The study was performed in the Paris Sud University simulation center (LabForSIMS) at the University Hospital Bicêtre, France, in a dedicated laboratory with a realistic, simulated delivery room with real medical equipment. The SimNewB™ simulator mannequin (Laerdal, Stavanger, Norway) was used for the study. PPV was provided by a neonatal mask and a T-piece ventilator (Neopuff™ Infant Resuscitator, Fisher & Paykel, Auckland, New Zealand). Sessions were video recorded by two cameras and sound amplified by ambient sound recorders and individual microphones worn by each trainee. According to the French national regulation, this type of study does not require any IRB approval or waiver, since it is not performed on patients' data. However, all trainees gave informed consent to session recordings and their use for scientific purposes.
The training sessions were part of the mandatory teaching of a newborn’s resuscitation for first-year pediatric residents of the Paris region and included a classroom-style course for one day followed by simulation-based training for a half-day (4 hours). The training sessions were organized during each academic year from January to June. The organization of the course was standardized and did not vary during the study period. A group of 9 to 11 students was enrolled in each session, overseen by 3 to 4 instructors who were both experienced in simulation and specialized in neonatal resuscitation. Instructors’ roles were allocated before each scenario: either as debriefer watching the scenario with the observers in the debriefing room, as a computer manager in the control room, or as a scenario facilitator (most often playing the role of the midwife). In case of an available fourth trainer, this trainer would be a co-debriefer and also watched the scenario in the control room.
A 20-minute briefing covering general teaching about European guidelines and the principles of simulation-based training preceded the sessions. The slideshow used during the briefing was overall the same during the whole study period, the only modifications in period 2 concerned one slide showing the duration of step A and another one showing that routine intubation should not be performed for tracheal suction before PPV start for non-vigorous infants born with meconium-stained amniotic fluid. Each session comprised of 5 or 6 scenarios. Each scenario began with a short oral presentation of the medical situation. The scenarios were designed to evaluate a specific educational objective, and all scenarios covered at least step A. A pair of trainees participated in each scenario, and videos were broadcast live in the debriefing room in which the other participants observed. A structured debriefing by trained instructors took place immediately after each scenario.
The same educational progression with specific learning objectives was maintained in all sessions (see Table in Supplemental Digital Content 1, which shows the structure of our sessions). In the first scenario, the baby was born tonic in clear amniotic fluid, and PPV was not required. In the second scenario, the baby was born non-vigorous in clear amniotic fluid, though it was always a relatively easy scenario requiring only mask ventilation. In the third scenario, the baby was born non-tonic in an amniotic meconium fluid. In the fourth and fifth scenarios, the baby was born non-vigorous (requiring PPV) but with increased medical complexity. The sixth one was optional and was not included in the study. To achieve a higher level of reproducibility in running a scenario for the different groups of residents, all the scenarios were preprogrammed.
We included the scenarios which had been performed by first-year pediatric residents during six consecutive academic years (2013-2018). We separated the training sessions into two periods, i.e., before (2013, 2014, 2015) and after (2016, 2017, 2018), the new guidelines were published in October 2015. As the course is organized between January and June for each academic year, 2015 scenarios were included in period 1.
To be able to assess resident performance in the same conditions in both periods, we included only the 2nd and 4th scenarios of each session (see Tables in Supplemental Digital Content 2 and 3, which show scenarios used for the study). We excluded all of the 1st scenarios because it did not require PPV use and all the 5th scenarios as at least one of the two participants had often previously participated in another scenario in the same session. Additionally, we excluded all the 3rd scenarios because step B was different for non-tonic infants born with meconium-stained amniotic fluid in period 1 with routine intubation for tracheal suction before PPV. Even if step B has been similar -- whatever the color of amniotic fluid in the current guidelines -- we also excluded the 3rd scenario in period 2 to limit the bias related to the change.
Because the mannequin could not move from the resuscitation table, the mannequin was covered on the table before birth. The scenario began when the scenario facilitator who played the role of the midwife came into the resuscitation room and removed the cover. The start time of PPV was defined when the resident occluded the T piece for the first time.
Before the study, we reported in a checklist the nine items required during step A for a term newborn according to national and international guidelines. Although the 2010 guidelines did not clearly recommend a method to assess heart rate (HR), we have been teaching the residents since 2012 to evaluate HR by ECG monitoring in our learning center, given the inaccuracy of clinical methods8,9 and the superiority of the ECG versus oximetry10. Since there was difficulty detecting differences between the activities of "stimulating the baby" and "drying the baby," we grouped these items such that eight tasks were evaluated for each scenario for both periods (Table 1).
Table 1:Checklist of initial assessment tasks of Step A as defined by the European and the French Guidelines, completed in by video reviewers for the study.
Date of the session (fill out only after reviewing)
Scenario number:
Name of the scenario:
Reviewer‘s name:
Tasks
|
Before PPV start
|
After PPV start
|
Not performed
|
Apgar clock
(Correct if it is the 1st task executed)
|
|
|
|
Cap
(Correct if the task was finished before PPV start)
|
|
|
|
Drying
Stimulating
(Correct if the task was finished before PPV start)
|
|
|
|
Oro pharyngeal suction
(Correct if the task was finished before PPV start)
|
|
|
|
Nose suction
(Correct if the task was finished before PPV start)
|
|
|
|
Temperature probe
(Correct if the task was finished before PPV start)
|
|
|
|
HR assessment
(3-lead ECG)
(Correct if the task was finished before PPV start)
|
|
|
|
Oximetry sensor
(Correct as long as the task was beginning at PPV start)
|
|
|
|
PPV start
|
Time : sec
|
|
|
For the sake of this study, two instructors reviewed all available videos of the scenarios. They did not know the date of the sessions, only a random assigned number. The instructors assessed each task of step A, and the time of PPV-start. The instructors filled out the checklist described (in table 1). In case of discordance between reviewers, the video was reviewed jointly to reach a consensus. The duration of scenarios and debriefings were evaluated and kept for further analysis.
Results were analyzed using the STATA statistical software (StataCorp LLC, Texas 77845-4512, USA). Gaussian distribution of data was evaluated by Shapiro-Wilk test. The Welsh’s t-test and Pearson’s chi-squared test were used to compare groups when appropriate. For multivariate analysis, linear regression model was used. All tests were two-sided, and a p value < 0.05 was considered significant.