Short-Term Outcomes for Robotic Colorectal Surgeryby Provider Volume

Deborah S Keller, MS, MD, Lobat Hashemi, MS, Minyi Lu, PhD,Conor P Delaney, MD, MCh, PhD, FRCSI, FACS, FASCRS

BACKGROUND: There has been a rapid increase in robotic colorectal surgery. Benefits of this technology areunclear and being investigated. However, differences in outcomes between centers have notbeen evaluated. Our aim was to evaluate outcomes for robotic colorectal procedures bysurgeon and hospital volume.

STUDY DESIGN: A national inpatient database was reviewed for robotic colorectal resections performed duringan 18-month period. Hospitals and surgeons were stratified into high, average, and low casevolumes based on a normal distribution scale. High, average, and low volume was defined as�10, 11 to 20, and >20, respectively, for hospitals, and �5, 6 to 15, and >15, respectively,for surgeons. Short-term outcomes and hospital cost were evaluated.

RESULTS: There were 1,428 robotic colorectal cases across 123 hospitals and 411 surgeons evaluated.Only 13% (n ¼ 16) of hospitals and 4.4% (n ¼ 18) of surgeons performed a high volumeof robotic colorectal cases. Lower volume was associated with significantly more overallcomplications (p < 0.001; p < 0.001), longer length of stay (p ¼ 0.005; p < 0.001), andhigher cost (p < 0.001; p < 0.001) at the hospital and surgeon level, respectively. High-volume hospitals and surgeons had significantly lower rates of postoperative bleeding(p < 0.001; p < 0.001) and ileus (p ¼ 0.003; p ¼ 0.0014).

CONCLUSIONS: Lower-volume providers, who are performing the majority of procedures, are generating morecomplications, longer hospital lengths of stay, and higher costs of care. These results havea negative impact on quality outcomes measures for those facilities. Although surgeons andhospitals continue to selectively explore robotics, this should be limited to high volumeand interested surgeons and hospitals to offer high-quality outcomes to patients. (J AmColl Surg 2013;217:1063e1069. � 2013 by the American College of Surgeons)

Robotics is an emerging technology with dramatic uptakein surgery. As of December 31, 2012, there were morethan 2,500 robotic surgical systems installed in morethan 2,000 hospitals. In 2012, approximately 450,000robotic-assisted procedures were performed, an increaseof 25% compared with 2011.1 During the last 10 years,

Disclosure Information: Ms Hashemi and Dr Lu are employed byCovidien, a medical device company that is not a competitor to therobotic industry. All other authors have nothing to disclose.

Abstract presented at the American College of Surgeons 99th AnnualClinical Congress, Washington, DC, October 2013.

Received May 6, 2013; Revised June 30, 2013; Accepted July 3, 2013.From the Division of Colorectal Surgery, Department of Surgery, Univer-sity Hospitals Case Medical Center, Cleveland, OH (Keller, Delaney) andBiostatistics, Healthcare Economics and Outcomes Research, Covidien,Mansfield, MA (Hashemi, Lu).Correspondence address: Conor P Delaney, MD, MCh, PhD, FRCSI,FACS, FASCRS, Division of Colorectal Surgery, Department of Surgery,University Hospitals Case Medical Center, Case Western Reserve Univer-sity, 11100 Euclid Ave, Cleveland, OH 44106-5047. email: Conor.Delaney@uhhospitals.org

1063ª 2013 by the American College of Surgeons

Published by Elsevier Inc.

robotics have become the predominant approach forminimal invasive prostatectomy.2,3 After achieving marketdominance in prostatectomy, the robot approach hadstarting penetrating colorectal surgery.The rapid growth continues despite multiple single-

institution reports, randomized trials, and systemicreviews demonstrating that robotic surgery substantiallyincreases the cost of care without any clear advantagesin short-term or long-term outcomes for colorectalsurgery.4-11 At a price of more than $1.7 million perrobot, $125,000 in annual maintenance costs, and upto $2,000 per case for the cost of single-use instruments,robotic surgery is the most expensive approach.12 Otherfields have reported the unnecessary expenditure onexpanded use of robotics for routine care. Wright andcolleagues13 reviewed nearly 265,000 women who under-went routine hysterectomy, finding each robotic caseadded $2,189 compared with traditional laparoscopicsurgery without any substantial benefit in outcomes orcomplications. The rate of robotic cases increased from

ISSN 1072-7515/13/$36.00

http://dx.doi.org/10.1016/j.jamcollsurg.2013.07.390

Table 1. Volume Statistics and Demographic Data

Characteristic

Surgeon level

Cases,* n 495 374 559

Surgeon, n (%) 346 (84.2) 47 (11.4) 18 (4.4)

Hospital level

Cases,y n 300 306 822

Hospital, n (%) 88 (71.5) 19 (15.5) 16 (13)

Age, y, mean (SD) 59.5 (15.1) 60.5 (14.0) 61.6 (14.2)

Sex, n (%)

Female 161 (20.9) 160 (20.7) 451 (58.4)

Male 139 (21.2) 146 (22.3) 371 (56.5)

Race, n (%)

Caucasian 214 (19.9) 180 (16.7) 682 (63.4)

African American 34 (29.1) 28 (23.9) 55 (47.0)

Hispanic 1 (1.2) 22 (26.2) 61 (72.6)

Other 51 (33.8) 76 (50.3) 24 (15.9)

APR-DRG severityscore, n (%)

1064 Keller et al Robotic Colorectal Surgery Outcomes by Volume J Am Coll Surg

0.5% to 9.5% during the study period; with the growthtrend, the authors found using robotics for all routinehysterectomies would add an unnecessary $1 to $1.9billion in unnecessary health care costs each year.13 Incolorectal surgery, the data are similar. Robotics hasbeen reported to add approximately $3,500 per proce-dure and $2.5 billion nationally per year withoutevidence of improved outcomes over standard laparos-copy.14 The data do not support the routine use of roboticcolorectal surgery as a cost-effective procedure.15 Aggres-sive direct-to-consumer marketing and an assertive salesforce can help explain how robotic technology continuesto expand despite data on quality outcomes andmounting complications.16-18 A systematic analysis of400 US hospital websites found 41% described roboticsurgery on the front page; many with manufacturer stockimages (78%) and direct links to the manufacturer’s web-site (33%). Unsupported claims of clinical advantages(86%) and improved cancer control (32%) were alsofound, and no sites mentioned the risks of roboticsurgery. The authors concluded that hospitals overesti-mate benefits and underestimate risks.19

With the rapid uptake and acquisition of robots,factors affecting outcomes have not been well described.Previous studies have proven that hospitals and surgeonsthat perform higher volumes of cases have lower compli-cation and mortality rates.20,21 There is little informationin the literature on complications and outcomes based oncase volumes in robotic surgery. As less than 20% ofhospitals perform more than 100 robotic procedures,and less than 3% performed more than 250 annually,22

data on robotic results by volume are essential.The objective of this study was to evaluate outcomes

for robotic colorectal procedures by surgeon and hospitalcase volume. Our hypothesis was that hospitals andsurgeons that perform higher-level robotic colorectalprocedures would have substantially better patient andfinancial outcomes.

Mild 129 (20.7) 115 (18.5) 378 (60.8)

Moderate 109 (19.6) 123 (22.2) 323 (58.2)

Severe 53 (27.3) 50 (25.8) 91 (46.9)

Extreme 9 (15.8) 18 (34.6) 30 (52.6)

Teaching institution,n (%)

No 176 (24.3) 184 (25.5) 363 (50.2)

Yes 124 (17.6) 122 (17.3) 459 (65.1)

Provider area, n (%)

Rural 33 (64.7) 0 (0) 18 (35.3)

Urban 267 (19.4) 288 (20.9) 822 (59.7)

*Low (<5) cases; average (6 to 15) cases; high (>15) cases.yLow (<10) cases; average (11 to 20) cases; high (>20) cases.APR-DRG, All Patient Refined Diagnosis-Related Groups.

METHODS

Data source

A retrospective analysis was performed using the PremierPerspective database (Premier, Inc.). Premier Perspectivecontains complete patient billing and hospital cost frommore than 500 hospitals in the United States. The databasecontains data onmore than 45million inpatient visits, rep-resenting, approximately 1 of every 5 inpatient dischargesin the United States. Patient discharge data from not-for-profit, nongovernmental, community and teaching hospi-tals, along with health systems, are included in this data-base. Data are drawn from de-identified UB92 discharge

forms, including demographic characteristics, dischargediagnoses, ICD-9 and CPT codes, and discharge status(including death, but not its cause).

Patient selection

All inpatient discharges with procedures codes for roboticcolectomy between January 1, 2010 and June 30, 2012were identified using a combination of ICD-9 and CPTcodes (Table 1). In addition to the ICD-9 and CPTcodes, detailed billing and instrument charge descriptionsfrom hospital billing systems were used to identifypatients undergoing robotic colectomy procedures bytext searches for robotic and/or DaVinci. All patientsincluded were discharged alive and were older than18 years of age. Patient demographic data, includingage, race, sex, All Patient Refined Diagnosis-RelatedGroups severity (for comorbidity), type of hospital the

Vol. 217, No. 6, December 2013 Keller et al Robotic Colorectal Surgery Outcomes by Volume 1065

procedure was performed at (teaching vs nonteaching),and rural vs urban setting, were evaluated.

Outcomes

The main outcomes measures were hospital length ofstay, cost of care, complication rate, and mortality rates.Complications were identified using ICD-9 codes andgrouped into 4 categories, intraoperative (inadvertententerotomy), postoperative (postoperative hemorrhage,peritonitis, ileus, sepsis/anastomotic leak), surgical siteinfection (hematoma/ seroma, wound infection, wounddehiscence), and medical (infection-pneumonia, urinarytract, gastrointestinal complication). Appendix 1 (onlineonly) lists the ICD-9 codes used to identifycomplications.

Hospitalization cost

Total hospital costs (including costs of surgery, costs ofmedication, supplies, anesthesia, room and board, andother costs), as well as length of hospital stay, were alsoassessed. In the Premier database, costs are defined asthe actual cost to treat the patient (which includes allsupplies, labor, and depreciation of equipment). Thiscost includes both fixed cost, which does not vary basedon the volume of procedures performed and variablecosts, which are the direct cost and can vary based onthe volume of procedures performed, based on HospitalCensus and Average Wholesale Price. Hospitalizationcost included (but was not limited to) cost associatedwith room and board (including ICU), surgery (includingoperation room cost), central supplies (including all lapa-roscopic and open staplers, instrumentation, and sutures),anesthesia, laboratory, pharmacy, emergency room,pathology, blood bank, and radiology charges. Costs areas reported by hospitals for >85% of reporting hospitals.The remaining hospitals have costs calculated by Premierusing cost to charge ratios as reported in their MedicareCost Report.

Surgeon and hospital volume

Hospitals and surgeons were stratified into high-, average-,and low-case volumes based on a normal distribution scale.High-, average-, and low-case volume was defined as>20,11 to 20, and<10 for hospitals, and>15, 6 to 15, and<5for surgeons. Short-term outcomes and hospital cost wereevaluated.

Statistical analysis

Statistical analysis was performed using SAS softwareversion 9.2 (SAS Institute Inc.). A 2-sided unpairedStudent’s t-test was used for normally distributedcontinuous variables, and 2-sided Wilcoxon rank sum

(Mann-Whitney U) tests were for nonparametric data.Chi-square tests and Fisher’s exact test were used to compareproportions. Statistical significant was defined as p< 0.05.

RESULTSPatient demographic data and case breakdown by hospitaland surgeon volume are detailed in Table 1. In total,1,428 robotic colorectal cases were evaluated across 123hospitals and 411 surgeons. Of the cases, 1,377(96.4%) were performed in an urban setting. The caseswere evenly split between teaching (49.4%) andnonteaching facilities (50.6%). Only 13% (n ¼ 16) ofhospitals performed a high volume of robotic colorectalcases. At the individual surgeon level, 4.4% (n ¼ 18) per-formed a high volume of robotic colorectal cases; 71.5%of robotic colorectal surgery cases were performed atlow-volume hospitals and low-volume robotic surgeonsperformed 84.2% of cases.Short-term outcomes and cost at the hospital level are

demonstrated in Table 2. A lower volume of robotic colo-rectal surgery cases was associated with significantly moreoverall complications (p < 0.0009), longer hospitallength of stay (p ¼ 0.0053), and higher cost for theepisode of care (p < 0.001). Low-volume hospitals alsohad significantly lower rates of postoperative complica-tions compared with high- and average-volume hospitals(p ¼ 0.0005), including rates of hemorrhage (p ¼0.0005) and ileus (p ¼ 0.0031). There were no signifi-cant differences in the rates of intraoperative, medical,or surgical site complications by hospital volume. Themortality rate was <1% for all hospitals surveyed(Table 2).Outcomes at the surgeon level paralleled the hospital

level. The low-volume robotic colorectal surgeons hadsignificantly more overall complications (p ¼ 0.0003).Low-volume surgeons experienced significantly higherrates of postoperative complications in general (p <0.0001), and postoperative bleeding (p < 0.001) andileus (p ¼ 0.0014), specifically. In addition, low-volume surgeons had longer length of stay (p < 0.001)and higher costs for the episode of care (p < 0.001)(Table 3).

DISCUSSIONWith the emphasis on increasing quality in surgery,methods to prospectively identify and reduce complica-tions are essential. It is logical that the highest-volumehospitals and the highest-volume surgeons will have lowermortality and complication rates. This has been supportedby previous studies in colorectal surgery for specific diag-noses and procedures. In a 7-year review of rectal cancer

Table 2. Hospital-Level Outcomes by High, Average, and Low Case Volume

Class Low (n ¼ 300) Average (n ¼ 306) High (n ¼ 822) p Value

Overall, n (%) 89 (29.7) 90 (29.4) 172 (20.9) 0.0009*

By type, n (%)

Postoperative 72 (24) 80 (26.1) 138 (16.8) 0.0005*

Hemorrhage 27 (9.0) 31 (10.1) 43 (5.2) 0.0058*

Postoperative ileus 55 (18) 42 (14) 88 (11) 0.0031*

Anastomotic leak 8 (2.7) 8 (2.6) 16 (2.0) 0.68

EC fistula 1 (<1.0) 6 (2.0) 4 (<1.0) 0.06

Intraoperative 5 (1.7) 5 (1.6) 15 (1.8) 0.97

Inadvertent enterotomy, n (%) 5 (1.7) 5 (1.6) 15 (1.8) 0.97

Medical, n (%) 28 (9.3) 23 (7.5) 45 (5.5) 0.06

Infection 12 (4.0) 8 (2.6) 17 (2.1) 0.197

GI complication 20 (6.7) 18 (5.9) 34 (4.1) 0.17

Surgical site, n (%) 22 (7.3) 19 (6.2) 51 (6.2) 0.78

Hematoma/ seroma 5 (1.7) 8 (2.6) 27 (3.3) 0.34

Wound dehiscence 4 (1.3) 1 (<1.0) 3 (<1.0) 0.13

Wound infection 17 (5.7) 12 (3.9) 26 (3.2) 0.15

Mortality, n (%) 2 (<1.0) 2 (<1.0) 2 (<1.0) 0.48

Length of stay, d, mean � SD 6.99 � 8.30 6.27 � 5.52 5.67 � 5.45 0.0053*

Patient cost, mean � SD (US $) 23,667 � 25,201 20,817 � 16,794 17,515 � 13,157 <0.0001*

*Statistically significant p value.EC, enterocutaneous; GI, gastrointestinal.

1066 Keller et al Robotic Colorectal Surgery Outcomes by Volume J Am Coll Surg

outcomes from the Surveillance, Epidemiology, andEnd Results database, Billingsley and colleagues foundpatients of very-high-volume surgeons had the lowestcomplication rate, and considerably lower odds of compli-cations compared with patients of low-volume surgeons.20

Table 3. Surgeon-Level Outcomes by High, Average, and Low

Class Low (n ¼ 495) A

Overall, n (%) 152 (30.7)

By type, n (%)

Postoperative 133 (26.9)

Postoperative hemorrhage 56 (11.3)

Anastomotic leak 15 (3.0)

Postoperative ileus 86 (17.3)

EC fistula 5 (1.0)

Intraoperative enterotomy 7 (1.4)

Medical, n (%) 40 (8.1)

Infection 14 (2.8)

GI complication 30 (6.1)

Surgical site, n (%) 38 (7.7)

Hematoma/ seroma 13 (2.6)

Wound dehiscence 5 (1.0)

Wound infection 26 (5.3)

Mortality, n (%) 4 (<1.0)

Length of stay, mean � SD 7.77 � 8.23

Patient cost (US $) 23,887 � 22,985

*Statistically significant p value.EC, enterocutaneous; GI, gastrointestinal.

Harmon and colleagues examined the association ofsurgeon and hospital case volumes with the short-termoutcomes in colorectal cancer patients during a 5-yearperiod.21 They found the majority of surgeons (81%)and hospitals (58%) were low-volume group; only a small

Volume

verage (n ¼ 374) High (n ¼ 559) p Value

86 (23.0) 113 (20.2) 0.0003*

69 (18.5) 88 (15.7) <0.0001*

26 (7.0) 19 (3.4) <0.0001*

8 (2.1) 9 (1.6) 0.295

40 (10.7) 59 (10.6) 0.0014*

4 (1.1) 2 (<1.0) 0.36

9 (2.4) 9 (1.6) 0.52

25 (6.7) 31 (5.6) 0.26

13 (3.5) 10 (1.8) 0.26

18 (4.8) 24 (4.3) 0.41

22 (5.9) 32 (5.7) 0.38

10 (2.7) 17 (3.0) 0.91

1 (<1.0) 2 (<1.0) 0.25

14 (3.7) 15 (2.7) 0.096

1 (<1.0) 1 (<1.0) 0.25

5.51 � 5.34 4.95 � 3.86 <0.0001*

19,392 � 15,974 15,726 � 9,778 <0.0001*

Vol. 217, No. 6, December 2013 Keller et al Robotic Colorectal Surgery Outcomes by Volume 1067

minority of surgeons performed more than 10 cases peryear. However, high-volume surgeons were associatedwith substantial improvement in mortality, length ofstay, and hospital costs. Higher-volume hospitals wereable to improve outcomes for medium- and low-volumesurgeons, but outcomes never equaled those of the high-volume surgeon cohort.21

Our findings are consistent with the previously pub-lished results. We found the majority of robotic colorectalsurgery is performed at low-volume centers by low-volumesurgeons. As expected, the overall complication rates arehigher for low-volume providers. In addition, there arehigher rates of postoperative complications, such as post-operative ileus, a major contributor to prolonged lengthof stay and associated increased health care costs.23,24

A learning curve is anticipated when any new tech-nology is introduced, during which an increase in compli-cations can be expected. We can assume the increase incomplications described with robotics is similar to thoseexperiences during the early laparoscopic period.However, as our results demonstrate, the premise of highvolume/high quality is currently not applied in roboticcolorectal surgery. With robotic colorectal surgery, thereis no standardized training program and no expertconsensus on howmuch training is needed, despite rapidlyexpanding use.12 Mastery has been estimated to require3 distinct phases, with the number of procedures varyingwidely, between 25 and 200 total patients.25-27 Trainingmight also not improve the complication rate. In a studyfrom the Mayo Clinic, Woelk and colleagues found 91robotic hysterectomies are needed for surgical competence,however, although the operative time and length of staycan decrease with experience, the complication rate didnot decrease substantially.28 It could also be assumed thatlow-volume robotic providers are also low volume for lapa-roscopic surgery; the data support the concept that there isa considerable learning curve for robotics borne by thepatient and payer. In addition, there is no standardizedreporting for complications. To evaluate complicationrates using a standardized and validated reporting method-ology, Johar and colleagues reassessed 939 robot-assistedradical cystectomies from the International RoboticCystectomy Consortium database.29 The authors foundcomplication rates of 41% at 3 days and 48% at90 days, with mortality rates of 1.3% and 4.2% at 30and 90 days, respectively. The study demonstrates whenusing a standardized reporting tool, complications afterrobot-assisted radical cystectomies are considerable, andunstandardized reporting can overestimate roboticoutcomes and underestimate morbidity.29

There are few complication data in the current litera-ture for comparison. Only one publication addresses

unique complications of robotics in colon and rectalsurgery; this report recognizes that complications haveyet to be fully realized as this new technology undergoesadditional examination.2 Complications are likely under-reported, as there is no standardized reporting system andthe majority of data stem from news media and investorreports.17,30 A review of adverse event reports sent to theFDA shows that robot injuries increased from 24 in2009 to at least 115 in 2012. The robots have been linkedto at least 70 deaths since 2009, with at least 30 in 2012alone.31 An industrial research report questioned the com-pany’s failure to disclose risks leading to surgical compli-cations.32 With these national associations and hospitalboards are publicly opposing robotic surgery for popularprocedures, such as radical prostatectomy and uncompli-cated hysterectomy.18,33 With our findings, providers canalso disfavor the robotic tool at low-volume hospitals orby low-volume robotic surgeons for routine colorectalsurgery.A solution to achieve the best outcomes might be

regionalization. Luft and colleagues described the valueof regionalization for certain operations after findingthe hospitals with high case volume had lower surgicalmortality rates.34 In our series, there was no significantdifference in mortality by hospital volume; however, therewere significant differences in short-term patientoutcomes that could improve by regionalizing roboticcolorectal surgery. Making regional centers for robot-assisted laparoscopic surgery might also be the best modelto make the technology cost effective. When the robot isused to its maximum potential in high-volume centerswith high-volume surgeons, market-driven competitioncan become affordable.35 In addition, the impact ofhospital factors on surgical outcomes has been docu-mented previously.36,37 Regionalization, where the entirehospital system is familiar with the procedures and path-ways, might also have the advantages of improving lengthof stay and short-term outcomes. In the current state, ourresults show that costs are considerably higher in low- andaverage-volume hospitals. Concentrating care in high-volume or regional centers could potentially improvehealth care spending on the new technology. Standard-izing complication reporting can also be a benefit. Arequirement for reporting data on complications,including ureter injuries, hemorrhage, anastomotic leaks,surgical site infections, and medical complications, canhelp guide our use of this new tool.We recognize limitations in the current study. There is

the risk that a large dataset can be manipulated to fit ananalytical model. However, this was not a weakness inour analysis. We did not use a logistic or linear regressionmodel for the analysis. In addition, the large numbers of

1068 Keller et al Robotic Colorectal Surgery Outcomes by Volume J Am Coll Surg

patients from different geographic regions increases thegeneralizability and statistical power of our results. Theresults could potentially be affected by coding error.However, with the large sample size, we would not expectthis to have a substantial effect on our outcomes.

CONCLUSIONSUse of robotic colorectal surgery is rapidly expanding, butthe majority of cases are performed at low-volume hospi-tals and by low-volume surgeons. Both surgeon andhospital volume are significant factors in postoperativeoutcomes. Lower-volume providers, which are perform-ing the majority of procedures, are generating morecomplications, longer hospital lengths of stay, and highercosts of care. These results have a negative impact onquality outcomes measures for those facilities. The tech-nology is evolving, and we continue to follow reports ofoutcomes to help decide the true role for this tool.Limiting robotic colorectal surgery to high-volume hospi-tals and high-volume surgeons can improve patientoutcomes and cost efficiency. In addition, prospectivestudies are needed to define an appropriate role forrobotic technology.

Author Contributions

Study conception and design: Keller, Hashemi, Lu,Delaney

Acquisition of data: Keller, Hashemi, LuAnalysis and interpretation of data: Keller, Hashemi, Lu,Delaney

Drafting of manuscript: Keller, Hashemi, LuCritical revision: Keller, Hashemi, Lu, Delaney

REFERENCES

1. Intuitive Surgical, Inc. Company Profile. 2013;PN 871072 RevM. Available at: http://www.intuitivesurgical.com/company/.Accessed May 2013.

2. Ramamoorthy S, Obias V. Unique complications of roboticcolorectal surgery. Surg Clin North Am 2013;93:273e286.

3. Wexner SD, Bergamaschi R, Lacy A, et al. The current statusof robotic pelvic surgery: results of a multinational interdisci-plinary consensus conference. Surg Endosc 2009;23:438e443.

4. Tyler JA, Fox JP, Desai MM, et al. Outcomes and costs asso-ciated with robotic colectomy in the minimally invasive era.Dis Colon Rectum 2013;56:458e466.

5. Ahmed K, Ibrahim A, Wang TT, et al. Assessing the cost effec-tiveness of robotics in urological surgery�a systematic review.BJU Int 2012;110:1544e1556.

6. Trastulli S, Farinella E, Cirocchi R, et al. Robotic resectioncompared with laparoscopic rectal resection for cancer: system-atic review and meta-analysis of short-term outcome. Colo-rectal Dis 2012;14:e134ee156.

7. deSouza AL, Prasad LM, Park JJ, et al. Robotic assistance inright hemicolectomy: is there a role? Dis Colon Rectum2010;53:1000e1006.

8. Memon S, Heriot AG, Murphy DG, et al. Robotic versus lapa-roscopic proctectomy for rectal cancer: a meta-analysis. AnnSurg Oncol 2012;19:2095e2101.

9. Park JS, Choi GS, Lim KH, et al. S052: a comparison ofrobot-assisted, laparoscopic, and open surgery in the treatmentof rectal cancer. Surg Endosc 2011;25:240e248.

10. Rawlings AL, Woodland JH, Vegunta RK, Crawford DL.Robotic versus laparoscopic colectomy. Surg Endosc 2007;21:1701e1708.

11. Delaney CP, Lynch AC, Senagore AJ, Fazio VW. Comparisonof robotically performed and traditional laparoscopic colo-rectal surgery. Dis Colon Rectum 2003;46:1633e1639.

12. BarbashGI,Glied SA.New technology andhealth care costsdthecase of robot-assisted surgery. NEngl JMed 2010;363:701e704.

13. Wright JD, Ananth CV, Lewin SN, et al. Robotically assistedvs laparoscopic hysterectomy among women with benign gyne-cologic disease. JAMA 2013;309:689e698.

14. Delaney CP, Senagore AJ, Ponsky L. Robot-assisted surgeryand health care costs. N Engl J Med 2010;363:2175. authorreply 2176.

15. Bertani E, Chiappa A, Biffi R, et al. Assessing appropriatenessfor elective colorectal cancer surgery: clinical, oncological, andquality-of-life short-term outcomes employing different treat-ment approaches. Int J Colorectal Dis 2011;26:1317e1327.

16. Schiavone MB, Kuo EC, Naumann RW, et al. The commer-cialization of robotic surgery: unsubstantiated marketing ofgynecologic surgery by hospitals. Am J Obstet Gynecol2012;207:174.e1e174.e7.

17. Citron. Intuitive Surgical’s bad day in court: trial proceeds in threeweeks. 2013. Available at: http://www.citronresearch.com/intuitive-surgicals-bad-day-in-court-liability-trial-proceeds-in-three-weeks/. Accessed May 2013.

18. Breeden JT. Statement on robotic surgery by theAmericanCollegeof Obstetricians and Gynecologists. 2013. Available at: http://www.acog.org/About_ACOG/News_Room/News_Releases/2013/Statement_on_Robotic_Surgery. Accessed May 2013.

19. Jin LX, Ibrahim AM, Newman NA, et al. Robotic surgeryclaims on United States hospital websites. J Healthc Qual2011;33:48e52.

20. Billingsley KG, Morris AM, Green P, et al. Does surgeon casevolume influence nonfatal adverse outcomes after rectal cancerresection? J Am Coll Surg 2008;206:1167e1177.

21. Harmon JW, Tang DG, Gordon TA, et al. Hospital volumecan serve as a surrogate for surgeon volume for achieving excel-lent outcomes in colorectal resection. Ann Surg 1999;230:404e411; discussion 411�413.

22. Office of Statewide Health Planning and Development.Patient discharge data file documentation. 2010; PDD (PublicFile) 2010. Available at: http://www.oshpd.ca.gov/HID/DataFlow/index.html. Accessed May 2013.

23. Basse L, Hjort Jakobsen D, Billesbolle P, et al. A clinicalpathway to accelerate recovery after colonic resection. AnnSurg 2000;232:51e57.

24. Senagore AJ. Pathogenesis and clinical and economic conse-quences of postoperative ileus. Am J Health Syst Pharm2007;64[Suppl 13]:S3eS7.

25. Bokhari MB, Patel CB, Ramos-Valadez DI, et al. Learningcurve for robotic-assisted laparoscopic colorectal surgery.Surg Endosc 2011;25:855e860.

26. Sng KK, Hara M, Shin JW, et al. The multiphasic learningcurve for robot-assisted rectal surgery. Surg Endosc 2013Mar 19 [Epub ahead of print].

Vol. 217, No. 6, December 2013 Keller et al Robotic Colorectal Surgery Outcomes by Volume 1069

27. Jimenez-Rodriguez RM, Diaz-Pavon JM, de la Portilla deJuan F, et al. Learning curve for robotic-assisted laparo-scopic rectal cancer surgery. Int J Colorectal Dis 2012;28:815e821.

28. Woelk JL, Casiano ER, Weaver AL, et al. The learning curveof robotic hysterectomy. Obstet Gynecol 2013;121:87e95.

29. Johar RS, Hayn MH, Stegemann AP, et al. Complicationsafter robot-assisted radical cystectomy: results from the Inter-national Robotic Cystectomy Consortium. Eur Urol 2013;64:52e57.

30. Carreyrou J. Surgical robot examined in injuries. Wall StreetJournal. 2010. Available at: http://online.wsj.com/article/SB10001424052702304703104575173952145907526.html.Accessed May 2013.

31. Langreth R. Intuitive surgical’s robot surgeons encounterhuman lawyers. Bloomberg Businessweek. March 14, 2013.Available at: http://www.businessweek.com/articles/2013-03-14/intuitive-surgicals-robot-surgeons-encounter-human-lawyers.Accessed May 2013.

32. Citron Research Reports on Intuitive Surgical: has thehalo been broken on intuitive surgical? 2012. Available at:http://www.citronresearch.com/has-the-halo-been-broken-on-intuitive-surgical/. Accessed May 2013.

33. Friedrich M, Steiner T, Popken G. [Rage against themachine�necessity of robotic assisted prostatectomy]. AktuelleUrol 2013;44:129e136 [article in German].

34. Luft HS, Bunker JP, Enthoven AC. Should operations beregionalized? The empirical relation between surgical volumeand mortality. N Engl J Med 1979;301:1364e1369.

35. Liberman D, Trinh QD, Jeldres C, Zorn KC. Is roboticsurgery cost-effective: yes. Curr Opin Urol 2012;22:61e65.

36. Holm T, Johansson H, Cedermark B, et al. Influence ofhospital- and surgeon-related factors on outcome after treat-ment of rectal cancer with or without preoperative radio-therapy. Br J Surg 1997;84:657e663.

37. Rogers SOJ, Wolf RE, Zaslavsky AM, et al. Relation ofsurgeon and hospital volume to processes and outcomes ofcolorectal cancer surgery. Ann Surg 2006;244:1003e1011.

1069.e1 Keller et al Robotic Colorectal Surgery Outcomes by Volume J Am Coll Surg

Appendix 1.Procedures by ICD-9 code by hospital volume

Cases by ICD-9 Code

Low (<10)Average(11e20)

High(>20)

Totaln % n % n %

17.32 Laparoscopic Cecectomy 3 15 3 15 14 70.0 20

17.33 Laparoscopic Right Hemicolectomy 59 20.3 42 14.4 190 65.3 291

17.34 Laparoscopic Transverse Colectomy 7 29.2 5 20.8 12 50.0 24

17.35 Laparoscopic Left Hemicolectomy 24 29.2 14 17.1 44 53.7 82

17.36 Laparoscopic Sigmoidectomy 80 18.1 77 17.4 285 64.5 442

45.72 Open/other Cecectomy 1 20.0 1 20.0 3 60.0 5

45.73 Open/other Right Hemicolectomy 13 23.6 17 30.9 25 45.5 55

45.74 Open/other Transverse Colectomy 2 22.2 2 22.2 5 55.6 9

45.75 Open/other Left Hemicolectomy 8 29.6 9 33.3 10.1 37.1 27

45.76 Open/other Sigmoidectomy 27 32.9 19 23.2 36 43.9 82

45.81 Laparoscopic Total Abdominal Colectomy 3 12.5 9 37.5 12 50.0 24

45.82 Open Total Abdominal Colectomy 1 25.0 0 3 75.0 4

48.42 Laparoscopic Pull-through resection of Rectum 0 1 33.3 2 66.7 3

48.43 Open Pull-through Resection of Rectum 1 100 0 0 1

48.51 Laparoscopic Abdominoperineal resection 25 24.1 33 31.7 46 44.2 104

48.52 Open Abdominoperineal resection 5 25.0 9 45.0 6 30.0 20

48.63 Other Laparoscopic Cases 41 17.4 65 27.7 129 54.9 235

Total 300 21.0 306 21.4 822 57.6 1428