Model development

The laparoscopic myomectomy model simulates a uterus with a 5 cm intramural fibroid, with its pseudocapsule already exposed. The design of the model serves to address the following 2 key components of laparoscopic myomectomy: fibroid enucleation and hysterotomy repair. The design and setup of the laparoscopic myomectomy model are shown in Figure 1 , and the construction of the model is demonstrated in Video 1 . The following materials were used to construct the model: a “cup turner” foam cylinder, pink felt, a 2-inch stress ball, 2-inch self-adhesive bandage wrap, multipurpose sealing wrap (Press ‘N Seal), red marker, and hook-and-loop fastener (Velcro). The model is reusable for multiple attempts and/or multiple learners.

Laparoscopic myomectomy simulation model

A, Completed model simulating a uterus with a 5 cm intramural fibroid, with its pseudocapsule exposed; B, View of the simulation model affixed to a laparoscopic box trainer and connected to a monitor for viewing and recording.

Schneyer et al. Laparoscopic myomectomy simulation model. Am J Obstet Gynecol 2022 .

The model was placed in a laparoscopic box trainer that allowed for ipsilateral, contralateral, or suprapubic port placement. The box trainer was connected to a monitor for viewing and recording ( Figure 1 ). The model was affixed to the trainer using a combination of hook-and-loop fastener and alligator clips.

Validity assessment

For the validation component of the study, we utilized the framework described by Messick. This contemporary framework describes validity as a process rather than an endpoint, involving the collection of evidence to support or refute a validity hypothesis. The potential sources of validity evidence include content, response process, internal structure, relationships with other variables, and consequences . Response process evidence and consequences evidence were not directly assessed in our study.

Validity evidence was primarily collected by evaluating expert and novice surgeons’ performances on the simulation model. The participants were recruited at a single quaternary care institution (Cedars-Sinai Medical Center) and at the Society of Gynecologic Surgeons 47th Annual Scientific Meeting. Eligible participants included OB/GYN residents (postgraduate year 2 and above), fellows, or attending surgeons who perform laparoscopic myomectomies as part of their training or routine practice. Participants who had performed greater than 20 laparoscopic myomectomies as the primary surgeon were categorized as experts, whereas those who had performed 20 or fewer were categorized as novices.

Upon enrollment, the participants were provided a randomly assigned, nonconsecutive study identification number generated on randomizer.org . They then completed a deidentified pre-task questionnaire that included demographic information and self-reported surgical and simulation experience. Next, the participants watched an instructional video outlining and demonstrating the simulation task ( Video 2 ). Part 1 of the task involved fibroid dissection and enucleation using laparoscopic instruments (scissors, grasper, and tenaculum) without the use of energy. Part 2 of the task involved two-layer closure of the hysterotomy using 12-inch, 0-caliber barbed suture on a GS-21 needle.

The participants were allowed to select their preferred trocar placement and could practice any component of the task for up to 10 minutes. While performing the task, they were allowed to ask the study coordinator to adjust the camera angle as needed. No technical guidance or assistance was otherwise provided. Each performance was recorded directly from the monitor without sound, and the recording was labeled with the participant’s study identification number. The time to completion of the task was also recorded. Part 1 began with the grasper contacting the model and ended when the fibroid was completely detached from its pseudocapsule, whereas Part 2 began with instrument entry into the camera view and ended with cutting of the suture.

Relationships with other variables evidence for the simulation model was collected by comparing expert and novice performances on the simulation task. The primary outcome of the study was to compare expert and novice performance scores for the task using the validated Global Operative Assessment of Laparoscopic Skills (GOALS) scale and a modified GOALS scale. The standard GOALS scale assesses technical performance in the following 5 domains deemed essential to laparoscopy: depth perception, bimanual dexterity, efficiency, tissue handling, and autonomy. The autonomy domain was omitted for this study, as the participants did not receive any guidance during the task. Each domain was graded on a 5-point Likert scale for a maximum composite score of 20 points. In addition, the following 3 novel metrics were created to assess the skills specific to minimally invasive myomectomy: fibroid dissection and enucleation, needle handling, and two-layer uterine closure ( Figure 2 ). These novel metrics were also graded on a 5-point Likert scale and added to create the modified GOALS score (maximum score of 35 points). Our novel metrics were modeled after the previously validated GOALS scale and customized for the laparoscopic simulation task with expert consensus. The needle handling metric was adapted from a similar modified GOALS scale developed by King et al.

Novel metrics incorporated into the modified GOALS scale for laparoscopic myomectomy

The 3 novel metrics were modeled after the previously validated GOALS scale. The needle- handling metric was adapted from a similar modified GOALS scale developed by King et al. The assessment tool utilizes a 5-point Likert scale, with 1, 3, and 5 representing anchor descriptors. Options 2 and 4 are selected if the participant’s skill falls between the described points.

GOALS , Global Operative Assessment of Laparoscopic Skills.

Schneyer et al. Laparoscopic myomectomy simulation model. Am J Obstet Gynecol 2022 .

Scoring of the deidentified videos was carried out by 2 blinded expert reviewers. For each of the domains, the 2 reviewer scores were averaged to calculate a single score for each participant. To avoid significant discrepancies in scoring using the assessment tool, the reviewers were first trained using 4 of the deidentified videos; the remaining videos were scored independently.

Content evidence was collected with a post-task questionnaire assessing the realism, utility, acceptability, and educational effect of the model on a 5-point Likert scale. This questionnaire was adapted from a previous gynecologic simulation study by Tunitsky-Bitton et al.

Statistical analysis

The Mann–Whitney U test was used to compare the task completion times and performance scores. The interrater reliability of scoring between the 2 reviewers (ie, internal structure evidence) was assessed using the interclass correlation coefficient. The sample size was calculated using the means and standard deviations of the GOALS scale using previously published studies evaluating similar simulation models as a reference. On the basis of these data, we powered our study to detect a difference of 3.0 points between the expert and novice groups on the GOALS scale, assuming a standard deviation of 1.5. ^, Given the limited number of expert-level surgeons compared with greater than 20 trainees eligible for participation, we planned for a 3:1 ratio of novices to experts. Assuming 90% power and a significance level of 0.05, a minimum of 4 experts and 12 novices was required.