Teaching teamwork during the Neonatal Resuscitation Program: a randomized trial

Abstract

The Neonatal Resuscitation Program (NRP) is the standard curriculum used to teach caregivers how to treat newborns in the delivery room. Worldwide, the NRP could improve outcomes of thousands of newborns per year.1 However, approximately 30% of NRP steps are not performed or performed incorrectly,2, 3 and pediatric residents often fail to intubate infants correctly.4 Neonatal resuscitation should be a team activity that involves at least two people who work together to achieve a shared goal.5 Breakdowns in teamwork may contribute to the quality problems noted above. For example, team behaviors are correlated with the quality of neonatal resuscitation,3 communication breakdowns are root cause of 72% of perinatal deaths and injuries,6 perceptions of effective teamwork are correlated with less burnout and fewer delays in labor and delivery,7 and there is broad consensus from expert groups and researchers that measuring and improving teamwork will help improve the quality of health care.8, 9, 10 However, none of the nine lessons in the NRP textbook1 includes instruction about teamwork, in part because no studies have shown that team training can improve either teamwork or the quality of health care.11, 12 Therefore, we conducted a study with two specific aims: (1) to incorporate teamwork skills and information about human error into the 1-day NRP training program for interns and (2) to randomize interns to this new version of NRP or the standard NRP and measure the effects on teamwork during the simulated resuscitations at the end of the day. We hypothesized that it would be feasible to add the teamwork training to the 1-day NRP course, and that interns who were randomized to the NRP course with team training would exhibit more teamwork behaviors during the simulated resuscitation than would the interns in the standard NRP course. All interns in pediatrics, combined internal medicine and pediatrics, family medicine, and obstetrics and gynecology who began training in June 2005 were eligible for the study. The study was approved by the IRB, and consenting subjects were randomly assigned (generated using a random number generator by ALW) to attend the standard NRP course or the NRP course with additional instruction in teamwork skills and human error. The course was conducted during their orientation in June. The following team behaviors were taught: inquiry, information sharing, assertion, evaluation of plans, workload management and vigilance (Figure 1). These behaviors were chosen and defined based upon focus groups with neonatal physicians and nurses, observations of actual neonatal resuscitations3, 13, 14 and behaviors found to prevent and manage error in commercial aviation.9, 15 The behaviors were defined during the lecture, they were illustrated during the role play and video clips, discussed during question and answer sessions, and instructors reminded students to use them during the skills stations. Observer training and reliability testing occurred in three steps over a 5-month period using video recordings from an earlier study.3 First, observers were oriented by an investigator (EJT) and they viewed a training video that demonstrated examples of teamwork behaviors during neonatal resuscitations. Second, the investigator and two observers independently viewed two resuscitations and then met to discuss differences in their observations of teamwork behaviors and to clarify definitions. Third, a total of 28 resuscitations were scored, divided among five phases. At the end of each phase, we calculated a Cohen's for each teamwork behavior, and the raters and investigator discussed ways to clarify the definitions of team behaviors and to improve reliability. s for the final phase of training ranged from moderate to excellent (workload management (=0.54), evaluation of plans (=0.57), vigilance (=0.59), information sharing (=0.66), inquiry (=0.82) and assertion (=0.87)). When viewing a video recording, the observers indicated the number of times each behavior was exhibited. Episodes of inquiry, information sharing, assertion or evaluation of plans were noted based upon verbalizations of team members. Vigilance and workload management were measured as percent time in that particular state. Teams were considered to be non-vigilant when any team member stopped watching the resuscitation for more than 3 s. Workload management assessed the appropriate distribution of tasks during the procedure. Teams were without workload management if any team member did not offer to assist with the resuscitation when the leader was handling two or more tasks simultaneously (for example, if the leader was performing bag-mask ventilation and trying to measure heart rate simultaneously). Of the 51 eligible subjects, 29 were from pediatrics and combined pediatrics-internal medicine, 12 from family medicine and 10 from obstetrics and gynecology. Fifty were asked to participate (one was not in town when consent was obtained) and there were five non-consenters from pediatrics and five from obstetrics and gynecology, so 40 interns were randomized. There were no differences in the distribution of types of interns in the two groups. The intervention group had 11 pediatric interns, 6 family medicine and 2 obstetrics and gynecology; the control group had 12, 6 and 3, respectively. Seven interns from the control group had an incomplete video recording because the camera ran out of tape or the camera was started after the event began (there was only one person responsible for monitoring six cameras). One intern from the intervention group did not take a turn as a team leader. This left 17 intervention and 15 control interns in the final analysis (Figure 2). The interns in the NRP with team training group exhibited more frequent information sharing, inquiry and assertion than the interns in the control group (Figure 3). Evaluation of plans was observed once in the NRP with team training group. It was not observed in...