Predictors of Major Complications after Laparoscopic Cholecystectomy: Surgeon, Hospital, or Patient?

Transcription

1 Predictors of Major Complications after Laparoscopic Cholecystectomy: Surgeon, Hospital, or Patient? Melissa M Murphy, MD, MPH, Sing-Chau Ng, MS, Jessica P Simons, MD, MPH, Nicholas G Csikesz, MD, Shimul A Shah, MD, FACS, Jennifer F Tseng, MD, MPH, FACS BACKGROUND: STUDY DESIGN: RESULTS: CONCLUSIONS: Regionalization of care has been proposed for complex operations based on hospital/surgeon volume mortality relationships. Controversy exists about whether more common procedures should be performed at high-volume centers. Using mortality alone to assess routine operations is hampered by relatively low perioperative mortality. We used a large national database to analyze the risk of major in-hospital complications after laparoscopic cholecystectomy (LC). Patients undergoing LC were identified in the Nationwide Inpatient Sample from states with surgeon/hospital identifiers. Previously validated major complications including acute myocardial infarction, pulmonary compromise, postoperative infection, deep vein thrombosis, pulmonary embolism, hemorrhage, and reoperation were assessed. Univariate and multivariable analyses were performed and independent risk factors of complications were identified. A total of 1,102,071 weighted patient discharges were identified, with a complication rate of 6.8%. Univariate analyses showed that advanced age, male gender, and higher Charlson Comorbidity Score were associated with higher complication rates (p ). Higher surgeon volume ( 36/year versus 12/year) and higher hospital volume ( 225/year versus 120/ year) were associated with fewer complications (6.7% versus 7.0%, 6.4% versus 7.0%, respectively; p ). Multivariable analysis showed that advanced age (65 years or older versus younger than 65 years; adjusted odds ratio [AOR] 2.16; 95% CI, ), male gender (AOR 1.14; 95% CI, ), and comorbidities (Charlson Comorbidity Score 2 versus 0; AOR 2.49; 95% CI, ) were associated with complications. Neither surgeon nor hospital volume was independently associated with increased risk of complications. Major in-hospital complications after LC are associated with individual patient characteristics rather than surgeon or hospital operative volumes. These results suggest regionalization of general surgical procedures might be unnecessary. Rather, careful patient selection and preoperative preparation can diminish overall complication rates. (J Am Coll Surg 2010;211: by the American College of Surgeons) Luft and colleagues first reported the inverse relationship between hospital operative volume and mortality in In the ensuing 30 years, this relationship has been Disclosure Information: Nothing to disclose. This work was supported by the Evans-Allen-Griffin Fellowship (Murphy), by the American Surgical Association Foundation, and a Howard Hughes Early Career Award (Tseng). Abstract presented at the American College of Surgeons 95 th Annual Clinical Congress, Surgical Forum, Chicago, IL, October Received 25 November 2009; Revised 23 February 2010; Accepted 23 February From the Department of Surgery, Surgical Outcomes Analysis and Research, University of Massachusetts Medical School, Worcester, MA. Correspondence address: Jennifer F Tseng, MD, MPH, Department of Surgery, Surgical Outcomes Analysis and Research (SOAR), University of Massachusetts Medical School, 55 Lake Ave North, Suite S3-752, Worcester, MA Jennifer.Tseng@umassmemorial.org further examined, with a particular emphasis on complex surgical procedures, including pancreatectomy, 2,3 esophagectomy, 4 and cardiac surgery. 5,6 The effect of volume metrics, specifically hospital and surgeon volume, on outcomes for more common surgical procedures has been less well-studied. Routine general surgery procedures, including laparoscopic appendectomy 7,8 and laparoscopic cholecystectomy (LC) 9 have associated perioperative mortality rates of 1%, leaving mortality a difficult marker to assess patient outcomes. Morbidity, specifically development of postoperative complications, can be a more useful method to evaluate outcomes after common surgical procedures. Additionally, investigating whether hospital and/or surgeon volume impacts patient outcomes for common surgical procedures would be a useful addition to our current un by the American College of Surgeons ISSN /10/$36.00 Published by Elsevier Inc. 73 doi: /j.jamcollsurg

2 74 Murphy et al Complications after Cholecystectomy J Am Coll Surg Abbreviations and Acronyms AOR adjusted odds ratio CCS Charlson Comorbidity Score LC laparoscopic cholecystectomy NIS Nationwide Inpatient Sample OC open cholecystectomy derstanding of the interplay between volume and outcomes. LC is currently the gold standard in operative management for the majority of gallbladder disease, with 80% of cholecystectomies performed laparoscopically in the United States. 10 Perioperative mortality associated with LC is 1%, 9 and reported morbidity ranges from 2% to 17% We sought to determine whether annual surgeon and/or hospital volume were associated with development of major postoperative in-hospital complications for patients undergoing LC using a large national database. METHODS A retrospective analysis was performed using discharge records from the Nationwide Inpatient Sample (NIS) from for all patients who underwent LC for which surgeon and hospital identifiers were available. The NIS, a part of the Healthcare Cost and Utilization Project, is a national, all-payer discharge database containing information for approximately 7 million hospital discharges annually. This represents a stratified sample of 20% of non- Federal US community hospitals from participating states, including academic and specialty hospitals. The NIS weighting strategy facilitates population-based estimates to be drawn at the national level. All statistical analyses were performed based on these survey weights; therefore, results are reported as either unweighted (actual) or weighted (national) frequencies. 3 This study was reviewed by and received exemption from the University of Massachusetts Institutional Review Board because no personal identifiers are listed in the NIS data. Cohort assembly The ICD-9-CM 14 procedure codes were used to identify patients undergoing LC (51.23, with concomitant V 64.4, V64.41). We excluded patients younger than 18 years of age or older than 95 years of age from additional analyses. Patients with newborn, trauma, or other admission type were excluded. Provider identifiers Hospital and surgeon identifiers were used to determine the number of LC per individual hospital and surgeon per year. Because record sampling across the NIS does not correlate across years, a continuous single surgeon identifier during the time period of the study was not possible. Therefore, all volume calculations were done on a year-toyear basis as described previously. 15 To evaluate hospital volume, we categorized groups based on approximately equal sizes for purposes of comparison before any volume outcomes data analysis, as described previously. 3,16,17 The learning curve for surgeons to master the technique of LC in the literature has been reported to be between 10 and 30 cases Based on this data, to evaluate surgeon volume, we categorized surgeons by whether they performed 12 LC annually (low volume), 12 to 35 LC annually (medium volume), and 35 LC annually (high volume). Variables Patient demographic characteristics compiled in the NIS were identified. Age was grouped into younger than 35 years of age, 35 to 64 years, and 65 years of age or older. Race was categorized as white, black, Hispanic, and other, including Asian, Pacific Islander, and Native American. Race was missing in 4.7% of the cohort (52,099 weighted patient discharges). Patient comorbidities were assessed by the Romano modification of the Charlson Comorbidity Score (CCS). 21,22 The CCS score uses ICD-9 diagnosis and procedure codes to give a weighted, risk-adjusted comorbidity index value for an individual patient. The CCS takes into consideration any comorbid conditions that occurred within a year before diagnosis and that met the inclusion criteria as defined by the index parameters. We collapsed the groups as follows: 0 (no pre-existing comorbidity), 1 (1 comorbid condition), 2 (2 comorbid conditions), and 3 ( 3 comorbid conditions). Patients were categorized into 2 broad groups representing severity of disease by whether biliary tract inflammation or cancer was present as described previously. 13 Biliary tract inflammation was defined by the presence of acute cholecystitis, chronic cholecystitis, cholangitis, or biliary malignancy using ICD-9 diagnoses codes (575.0, , , , 574.8, 574.6, 574.3, 576.1, 577.0, 577.1, 156.0, 156.1, 156.2, 156.8, 156.9, 155.0, 155.1, 571.6, 574.1, 575.1, 574.7, 574.4, 575.5, 576.4, 576.3). Cholelithiasis without the presence of cholecystitis was defined as having no biliary tract inflammation (ICD-9 diagnoses codes; 574.2, 574.9, 574.5). Other diagnosis/indications for LC were excluded from additional analyses. Hospital admission status was categorized as either elective or emergent and hospital teaching status was categorized as teaching or nonteaching institutions.

3 Vol. 211, No. 1, July 2010 Murphy et al Complications after Cholecystectomy 75 Outcomes The primary outcomes measure was development of a major postoperative in-hospital complication. The specific diagnoses and codes were chosen based on their validation as true complications rather than comorbidities using ICD-9 diagnoses and procedure codes by the work of Lawthers and colleagues. 23,24 Postoperative complications were defined by secondary diagnoses, including postoperative infection (except wound and pneumonia) (008.45, , 510.0, 510.9, 513.1, 519.2, , , 683); acute MI ( ); aspiration pneumonia (507.0); deep venous thrombosis and pulmonary embolism (415.1, , , 451.2, , 453.8, , 453.4, , , ); postoperative pulmonary compromise (514, 518.4, 518.5, , ); postoperative gastrointestinal hemorrhage (530.82, , , , , , , , , , , , , , , , , , , , 578.9; reopening of a surgical site (01.23, 03.02, 06.02, 34.03, 35.95, 39.49, 54.12, 54.61); and procedure-related lacerations or perforations (530.4, , 575.4, 29.51, 31.61, 33.41, 33.43, 42.82, 44.61, 46.71, 46.75, 48.71, 50.61, 51.91, 55.81, 56.82, 57.81, 58.41, 69.41). Inpatient mortality was included as a postoperative complication. Conversion from laparoscopic to open cholecystectomy (OC) was evaluated as one of the secondary outcomes. Statistical analysis All data were analyzed using the advanced survey procedures in the SAS statistical software (version 9.1; SAS Institute Inc). Univariate analyses of categorical variables were performed using Rao-Scott chi-square tests, with p 0.05 considered statistically significant. Trend tests were performed using the Cochrane-Armitage trend test. Multivariate logistic regression was constructed with the occurrence of a major in-hospital complication as the dependent variable. Independent variables included patient age group (younger than 35, 65 to 64, 65 year of age or older), gender, insurance status, comorbidities (CCS), and indication (inflammation, no inflammation). Hospital-level characteristics included teaching status (teaching, nonteaching), admission type (elective, emergent), annual hospital volume of LC ( 120, 121 to 224, 225), and annual surgeon volume ( 12, 12 to 35, 36) were included in the multivariate logistic regression analyses. Adjusted odds ratios (AOR) were calculated to determine the effects of the identified covariates on development of a major postoperative in-hospital complication. RESULTS Study cohort characteristics Querying the NIS for the years , we identified 1,102,071 weighted (228,171 unweighted) patient-discharge records with a procedure code for LC (Table 1). The majority of patients were female (69%, n 758,680), with 20% (n 224,505) younger than 35 years of age, 47% (n 513,423) were 35 to 64 years of age, and 33% (n 364,143) were 65 years of age or older. White patients constituted 68% (n 717,229) of the cohort, followed by Hispanics (16%, n 176,152), blacks (10%, n 104,052), and other (5%, n 52,539). Race information was missing for 52,099 (4.7%) of the cohort.the majority of patients had a CCS of 0 (69%, n 758,145) and 4% (n 44,245) with a CCS 3. Most LC were performed urgently (75%, n 830,472) at nonteaching hospitals (63%, n 696,877) (Table 2). Inflammation was present in the majority of cases (97%, n 1,102,071). One-third of LCs were performed at hospitals performing 120 LCs annually, with additional thirds at hospitals performing 120 to 224 LC/annually, and 225 LC/annually. Most surgeons performed between 12 and 35 LC annually (49%, n 542,624), with 29% (n 323,609) performing 36 annually, and 22% (n 235,838) performing 12 annually. Overall postoperative complication rate for patients undergoing LC was 6.8% (n 75,414) (Table 3). The most frequent complications included gastrointestinal hemorrhage (3.5%, n 39,021), pulmonary compromise (1.3%, n 13,780), perforation/laceration (0.5%, n 5,320), infection (0.5%, n 5,198), and death (0.5%, n 5,200). Of note, conversion to OC occurred in 9% (n 99,601) of cases. After surgery, the majority of patients were discharged home (91%, n 1,000,473), with an additional 4% (n 46,041) discharged home with visiting nurse services, and 4% (48,353) sent to an acute rehabilitation center/skilled nursing facility/short-term hospital. Univariate analysis Advanced age was associated with an increased risk of a major postoperative complication developing (younger than 35 years, 3.4% complication risk; 35 to 64 years, 5.9% complication risk; 65 years of age or older, 9.8% complication risk; p ). Complications were more likely to develop in male patients (male, 8.6% versus female, 6.0%; p ). Increasing patient comorbidities correlated with increased complication risk (CCS 0, 4.7%; CCS 1, 9.7%; CCS 2, 14.0%; CCS 3, 16.7%; p ). Presence of inflammation was associated with an increased risk of complications (6.9%, n 73,787) when

4 76 Murphy et al Complications after Cholecystectomy J Am Coll Surg Table 1. Characteristics for Patients Who Underwent Laparoscopic Cholecystectomy with Associated Univariate Analysis of Postoperative Complications, Nationwide Inpatient Sample Overall weighted Complication rate Characteristic n % n % p Value Total 1,102, , Age, y , , , , , , Gender Male 343, , Female 758, , Race White 717, , Black 104, , Hispanic 176, , Other 52, , Charlson Comorbidty Score 0 758, , , , , , , , Indication Inflammation 1,070, , No inflammation 31, , Conversion Yes 99, , No 1,002, , Disposition Home 1,000, , Home with VNA 46, , Rehabilitation/SNF 48, , AMA 1, Died 5, , Race missing for 51,922 weighted patient discharges. By definition, all patients who died were considered to have a postoperative complication. AMA, against medical advice; SNF, skilled nursing facility; VNA, Visiting Nurse Association. compared with no inflammation (5.2%, n 1,627). Higher postoperative complications were associated with LC performed urgently (urgent, 7.5% versus elective, 4.7%; p ). Patients undergoing LC in nonteaching hospitals had higher complication rates (teaching hospital, 5.8% versus nonteaching hospitals, 7.5%; p ). Higher-volume hospitals ( 225 LC/year) had a lower complication rate of 6.4%, when compared with 7.0% for medium-volume (121 to 224 LC/year) and low-volume ( 120 LC/year) hospitals (p ). Low-volume surgeons ( 12 LC/year) had higher complication rates (7.0%) than medium (12 to 35 LC/year) (6.8%) or highvolume ( 36 LC/year) surgeons (6.7%) (p ). Multivariate analysis Using a multivariable logistic regression model, independent predictors of developing major in-hospital complications were evaluated. Significant factors identified on univariate analyses and those with biologic plausibility were included in the model, with results shown in Table 4. Patient factors, including increasing age (younger than 35 years referent; 35 to 64 years, AOR 1.53; 95% CI, ; 65 years of age or older; AOR 2.16; 95% CI, ), male gender (female referent; male, AOR 1.14; 95% CI, ), and the presence of comorbidities (CCS 0 referent; CCS 1, AOR 1.78; 95% CI, , CCS 2, AOR 2.49; 95% CI, ; CCS 3, AOR 3.02; 95% CI, ) independently

5 Vol. 211, No. 1, July 2010 Murphy et al Complications after Cholecystectomy 77 Table 2. Hospital Characteristics for Laparoscopic Cholecystectomy Associated Univariate Analysis of Postoperative Complications, Nationwide Inpatient Sample Overall weighted Complication rate Characteristic n % n % Teaching status Teaching 405, , Nonteaching 696, , Hospital volume 120/y 372, , /y 361, , /y 368, , Admission type Urgent 830, , Elective 271, , Surgeon volume , , , , , , p Value predicted development of a major postoperative complication. The presence of inflammation/cancer did not predict development of postoperative complications (no inflammation referent; inflammation, AOR 1.08; 95% CI, ). Urgent LC (elective referent; urgent, AOR 1.69; 95% CI, ) and LC performed at nonteaching hospitals were predictive of developing postoperative complications (teaching referent; nonteaching, AOR 1.22; 95% CI, ). Neither annual hospital volume ( 225 LC/year referent; 120 LC/year, AOR 0.96; 95% CI, ; LC/year, AOR 1.01; 95% CI, ), nor annual surgeon volume ( 36 LC/year referent; 12 to 35 LC/year, AOR 1.00; 95% CI, ; 12 LC/year, AOR 1.06; 95% CI, ) predicted development of complications on multivariable analysis. A second multivariable logistic regression model was constructed evaluating predictors of conversion from LC to OC (Table 5). Similar to complications, increasing age (younger than 35 years referent; 35 to 64 years, AOR 1.95; 95% CI, , 65 years of age or older, AOR 2.61; 95% CI, ), and male gender (female referent; male, AOR 1.78; 95% CI, ) were independent predictors of conversion from LC to OC. The presence of increasing patient comorbidities did not predict intraoperative conversion (CCS 0 referent; CCS 1, AOR 1.04; 95% CI, , CCS 2, AOR 0.91; 95% CI, , CCS 3, AOR 1.06; 95% CI, ). Table 3. Major Postoperative In-Hospital Complications and In-Hospital Mortality after Laparoscopic Cholecystectomy, Nationwide Inpatient Sample Complication n, weighted % of total No. of complications 75, Postoperative infection 5, Myocardial infarction 3, Aspiration pneumonia 2, Deep venous thrombosis/ pulmonary embolism 3, Pulmonary compromise 13, Gastrointestinal hemorrhage 39, Reopening surgical site Perforation/laceration 5, Mortality 5, LC performed at low-volume hospitals was predictive of conversion to OC ( 225 LC/year referent; 120 LC/ year, AOR 1.32; 95% CI, ). LC performed by low- and medium-volume surgeons independently predicted conversion to OC ( 36 LC/year referent; 12 LC/year, AOR 1.99; 95% CI, , LC/ year, AOR 1.61; 95% CI, ). Complication rates increased slightly during the time period studied from 6.6% in 1998 to 7.7% in 2006 (trend test; p ) (Fig. 1). DISCUSSION We investigated whether annual surgeon and/or hospital volume were predictive of development of major postoperative in-hospital complications after LC. We found neither surgeon nor hospital volume to be associated with an increased risk of complications. Rather, individual patient factors including advanced age, presence of comorbidities, and male gender were predictive. Conversion from LC to OC was higher for both low-volume surgeons and lowvolume hospitals, with no differences in complication rates. The number of complications increased during the time period studied. Reported morbidity for patients undergoing LC varies in the literature. In a meta-analysis, Shea and colleagues described the tremendous variability in the number and types of reported complications among LC case series. 25 LC-associated complications have been reported to range between 2.1% and 17%. 26,27 None of these published reports studied the role of hospital and/or surgeon volume in development of postoperative complications. To better evaluate individual patient risk, our group recently published a simple risk score to predict development of major complications in patients undergoing LC using factors available preoperatively. Male sex, increasing patient age,

6 78 Murphy et al Complications after Cholecystectomy J Am Coll Surg Table 4. Multivariable Logistic Regression Analyzing Risk of Major In-Hospital Complications after Laparoscopic Cholecystectomy, Nationwide Inpatient Sample Factor Age, y Gender Male Female Charlson Comorbidity Score Indication Inflammation No inflammation Admission type Urgent Elective Teaching status Teaching Nonteaching Hospital volume 120/y /y /y Surgeon volume Table 5. Multivariable Logistic Regression Analyzing Risk of Conversion from Laparoscopic to Open Cholecystectomy, Nationwide Inpatient Sample Factor Age, y Gender Male Female Charlson Score Indication Inflammation No inflammation Admission type Urgent Elective Teaching status Teaching Nonteaching Hospital volume 120/y /y /y Surgeon volume patient comorbidities, gall bladder inflammation, emergent surgery, and surgery at nonteaching hospitals were associated with increased risk of postoperative complications developing. The volume metrics literature is largely centered on complex surgical procedures and major cancer operations. Improved outcomes, in general limited to perioperative mortality, have been described for patients undergoing highly specialized surgical procedures, including pancreatectomy, 2,3 cardiac surgery, 5,6 and esophagectomy. 28 For more common general surgery procedures, including LC, the impact of surgeon and hospital volume has been less well-studied. Using the Veterans Affairs National Surgical Quality Improvement Program, Khuri and colleagues investigated whether there was an association between 30- day mortality and hospital volume for 8 surgical procedures of intermediate complexity, including laparoscopic cholecystectomy. 29 For all 8 operations, no association between volume and 30-day mortality was demonstrated. This study differs from ours in that 90% of subjects were male and 5% of LC cases were considered emergent. Perhaps most importantly, the relatively infrequent result of mortality was used and postoperative complications were not evaluated. Using the NIS database, Csikesz and colleagues recently reported patients with acute cholecystitis underwent fewer conversions and had decreased lengths of stay if operated on by high-volume surgeons as defined by performing 15 LC/year. 30 No difference in mortality or bile duct injury was found based on surgeon volume metrics; additional postoperative complications were not evaluated. The learning curve for LC has been described to exist and last for the first 30 cases As LC is now the gold standard for the management of the majority of gallbladder disease, surgeons trained in the current era should reach this learning curve early in their surgical training. Perhaps as a result of adequate LC training, our current work dem-

7 Vol. 211, No. 1, July 2010 Murphy et al Complications after Cholecystectomy % Postoperative Complications Complications Year Trend test, p> Figure 1. Complications after laparoscopic cholecystectomy over time. onstrates that patients undergoing LC by attending surgeons performing an average of 1 LC per month do not have an increased risk of a major postoperative complication developing. Additionally, the fact that the number of postoperative complications rose slightly during the time period studied, might support the notion that the national learning curve for LC might have been surmounted. Conversion from LC to OC occurs for several reasons, including dense adhesions and difficulty dissecting Calot triangle. However, the authors are aware that conversion should not be seen as a complication, but rather an attempt to perform the operation with the greatest patient safety. 31 We found that patients are more likely to undergo conversion if surgery is performed by a low-volume surgeon or in a low-volume hospital, however, no difference in the number or types of reported complications was found by surgeon/hospital volume. Several limitations of the current work must be acknowledged. The NIS is an administrative database and therefore lacks certain clinical information, including patient-level factors (eg, laboratory values, ultrasound reports, preoperative performance status), operative data (eg, blood loss, transfusions, operative time), and long-term follow-up/ readmission information. We evaluated the development of major postoperative complications using a validated set of ICD-9 codes; however, the true complication rate can be underestimated, as complete individual medical records cannot be reviewed. Additionally, complications occurring after patient discharge were unable to be assessed secondary to database limitations. Because of insufficient coding specifications in NIS, we were unable to accurately assess the important complication of bile duct injury or leak because the use of ICD-9 codes to evaluate bile duct injuries has been demonstrated to substantially underestimate the occurrence of bile duct injuries. 32,33 Individual surgeon volume could not be followed continuously through the years, but rather annually, which can lead to an underestimation of actual surgeon volume. Despite these limitations, for LC, a commonly performed general surgery procedure, neither annual surgeon nor hospital volume independently predicts development of major postoperative in-hospital complications. Rather, individual patient factors, including advanced age, male gender, and the presence of medical comorbidities, are predictive of development of major complications. Therefore, assessing patients preoperatively based on individual patient factors can allow for preoperative risk stratification. As individual factors are associated with the development of major postoperative complications, preoperative patient optimization, and tailored postoperative management based on individual patient risk can improve patient outcomes. Based on our current work, drastic measures, such as the regionalization of general surgical procedures including LC, might not be necessary. As the armamentarium of biliary tract surgical procedures continues to expand, as evidenced by the recent reports of single-port laparoscopic cholecystectomies and natural orifice transluminal endoscopy surgery procedures, we must be vigilant and thorough in our review of these procedures, their associated learning curves, and impact on both patient morbidity and mortality. Acknowledgment: The contributions of Dr Fred Anderson in database provision and statistical analyses are gratefully appreciated.

8 80 Murphy et al Complications after Cholecystectomy J Am Coll Surg REFERENCES 1. Luft HS, Bunker JP, Enthoven AC. Should operations be regionalized? The empirical relation between surgical volume and mortality. N Engl J Med 1979;301: Birkmeyer JD, Warshaw AL, Finlayson SR, et al. Relationship between hospital volume and late survival after pancreaticoduodenectomy. Surgery 1999;126: McPhee JT, Hill JS, Whalen GF, et al. Perioperative mortality for pancreatectomy: a national perspective. Ann Surg 2007;246: Patti MG, Corvera CU, Glasgow RE, Way LW. A hospital s annual rate of esophagectomy influences the operative mortality rate. J Gastrointest Surg 1998;2: Birkmeyer JD, Stukel TA, Siewers AE, et al. Surgeon volume and operative mortality in the United States. N Engl J Med 2003; 349: Birkmeyer JD, Siewers AE, Marth NJ, Goodman DC. Regionalization of high-risk surgery and implications for patient travel times. JAMA 2003;290: Golub R, Siddiqui F, Pohl D. Laparoscopic versus open appendectomy: a metaanalysis. J Am Coll Surg 1998;186: Kazemier G, de Zeeuw GR, Lange JF, et al. Laparoscopic vs open appendectomy. A randomized clinical trial. Surg Endosc 1997; 11: Dolan JP, Diggs BS, Sheppard BC, Hunter JG. The national mortality burden and significant factors associated with open and laparoscopic cholecystectomy: J Gastrointest Surg 2009;13: Wiesen SM, Unger SW, Barkin JS, et al. Laparoscopic cholecystectomy: the procedure of choice for acute cholecystitis. Am J Gastroenterol 1993;88: Stoker ME, Vose J, O Mara P, Maini BS. Laparoscopic cholecystectomy. A clinical and financial analysis of 280 operations. Arch Surg 1992;127: Baird DR, Wilson JP, Mason EM, et al. An early review of 800 laparoscopic cholecystectomies at a university-affiliated community teaching hospital. Am Surg 1992;58: Murphy MM, Shah SA, Simons JP, et al. Predicting major complications after laparoscopic cholecystectomy: a simple risk score. J Gastrointest Surg 2009;13: International Classification of Diseases. 9th Revision. Clinical modification. Salt Lake City: Medicode Publications; Eppsteiner RW, Csikesz NG, Simons JP, et al. High volume and outcome after liver resection: surgeon or center? J Gastrointest Surg 2008;12: Finlayson EV, Goodney PP, Birkmeyer JD. Hospital volume and operative mortality in cancer surgery: a national study. Arch Surg 2003;138: Shah SA, Bromberg R, Coates A, et al. Survival after liver resection for metastatic colorectal carcinoma in a large population. J Am Coll Surg 2007;205: Sariego J, Spitzer L, Matsumoto T. The learning curve in the performance of laparoscopic cholecystectomy. Int Surg 1993; 78: Hawasli A, Lloyd LR. Laparoscopic cholecystectomy. The learning curve: report of 50 patients. Am Surg 1991;57: Moore MJ, Bennett CL. The learning curve for laparoscopic cholecystectomy. The Southern Surgeons Club. Am J Surg 1995;170: Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40: Romano PS, Roos LL, Jollis JG. Adapting a clinical comorbidity index for use with ICD-9-CM administrative data: differing perspectives. J Clin Epidemiol 1993;46: Lawthers AG, McCarthy EP, Davis RB, et al. Identification of in-hospital complications from claims data. Is it valid? Med Care 2000;38: Simons JP, Shah SA, Ng SC, et al. National complication rates after pancreatectomy: beyond mere mortality. J Gastrointest Surg 2009;13: Shea JA, Healey MJ, Berlin JA, et al. Mortality and complications associated with laparoscopic cholecystectomy. A metaanalysis. Ann Surg 1996;224: Csikesz N, Ricciardi R, Tseng JF, Shah SA. Current status of surgical management of acute cholecystitis in the United States. World J Surg 2008;32: Brune IB, Schonleben K, Omran S. Complications after laparoscopic and conventional cholecystectomy: a comparative study. HPB Surg 1994;8: Birkmeyer JD, Siewers AE, Finlayson EV, et al. Hospital volume and surgical mortality in the United States. N Engl J Med 2002; 346: Khuri SF, Daley J, Henderson W, et al. Relation of surgical volume to outcome in eight common operations: results from the VA National Surgical Quality Improvement Program. Ann Surg 1999;230: Csikesz N. Singla A, Murphy MM, et al. Surgeon volume metrics in laparoscopic cholecystectomy. Dig Dis Sci 2009;Nov 13. [Epub ahead of print.] 31. Shamiyeh A, Danis J, Wayand W, Zehetner J. A 14-year analysis of laparoscopic cholecystectomy: conversion when and why? Surg Laparosc Endosc Percutan Tech 2007;17: Valinsky LJ, Hockey RL, Hobbs MS, et al. Finding bile duct injuries using record linkage: a validated study of complications following cholecystectomy. J Clin Epidemiol 1999;52: Taylor B. Common bile duct injury during laparoscopic cholecystectomy in Ontario: does ICD-9 coding indicate true incidence? CMAJ 1998;158: