing check, first rescue breathing, adequate ventilation volume, shorter response to compression time, hand positioning, better compression rates, and higher confidence and willingness to perform CPR [32,34,35,38]. However, instructor-led trainings had higher compression depth scores and higher scores in calling 911 when compared to online CPR training [26,38]. With regards to CPR quality, the nonstandard face to face CPR training methodology outperformed in the CPR algorithm, had higher "shout for help" rates, had better rate and quality of compressions, and had shorter hands-off time when compared with the standard training [19,22,23]. However, instructor-led groups took less time to start chest compressions [22]. The hybrid training groups and the instructor-led groups showed no statistically significant difference in the total obtained scores regarding CPR quality [25,26]. When compared to standard CPR training, online instructional methods showed better hand position, better chest compression rates, shorter hands-off time, and more frequency of calling for help [29,30,39]. However, correct hand placement and adequate depth of chest compression was better in the instructor-led group [26,30]. Lastly, when compared to standard CPR training, alternative instructional methods either had similar [24,29,31] or better knowledge retention [23,27,36,38].
The results of our study can be explained by certain determining factors. Due to access to better technology and readily available training material nowadays, numerous alternative training methodologies are being tested and compared with the standard training to assess their efficacy. This constant testing and repetition of training results in constant improvement in these alternate training methodologies resulting in better outcomes among participants. However, the quality of CPR, particularly the adequacy of chest compressions, is still better among instructor-led group as technology to effectively monitor chest compression depth remotely is not widely available currently.
Our systematic review has certain implications. First, since the studies included in this review had a moderate or strong global rating, comparisons made between standard and alternative CPR instructional methods can be used for future trainings. Secondly, standard CPR training is resource intensive driven by availability of instructors and therefore has limited scalability. This is especially true in low resource settings where creating an organizational structure and large cadre of instructors to deliver courses may take longer times and require more resources. Our study highlights the feasibility of utilizing instructional technologies and also recognizes the shortcomings of using technology-only solutions. Thirdly, "standard" CPR training had significant variability in both the duration and the standard of content among different studies. It is important to create standards so that future methodologies can be measured and further innovative solutions can be developed. Given the risk of infection spread due to pandemics such as COVID-19, we believe that alternative to face-to-face teaching methodologies have significant promise and can be implemented safely and effectively to increase the rate and effectiveness of bystander CPR and in turn save more lives by strengthening the first component of the chain of survival. Future alternatives to face-to-face instruction including possibly remote monitoring of students may improve correct hand placement and adequate depth of chest compression.

Limitations of the study
This article has some limitations. Most of the studies included in this review were conducted in developed countries and therefore, effective adaptability of alternate training methods in the local setting cannot be ascertained. Moreover, no study looked at CPR performance during an actual cardiac arrest event and none of the conducted studies measured the impact of different teaching methodologies on a population level. Furthermore, potential bias towards a particular CPR teaching methodology among trainers cannot be ruled out. Lastly, since no uniformity existed in the duration and content of standard CPR training, the outcomes cannot be compared with alternate training methods concretely enough.

Conclusion
This review outlines that alternative CPR training methodologies are as effective or even possibly better when compared to standard in-person classroom CPR training in CPR performance and knowledge acquisition. However, effective CPR quality still largely depends on some in-person training. Due to promising results seen in alternate training methodologies and non-uniformity seen in standard instructional techniques, these instructional methods can be adopted as an alternative, particularly during this time of the COVID-19 pandemic. Moreover, future research should aim to develop uniformity in standard CPR training methodology, which will make comparison with alternative CPR instructional techniques more plausible.