Study design, setting and population
We studied ETI performance by third-year allopathic MD and combined MD/graduate degree (PhD, MPH, etc.) students completing a required two-week anesthesiology clerkship at the University of Pittsburgh School of Medicine. At the time of this study, the existing clinical anesthesiology clerkship included ‘hands-on’ experience with ETI and other anesthesia procedures. This study evaluated results of the existing educational curriculum; there were no changes to the educational program, procedures or clinical interventions. Because of the observational nature of this study, the University of Pittsburgh Institutional Review Board approved the study without requiring informed consent from participating students or patients.
Components of the educational curriculum
The medical student anesthesiology clerkship was an existing component of a required eight-week combined surgery and anesthesiology clerkship. The structured clerkship curriculum consisted of assigned readings and human-simulator-based training exercises as well as participation in clinical anesthesia care.
On the first day of the rotation, all students received basic instruction in airway management procedures, including human-simulator-based training in mask ventilation, laryngeal mask airway insertion and ETI. Under the direction of five faculty instructors, four simulator-based ETI training sessions were held over the two-week clerkship at the university's simulation center. The first session occurred prior to any clinical exposure, and included a review of basic ETI technique, two to three simulated ETIs per student and simulated induction of general anesthesia. The other sessions took place throughout the clerkship at weekly intervals. All students in each clerkship group attended the simulation sessions at the same time.
After the introductory session, students participated in the clinical care of operating room patients under the direct supervision of attending anesthesiologists, including performing clinical ETI. Institutional policy required attending anesthesiologists to be physically present during all ETI. Students received ETI opportunities strictly at the discretion of the attending anesthesiologist. Students rotated with anesthesiology departments at one of seven hospitals, including university, community and pediatric centers.
Methods of measurement and outcomes
As a clerkship requirement, students recorded all attempted clinical procedures (endotracheal intubation, mask ventilation and other procedures) in a structured logbook. We informed students of the research study on the first day of the clerkship. Although logbook completion was a requirement for the clerkship, student submission of the logbook for the research study was voluntary. An independent third party de-identified all logbooks prior to analysis by the study team. The University Institutional Review Board approved this approach.
For each attempted ETI, students reported pertinent data, including the date of procedure, patient age in years, gender, Mallampati score, the number of direct laryngoscopies and their procedural success.
The Mallampati score (classes I–IV) characterizes anticipated ETI difficulty based upon the oropharyngeal structures visible on pre-operative evaluation (8
). For example, an individual in whom the tonsillar pillars and uvula are fully visible would be considered Mallampati class I (anticipated easy intubation). In contrast, an individual in whom none of the oropharyngeal structures can be seen would be considered a Mallampati class IV (anticipated very difficult intubation). Anesthesiologists use the Mallampati score to predict the difficulty of ETI efforts. Students reported Mallampati scores based upon pre-operative assessment by themselves or attending anesthesiologists (9
Conventional oral ETI involves direct laryngoscopy (insertion of the metal laryngoscope into the mouth to expose the vocal cords) and intubation (placement of tube through vocal cords). We defined an ETI attempt as one direct laryngoscopy with intent to place an endotracheal tube, regardless of the success of endotracheal tube placement. We defined a successful ETI attempt as the combination of successful direct laryngoscopy and successful placement of an endotracheal tube into the trachea by the student. If the student's ETI attempts failed but another individual subsequently performed successful ETI, the student recorded the encounter as a failed student ETI. We excluded cases where the student performed only laryngoscopy or only endotracheal tube insertion but did not attempt to perform both. We excluded laryngeal mask airway insertions or instances with mask ventilation only. We did not include data from simulated mannequin or human-simulator ETI in the analysis.
Supervising attending anesthesiologists verified all logbook entries. As this was an observational study, we did not control the types of patients, ETI techniques used (for example, blade size or type), number or duration of direct laryngoscopies, number of clinical encounters, nor number of supervising attending physicians. We did not have information on the characteristics of participating students or attending anesthesiologist preceptors.
We studied ETIs performed by medical students during the period 1 January 2006 to 30 June 2007.
We analyzed the data using descriptive statistics, identifying the median, interquartile range (IQR) and overall range of procedures attempted by each student. We examined overall ETI success rate as well as the success rate for each sequential ETI encounter.
To assess the effect of procedural experience on ETI skill acquisition, we examined the ‘learning curve’ for ETI success. A learning curve describes the relationship between repeated efforts and operator proficiency (10
). In this case the learning curve characterized the improvement in student ETI success with successive clinical ETI efforts. We modeled the relationship between ETI success and cumulative ETI encounters using mixed-effects logistic regression, using ETI success as the dependent variable, cumulative ETI experience as the primary fixed effect and each individual student as a random (clustering) effect. Under this approach, each patient served as the unit of analysis, and the random effect accounted for multiple observations per student.
Because of their potential influence on ETI success, we adjusted the models for patient age, sex, Mallampati score and number of laryngoscopies. We defined ETI success on a per-patient basis, not on a per-laryngoscopy basis. In a separate model, we also controlled for hospital location differences. The general form of the regression model was:
Prior studies defined ETI proficiency as 90% predicted ETI success (4
). We similarly attempted to identify the region of the learning curve approximating 90% predicted ETI success. While applied in prior studies, we elected not to use cumulative sum techniques, which describe performance trends for individual subjects only (4
). We conducted all analyses using Stata v.10.0 (Stata Corporation).