The two focus groups featured a total of ten participants (participant nos. 1 to 10), and six individual interviews were conducted (participant nos. 11 to 16). Table 1 presents the characteristics of the participants. The first focus group lasted 1 h 25 min, and the second lasted 1 h 50 min. The single motive indicated for nonparticipation pertained to scheduling issues. Individual interviews were conducted, which lasted from 30 to 54 min. Table 2 provides an overview of thematic categories and representative quotations for the deductive and inductive coding.
The qualitative analysis comprised 203 codes in the focus groups and 206 codes in the individual interviews, which were divided into five groups based on the components of reflection in the Nguyen model: thoughts and actions; attentive, critical, exploratory, and iterative process (ACEI); underlying conceptual frames component; change and self (see Table 3).
The “thoughts and actions” component
Participants felt the need to train themselves to read ECGs based on their upcoming confrontation with ECGs during the internship and their fear of error.
“So I said to myself that for the emergency unit internship, I had to reread it beforehand because if something happens in the middle of the night, I would finally be able to handle the situation” (participant no. 12).
“About the ECG, since it’s normal, I don’t do anything, but obviously you reassure [the patient]. Of course, right? Now [during the internship], obviously, you always reassure people, you don’t even realize it, actually” (participant no. 3).
The ACEI component
Many codes derived from the analysis of the interview transcripts correspond to an attentive, critical, exploratory, and iterative thought process:
“I would put myself in the situation; I’d tell myself, I’m the doctor, this patient is my patient, and yeah, I was truly totally in the case” (attentive interaction – participant no. 8);
“As I saw that I was missing something each time, I tried to be more systematic and to read them properly” (critical interaction – participant no. 12),
“I found it interesting that normal ECGs were included as well, especially to make us think about the limits of normality, which is something we do not see enough of in training courses” (exploratory interaction – participant no. 11).
“Well, I adjusted to the answers, anyway – for instance, about reassurance, there was a first case where I hadn’t written it down, and in the answer, he put reassurance, and so in the last case where it was anxiety and all that, I put reassurance” (iterative interaction – participant no. 9).
Some aspects could be improved to foster the process of reflection. In particular, the long duration and cognitive overload may have had negative impacts on learning: “In fact, it was very long, and at one point I could not manage; I was confusing all the ECGs I saw” (attentive interaction – participant no. 12) – “I believe that the training is planned to last for three hours; I find that three hours, I find that it is tight if we want to pause on the resources to go deeper” (exploratory interaction – participant no. 11).
Critical interaction was sometimes limited by the difficulty of understanding the panel answer: “when it is a little too technical [the cardiologist interpretation of the ECG], I can quite easily let it go and say to myself that it is too much for me” (critical interaction – participant no. 15).
The change component
Several utterances made by participants illustrated a desire to change their thought processes: “for instance, you see [atrial fibrillation] […] I told myself, ‘Okay, we have the right to handle that ourselves’” (participant no. 1).
“I made sure to do things a little bit more seriously, because, you know, there are actually important things to monitor, so yeah, that and being thorough when I read my ECGs” (participant no. 4) – “Same here, yeah, it’s true that we often tend to just glance at the thing; maybe we don’t necessarily, you know [laughs], do it as well as we did when we were students” (participant no. 7) – “do a better job of reading the ECG” (participant no. 5).
The underlying conceptual frames component
Our analysis showed elements in the individual interviews that were related to the underlying conceptual frames of the participants. Participants noted that the program caused them to consider their clinical reasoning processes in light of the clinical cases. They focused on their nonanalytical clinical reasoning strategies: “it is true that for hyperkalemia, it is a typical picture to recognize directly without reasoning too much like that for hyperkalemia; I saw it is typical, and I knew what he had/in that sense it is easier” (participant no. 14). The more typical the clinical picture was, the more oriented the ECG reading was. The participants sought specific abnormalities on the ECG and then ruled out some hypotheses based on their findings. These nonanalytical clinical reasoning strategies were often the source of their errors. Therefore, participants decided to rely more on analytical clinical reasoning strategies: “To interpret […] the first ones I did like that [a global interpretation], in fact, I checked whether there was something obvious afterward as I saw that I was missing something each time. I tried to be more systematic and to read them well” (participant no. 12).
Regarding analytical clinical reasoning strategies, some participants used a systematic reading method: “I remember in the cardio teaching when I was a graduate student, he [the teacher] said that you had to look at all the [ECG leads] because you could be surprised, and it pushed me back to looking properly to be more regular [during the LbC ECG training]” (participant no. 12). In this type of reading, the participants relied on their clinical scripts: “here, there was precordialgia = thoracic pain = PIED pericarditis-infarction-pulmonary embolism-dissection; [there was a] ST segment anomaly, so it can only be infarction or pericarditis” (participant no. 16).
When participants faced issues related to clinical reasoning, they used resources during the training to develop or create new prototypes and clinical scripts: “When I saw the answer to each question, I looked again on the internet […] I looked to see whether there were more examples of the same type to have additional images in head” (participant no. 13). Some participants kept these resources to foster a change in their future practice. The main limitation to such change mentioned by participants was the lack of explicitness in the interpretation of the panelists: “I thought it was good to have the opinion of the cardiologist or the general practitioner each time, but I found that it was not always detailed in relation to how we see that it is a Bouveret or a Brugada; I found that it was a little lacking [in details]” (participant no. 13).
Finally, clinical reasoning strategies were also guided by emotional factors and uncertainty: “In this situation, for example, I would send [hesitates a lot] […] it would have been about the complaint […] I would have sent him to the emergency room, with regard to the ECG; I do not know […] We are still worried” (participant no. 12).
The self component
The participants easily connected their thoughts to their ‘selves’ both during and after the training program: “My way of reasoning is fairly logical compared to the others; I’m a little bit reassured, and I’ve gained a little bit of trust in myself” (participant no. 3). This approach may have encouraged some of them to use the ECG in their future practice: “honestly, I didn’t think I would need to read ECGs afterward [after finishing his or her studies] but in fact there are plenty of GPs who have an ECG, and it’s not bad; it allows a first screening” (participant no. 12).
The absence of grades and the purely learning-oriented goal of the training program was important with regard to this component: “I think it isn’t bad doing it like this because in the end, when I did it, I wasn’t under any pressure, and so you know, you’re not stressed out. When you’re being graded or evaluated, you get stressed out, whereas this is truly just for us” (participant no. 5).
Inductive coding revealed interesting thoughts that emerged in this LbC format concerning the impact of the identity of the panel members, the absence of a scoring system and the question of uncertainty in ECG reading.
Participants were unsure of what value they should attribute to the answers offered. This uncertainty seems to foster a critical distance from the role model that recognized panel members could provide: “they tell us cardiology specialists, but we don’t know who they are at all [in the ECG LbC training program panel]” (participant no. 2) – “The thing is, we know cardiologists, who… you know… [do a bad job] […] so, people we don’t trust completely [the cardiologists panel]”. These parts were also coded as critical thoughts in the ACEI component.
The absence of a scoring system and the assessment of the learning goal of the training program was important to the participants: “I think it isn’t bad [the ECG LbC training program] doing it like this, because in the end, when I did it, I wasn’t under any pressure, and so you know, you’re not stressed out. When you’re being graded or evaluated, you get stressed out, whereas this is truly just for us” (participant no. 5). This part was also coded as an attentive thought associated with the ACEI component.
Uncertainty is often a source of stress in the context of ECG interpretation in real-life situations for participants and a potential obstacle to the use of ECGs in their future practice. The participants highlighted three main elements of this LbC program that facilitated more effective management of this uncertainty. First, the existence of a panel of experts, whose interpretations varied slightly but which led in all cases to an adapted management of the situation: “they do not all have the same opinion, and that does not mean that if we do not all do the same […] it does not necessarily mean that we’re wrong. I found that these moments when there were different opinions among the cardiologists as well as among the general practitioners were a little less demeaning than the MCQ [multiple choice question] methods, which are it is right or it is wrong and that’s it” (participant no. 11).
Second, situations featuring normal or subnormal ECGs are an important source of uncertainty for the participants, and they felt the need to train on these situations: “finally, when I’m least sure of myself is when it is normal/and in that sense, it would be interesting to do more subnormal or physiological ECG” (participant no. 15).
Finally, participants emphasized the use of transversal skills to manage a patient situation based on the suggestions made by the general practitioners panel: “it allowed me to better conceptualize the call to the cardiologist, for example, to better understand when we let the patient make an appointment or when I must call. It allowed me to touch upon this notion of temporality, a short term, a long term” (participant no. 11).