Abstract
Objective:
This research assessed if attending surgical volume impacted the relationship between resident participation and operative time in robotic surgery for benign gynecologic conditions.
Materials and Methods:
This retrospective cohort study included 1 academic and 1 community hospital, where a total of 329 patients underwent robotic surgery for benign gynecologic conditions from January 2018 to March 2019. The primary outcome was total operative time. The primary exposure was resident participation versus nonparticipation in robotic cases. The patients were operated on by 3 surgeons from the same group. The lifetime robotic hysterectomy volumes were 1250 (over 11 years), 146 (over 8 years), and 107 (over 8 years), respectively for the 3 surgeons.
Results:
The unadjusted mean total operative time was longer for cases involving residents (130.3 minutes [95% confidence interval (CI): 122.8–137.7 minutes] versus 109.6 minutes [95% CI: 104.0–115.3 minutes]; p < 0.001). Among hysterectomy cases, the adjusted mean total operative time was 22.3 minutes longer for cases with resident participation versus no participation (95% CI: 11.9–32.7 minutes longer; p < 0.001). Among hysterectomy cases, the mean total operative time for the high-volume surgeon increased by 14.9 minutes with resident participation, compared to an increase of 45.3 minutes for his 2 low-volume partners.
Conclusions:
This retrospective chart review's results suggest that attending surgeon robotic volume and surgery type appear to mediate the relationship between resident participation in robotic surgery and total operative time. More research is needed to confirm these findings. (J GYNECOL SURG 37:303)
Introduction
There is no dispute that minimally invasive gynecologic surgery is the current standard of care, including both vaginal and laparoscopic approaches. For hysterectomies, the American College of Obstetricians and Gynecologists (ACOG) recommends using minimally invasive rather than open abdominal approaches whenever feasible. The ACOG also recognizes that, among minimally invasive techniques, robotic-assisted gynecologic procedures have shown the greatest increase in overall use since U.S. Food and Drug Administration approval in 2005. 1 A 2013 study by Wright et al. reported a near 1000-fold increase in robotically assisted hysterectomies between 2007 and 2010, from 0.5% to 9.5%, respectively. 2
As the robotic approach becomes more popular, the tendency and or pressure to include robotics into residency training increases. In the case of for-profit community hospitals in the United States that have started up obstetrics and gynecology (OBGYN) residency programs, the time it takes to train residents in robotics is of financial interest, because cases that take longer reduce the economic efficiency of the operating room. Private physicians who also function as surgical teachers of residents, have obligations to be financially efficient and to maintain a certain surgical flow. One of the obstacles at this institution to allowing residents to sit on the robotic console during real cases is the concern for prolonged operative time. Another obstacle is that not all hospitals have dual-console platforms where both the surgeons and trainees can both be utilize the robot at the same time and the attending surgeon can take over swiftly at any point if the trainee encounters difficulty.
One study by Freeman et al. demonstrated that resident participation in gynecologic oncology cases performed robotically was associated with an increase in mean operative time of 31 minutes—a 23% increase. 3 Several studies have been done in the urology specialty looking at resident involvement in robotic prostatectomy and these studies have shown an increase in median operative of time of between 17 and 44 minutes associated with residents' participation.4–6 One study of resident participation in robotic-assisted sacrocolpopexy showed no statistically significant increase in both total operative time and total procedure time. 7 Another study of resident participation in robotic thoracic surgery showed no statistically significant increase in operative time. 8 The impact of resident participation on total robotic-console time in benign, nonreconstructive gynecologic cases, such as total hysterectomy, is unknown.
The primary aim of this retrospective chart review was to assess if attending-surgeon volume influences the relationship between resident participation and total operative time in robotically assisted hysterectomy cases.
Materials and Methods
This was a retrospective cohort study. A total of 329 patients undergoing robotic surgery for benign conditions at 2 hospitals—one having an Obstetrics & Gynecology (OBGYN) residency program and one with no residency programs.
Two outcomes were looked at: total operative time and total console time. Total operative time is the time from skin incision to skin closure. Console time was defined as the time interval from the point the surgeon or resident first sat on the console after docking was complete to the point where the surgeon or resident permanently got off the console to begin undocking. The primary “exposure” variable was whether a resident participated in the case or not.
Ethical standards
This study was approved by the institutional review board at Touro University of Nevada on April 10, 2019 by expedited review (category 5).
Informed consent
Each patient gave full informed consent to the surgery she underwent. Dr. Swainston's prospective collection of data, such as operative time for every surgery his group performs, was done for his own internal quality-improvement purposes and did not require an additional consent by each patient. Patient identifiers were removed prior to the dataset being shared with the coauthors of this study for the purposes of analysis and drafting of this article.
The hospital setting
One of the 2 hospitals was a busy community hospital with no graduate medical education programs. Neither of the two OBGYN residency programs in the city had their residents rotate at this hospital. The other hospital was also a community hospital, owned by a national for-profit corporation, but it housed both an OBGYN and a general-surgery residency program. This hospital had dual console platforms and they were used for cases when a resident participated.
The residency program
The OBGYN residency program was started in 2017 with 4 new interns and 4 second-year residents who had transferred from other programs. At the beginning of the study period, in January 2018, there were only 2 classes of residents (postgraduate year [PGY]-1 and PGY-2). By the end of the study period there were only 3 classes (PGY-1, -2 and -3). The OBGYN residency program had a robust robotic-simulation curriculum and a strict policy that, for a resident to sit on the console in the operating room that resident needed to complete all 30 required robotic-simulation modules with an average score of no less than 85%. The residents also had to pass a dry-laboratory test on the robot and be certified by the director of robotics before sitting on the console during a real case. Because of all the requirements that needed to be met prior to using the robot during a real case, all residents who participated in the robotic cases in this study were either PGY-2 or PGY-3.
The attending surgeons
The 3 attending surgeons who performed these 329 robotic cases all belonged to the same group. The surgeon who performed the overwhelming majority of the cases was the senior partner in the group. He had developed a data-collection sheet to track a wide array of variables related to each robotic-surgery case that the members of his group performed. These were manually recorded at the time of each procedure and then quality checked by him after the case. One such variable was whether or not a resident participated in the operation. During the study period, resident participation was categorized in a binary fashion: The resident either sat on the console or did not do so. The specific percentage of cases done by the resident was not estimated and recorded.
The data set
The data used in this study were collected prospectively. Data were collected and analyzed from various types of procedures, including lysis of adhesions, resection and fulguration of endometriosis, ovarian cystectomies, hysterectomies, and sacrocolpopexies. The preprocedural diagnoses were limited to benign conditions, most commonly abnormal uterine bleeding, pelvic pain, endometriosis, uterine leiomyomas, pelvic organ prolapse, and others. The variables collected by the surgeons on their robotic surgeries are shown in Table 1
Baseline Characteristics of the Study Cases (All Robotically Assisted)
BMI, body mass index; SD, standard deviation; min, minutes.
Statistical analysis
The basic characteristics of patients whose robotically assisted surgeries involved resident participation were first compared (with respect to mean operative times) to those for which there was no resident participation (Table 2).
Unadjusted Mean Total Operative Time for Robotically Assisted Hysterectomy Cases Stratified by Surgeon Based on Resident Participation Versus Nonparticipation in the Cases
Models adjusted for patient body mass index (continuous variable) and uterine weight (continuous variable).
CI, confidence interval; min, minutes.
The next set of analyses involved construction of linear-regression models to model mean total operative time as a function of resident participation while adjusting for potential confounders. The confounders that were adjusted for included body mass index (BMI), specimen weight (in g), and a variable indicating which surgeon was performing the case. BMI was modeled as a continuous variable. The analyses were limited our to hysterectomy cases. For these cases, the analyses were stratified further into cases for which Surgeon 1 was the teaching attending and cases for which either Surgeon 2 or 3 was the teaching attending.
Finally, a linear regression model, focused on hysterectomy cases, was created that contained an interaction term (resident participation × surgeon). The statistical significance of the coefficient on this interaction term was used to determine formally if there was a statistically significant mediating impact of surgeon volume on the relationship between resident participation and mean total operative time.
All analyses were performed using STATA version 15 (StataCorp, College Station, TX).
Results
After completion of data collection, a total of 329 robotic operations for benign gynecologic conditions were analyzed, 41.6% of which included resident participation. Table 1 shows the basic characteristics of the cases in which a resident participated, compared to those in which a resident did not participate. One surgeon accounted for 86% of the cases in which a resident participated and 77.7% of the cases in which a resident did not participate. This surgeon's robotic-hysterectomy volume was notably higher than his 2 partners. In calendar year 2018, Surgeon 1 performed 167 robotic-assisted hysterectomy procedures. In the same year, Surgeon 2 performed 30, and Surgeon 3 performed 13 robotic-assisted hysterectomy procedures. Surgeon 1 performed his first robotic case in November 2009 and had performed 1250 robotic-assisted hysterectomy procedures as of October 2020. Surgeon 2 performed a first robotic case in November 2012 and had performed 146 robotic-assisted hysterectomies as of October 2020. Surgeon 3 performed a first robotic case in May 2012 and had performed 107 robotic-assisted hysterectomies as of October 2020.
The unadjusted mean total operative time was 20.7 minutes longer for cases involving residents (130.3 minutes [95% confidence interval (CI): 122.8–137.7 minutes] versus 109.6 minutes [95% CI: 104.0–115.3 minutes]; p < 0.001). Cases with resident participation were more likely to include a hysterectomy and were more likely to be performed on an obese patient. There was no statistically significant difference in patient age between resident cases and nonresident cases (Table 1).
Total operative time
Among hysterectomy cases, the adjusted mean total operative time was 22.3 minutes higher for cases in which a resident participated versus for cases in which a resident did not participate (95% CI: 11.9–32.7 minutes; p < 0.001).
When Surgeon Number 1, who averages more than 100 robotic hysterectomies per year, performed a robotic hysterectomy, the adjusted difference in total operative time when the surgeon had a resident versus not having a resident was 14.9 minutes (Table 2). When Surgeon Numbers 2 and 3 performed robotic hysterectomies, the adjusted difference in operative time when they had residents versus having no residents was 45.3 minutes. An interaction effect was formally tested to determine if the surgeon significantly influenced the association between resident participation and operative time. In this regression model, the coefficient on the interaction term (resident participation × surgeon) was statistically significant, indicating an interaction effect.
Total console time
Among hysterectomy cases, the adjusted mean console time was 17.5 minutes higher for cases in which a resident participated versus cases for which a resident did not participate (95% CI: 9.5–25.5 minutes; p < 0.001).
When Surgeon Number 1 performed a robotic hysterectomy, the adjusted difference in console time when he had a resident versus not having a resident was 15.5 minutes (95% CI: 6.6–24.3). For Surgeon Numbers 2 and 3, when they performed a robotic hysterectomy, the adjusted difference in console time when they had residents versus not having residents was 26.2 minutes (95% CI: 6.5–45.9). An interaction effect was formally tested to determine if the surgeon influenced the association between resident participation and console significantly. In this regression model, the coefficient on the interaction term (resident participation × surgeon) was not statistically significant.
Discussion
This was a retrospective chart review of more than 300 consecutive robotic surgeries performed by 3 surgeons who belong to the same private practice and who all teach residents at a community hospital in Las Vegas, NV. There were several key findings.
The impact of resident participation in robotic gynecologic cases appears to be influenced by the type of surgeon (high-volume versus low-volume). With the high-volume surgeon, the increase in adjusted mean total operative time with resident participation was only 14.9 minutes versus an increase of 45.3 minutes with his low-volume partners.
This current study suggests that high-volume surgeons can teach with a smaller increase in their total operative time, compared to low-volume surgeons. It is interesting to note that when the focus was on just the console time, the two low-volume partners experienced an increase of 26 minutes when they had residents in their robotic-assisted hysterectomy cases, compared to 15.5 minutes for the high-volume surgeon when he had a resident. This difference of differences was not statistically significant in the dataset. In contrast, the impact of resident participation on total operative time, for hysterectomy cases, was significantly different for the high-volume versus the 2 low-volume partners. This suggests that the biggest difference in operative efficiency between the low-volume partners and the high-volume surgeon, when they had residents, was in the nonrobotic portion of the hysterectomy cases.
One prior study of 745 robotic cases looked at resident participation in hysterectomies for endometrial cancer performed robotically and found an increase of 31 minutes in the median operative time associated with resident participation. 3 The researchers found that, even in hysterectomy cases of varying complexity, there was a continued association between resident participation and operative time. That study defined “operative time” as the total time from skin incision to skin closure. Broadly speaking, the current study, which focused exclusively on surgeries for benign indications, was consistent with this prior study. It was found that, for hysterectomy cases, resident participation was associated with a clinically meaningful increase in total operative time (22.3 minutes). Even when stratified by surgeons, there was still an increase in adjusted mean total operative time (14.9 minutes and 45.3 minutes, respectively) and total console time (15.5 and 26.2 minutes, respectively). In contrast to the current study's findings, and several others,4–6 3 other studies found no difference in operative time between robotic cases performed with resident involvement versus robotic cases without resident involvement.7–9
One of the strengths of the current study is that, to the best of the authors' knowledge, this study is the first detailed examination of the impact of the volume or experience of the surgical teacher on the relationship between resident participation in robotic surgery and total operative time. None of the prior studies the current authors reviewed examined the impact of the attending surgeon's characteristics on the association between resident involvement and operative time or clinical outcomes.
Although this was a retrospective cohort study, the data set used was assembled prospectively. The high-volume surgeon (Surgeon Number 1) had intentionally created a data-collection sheet to be completed on all his robotic cases. He designed this data-collection sheet prospectively to include a field for resident participation versus no resident participation and intentionally designed this data-collection sheet to be used for quality-improvement purposes. This was not an administrative data set. Administrative data sets are typically not designed for research purposes and, when used for research purposes, have inherent limitations.
The main limitation of this retrospective chart review was that it did not examine postoperative complications. Postoperative complications were recorded, but the overall incidence was too low to permit any meaningful analyses based on resident participation versus resident nonparticipation. Another limitation was that all of the 329 patients in this study were operated on by just 3 surgeons. The current authors admit readily that more research, performed in different settings, is necessary to validate the generalizability of the current findings. Of note, most studies examining the impact of resident participation in robotic surgery on operative time and clinical outcomes have been single-institution studies. Large administrative databases built before the proliferation of robotic surgery do not allow a researcher to tell whether a surgery coded as a laparoscopic surgery was performed via conventional laparoscopy versus a robotically assisted laparoscopy. Finally, the specific percentages of cases done by the resident was not estimated and recorded.
Conclusions
In this retrospective chart review, resident participation in benign hysterectomy cases performed robotically resulted in an increase in total operative time. Among hysterectomy cases, this increase in total operative time was longer if the attending surgeon was high-volume versus being low-volume. If the current results are replicated by other studies, there would be implications for the way that robotic cases are scheduled at community teaching hospitals, where some surgeons teach residents while other surgeons operate without residents.
Footnotes
Acknowledgments
The authors wish to thank Ms. Jasmine L Hankey for her editorial assistance.
Author Disclosure Statement
Dr. Swainston is a paid speaker for Intuitive, which makes the da Vinci® robot. He did not receive compensation for this study. This study was not funded by Intuitive. All other coauthors (S.M., A.M., and D.L.H.) have no financial conflicts of interest.
Funding Information
This research was supported (in whole or in part) by Hospital Corporation of America (HCA) and/or an HCA-affiliated entity. The views expressed in this publication represent those of the author(s) and do not necessarily represent the official views of HCA or any of its affiliated entities.