This study was approved by the institutional review board of the National Institute of Health and Medical Research (reference IRB00003888). All data were collected anonymously, and no personal identifiers were recorded at any stage. Informed consent was obtained from all participants prior to study participation. Participation was voluntary, and students could decline or withdraw at any time without any academic or clinical consequences. No financial or other compensation was provided for participation.

This study was conducted in hospitals affiliated with the Faculty of Health, Sorbonne University, one of the medical schools in the Paris area.

All third-year medical students (General Medical Sciences Diploma) completing a 3-month clinical rotation in pulmonology were eligible for inclusion if they performed their first arterial puncture on a patient and had not previously received simulation-based training for this procedure. Each eligible student was invited to complete the online questionnaire within 24 hours of performing the procedure (Multimedia Appendix 1). At the time of rotation assignment, students selecting a pulmonology placement were verbally informed that they would be invited to complete a short online questionnaire after performing their first arterial puncture. This information was reiterated at the beginning and end of the rotation and again during the end-of-semester pulmonology examination. At the start of each rotation, senior physicians in the participating departments were informed of the study via email and provided with a link to the questionnaire for distribution to their students. Data were collected using an online questionnaire specifically developed for this study and hosted on the SurveyMonkey platform (SurveyMonkey Inc; Multimedia Appendix 1). The initial draft was developed by the study investigators and reviewed for face and content validity by a panel of 5 academic pulmonologists and medical educators from Sorbonne University. The instrument was pilot-tested among medical students before implementation.

Students in group 2 received the usual teaching routinely provided in the departments to which they had been assigned. Teaching practices varied and could include bedside instruction and/or verbal explanations given before entering the patient’s room. Instruction could be provided by a nurse, a senior medical student, a resident, or an attending physician. In some cases, students performed the procedure after first observing it on another patient. No intervention was introduced to modify these usual teaching practices.

Evaluation criteria included several dimensions of student and patient experience. For students, the following variables were assessed: global anticipatory apprehension, apprehension about causing pain, apprehension about procedural failure, satisfaction with the procedure, apprehension about repeating the procedure, and confidence in performing it again. The decision to use topical anesthesia was also recorded, as well as the success rate of the arterial puncture. For patients, perceptions of the procedure were evaluated through 2 measures: anticipatory apprehension and pain experienced during the puncture. Both students and patients rated their perceptions using a 10-cm visual analogue scale (VAS) ranging from 0 (“none at all”) to 10 (“maximum imaginable”).

Statistical analyses were performed using Prism (GraphPad Software). Categorical variables were reported as counts and percentages, whereas continuous variables were presented as means and SDs or medians and IQRs depending on their distribution. Given the limited sample size and the ordinal nature of several variables, nonparametric statistics were used. The primary analysis compared group 1 and group 2. Group comparisons were conducted as follows: (1) categorical variables were compared using the Fisher exact test, and (2) continuous variables were compared using the Mann-Whitney U test. A sensitivity analysis was performed to compare group 1 with the group 2 subgroups: group 2+ (students who received prior training) and group 2− (students who did not receive prior training). All tests were 2 sided, with the type I error rate set at α=.05.

During the study period, 135 third-year medical students were assigned to the participating departments. Of these 135 students, 71 (52.6%) were entirely naive to arterial puncture and performed their first procedure during their rotation (Figure 2). The cohort included 78.9% (56/71) women and 21.1% (15/71) men, with a median age of 21 (IQR 20‐21) years; 90.1% (64/71) were right-handed. A total of 83.1% (59/71) had previously observed an arterial puncture performed by a nurse or physician. All participants began completing the survey immediately after the procedure, but 21.1% (15/71) submitted incomplete responses. These included 16.7% (4/24) of the students from group 1, a total of 16.3% (7/43) from group 2, and 5.6% (4/71) whose group affiliation could not be determined. These cases were excluded from analysis, yielding a final sample of 67 students. Group 1 included 35.8% (24/67) of the students. Among the 43 students in group 2, a total of 19 (44.2%) reported having received no prior training (group 2−), and 24 (55.8%) reported some form of training (group 2+), which typically consisted of theoretical explanations and observation of procedures performed by nurses, senior students, or physicians.

In total, 92.9% (52/56) of the students were accompanied during their first arterial puncture (19/19, 100% in the group 2− subgroup). Attendants were nurses in 52.2% (35/67) of cases, medical trainees in 31.3% (21/67) of cases (including fourth- to sixth-year medical students and residents more advanced in training than the study participants, who were all in their third year), and senior physicians in the remaining 16.4% (11/67) of cases.

No statistically significant differences were observed between group 1 and group 2 in terms of global anticipatory apprehension, apprehension about causing pain, apprehension about procedural failure, satisfaction with the procedure, apprehension about repeating the procedure, or confidence in repeating the procedure (Table 1). However, anticipatory apprehension about causing pain was significantly lower in group 1 compared to group 2− (median 4.0, IQR 2.0‐7.0 vs median 6.4, IQR 4.2‐8.5; P=.04; Table 2 and Figure 3). Conversely, postprocedure satisfaction was significantly higher in group 2− than in group 1 (median 8.2, IQR 2.3‐10.0 vs median 3.6, IQR 1.8‐7.2; P=.03; Table 2 and Figure 3).

Overall, 62.5% (35/56) of the students elected to use anesthesia before performing the arterial puncture. Among them, 45.7% (16/35) used an EMLA 5% patch, 51.4% (18/35) used EMLA 5% cream, and 2.9% (1/35) used subcutaneous lidocaine. Anesthesia use was significantly less frequent in group 1 compared to group 2 (7/20, 35% vs 28/36, 77.8%; P=.003). The use of anesthesia did not lead to statistically significant differences in students’ perceptions or patients’ reported experiences, although procedural pain and self-reported confidence showed nonsignificant trends (Table 3).

The first-attempt success rate was 23.1% (3/13) in group 1 and 48.3% (14/29) in group 2 (P=.18). Within group 2, students with no prior training (group 2−) achieved a notably higher first-attempt success rate than those with some prior training (group 2+): 71.4% (10/14) vs 28.6% (4/14), respectively (P=.06). The overall success rate within the first 3 attempts was 65% (13/20) in group 1 and 80.6% (29/36) in group 2 (P=.20). The median number of attempts required for success was significantly higher in group 1 than in group 2 (median 1, IQR 0.5‐2 vs median 1, IQR 0‐1; P=.02). The number of procedures requiring supervisor intervention after 3 failed attempts was identical in both groups (7/20, 35% and 7/36, 19.4%).

Arterial punctures were performed by students on 61 patients (median age 65, IQR 57‐76 years; n=38, 62.3% women and n=23, 37.7% men). The number of patients (n=61) was lower than the number of participating students because some students performed their first arterial puncture on the same patient for logistical reasons. A total of 26.2% (n=16) of the patients were receiving oral analgesics, primarily paracetamol. Most patients (n=41, 67.2%) had previously undergone multiple radial arterial punctures.

Overall, anticipatory apprehension was rated at a median of 3.5 (IQR 2.0‐6.0) on the VAS, with no statistically significant difference between patients who did not receive anesthesia and those who did (median 4.0, IQR 2.0‐6.2 vs median 3.0, IQR 1.5‐5.7; P=.50). Overall procedural pain was rated at a median of 2.7 (IQR 1.2‐5.4). Pain scores were higher in the group without anesthesia than in the group with anesthesia (median 4.8, IQR 2.1‐6.4 vs median 2.1, IQR 1.2‐4.0), although the difference did not reach statistical significance (P=.08).

In a context where the training provided to French medical students for arterial puncture appears highly variable and generally brief—typically lasting less than 30 minutes—and where unsupervised students without specific training demonstrate frequent procedural errors (Multimedia Appendix 1), our results show that a structured training program can meaningfully influence behavior. Specifically, students who completed the structured program (group 1) were significantly less likely to choose local anesthesia before the procedure compared to those who received only informal or no training (group 2; 7/20, 35% vs 28/36, 77.8%; P=.003) despite having similar levels of self-reported anticipatory apprehension. This occurred even though the rate of successful punctures did not differ significantly between groups. From the patients’ perspective, no statistically significant differences were found in anticipatory apprehension or procedural pain based on anesthesia use. Nonetheless, the difference in pain ratings exceeded the minimal clinically important difference on a VAS—typically estimated at 13 mm (1.3 cm) [18]—suggesting a clinically meaningful effect even in the absence of statistical significance. These findings are consistent with those of prior studies that report limited impact of topical anesthetics and highlight the role of operator anxiety in shaping patient experience [4,5,9].

Recent reviews in respiratory care education highlight that simulation—whether of low or high fidelity—enhances clinical reasoning, adherence to safety standards, and learner engagement [14]. To our knowledge, this study is among the first to examine the impact of structured training on medical students performing their first arterial puncture. While no direct parallels exist in this population, similar findings have been reported in other health care student groups. For example, Collins et al [13] implemented a 5-step curriculum for respiratory care students combining theory, simulation, and peer assessment, which resulted in high interrater performance scores. In nursing education, structured programs using artificial arm simulators have improved procedural success rates and reduced errors [11], and a competency-based model allowed nurses to reach proficiency levels typically associated with physicians [19]. Given these precedents—and the findings from our preliminary observations (Multimedia Appendix 1)—our study focused not on procedural skill acquisition but on the impact of structured training on students’ apprehension and confidence. In other domains, simulation-based training and structured debriefing have been shown to reduce anxiety and enhance self-confidence. For instance, Schmidt et al [20] demonstrated reduced anxiety in students presenting clinical cases after receiving video-based feedback, and Toy et al [21] reported improved confidence and competence following simulation-based training in core invasive procedures. Similar outcomes have been reported with academic detailing, an approach that typically refers to individualized educational outreach for practicing clinicians [22] that can influence clinical decision-making and has also been shown to improve confidence, competence, and critical thinking [23]. In our study, students who completed the structured training program (group 1) reported levels of apprehension and confidence similar to those of students with informal or no training (group 2) both before and after the procedure. While these results might initially appear less favorable, group 1 students were significantly less likely to use local anesthesia (7/20, 35% vs 28/36, 77.8%; P=.003), revealing a notable divergence between self-reported perception and actual behavior. This suggests that the training influenced confidence in ways not fully captured by subjective ratings. Choosing to perform a first arterial puncture without anesthesia may reflect a stronger sense of preparedness despite lingering concerns about causing pain. Supporting this interpretation, group 1 students showed significantly less apprehension about causing pain than students with no prior training (group 2−). Group 1 students also reported lower postprocedure satisfaction than group 2− students, which may reflect more realistic expectations shaped by the training and a narrower gap between anticipation and experience. The lower use of local anesthesia by group 1 students may also be due to the educational content of the program, which did mention the limited effectiveness of topical anesthetics for arterial puncture according to the literature. However, group 1 students were not advised against their use and were taught how to administer them. This leads us to believe that increased self-confidence did play a role in the observed difference, possibly in addition to or in synergy with adherence to the course content.

This study has several limitations. First, the structured training was not delivered by certified simulation instructors as recommended by the French National Authority for Health. This may have affected its quality—particularly the debriefing, widely regarded as the most critical element of simulation-based learning. Second, the educational interventions received by students in both groups were inherently heterogeneous. Teaching practices varied across hospital sites, ranging from bedside instruction by a nurse, senior medical student, resident, or attending physician to situations in which students performed the procedure with minimal prior observation. This design intentionally reflected the authentic diversity of clinical training conditions rather than an artificially standardized environment to assess the added value of a structured program under real-world conditions. Such heterogeneity may have introduced uncontrolled differences in baseline exposure. Future studies could address this by comparing structured training with a standardized “usual practice” curriculum to more precisely isolate the effects of the structured training. Third, we relied on nonvalidated VASs to assess students’ apprehension and confidence. While these were practical and easy to administer, their psychometric properties are uncertain, limiting the reliability and generalizability of the findings. Future studies should consider using validated instruments for anxiety and procedural self-efficacy, as well as integrating frameworks such as entrustable professional activities to better capture learners’ readiness for unsupervised clinical practice. Fourth, contextual and patient-related factors could not be controlled across hospital sites. Differences in patient characteristics or case complexity may have influenced the technical difficulty of arterial puncture and contributed to variations in first-attempt success rates. Additional limitations include the potential demotivating effect of a lengthy questionnaire, which may have contributed to incomplete responses and the absence of data on the time interval between training and the procedure. Although students had continuous access to theoretical materials, we cannot confirm that they reviewed them prior to performing the puncture, which may have attenuated the effects of the training.

Structured training for arterial puncture influences students’ behavior in meaningful ways even when self-reported confidence and apprehension appear unchanged. The reduced reliance on local anesthesia among trained students may reflect greater readiness to take responsibility for a potentially painful procedure. These findings support the value of formalized training and highlight the need for validated tools and long-term evaluation to fully capture its educational and clinical impact. Although exploratory and limited by methodological constraints, this study provides initial evidence that such pedagogical frameworks could be meaningfully integrated into undergraduate medical curricula in pulmonology and related disciplines.