We conducted a randomized cross over study at Maisonneuve Rosemont Hospital, a Level III NICU in Montreal, Canada. The institutional scientific and ethics review board approved the study. Every participant gave written informed consent before enrolment. Participants were informed of their right to discontinue participation at any time.
Participants
Residents were recruited during their pediatrics or neonatology rotation in our hospital. All first-year trainees were approached sequentially by the main investigator to participate during the recruitment period. Trainees with prior experience in live newborn intubation were excluded in order to select novice learners. Students with a history of beta blocker use within the past year, and students with a history of antidepressant medication were excluded. All trainees had completed the neonatal resuscitation program (NRP) training course during which they had the chance to practice NMEI on a mannequin at the start of their rotation block. For our study, participants were told they would perform NMEI and that their performance during two NMEI scenarios would be evaluated in the context of scientific investigation. The true goal of the study was not revealed until the end of the experience. Participants were asked to keep their experience confidential.
Intervention
A full body neonatal mannequin approximating a term newborn (Newborn Anne, Laerdal Medical Corporation, Wappinger Falls, NY) was setup in a hospital delivery room suite on a radiant warmer. Subjects were asked to wear a cardiac monitor, and 5 min later after a period of rest, they were then called to the delivery room. Heart rate was assessed continuously, starting 5 min before NMEI and stopping 5 min after cessation of the procedure, using a wireless heart rate monitor (Polar M400). Five minutes after their arrival to the delivery room, they were given a stylet and a 3.5 ET tube and asked to intubate orally a neonatal mannequin. No practice attempts were allowed. After 30 s, the residents were reminded verbally that they had 15 remaining seconds. Each individual attempt was recorded as the time in seconds from insertion to the removal of the laryngoscope, using a chronometer operated by the main investigator. Success or failure of the attempt was assessed by the investigator after removal of the laryngoscope. If the resident failed the intubation, he was given the laryngoscope and the tube for a subsequent attempt and the chronometer was restarted from the time it was stopped. The duration of all attempts was limited to a total of 45 s. A participant who never succeeded, was assigned an intubation time of 45 s. A maximum time limit of 45 s was chosen, to best approximate the physiological tolerance of a live infant. The NRP recommends limiting each attempt to 30 s. However, previous studies suggest a mean intubation time of 38 s in trainees [8].
In this randomized crossover study, each volunteer participated in two NMEI scenarios under two different conditions, A and B, and served as his/her own control. The sequence, A then B, or B then A, was assigned randomly.
Condition a
Only one staff neonatologist was present with the participant during NMEI.
Condition B
An audience of 5 people were present to watch the participant during NMEI, with at least 2 clearly identified neonatologists, one being the main investigator. The other 3 observers were health care professionals (nurses or other trainees) from the perinatal service. Audience members stood within 2 m of the subject, but remained silent, neutral and did not interact with the trainee.
The residents who performed NMEI in condition A on day one were crossed-over to NMEI the next day under condition B, and vice versa. The time and the location of the 2 procedures were the same.
Primary outcome
We chose time to successful intubation as the primary outcome. We did not use success or failure of intubation as the primary outcome, since we expected all trainees to eventually be able to intubate a static mannequin with no vital signs given enough time. The mean difference in total intubation time between condition A and B was compared.
Secondary outcomes
The increase in heart rate from baseline (5 min before NMEI) to peak (during laryngoscopy) as a percentage value was calculated for each subject, and the mean difference between condition A and B was compared.
The rate of success on first attempt as well as the rate of success across all attempts was reported under each condition. Successful intubation was defined as tube placement below the vocal cords within 45 s.
Sample size
To estimate the required sample size, we used the study by O’Donnel et al. who gave an approximate intubation time for residents of 38 s in live infants, with a standard deviation of 20 s [8]. Using a two tailed alpha threshold of 0.05 and power of 80%, a sample size of 50 subjects was estimated to detect an 8 s difference time between the two different groups. Eight seconds represents a 25% change in time to intubate and is clinically significant for a newborn.
Randomization
A simple randomization scheme was used, and the sequence of envelopes was generated by a random coin flipper (www.random.org). Allocation was revealed by opening a pre-prepared sealed envelope before participants were called to the delivery room. Neither the participants nor the investigator was blinded to allocation. The audience present under condition B did not know if the resident had already performed under condition A.
Statistics
The difference in NMEI time and heart rate increase was calculated between conditions A and B for each candidate, and the mean difference within subjects was compared to the null value of 0 using a two tailed paired t-test. The rate of success of intubation between both conditions was compared with the McNemar chi-squared test. An alpha level of 0.05 was considered significant. All calculations were done with SAS 9.4 (Toronto, ON).