A Comparative Direct Cost Analysis of Pediatric Urologic Robot-Assisted Laparoscopic Surgery Versus Open Surgery: Could Robot-Assisted Surgery Be Less Expensive?

Introduction

T he cost of healthcare in the United States is a concern for many in the medical community. National health expenditures reached $2.3 trillion in 2008 and are expected to nearly double by 2019. ¹ While there is continuous debate regarding solutions for these rising expenses, there is no question that costs must be addressed and that institutions should seek interventions that reduce expense without compromising care. With this increasing concern for cost, many have questioned the value of the da Vinci^® Surgical System (Sunnyvale, CA), a surgical robot with a reputation for high costs and unclear clinical benefits. ^2,3 Cost comparison studies across numerous disciplines seem to support this characterization, yet the number of hospitals with robotic surgical systems continues to rise. ⁴

There are hints in the literature that robot-assisted surgery is not always a poor choice for the cost conscious. Scales and colleagues ⁵ created a theoretical model of robot-assisted and open prostatectomy costs and determined that altering certain factors, such as caseload and practice setting, could make robot-assisted prostatectomy less expensive than its open counterpart. Several small empiric urology and general surgery studies have demonstrated reductions in direct costs when comparing robot-assisted surgery with open surgery. ^6,7 We, therefore, sought to undertake a more detailed cost comparison of a large number of matched robot-assisted and open pediatric urologic surgeries at our institution.

Materials and Methods

A series of 73 consecutive urologic robot-assisted laparoscopic cases were identified for the fiscal years 2005 to 2009 (October 2004 to September 2009). These included 43 pyeloplasties, 26 ureteral reimplantations, three creations of a catheterizable urinary conduit, and one excision of a urachal cyst. Each case was randomly matched for paired analysis by computer to a similar open surgical case based on procedure, patient age, and fiscal year for a total of 146 patients. No comparisons were made with laparoscopic surgery without the aid of the surgical robot; the strictly laparoscopic approach for the above procedures has been abandoned at our institution because of the increased time and difficulty involved. Patients whose urologic surgery was at the time of renal transplant were excluded. One patient who underwent open surgery and subsequently had a hospital stay of 182 days as a result of medical problems not related to surgery was also excluded because his case was thought to be an anomaly. Detailed demographic and outcome data were collected after matches were completed so that study staff remained blinded to outcomes during the matching phase.

Demographic information included age, sex, and American Society of Anesthesiologists (ASA) physical status. Operating room use cost was ascertained from the hospital's fiscal database; this cost was calculated as the sum of an initial setup fee and a charge based on the actual amount of time that the operating room was dedicated to a particular patient. Additional operative features gathered included whether fellows or residents were being taught during the case, and whether an additional procedure separate from the main operation was performed while the patient was under anesthesia.

The decision to perform surgery via the open or robotic approach was based on patient anatomy and pathology, as well as surgeon and family preference. All robotic cases were performed using the standard model da Vinci robot with four arms. Uncomplicated open and robot-assisted ureteral reimplantation surgeries followed an identical standardized postoperative clinical practice guideline. For the other surgeries, postoperative care was directed by the attending surgeon. Children were discharged when their pain was well controlled with oral medications, their oral intake was acceptable, any postoperative nausea or vomiting had dissipated, there was minimal output from any surgical drains, and they were voiding without difficulty.

Our institution's manager of business planning and analysis pulled direct costs from the hospital's internal fiscal records. Direct costs represent internal costs generated by each department related to a patient's care. These are more specific to a particular patient than indirect costs, which cover hospitalwide expenses and are allocated to all patients regardless of use. Direct costs are also more reflective of actual resource allocation than insurance charges or payments, which are dependent on negotiations between hospitals and insurance companies. ^8,9

Direct costs included operating room expenses as well as those related to the postoperative stay. Operating room costs included anesthesia care, operating room use, and robotic instruments if used, plus per-minute costs including surgical suite costs, nursing and operating room staff salaries, and basic surgical supplies. Surgeon and anesthesiologist charges were not included. Postoperative costs included expenses related to postanesthesia recovery room care, hospital room and board, radiologic imaging, laboratory and blood bank use, pathology, pharmacy costs, and general hospital supplies. Room and board costs were based on 24-hour blocks starting on the admission date. Other costs were recorded based on use. No adjustment was made for inflation; however, cases and controls were matched by fiscal year.

All statistical analyses were performed using IBM SPSS Statistics, Version 19. Differences in categorical values were compared using the Pearson chi-square test, while differences in continuous values were compared using the independent samples t test. A value of P<0.05 indicated statistical significance.

Results

The clinical and surgical data for our matched patients are shown in Table 1A. There was no significant difference between the open and robotic cohorts in terms of age, ASA status, percent of teaching cases, or percent of surgeries with additional procedures. Although there were more male patients in the robotic cohort (P=0.01), there were no other significant sex-based differences in patient, operating room, or postoperative characteristics.

Operating room time was similar between open and robot-assisted surgeries, averaging 262 minutes for open cases, and 255 minutes for robotic. Length of stay was significantly longer for open surgery patients (3.8 days vs 1.6 days, P<0.001). While many of these surgeries were performed on medically complex patients, the surgical times and lengths of stay were consistent with national data for this period from the Pediatric Health Information Service database. There were no differences in follow-up times or complications, and most complications were Clavien grade I or II (Tables 1A and 1B). For the two Grade III complications in the open cohort, one patient had recurrent obstruction and one had a ureteral leak postoperatively. In the robotic cohort, one patient had an intraoperative vascular injury, and two patients had stent migration.

Robot-assisted laparoscopic surgery direct costs were on average 11.9% lower ($8795 vs $9978, P=0.03) than open surgery direct costs overall, and lower for each fiscal year analyzed (Table 2). This difference was primarily because of the increased room and board cost of open surgery ($4197 vs $1707). Room and board costs were the largest contributors to open surgery direct costs, representing 42% of costs. The break-even point for room and board costs was $597 a day. If per-day costs were equal to or less than $597, open surgery would be equal to or less expensive than robot-assisted surgery (Fig. 1A). Alternatively, the break-even point for length of stay was 2.8 days, given our hospital's daily cost.

OPEN IN VIEWER

FIG. 1.

(A). Direct cost of open and robotic surgery with changing room and board costs per day. (B) Annual caseload needed to cover indirect robot costs using direct cost savings alone.

Robotic surgery costs were similar to those of open except for the discrepancy in room and board costs described above and the cost of the robotic equipment. Although it was not the largest percentage of robotic surgery direct costs, robot equipment at 16% of total costs represented the largest deviation from open surgery (Table 2). The average robotic equipment cost was $1407 (16% of total robotic surgery cost), which included on average four instruments as well as robot-specific disposables.

The purchase and maintenance costs of the robot were not included in the direct cost analysis, because these are indirect costs at our institution. Because our hospital does not amortize indirect costs, we calculated these additional expenses based on historical data. This included the 10-year lifetime for the standard model robot at our institution, the purchase cost of $1,200,000, the trade-in value of $300,000, the annual maintenance cost of $100,000, and an average of 67 cases per year. This results in an indirect cost of $1343 per case for robot purchase, plus $1492 per case for robot maintenance. Of note, no other indirect costs were examined. Potentially relevant indirect costs for both open and robotic surgery would include surgical instrument purchase and maintenance, operating room beds, hospital housekeeping, laundry, and medical recordkeeping.

If estimations of the indirect costs of robot purchase and maintenance were included, robotic surgery costs per case would be $1652 (17%) greater than those of open surgery. When including these indirect robot costs, the additive cost of using the surgical robot comes from annual maintenance (35%), purchase price (32%), robotic instruments (23%), and robot-specific disposables (10%). This represents the maximum cost difference, because it does not include the indirect costs that are specific to open surgery, such as separate instrument, maintenance, and extra physician time resulting from increased length of stay.

While the above numbers represent our historical experience with the da Vinci model used during the time frame analyzed in our matched cohort, a similar calculation can be performed for our institution's current robot using a combination of historical and expected costs. We can use an estimated 10-year lifetime for the new SI model robot, with the actual purchase cost of $2,100,000, an expected trade-in value of $500,000, an actual annual maintenance cost of $150,000, and an expected average of 120 cases per year. This results in an expected indirect cost for the current system of $1333 per case for robot purchase plus $1250 per case for robot maintenance. All else being equal, the break-even point would be a surgical volume of 262 cases per year.

Discussion

This study shows that robot-assisted surgery direct costs can be less than comparative open surgery costs under the proper circumstances. The cost contributor with the greatest impact for open surgery is room and board, because of the statistically significant increased length of stay for these patients. Other studies have noted a similar trend in hospital stays, most pronounced in pediatric and cardiac surgery patients, raising the possibility that patients with increased postoperative care needs benefit the most from a minimally invasive approach (Table 3). Our per-day room and board costs are higher than those at many adult hospitals, so the magnitude of savings from robot-assisted surgery may be greater than at other institutions. Another significant factor is choice of patients and procedures for our robot-assisted surgery cases. Older pediatric patients especially may have shorter hospital stays or decreased postoperative pain medication use after robotic surgery. ¹⁰ Procedure choice also affects our results; the robotic approach is only chosen for surgeries that are not performed laparoscopically at our hospital because of the difficulty or length of time needed. In these cases, the improved technical precision of the da Vinci robot allows for a minimally invasive option that would not otherwise be practical. The clinical and financial success of our institution's robotic program is at least partially from this precise selection of patients and procedures.

The driving factor for cost in robot-assisted surgery is the cost of robotic equipment. Our direct equipment costs are slightly lower than those in previous studies, ^7,11 likely because of the effort by our robotic surgeons to minimize robotic instrument usage. On average, our cases use four instruments, compared with the seven per case published by Bolenz and colleagues. ¹¹ While increased operating room times for robotic cases were a factor in other publications, this was not found in our study. Our robotic surgery program has been in operation since 2001, so the surgeries in this analysis were performed by or overseen by experienced surgeons who had completed the robotic surgery learning curve.

In addition, our institution maintains a consistent, specifically trained robotic team. A primary robotic surgeon supervises setup for all cases and performs 78% of operations. A core group of operating room nurses assist with robotic cases, having undergone a training course designed to minimize setup and docking times. Finally, selected robotic anesthesiologists use a specific protocol for robotic surgery to minimize patient wake-up time. ¹⁰ While staffing levels are similar, robotic cases have the advantage of a consistent team whose members have worked together for years, an opportunity that is not available for open surgeries at our institution.

Of note, our analysis did not include physician billing charges for surgeons or anesthesiologists. Because physicians at our institution are salaried, their cost is related to the opportunity cost of the time spent at that surgery. Because operating room times were equivalent between open and robot-assisted surgery, we did not feel the need to include a cost for the surgeon or anesthesiologist, because these would be equivalent for our two cohorts as well. There are other opportunity costs missing from our analysis, including the additional time needed for physicians to see patients while they are hospitalized as well as the expenses for the pain treatment service physicians to manage patients if they received an epidural or patient-controlled analgesia. These cost differences were thought to be small and to more heavily affect the open surgery cohort because of their longer length of stay. Because they were unlikely to significantly change our results, we chose not to include them.

One potential limitation of this study is that a single surgeon was involved in all of the robotic cases, either as the attending surgeon or in an assisting role. There were no significant differences, however, in the length of stay or direct costs for robotic surgery between patients who had this surgeon as an attending and those who had a different surgeon attending (with the previously mentioned surgeon assisting). The same surgeon also cared for several of the patients undergoing open surgery. Despite this, there is no way to ensure that this surgeon's involvement does not influence postoperative management and contribute to the variation in length of stay between open and robotic patients.

Another limitation is our choice to only look at direct costs. Because charges and reimbursements often depend on insurance companies rather than the details of surgery or hospital stay, we thought direct costs allowed for a more realistic comparison between our open and robotic cohorts. This could also make the findings more generalizable because hospital reimbursement is globally heterogeneous. Focusing on direct costs, however, ignores the large indirect costs of robot purchase and maintenance. Although robot purchase, maintenance, and equipment costs are set by Intuitive Surgical and therefore are not currently candidates for interventions that could improve cost-effectiveness, they are a looming concern for many hospitals.

When these costs are included, robotic surgery appears more expensive per case than open surgery. But this is a difficult comparison to make, because no indirect costs for open surgery have been accounted for. Because open surgery patients have a longer stay in the hospital, their contribution to indirect costs could be greater than that of robotic surgery patients. Unfortunately, billing structure at our institution prevents this type of analysis. In addition, surgical robots are often paid for with separate specific endowments or grants making straightforward cost analysis challenging.

Despite the difficulty in comparing these costs, including indirect costs does allow some observations to be made about robotic surgery cost generally. Annual maintenance fees were the most significant contributor to the additive cost of using the surgical robot at our institution, partially because of the long lifespan of the robot. It is clear that such high annual maintenance fees delay the recovery of indirect robot costs despite the direct cost savings. Although indirect costs such as robot purchase and maintenance would in actuality be paid for using insurance payments or allocated funding, it is possible to evaluate payment using the direct cost savings from performing robotic procedures. Using this comparison, it is clear that the added cost of annual maintenance substantially increases the number of cases needed per year to break even in a set time (Fig. 1B).

At our institution, robot-assisted laparoscopic surgery results in lower direct costs than matched open surgery. While these results are specific to a single hospital, they demonstrate that the minimally invasive nature of robotic surgery translates into reduced costs when the expense reduction from decreased length of stay is greater than the cost of the robotic equipment. As other institutions examine the cost-effectiveness of their robotic programs, our analysis would lead us to encourage them to focus their efforts on those patients and procedures in which the robotic approach can have the greatest impact on both cost and clinical outcome. This includes appropriate selection of cases as well as maintaining a consistent robotic team and resisting the temptation to upgrade robotic models.

Despite the common perception that surgery using the da Vinci robot inherently incurs higher costs than open surgery, ⁸ robot-assisted surgery could be more cost-effective under the proper circumstances. Even with current indirect purchase and maintenance costs included, the difference can be overcome with increased volume and intelligent case selection. A promising example is pediatric urologic surgery at our institution, where our specific lengths of stay, room and board costs, and operative times are such that direct costs are lower for robotic surgery than open surgery, resulting in savings for our hospital.

Conclusions

In the past decade, studies analyzing the cost of the da Vinci robot asked the question, “Is robotic surgery cost-effective?” As robotic surgery increases in popularity with patients and the da Vinci robot saturates hospitals in this country and others, however, a new question needs to be asked: “What can we do to make robot-assisted surgery more cost-effective?” Our comparative analysis of direct cost data identifies the most important factors contributing to cost-effective robot-assisted surgery at our institution: Shortened length of hospitalization, prudent instrument usage, equivalent operating room times, appropriate patient and procedure selection, and a consistent robotic team. While robotic surgery remains relatively more expensive than open surgery when indirect costs are considered, there remains ample potential for expense reduction. Although robot purchase price, annual maintenance fees, and instrument costs are currently fixed by Intuitive Surgical, there are numerous potential alternative robotic surgery systems on the horizon; specifically analyzing overall robotic surgery expenditures will help institutions to evaluate these systems in a more balanced and comprehensive way.