Complications and Failure to Rescue After Laparoscopic Versus Open Radical Nephrectomy

Complications and Failure to Rescue After Laparoscopic Versus Open Radical Nephrectomy

Complications and Failure to Rescue After Laparoscopic Versus Open Radical Nephrectomy Hung-Jui Tan, J. Stuart Wolf, Jr., Zaojun Ye, John T. Wei and D...

521KB Sizes 0 Downloads 54 Views

Complications and Failure to Rescue After Laparoscopic Versus Open Radical Nephrectomy Hung-Jui Tan, J. Stuart Wolf, Jr., Zaojun Ye, John T. Wei and David C. Miller* From the Dow Division of Health Services Research (HJT, ZY, JTW, DCM) and Michigan Center for Minimally Invasive Urology (JSW), Department of Urology and Center for Healthcare Outcomes and Policy (JTW, DCM), University of Michigan, Ann Arbor, Michigan

Abbreviations and Acronyms ARF ⫽ acute renal failure excluding chronic dialysis FTR ⫽ failure to rescue FTR-M ⫽ FTR after major complication GI ⫽ gastrointestinal GU ⫽ genitourinary LRN ⫽ laparoscopic radical nephrectomy ORN ⫽ open radical nephrectomy PSI ⫽ Patient Safety Indicator RCC ⫽ renal cell carcinoma VTE ⫽ postoperative pulmonary embolism or deep vein thrombosis Submitted for publication February 8, 2011. Supported by the Edwin Beer Research Fellowship in Urology and Urology-Related Fields from the New York Academy of Medicine, University of Michigan Comprehensive Cancer Center, and Agency for Healthcare Research and Quality (K08 HS018346-01A1) (DCM). * Correspondence: Department of Urology, University of Michigan, North Campus Research Complex, 2800 Plymouth Rd., Building 520, 3rd Floor, #3172, Ann Arbor, Michigan 48109-2800 (telephone: 734-936-0054; FAX: 734-232-2400; e-mail: [email protected]).

Purpose: Since to our knowledge the population level impact of laparoscopy on postradical nephrectomy morbidity and mortality remains unknown, we compared the rates of postoperative complications and failure to rescue (the fatality rate in patients with a complication) in patients treated with laparoscopic vs open radical nephrectomy. Materials and Methods: Using linked SEER (Surveillance, Epidemiology and End Results)-Medicare data we identified patients with kidney cancer who were treated with laparoscopic or open radical nephrectomy from 2000 through 2005. After measuring the frequency of postoperative complications and failure to rescue we fit multivariate logistic regression models to estimate the association of these outcomes with surgical approach, adjusting for patient characteristics, cancer severity and surgery year. We also assessed the relationship between case volume, complications and failure to rescue. Results: We identified 2,108 (26%) and 5,895 patients (74%) treated with laparoscopic and open radical nephrectomy, respectively. The overall rates of complications and failure to rescue were 36.9% and 5.3%, respectively. The predicted probability of any, major, medical and surgical complications was 15%, 12%, 13% and 23% lower, respectively, after laparoscopic than after open radical nephrectomy (each p ⬍0.05). Despite less frequent complications patients treated with laparoscopic radical nephrectomy had a greater probability of failure to rescue (7.6% vs 4.6%, p ⫽ 0.010). Higher volume surgeons and hospitals had a lower rate of failure to rescue in patients treated with radical nephrectomy (each p ⬍0.05) but not with open radical nephrectomy. Conclusions: Supporting the decreased morbidity of laparoscopy, patients treated with radical nephrectomy had fewer complications than those who underwent open radical nephrectomy. However, failure to rescue was more common in patients with a complication after radical nephrectomy, suggesting that these events may be more difficult to recognize and manage successfully, especially among less experienced surgeons and hospitals. Key Words: kidney, nephrectomy, laparoscopy, mortality, complications

1254

www.jurology.com

LAPAROSCOPIC radical nephrectomy is the preferred surgical approach in most patients who require total kidney removal for RCC.1 Compared with ORN, LRN provides easier, more rapid convalescence while achieving equivalent cancer control.2,3 Given these

patient benefits, LRN has been adopted steadily during the last 2 decades and is now performed by most urologists in the United States.4 Although early adopters of LRN reported complication rates equivalent to those of the open technique,2,3,5 to

0022-5347/11/1864-1254/0 THE JOURNAL OF UROLOGY® © 2011 by AMERICAN UROLOGICAL ASSOCIATION EDUCATION

Vol. 186, 1254-1260, October 2011 Printed in U.S.A. DOI:10.1016/j.juro.2011.05.074

AND

RESEARCH, INC.

FAILURE TO RESCUE AFTER LAPAROSCOPIC VERSUS OPEN RADICAL NEPHRECTOMY

our knowledge it remains unknown whether the widespread dissemination of this groundbreaking procedure changed the frequency and nature of postradical nephrectomy complications at the population level. Examples show that the dissemination of a complex new technology has yielded intended benefits and unintended consequences. For instance, during its rapid incorporation into general surgery practice laparoscopic cholecystectomy was unexpectedly associated with a higher rate of common bile duct injury.6 While these injuries were initially accepted as part of the learning curve, they have since been linked to decreased patient survival.7 Complications similar in severity, eg ligation of the superior mesenteric artery and bowel fistula formation,8 have been reported early in the LRN experience but additional data are needed to clarify the impact of the widespread adoption of laparoscopy on population level morbidity and mortality in patients who undergo radical nephrectomy for RCC. In this context we used linked SEER-Medicare data to compare the frequency of complications and the case fatality rate after a complication, so-called FTR, in patients treated with LRN vs ORN. By defining more precisely the prevalence and consequences of complications after LRN vs ORN these data will facilitate efforts to optimize patient safety and better understand the comparative effectiveness of LRN at the population level.

MATERIALS AND METHODS Data Source, Cohort Identification and Surgery Assignment We used linked data from the National Cancer Institute SEER Program, and the Centers for Medicare and Medicaid Services to identify 12,031 patients diagnosed with nonurothelial, nonmetastatic kidney cancer from 2000 through 2005, which was a period of widespread urologist adoption of LRN.4,9 Using a validated algorithm9 we determined the specific procedure in each patient treated surgically. We then limited our sample to the 8,003 patients treated with unilateral LRN or ORN for localized or regional kidney cancer.

Patient Level Covariates For each patient we used SEER data to determine demographic and cancer specific information, including age, gender, race/ethnicity, marital status and cancer severity, ie tumor size and stage. We assigned patients to 1 of 3 socioeconomic strata based on patient level Zip Codes10 and measured preexisting comorbidity using a modification of the Charlson index.11

Primary Outcomes For patients treated with LRN or ORN we assessed 2 outcomes during the index hospitalization or within 30 days of surgery, including 1) postoperative complications and 2) FTR. Guided by the published literature12–14 we identified specific ICD-9 codes for GI complications, car-

1255

diac complications, ARF,14 GU complications, postoperative hemorrhage, postoperative infection (eg pneumonia or Clostridium difficile), wound complications, pulmonary failure, sepsis, neurological complications and miscellaneous technical complications related to surgery. Each measure was described previously by the Complications Screening Program and validated through chart review.12,13 Our catalogue of complications also included PSIs, version 4.2, developed at the Agency for Healthcare Research and Quality, including 1) foreign body left during the procedure, 2) iatrogenic pneumothorax, 3) VTE and 4) accidental puncture or laceration, most often injury to the GI tract, bladder or blood vessel, and referred to in this study as iatrogenic injury.15,16 We selected these PSIs based on previous validation studies and compatibility with our data set.15–17 For analytical purposes we grouped specific complications and PSIs into certain categories, including 1) any complication, 2) major complications (ie pulmonary, cardiac, ARF, VTE, GI, sepsis, wound, iatrogenic injury and hemorrhage), 3) medical complications (ie pulmonary, cardiac, ARF, VTE, GI, sepsis, infection, neurological and infection) and 4) surgical complications (ie iatrogenic injury, wound, GU, hemorrhage, miscellaneous, foreign body left during the procedure and iatrogenic pneumothorax). Finally, we determined procedure specific complication rates by dividing the number of cases in each complication category by the number of patients who underwent LRN or ORN. FTR is accepted widely as a measure of the timely recognition of and successful management for complications. Measurement of this outcome is now used by various stakeholders, including Agency for Healthcare Research and Quality policy makers, to assess the quality of inpatient care.15,18 Consistent with the existing literature, we defined FTR as death during the index hospitalization or within 30 days of surgery in a patient who experienced any of the defined complications.18 In other words, FTR represents the case fatality rate among patients with a postoperative complication. We calculated procedure specific rates of FTR by dividing the number of patients who died after a complication by the total number who experienced a complication. Since minor complications that are unlikely to lead to death may falsely increase the overall FTR rate, we also evaluated and calculated the FTR rate in patients with a major complication, referred to as FTR-M.19

Primary Statistical Analysis We used chi-square tests to evaluate the association between surgical approach (LRN vs ORN) and patient level covariates. We then used the chi-square test or Fisher exact test as appropriate to compare unadjusted rates of complications and FTR in patients treated with LRN vs ORN. We fit multivariate logistic regression models to estimate the association between surgical approach and our primary outcomes. We specified each outcome (complications and FTR) as a binary (yes/no) variable. We implemented generalized estimating equations to account for patient outcome clustering at hospitals and adjusted our

1256

FAILURE TO RESCUE AFTER LAPAROSCOPIC VERSUS OPEN RADICAL NEPHRECTOMY

models for patient characteristics, cancer severity and surgery year. We then calculated the covariate adjusted predicted probability of each outcome in patients treated with LRN or ORN. We also refit our models after excluding certain complications, including 1) ARF since some degree of renal insufficiency is expected after nephrectomy, and 2) GU complications since to our knowledge its claims based definition has not been previously validated. Finally, using previously described methods20 we determined the average annual nephrectomy case volume (open and laparoscopic) in the SEER-Medicare population for each surgeon and hospital. Using the Mantel-Haenszel chi-square test we examined the frequency of complications and FTR stratified by case volume quartiles and according to surgical approach. All statistical testing was 2-sided, done using SAS®, version 9.2 and performed at the 5% significance level. This study was deemed exempt by the University of Michigan institutional review board.

RESULTS We identified 2,108 patients (26%) treated with LRN and 5,895 (74%) treated with ORN from 2000 through 2005. Patients who underwent LRN were more often white, female and of higher socioeconomic position (each p ⬍0.05, Table 1). Patients treated with LRN were also more likely to have a tumor 4 cm or less and surgery performed after 2002 (each p ⬍0.001). The overall rate of complications and FTR was 36.9% and 5.3%, respectively. Compared to the open approach, patients treated with LRN had a lower unadjusted rate of any postoperative complication (31.7% vs 38.8%, p ⬍0.001). Figure 1 shows the frequency of specific types of complications by surgical approach. Overall rates of postoperative sepsis, ARF, infection, pulmonary failure, GU complications and iatrogenic injury were lower in patients treated with LRN than with ORN (each p ⬍0.05). Although most complications were less common after LRN, patients treated laparoscopically had a higher unadjusted FTR rate (7.0% vs 4.8%, p ⫽ 0.031). This relationship persisted in the subgroup with a major complication (8.0% vs 5.5%, p ⫽ 0.030). Table 2 lists the rates of specific complications in the 49 patients with LRN and the 126 with ORN who died, providing context for these observations. Neurological complications were more common in those who died after ORN (p ⫽ 0.021). Conversely GI complications and iatrogenic injury were more common in patients who died after LRN, although these associations did not quite attain statistical significance (p ⫽ 0.057 and 0.056, respectively). After adjusting for patient characteristics, cancer severity and surgery year the likelihood of any (OR 0.78, 95% CI 0.69 – 0.88), major (OR 0.83, 95% CI 0.73– 0.94), medical (OR 0.82, 95% CI 0.73– 0.93) and surgical (OR 0.76, 95% CI 0.64 – 0.90) complica-

Table 1. Patient characteristics by surgical approach

Age: 65–69 70–74 75–79 80–84 Greater than 84 Race: White Black Hispanic Other Gender: M F Married: Yes No Socioeconomic status tertile: High Intermediate Low Charlson comorbidity index: 0 1 2 or Greater Tumor size (cm): 4 or Less 4–7 Greater than 7 Tumor stage: Local or in situ Regional Surgery yr: 2000–2002 2003–2005

% LRN

% ORN

p Value

25.8 25.6 27.9 14.5 6.2

25.5 28.5 24.8 15.5 5.7

0.014

84.3 7.4 4.0 4.3

82.0 7.4 7.1 3.5

⬍0.001

55.0 45.0

58.0 42.0

0.018

61.3 38.7

61.7 38.3

0.761

40.2 30.2 29.7

31.3 32.5 36.3

⬍0.001

57.6 25.2 17.2

57.8 25.0 17.2

0.982

45.8 39.2 15.0

36.1 36.4 27.5

⬍0.001

80.1 19.9

72.5 27.5

⬍0.001

25.7 74.3

55.3 44.7

⬍0.001

tions was significantly lower in patients treated with LRN. The predicted probability of any, major, medical and surgical complications was 15%, 12%, 13% and 23% lower, respectively, after LRN than after ORN (fig. 2). Although it was less common, patients treated with LRN were more likely to die if they experienced any complication (FTR OR 1.65, 95% CI 1.13–2.42) or a major complication (FTR-M OR 1.65, 95% CI 1.13–2.41). The predicted probability of FTR and FTR-M was 62% and 59% greater, respectively, in patients treated with LRN than with ORN (fig. 2). Our findings did not change substantively when we excluded ARF and GU complications. Figures 3 and 4 show the volume-outcome relationships for postoperative complications and FTR, respectively. For LRN and ORN higher volume hospitals had lower rates of any, major, medical and surgical complications for each surgical approach. Although it was not quite statistically significant (LRN and ORN p ⫽ 0.057 and 0.097, respectively), we observed a similar relationship between surgeon volume and complication frequency. Higher volume

FAILURE TO RESCUE AFTER LAPAROSCOPIC VERSUS OPEN RADICAL NEPHRECTOMY

1257

Figure 1. Specific complication frequency by surgical approach. Miscellaneous complications include miscellaneous technical complications, foreign body left during procedure and iatrogenic pneumothorax. Asterisk indicates significant difference between treatment groups (chi-square test p ⬍0.05).

surgeons and hospitals had lower FTR and FTR-M rates in patients treated laparoscopically than with ORN (each p ⬍0.05).

DISCUSSION In a nationally representative sample of Medicare beneficiaries patients treated with LRN vs ORN had a lower likelihood of postoperative complications, including lower sepsis, pulmonary failure, iatrogenic injury, infection, ARF and GU complication rates. Overall these data provide population level support for LRN as the standard of care in patients who require radical nephrectomy for RCC. Table 2. Complication rates in patients who died after LRN or ORN Complication Category

% LRN

% ORN

p Value

Any Major Medical Surgical GI Sepsis ARF Postop infection Postop hemorrhage Cardiac Pulmonary failure GU Neurological Wound VTE Accidental puncture or laceration Miscellaneous technical*

95.9 95.9 91.8 34.7 28.6 30.6 40.0 26.5 22.5 28.6 49.0 2.0 — 6.1 10.2 14.3 4.1

87.3 84.9 84.9 31.0 15.9 31.8 37.3 31.8 18.3 27.0 46.8 7.9 10.3 4.0 5.6 5.6 2.4

0.105 0.067 0.320 0.634 0.057 0.885 0.750 0.500 0.529 0.833 0.798 0.295 0.021 0.688 0.320 0.056 0.620

* Includes foreign body left during procedure and iatrogenic pneumothorax.

However, at the same time patients treated with LRN who experienced a complication were more likely to die as a consequence of this adverse event, ie less likely to be rescued after a complication. The finding that GI complications and iatrogenic injuries were almost twice as common in patients who died after LRN suggests that challenges associated with recognizing and managing bowel and/or vascular injury may account for the higher FTR rates seen with the laparoscopic approach. Moreover, since FTR rates were highest for low volume surgeons and hospitals, these diagnostic and therapeutic challenges may be most pronounced early in the laparoscopic experience, reflecting the substantial learning curve associated with this surgical innovation.3,8 Although our series is 1 of the only studies of FTR in urological surgery, others have used population level assessments to examine the frequency of postradical nephrectomy complications.21,22 Distinct from this prior investigation, we report postoperative complication rates stratified by surgical approach and also observed a relatively higher overall frequency of complications. While the latter discrepancy could be due to an older and perhaps less healthy patient population, it may also reflect methodological differences, including our efforts to assess a wider range of complications and ascertain adverse events that developed after discharge home but during the immediate postoperative period. The complication rates for LRN that we report are similar to those noted by early adopters at our and other institutions.3,5,8 Moreover, while to our knowledge no prior empirical study has compared FTR for open and laparoscopic surgery, the higher FTR rate observed

1258

FAILURE TO RESCUE AFTER LAPAROSCOPIC VERSUS OPEN RADICAL NEPHRECTOMY

Figure 2. Predicted probability of complications (A) and FTR (B) by surgical approach adjusted for patient characteristics, cancer severity and surgery year.

for LRN is in many ways consistent with previous studies showing that bowel and vascular injuries, which are 2 of the most common major complications after LRN, are the most lethal complications after laparoscopic cholecystectomy.3,6,8

Our findings should be considered in the context of several limitations. 1) Like those of any observational study, our conclusions are susceptible to selection bias due to unobserved confounders. For instance, the selection of more favorable patients for

Figure 3. ORN (A and B) and LRN (C and D) complication frequency by hospital (A and C) and surgeon (B and D) case volume quartiles. Surgeon volume quartile thresholds were bottom—mean 1 case or fewer, 2nd—1.1 to 1.4, 3rd—1.5 to 2.0 and top— greater than 2 annually. Hospital volume quartile thresholds were bottom—mean 2 or fewer cases, 2nd—2.1 to 3.7, 3rd—3.8 to 5.9 and top— 6 or greater annually. Association with volume was assessed by Mantel-Haenszel chi-square test.

FAILURE TO RESCUE AFTER LAPAROSCOPIC VERSUS OPEN RADICAL NEPHRECTOMY

Figure 4. FTR and FTR-M by hospital (A) and surgeon (B) case volume quartiles, stratified by surgical approach. Surgeon volume quartile thresholds were bottom—mean 1 case or fewer, 2nd—1.1 to 1.4, 3rd—1.5 to 2.0 and top— greater than 2 annually. Hospital volume quartile thresholds were bottom—mean 2 or fewer cases, 2nd—2.1 to 3.7, 3rd—3.8 to 5.9 and top— 6 or greater annually. Association with volume was assessed by Mantel-Haenszel chi-square test.

LRN may explain their lower complication rates. On the other hand, it can be argued that, if present, the same concern would have biased our results toward lower rather than higher FTR rates in patients treated laparoscopically. 2) Although using claim data to identify complications was validated previously,12,13 our findings are subject to the accuracy of coding practices. However, any measurement bias due to this concern would most likely be nondifferential with respect to the surgical approach. 3) Many recent evaluations have incorporated standardized methods (eg the Clavien-Dindo system) to quantify surgical complication severity.23 Although we were unable to apply such a system explicitly, the defini-

1259

tion of FTR is analogous to that of a Clavien grade V complication, providing at least 1 point of reference to assess severity.18,23 4) Since this study was limited to Medicare beneficiaries, our results are not necessarily applicable to younger patients with RCC. These limitations notwithstanding, our findings have several implications for the care of patients with RCC. In addition to easier convalescence, patients treated with LRN appear to have fewer postoperative complications. Also, given the observed relationships with case volume and the substantial number of new adopters during the study interval,4 continued experience with this advanced surgical technique will most likely yield further decreases in the frequency of complications and FTR after LRN. However, simultaneously the increased risk of death after a complication in patients treated with LRN, especially for lower volume surgeons and hospitals, highlights the importance of surgeon preparation and hospital readiness to adopt new surgical technology, including efforts aimed at improving the timely recognition of and management for common and rare complications. For newly adopting surgeons the ability to prevent, recognize and address adverse events may be achieved through formalized fellowships that emphasize the cognitive and technical skills required for safe implementation.24 Alternatively sufficient skill may be acquired through more condensed training protocols based on paired mentoring and/or high intensity educational programs.25,26 At the specialty level the establishment of practice guidelines, certification requirements and/or standardized credentialing may further encourage new adopters to pursue formalized training programs.27 For hospitals, ensuring adequate resource availability and ancillary support (eg nurse staffing levels and education, and access to technology/services) may increase the rescue rate for patients with complications related to new surgical technologies.18,19,28 Beyond these structural components, implementing specific processes of care may improve patient outcomes across a wide spectrum of surgical interventions, including those involving complex new technologies. For instance, the Surviving Sepsis Campaign is associated with decreased mortality at participating hospitals.29 Moreover, comprehensive, multidisciplinary surgical safety systems that span the entire episode of care decrease postoperative complications and mortality by directly preventing adverse events, optimizing processes of care and enhancing collaboration among caregivers.30 By incorporating similar systems hospitals may attenuate the risk of untoward events during the introduction of future innovations in urology (eg natural orifice transluminal endoscopic surgery) and other surgical disciplines.

1260

FAILURE TO RESCUE AFTER LAPAROSCOPIC VERSUS OPEN RADICAL NEPHRECTOMY

CONCLUSIONS In a nationally representative sample of Medicare beneficiaries patients treated with LRN had a lower likelihood of postoperative complications than those treated with the open approach. However, these patients were more likely to die after a complication, suggesting that adverse events may

be more difficult to recognize or manage successfully. Together these data support LRN as the standard of care in patients with RCC who require radical nephrectomy while at the same time underscoring the need for efforts aimed at maximizing patient safety during the diffusion of surgical innovations.

REFERENCES 1. Novick AC, Campbell SC, Belldegrun A et al: Guideline for Management of the Clinical Stage 1 Renal Mass, 2009. Available at http://www. auanet.org/content/guidelines-and-quality-care/ clinical-guidelines.cfm. Accessed March 25, 2011. 2. Wolf JS, Merion RM, Leichtman AB et al: Randomized controlled trial of hand-assisted laparoscopic versus open surgical live donor nephrectomy. Transplantation 2001; 72: 284. 3. Dunn MD, Portis AJ, Shalhav AL et al: Laparoscopic versus open radical nephrectomy: a 9-year experience. J Urol 2000; 164: 1153. 4. Filson CP, Banerjee M, Wolf JS et al: Surgeon characteristics and long-term trends in the adoption of laparoscopic radical nephrectomy. J Urol 2011; 185: 2072. 5. Hollingsworth JM, Miller DC, Dunn RL et al: Surgical management of low-stage renal cell carcinoma: technology does not supersede biology. Urology 2006; 67: 1175. 6. Deziel DJ, Millikan KW, Economou SG et al: Complications of laparoscopic cholecystectomy: a national survey of 4,292 hospitals and an analysis of 77,604 cases. Am J Surg 1993; 165: 9.

11. Deyo RA, Cherkin DC and Ciol MA: Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol 1992; 45: 613. 12. Iezzoni LI, Daley J, Heeren T et al: Identifying complications of care using administrative data. Med Care 1994; 32: 700. 13. Weingart SN, Iezzoni LI, Davis RB et al: Use of administrative data to find substandard care: validation of the complications screening program. Med Care 2000; 38: 796. 14. Waikar SS, Wald R, Chertow GM et al: Validity of International Classification of Diseases, Ninth Revision, Clinical Modification codes for acute renal failure. J Am Soc Nephrol 2006; 17: 1688. 15. Guide to the Patient Safety Indicators: Agency for Healthcare Research and Quality, 2007. Available at http://www.qualityindicators.ahrq.gov/downloads/ psi/psi_guide_v31.pdf. Accessed April 1, 2011. 16. Utter GH, Zrelak PA, Baron R et al: Positive predictive value of the AHRQ accidental puncture or laceration patient safety indicator. Ann Surg 2009; 250: 1041.

21. Joudi FN, Allareddy V, Kane CJ et al: Analysis of complications following partial and total nephrectomy for renal cancer in a population based sample. J Urol 2007; 177: 1709. 22. Mitchell RE, Lee BT, Cookson MS et al: Radical nephrectomy surgical outcomes in the University HealthSystem Consortium Data Base: Impact of hospital case volume, hospital size, and geographic location on 40,000 patients. Cancer 2009; 115: 2447. 23. Dindo D, Demartines N and Clavien PA: Classification of surgical complications: a new proposal with evaluation in a cohort of 6336 patients and results of a survey. Ann Surg 2004; 240: 205. 24. Cadeddu JA, Wolf JS, Nakada S et al: Complications of laparoscopic procedures after concentrated training in urological laparoscopy. J Urol 2001; 166: 2109. 25. Jones A, Eden C and Sullivan ME: Mutual mentoring in laparoscopic urology— a natural progression from laparoscopic fellowship. Ann R Coll Surg Engl 2007; 89: 422. 26. Kolla SB, Gamboa AJ, Li R et al: Impact of a laparoscopic renal surgery mini-fellowship program on postgraduate urologist practice patterns at 3-year followup. J Urol 2010; 184: 2089.

7. Flum DR, Cheadle A, Prela C et al: Bile duct injury during cholecystectomy and survival in Medicare beneficiaries. JAMA 2003; 290: 2168.

17. White RH, Sadeghi B, Tancredi DJ et al: How valid is the ICD-9-CM based AHRQ patient safety indicator for postoperative venous thromboembolism? Med Care 2009; 47: 1237.

8. Gill IS, Kavoussi LR, Clayman RV et al: Complications of laparoscopic nephrectomy in 185 patients: a multi-institutional review. J Urol 1995; 154: 479.

18. Silber JH, Williams SV, Krakauer H et al: Hospital and patient characteristics associated with death after surgery. A study of adverse occurrence and failure to rescue. Med Care 1992; 30: 615.

9. Miller DC, Saigal CS, Warren JL et al: External validation of a claims-based algorithm for classifying kidney-cancer surgeries. BMC Health Serv Res 2009; 9: 92.

19. Ghaferi AA, Osborne NH, Birkmeyer JD et al: Hospital characteristics associated with failure to rescue from complications after pancreatectomy. J Am Coll Surg 2010; 211: 325.

29. Levy MM, Dellinger RP, Townsend SR et al: The Surviving Sepsis Campaign: results of an international guideline-based performance improvement program targeting severe sepsis. Intensive Care Med 2010; 36: 222.

10. Diez Roux AV, Merkin SS, Arnett D et al: Neighborhood of residence and incidence of coronary heart disease. N Engl J Med 2001; 345: 99.

20. Miller DC, Saigal CS, Banerjee M et al: Diffusion of surgical innovation among patients with kidney cancer. Cancer 2008; 112: 1708.

30. de Vries EN, Prins HA, Crolla RM et al: Effect of a comprehensive surgical safety system on patient outcomes. N Engl J Med 2010; 363: 1928.

27. Zorn KC, Gautam G, Shalhav AL et al: Training, credentialing, proctoring and medicolegal risks of robotic urological surgery: recommendations of the society of urologic robotic surgeons. J Urol 2009; 182: 1126. 28. Aiken LH, Clarke SP, Cheung RB et al: Educational levels of hospital nurses and surgical patient mortality. JAMA 2003; 290: 1617.