The past 30 years have seen a significant change in the surgical treatment of women with cancer.1,2 A recent large randomized trial for endometrial cancer showed a decrease of 33% in serious complications, a 50% reduction in hospital stay, and an improved quality of life when laparoscopy was compared with laparotomy.3 Despite these compelling findings, only a minority of cancer patients benefit from the laparoscopic technique.4–6 The U.S. Food and Drug Administration approved the da Vinci Surgical System as the only computer-assisted (robotics) system for use in gynecologic procedures in 2005. It confers advantages of increased dexterity with the operating instruments, three-dimensional immersion views, and improved ergonomics.7,8
The main barriers to the acquisition of this technology include the cost and the lack of research evidence showing improved outcomes that could influence health care administrators and policymakers. The objectives of this study are to describe the effect of introducing a robotic surgical system and to evaluate its cost-effectiveness and clinical effectiveness.
MATERIALS AND METHODS
All consecutive women undergoing surgery for endometrial cancer since the initiation of the division of gynecologic oncology in a tertiary cancer center in March 2003 were included in this study. Two cohorts were included as follows: 1) all women who underwent either laparotomy or laparoscopy for the surgical treatment of endometrial cancer from March 2003 until December 2007, just before the introduction of robotic surgery at the hospital, and 2) all women who had robotic-assisted surgery for treatment of their endometrial cancer from the beginning of the robotics program in gynecologic oncology in December 2007 until March 2010. The latter patients were universally informed about the evaluation of this new surgical technology and all signed the regular hospital informed consent for surgery that described the robotic approach; in addition, patients signed a second independent informed consent form for the prospective evaluation of the robotic program. The hospital institutional review board approved the study protocol and the informed consent to evaluate the outcomes of robotic surgery. Surgical training for robotics is described in Appendix 1. All women found suitable to undergo surgery for the treatment of their endometrial cancer were offered robotic surgery regardless of body habitus or previous medical or surgical history. Both cohorts represent unselected consecutive women with endometrial cancer. Primary end points were to evaluate the clinical effectiveness and cost-effectiveness of introducing a robotics program for the treatment of endometrial cancer.
Study variables and their categories were designed objectively and not as open-ended explorations of chart data. All robotic surgeries were recorded in their entirety on DVD for skills evaluation and teaching purposes. Information was collected prospectively and a database was created for the purpose of documenting and evaluating the experience with this new minimally invasive technique. Data collection for the prerobotic era was based on data retrieval using patient charts maintained with the same level of clinical detail. Charts are electronic and rigorously maintained and intraoperative and postoperative complications are always documented. The clinical research staff was extensively trained to ensure that the data collection was performed systematically and uniformly, regardless of study era.
We used t tests and Wilcoxon rank-sum test to compare means and medians for interval-scaled variables and χ2 or Fisher exact tests for categorical or ordinal variables. Postoperative complications, hospital stay, and overall hospital costs were used as outcome variables in logistic regression models. Hospital stay and overall hospital costs were dichotomized above and below the median (Can$7,224 [Canadian dollars]) based on their distributions in the control cohort (laparotomy and laparoscopy).
Covariates adjusted for in each of the six models (shown in parentheses) were selected empirically based on a 5% change in estimate for the odds ratio of the specified outcome given the choice of robotic compared with historic cohort. Original cohorts were complications (age, blood loss, body mass index [BMI, calculated as weight (kg)/[height (m)]2], pelvic or periaortic lymphadenectomy, uterine volume), hospital stay (age, blood loss, American Society of Anesthesiologists [ASA] score, BMI, preoperative gastrointestinal disease, preoperative cardiovascular disease, histologic grade, periaortic lymphadenectomy, uterine volume), and cost (age, ASA score, blood loss, BMI, periaortic lymphadenectomy, histologic grade, surgical stage, uterine volume). Restricted cohorts were complications (age, ASA score, number of preoperative comorbidities, blood loss, BMI, smoking, histologic grade, periaortic lymphadenectomy, uterine volume), hospital stay (age, ASA score, blood loss, BMI, preoperative cardiovascular disease, preoperative gastrointestinal disease, histologic grade, periaortic lymphadenectomy, uterine volume), and cost (age, number of preoperative comorbidities, blood loss, BMI, periaortic lymphadenectomy, uterine volume).
Between December 2007 and March 2010, 143 consecutive women had robotic-assisted total hysterectomy and staging for endometrial cancer. Our purpose was to compare these patients to the historical cohort of 160 consecutive women who were surgically treated and staged for their endometrial cancer in the years 2003 to 2007, before the real-life initiation of the robotics program, and underwent either laparotomy (n=133) or laparoscopy (n=27), without further comparing between these techniques. Staging procedures were similar between both periods (Appendix 2). The preoperative patient characteristics of the two cohorts were similar (Table 1). The mean and median BMI were higher in the robotics cohort compared with the historical cohort (mean 31.5 compared with 28.8; median 29.8 compared with 27.6, respectively; both P<.005). Tumor characteristics did not differ significantly between the two cohorts (Table 2). The rate of completion of a pelvic lymphadenectomy was higher in the robotics group (Table 2), whereas the number of lymph nodes retrieved was similar in both groups, and no differences were seen in the rate of periaortic lymph node dissections between the two groups (Table 2).
The overall incidence of complications of moderate severity or more was significantly lower in the robotics group compared with the historical cohort (13% compared with 42%; P<.001), and intraoperative complications were similarly low in both cohorts (Table 3). The most common adverse event was superficial wound separation or infection, which occurred in 15.6% of the historical cohort and in 3.5% of the robotic cohort (P<.001). Cardiovascular morbidity, including exacerbation of heart failure (five cases postlaparotomy, one patient died), infarction (three cases postlaparotomy), and atrial fibrillation (two cases postlaparotomy and one case postrobotic), were more frequent in the historical cohort (6.3% compared with 0.7%; P=.01). Gastrointestinal complications were only seen in the historical group (P=.007).
The median operative time was significantly lower in the historical cohort (207 minutes) compared with robotics (241 minutes; Table 4). The median blood loss was significantly higher in the historical group (266 mL) than in the robotics cohort (73 mL). Only two patients (1.4%) underwent transfusion in the robotics group as compared with 10 (6.3%) in the historical group. Uterine weight, which is often higher in laparotomy cohorts, was higher in the robotics cohort (Table 4; P=.003). The mean hospital stay was substantially shorter in the robotics group compared with the historical group (2.2 days compared with 5.5 days; P<.001). The combination of reduced hospital stay and fewer complications in the robotics group contributed to an overall lower cost for the procedure. Excluding the amortization cost, the mean cost in Canadian dollars to treat a patient with endometrial cancer with robotic surgery was Can$7,644 compared with Can$10,368 treated in the historical cohort (P<.001; Table 5). When the average perioperative cost was calculated including the acquisition and yearly maintenance cost for workloads of 10 cases per week, the average cost for robotic surgery remained significantly lower at Can$8,370 (P=.002; Appendix 4).
The gains with robotic surgery are even more pronounced in the multivariable analysis when we consider both the entire cohorts and the restricted (matched) subcohorts (Table 6). Compared with the conventional procedures, the odds of any complication grade two or higher significantly decreased by threefold to fourfold in the robotics era when adjusted for all empirical confounders. Likewise, a nearly 20-fold decline in hospital stay was observed in the robotics cohort. As a result, overall cost was also reduced fourfold during the robotics period. The analyses on the restricted subcohorts with age-matched and BMI-matched characteristics showed even more pronounced reductions in complications, hospital stay, and cost.
A key factor in evaluating any cancer treatment procedure is oncologic outcome. At present, the length of follow-up of 2 years after robotic surgery indicates a lower recurrence rate compared with the historical cohort (Fig. 1; Wilcoxon P<.001; log-rank P<.001).
In this study, we are not comparing robotics to either laparotomy or laparoscopy, but rather we are evaluating the clinical and economic implications of introducing a robotics program in a tertiary care institution. Similar to others,7,8,11 we have seen an increase in minimal invasive surgery rates from 17% performed by laparoscopy to 98% by robotics. In the LAP2 study comparing laparoscopy to laparotomy, the conversion rate from laparoscopy to laparotomy was 25.8%,3 contrasting with 4.2% in the present study, in which all were performed at the completion of the robotic surgery through a Pfannenstiel incision solely to remove enlarged uteri that were not deliverable intact through the vagina. In August 2009, we adapted our technique and since then, no “conversions” were necessary.12 Other institutions also cite low rates of conversion from robotics to laparotomy, ranging from 2.9% to 4.0%.8,11,13
We designed our study conservatively to ensure that any differences in outcomes between eras would be ascribed primarily to the treatment intervention. Several study design characteristics add to this conservatism. The spans of both eras (prerobotic and postrobotic) captured in the study covered a modern period in terms of patient accrual and information management in our McGill teaching hospital. There were no variations in catchment area during the study period. The variations we observed between eras were mostly in the direction of a conservative bias, ie, more “difficult to treat” patients in the robotic era. As a result of the improved perception of security and confidence to treat more complex cases with the robot, characteristics such as age, BMI, ASA level, and previous surgeries were not determinants for patient selection for minimal invasive surgery procedures.
Both the lower complication rate13–15 and shorter hospitalization, mainly in the elderly and obese populations, translated into a significant decline in the cost of the surgical intervention per patient, even after accounting for the acquisition and yearly maintenance costs of the robot. Nevertheless, the economic advantage presented in this article may underestimate the actual value because only direct and indirect hospital costs were included, without calculating the overall gains to the patient and society related to quicker recovery and return to normal activities.
A limitation of this study relates to the retrospective nature of the data collection for the historical cohort. However, the study variables and their categories were designed objectively and not as open-ended explorations of chart data to minimize any possible biases in data retrieval. The fact that we collected data in the postrobotic era prospectively only indicates the direction of the study inquiry in terms of time and in no way suggests that the level of detail in data collection varied between phases. Furthermore, all endometrial cancer patients treated before the robotics program were included for completeness.
The strengths of this study include the analysis of real-life conditions in a university-based gynecologic oncology division with unselected laparotomy and laparoscopy patients in the historical cohort. In addition, the results remained robust whether we included the laparoscopy cases. For the main comparisons between surgical eras, we conducted analyses that adjusted conservatively for all empirical confounders and repeated these analyses in subcohorts selected randomly from joint strata based on frequency-matched age and BMI. These epidemiologic analyses yielded unbiased estimates for the primary outcomes, thus lending credibility to our results.
Clearly, introduction of robotics for endometrial cancer surgery in our tertiary Canadian center has increased the proportion of patients benefitting from minimally invasive surgery, has improved short-term outcomes without oncologic compromise, and has resulted in lower hospital costs.
1. Querleu D, Leblanc E, Castelain B. Laparoscopic pelvic lymphadenectomy in the staging of early carcinoma of the cervix. Am J Obstet Gynecol 1991;164:579–81.
2. Nezhet CR, Burell MO, Nezhat FR, Benigno BB, Welander CE. Laparoscopic radical hysterectomy with paraortic and pelvic lymph node dissection. Am J Obstet Gynecol 1992;166:864–5.
3. Walker JL, Piedmonte MR, Spirtos NM, Eisenkop SM, Schlaerth JB, Mannel RS, et al.. Laparoscopy compared with laparotomy for comprehensive surgical staging of uterine cancer: Gynecologic Oncology Group Study LAP2. J Clin Oncol 2009;27:5331–6.
4. Frumovitz M, Ramirez PT, Greer M, Gregurich MA, Wolf J, Bodurka DC, et al.. Laparoscopic training and practice in gynecologic oncology among Society of Gynecologic Oncologists members and fellows-in-training. Gynecol Oncol 2004;94:746–53.
5. Mabrouk M, Frumovitz M, Greer M, Sharma S, Schmeler KM, Soliman PT, et al.. Trends in laparoscopic and robotic surgery among gynecologic oncologists: A survey update. Gynecol Oncol 2009;112:501–5.
6. Naumann RW, Coleman RL. The use of adjuvant radiation therapy in early endometrial cancer by members of the Society of Gynecologic Oncologists in 2005. Gynecol Oncol 2007;105:7–12.
7. Hoekstra AV, Morgan JM, Lurain JR, Buttin BM, Singh DK, Schink JC, et al.. Robotic surgery in gynecologic oncology: Impact on fellowship training. Gynecol Oncol 2009;114:168–72.
8. Peiretti M, Zanagnolo V, Bocciolone L, Landoni F, Colombo N, Ming L, et al.. Robotic surgery: Changing the surgical approach for endometrial cancer in a referral cancer center. J Minim Invasive Gynecol 2009;16:427–31.
9. Pecorelli S. Revised FIGO staging for carcinoma of the vulva, cervix, and endometrium. Int J Gynaecol Obstet 2009;105:103–4.
11. Holloway RW, Ahmad S, DeNardis SA, Peterson LB, Sultana N, Bigsby GE IV, et al.. Robotic-assisted laparoscopic hysterectomy and lymphadenectomy for endometrial cancer: Analysis of surgical performance. Gynecol Oncol 2009;115:447–52.
12. Peeters F, Vaknin Z, Lau S, Deland C, Brin S, Gotlieb WH. Technical modifications in the robotic-assisted surgical approach for gynaecologic operations. J Robotic Surg 2010;4:253–7.
13. Boggess JF, Gehrig PA, Cantrell L, Shafer A, Ridgway M, Skinner EN, et al.. A comparative study of 3 surgical methods for hysterectomy with staging for endometrial cancer: robotic assistance, laparoscopy, laparotomy. Am J Obstet Gynecol 2008;199:360.e1–9.
14. Bell MC, Torgerson J, Seshadri-Kreaden U, Suttle AW, Hunt S. Comparison of outcomes and cost for endometrial cancer staging via traditional laparotomy, standard laparoscopy and robotic techniques. Gynecol Oncol 2008;111:407–11.
© 2012 The American College of Obstetricians and Gynecologists
15. DeNardis SA, Holloway RW, Bigsby GE IV, Pikaart DP, Ahmad S, Finkler NJ. Robotically assisted laparoscopic hysterectomy compared with total abdominal hysterectomy and lymphadenectomy for endometrial cancer. Gynecol Oncol 2008;111:412–7.