Costs and Utilization of Intraoperative Cholangiography

J Gastrointest Surg (2007) 11:1162–1167 DOI 10.1007/s11605-007-0209-9 Costs and Utilization of Intraoperative Cholangiography Edward H. Livingston & Jordan A. G. Miller & Brian Coan & Robert V. Rege Received: 1 June 2007 / Accepted: 1 June 2007 / Published online: 30 June 2007 # 2007 The Society for Surgery of the Alimentary Tract Abstract Background Routine intraoperative cholangiography (IOC) has been advocated as a viable strategy to reduce common bile duct injury (CDI) during cholecystectomy. This is predicated, in part, on the low cost of IOC, making it a cost-effective preventive strategy. Using billed hospital charges as a proxy for costs, we sought to estimate costs associated with the performance of IOC. Methods The 2001 National Inpatient Survey (NIS) database was assessed for IOC utilization and associated charges. Average charges for hospital admission where the primary procedure was laparoscopic cholecystectomy were compared for those associated with and without the performance of IOC. Results Eighteen percent of cholecystectomies were performed in facilities that never perform IOC. Routine IOC (defined as >75% of cholecystectomies performed in any one hospital having a concomitant IOC) was performed in only 11% of hospitals. In the remaining 71% of hospitals, selective IOC was performed. IOCs were associated with US $706–739 in additional hospital charges when performed in conjunction with laparoscopic cholecystectomy. We project a cost of US $371,356 to prevent a single bile duct injury by using routine cholangiography. Conclusion We conclude that only a minority of hospitals performs cholecystectomies with routine IOC. Because of the significant amount of hospital charges attributable to IOC, routine IOC is not cost-effective as a preventative measure against bile duct injury during cholecystectomy. Keywords Intraoperative cholangiography . Choledocholithiasis . National inpatient survey . Comorbidity . Cost effectiveness analysis Abbreviations CDE common duct exploration IOC intraoperative cholangiography E. H. Livingston (*) : J. A. G. Miller : B. Coan : R. V. Rege Division of Gastrointestinal and Endocrine Surgery, University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd Room E7-126, Dallas, TX 75390-9156, USA e-mail: [email protected] E. H. Livingston : J. A. G. Miller : B. Coan : R. V. Rege Veterans Administration North Texas Health Care System, Dallas, TX, USA E. H. Livingston Biomedical Engineering, University of Texas, Arlington, Texas, USA CDI LOS common bile duct injury length of stay Introduction Whether routine intraoperative cholangiography (IOC) is necessary or not during the performance of cholecystectomy remains unknown. Many factors contribute to the decision to perform IOC. There is little disagreement that IOC is indicated when there are suspected common bile duct (CBD) stones or if the biliary anatomy is unclear during the operation. Consensus is lacking regarding the need to routinely perform IOC with every cholecystectomy. IOC cost effectiveness factors into the decision to perform routine IOC. IOC introduces some extra cost when performing cholecystectomy; however, the extent of that cost is unclear. Several studies have presented IOC cost data and cite figures ranging from US $77 to 738 per IOC. In cost effectiveness analysis, the costs of a procedure are multiplied by the number of times the procedure is performed and compared to the expense associated with J Gastrointest Surg (2007) 11:1162–1167 caring for the complications that would be avoided had the procedure been performed. Given the rarity of CBD injury, very large numbers of IOCs would be required to avoid a single CBD injury. One conservative estimate is that 526 IOCs would be required to prevent a single CBD injury1. With this large multiplication, substantial errors in costbenefit calculations will result if the IOC cost data is inaccurate. For this reason, it is important to establish as accurately as possible the true cost of IOC. Determining IOC cost is not a simple matter. First and foremost is separating the actual costs from hospital charges. Most studies of IOC costs have actually examined hospital charges. Although charges are proportionate to costs, they generally exceed actual costs by some unknown degree. Because of contracting issues, true cost data is considered proprietary and rarely shared. However, when performing cost effectiveness analysis, use of hospital charges will result in fair comparisons as long as hospital charges are used for all the treatments assessed. The true cost of IOC has many components. There are obvious expenses such as the cost of the IOC catheter, dyes, and other supplies necessary for performing IOC. Although most series report that an IOC can be performed in 15 min; operating rooms cost over $1,000 per hour to run such that every minute of OR utilization adds greatly to operative costs. The procedure requires fluoroscopy so that there are costs associated with the technician operating the fluoroscope, as well as costs associated with the machine use. Less obvious are fluoroscopy costs attributable to the equipment’s depreciation, maintenance, and replacement expenses. Additional expenses are accrued resulting from a radiologist’s review of the IOC images. Rarely captured in any analysis is the time lost or additional professional fees for the surgeons performing the cholecystectomy and cholangiogram. Most IOC cost studies determined IOC charges as they appeared on hospital bills. Several of them assessed only the most obvious expenses such as charges attributable to the IOC supplies used. Having not accounted for the OR time, equipment use, radiologist billing, etc., these studies may underestimate the true cost of IOC. When hospital rates are established they generally account for the totality of costs. Projections for equipment replacement, depreciation, facility costs, etc. are prorated into the overall charge structure. Thus, estimation of procedure costs should be made from charges derived from the overall hospital bill to account for these indirect expenses. Comparison of hospital charges for similar patients matched for the presence or absence of the procedure of interest will result in the best cost estimates. This has been done for IOC, but only for a single institution experience. Because rate structures may vary considerably between hospitals, hospital charge data used for establishing policy must be obtained from multiinstitution analyses. To date, this has not been done. 1163 We sought to determine the best estimate for IOC expenses, using hospital charges as a proxy for cost. To do so, we analyzed the 2001 National Inpatient Survey. The NIS acquires annual clinical and hospital charge information from 20% of all hospitalizations in the USA. The sample is population adjusted from US Census information such that national-level estimates of hospital utilization, disease, and procedure incidence can be made. This extensive database lends itself to providing the best estimate of hospital charges attributable to IOC. Given that it is population representative, IOC charge data derived from the NIS will be applicable to the nation as a whole, facilitating cost–benefit estimates that will apply not only to academic but community-based surgical practices. Methods A subset of the 2001 NIS2 was created for all those with any diagnosis of gallbladder disease. These included any patients with a diagnostic code of 574.XX (cholelithiasis), 575.X (other disorders of gallbladder), or 576.X (other disorders of biliary tract). Only those patients having a primary diagnostic code of cholelithiasis (574.XX) or other diseases of the gallbladder (575.XX) were included in the analysis. Patients undergoing laparoscopic cholecystectomy were identified if they had a procedure code of 51.23. Those having open cholecystectomies were identified with procedure code 51.22 but not being simultaneously encoded with 51.23. Conversion from laparoscopic to open cholecystectomy was assumed if procedure codes 51.22 and 51.23 were present for the same patient or if a diagnostic code of v64.4 (laparoscopic procedure converted to open) was present. Disease burden was estimated by calculation of Charlson comorbidity scores3 using the Romano modification4. Patients with cholelithiasis were identified if they had a diagnostic code of 575.20 (calculus of gallbladder without mention of cholecystitis). If any of the diagnostic codes were 574.0 (calculus of gallbladder with acute cholecystitis), 574.1X (calculus of gallbladder with other cholecystitis), or 575.00 to 575.12 (acute or chronic cholecystitis without mention of calculus), the patient was assumed to have cholecystitis. Teaching status for a hospital was provided in the NIS database. The hospital size is classified as small, medium, or large, based on a complex assessment of the hospitals’ bed capacity and resources in its immediate region such that this categorization is relative to hospital capacities in its immediate vicinity. The size is established such that one third of the hospitals in any region are categorized as small, medium, or large. Thus, small hospitals have anywhere from 1–200, medium 25–550, and large >45 beds2. 1164 J Gastrointest Surg (2007) 11:1162–1167 Table 1 Characteristics of Patients Undergoing Cholecystectomy With and Without Intraoperative Cholangiography Cholangiogram Age % Female LOS (+/− SD) Only cholelithiasis Calculus of gallbladder with acute cholecystitis Calculus of gallbladder with other cholecystitis Calculus of bile duct with acute cholecystitis Calculus of bile duct with other cholecystitis Calculus of bile duct without mention of cholecystitis Acute cholecystitis (only) Charlson score (%) 0 1 2 3 4 5 Yes No p value 54±20 72 2.6±1.5 3 31 57 1.2 1.9 0.4 8 51±19 73 2.5±1.5 4 35 58 0.7 0.9 0.2 10 <0.0001 0.045 <0.0001 <0.0001 <0.0001 0.0311 <0.0001 <0.0001 0.0003 <0.0001 79.0 17.0 3.2 0.5 0.2 0.0 79.1 16.9 3.2 0.6 0.2 0.1 NS NS NS NS NS NS Continuous data are presented as the mean±SD. Statistical significance for mean differences were determined by Student’s t testing and for proportions by chi-square analysis. Their proportions of patients with Charlson comorbidities scores that range from 0 (none) to 5 (extensive comorbid disease) are presented with the statistical significance for the difference between IOC groups determined by chi-square analysis. Population estimates for the number of procedures performed were estimated from the discharge-weighting statistic provided by NIS2001. This weighting factor accounts for the multi-stage stratified design of the NIS, allowing for the information contained in the NIS database to be used for estimating the incidence of hospitalizations, procedures, and diseases in the USA. This weighting factor is not valid for hospital charges information and, therefore, was not used for calculating charges. Charges attributable to IOC were estimated from the corrected total charges provided by the NIS and only for admissions with length of stay (LOS)≤2 times the median LOS. This was performed to minimize the effect of outliers on IOC cost data. Costs were recalculated to minimize the effect of associated illnesses, concurrent procedures, or charge variability between hospitals. This was accomplished by grouping the hospitals together based on the NIS hospital identifier code. Once grouped, admissions with the same set of diagnostic and procedure codes were compared with the only difference being the presence or absence of IOC. Hospitals were included in this analysis only if they had at least one match for a set of diagnoses and procedures differing only by the presence or absence of IOC. The mean values for the matched sets were calculated and the mean for charges for admissions with IOC was subtracted from the mean for those without IOC. We reviewed the studies reporting IOC costs that are commonly cited5, as well as those we found from literature searches. These studies were systematically evaluated for the type of institution and methodology used for reporting costs. Statistical analysis and database extractions were performed using the SAS V.9 package (Statistical Analysis Software, Cary, NC). Proportions were compared by chisquare analysis and means by Student’s t test. Statistical significance was established if the p values were <0.05. Table 2 Differential Charges Attributable to Intraoperative Cholangiography Laparoscopic cholecystectomy Cholangiogram n Mean Charges IOC Cost % Increase No Yes 154,862 73,550 11,899 12,638 739 6 The percent increase refers to the percent increase in charges relative to charges from hospitalizations where cholecystectomy is performed without intraoperative cholangiography. 28 21,721 57,195 35 12,336 11,688 648 51,829 97,667 33 12,729 12,017 712 47,017 93,615 30 14,644 13,520 1124 18,866 44,608 32 14,593 13,745 848 n Charges n n 7,667 16,639 13,347 13,800 -453 Charges Charges Large Table 3 Effect of Hospital Size and Teaching Status on IOC Utilization and Cost With IOC No IOC IOC cost % Cases with IOC Charges Not Teaching Medium Small Charges Teaching Hospital Bed Size n Teaching n J Gastrointest Surg (2007) 11:1162–1167 1165 Results In 2001, there were 730,048 admissions for gallbladder disease. Of these, 337,729 (75%) underwent laparoscopic cholecystectomy and 112,874 (25%) had open operations. The median LOS for those undergoing laparoscopic cholecystectomy was 3 days with the mean±SD being 4.0± 5.0. To avoid the effect outliers might have on hospital charges, only those patients with LOS less than two times the median were considered for further analysis. For laparoscopic cholecystectomy, this was six or fewer days. Table 1 presents the clinical features of those undergoing cholecystectomy. Data were stratified by the presence or absence of IOC. Although many of these characteristics demonstrated statistical significance, mean values and proportions were very similar because of the extremely large numbers of patients in each group. Despite the statistical significance, patients were reasonably similar in each group with the exception of more patients having diagnoses of choledocholithiasis and being slightly older in the IOC groups. Although patients were slightly older in the IOC group, the Charlson comorbidity scores were the same between the groups, suggesting that patients having an IOC did not have a greater disease burden than those not undergoing IOC. When considering the entire cohort of patients, the excess charges attributable to IOC when performed in conjunction with laparoscopic cholecystectomy were US $739 or a 6% increase in hospitalization costs (Table 2). Further refinement of this analysis was achieved by examining the charges from single hospitals with admissions grouped by having the same diagnostic and procedure codes differing only by the presence or absence of IOC. By comparing charges with and without IOC from the same hospital, we minimized the effect of differing rate structures at various hospitals. We also minimized the potential confounding effects of comorbid disease or collateral operations on patient’s hospital course and consequent billed charges by assessing a patient cohort that only included the same primary diagnosis (cholecystitis or cholelithiasis) and primary procedure (laparoscopic cholecystectomy). We were able to complete these matches for 894 facilities. This analysis yielded laparoscopic cholecystectomy IOC costs of US $706 per case, similar to the US $739 estimated from the much larger cohort where these exclusions were not made. An appreciable number of hospitals, 18%, did not report IOCs being performed for any cholecystectomies. There was a trend for these to be smaller, non-teaching hospitals. Fifty-eight percent of cholecystectomies performed in the USA have an IOC rate ranging between 1– 50%. A relatively small number of hospitals, 11%, performed IOCs on 75% or more of all cholecystectomies. 1166 J Gastrointest Surg (2007) 11:1162–1167 Table 4 IOC Costs Previously Reported in the Literature Reference Type of Institution Number of Patients Study Period Cost Analysis IOC Cost Traverso 7 Philips 11 Berci 14 Fletcher 6 Ladosci 10 Soper 12 Flowers 8 Podnos 9 Community Community Community Academic Community Academic Academic Academic 55 840 2,400 40 734 164 364 NSc 8/1990–5/1994 1/1991–7/1992 NSb NSb 1/1991–12/1993 4/1991–2/1992 9/1989–1/1991 1/1996–12/2000 Selective Total Bill Selective Selective Selective Total Bill Selective Total Bill NSa US $500–725 US $315 NSa US $738 US $700 US $299 US $675 Two of the studies had been cited as providing IOC cost data but no specific reference could be found. IOC costs not explicitly stated b Study period not explicitly stated c Number of patients assessed not explicitly stated a Thus, routine cholangiography is infrequently practiced in the USA. We also assessed the data as a function of hospital size and teaching status (Table 3). Small hospitals did not have increased costs attributable the performance of IOC. With progressively increased size, IOC costs increased. Teaching status did not impact IOC costs. Notably, fewer cholecystectomies were performed with IOC in teaching relative to non-teaching facilities. Discussion Although there was some variability between facilities based on size and volume of IOCs performed, IOCs done in conjunction with laparoscopic cholecystectomy cost approximately US $700. Arguments supporting the performance of routine cholangiography rely on minimal costs associated with this procedure because common bile duct injury (CDI) is so rare. The most recent population-based studies from the Medicare database have shown that CDI occurs in 1 of every 172 cholecystectomies when cholangiography is not performed. When IOC is done in conjunction with cholecystectomy there is 1 CDI for every 256 cholecystectomies performed. From these data, one can conclude that 526 routine cholangiograms must be performed to prevent a single CBD injury1,5. Under these circumstances, if IOC costs US $706, the costs attributable to preventing a CBD injury would be $371,356. Reduced CBD injury has not been definitely linked to routine cholangiography, making it difficult to justify, especially in light of the fact that CBD injury is rare. Additionally, the surgical community remains unconvinced of the need for routine IOC; as indicated by our finding, only 11% of facilities perform routine IOC. The impact of variation in published IOC cost estimates on the apparent cost-effectiveness of routine IOC use to avoid CDI has been examined5. With estimated IOC costs ranging from US $77 to 738, costs attributable to routine IOC were $57,846 to 554,417 per CDI avoided. This tenfold difference in cost-effectiveness estimation illustrates the sensitivity of cost–benefit analysis to the estimated IOC cost. If IOC is inexpensive, the conclusion would be that routine IOC is a cost-effective prevention strategy contrasted to a lack of effectiveness if IOC costs are high. The same study cited costs US $13,612 to 300,000 per patient, attributable to the treatment of complications or deaths caused by CDI. Based on Washington State data, it was estimated that 714 IOCs are required to prevent one CDI and estimated IOC costs at US $122 per study. The authors concluded that IOC is cost effective in preventing CDI because $87,100 was expended to prevent one CDI. Our population-based data from a national rather than a regional sample concluded that IOC was much more costly than US $122 per case. We estimated that IOCs cost US $706 per case such that US $504,084 must be expended to avoid a complication that costs approximately US $300,000 to treat. These figures argue against routine IOC as a costeffective strategy for CDI risk reduction. We reviewed the previously published analyses of IOC cost. Among them, two studies previously cited5 as reporting costs had not explicitly done so6,7. These were eliminated from further consideration. Of the remaining studies, there was a wide range of costs attributed to IOC ranging from US $299 to 738 (Table 4). Both academic and community hospitals reported IOC costs in the same range. Although costs at academic medical centers are often thought to be higher than their community counterparts, this was not observed in our analysis. IOC cost data has also been reported in studies from as early as 1989 to 20008,9. There was no consistent trend of increased costs J Gastrointest Surg (2007) 11:1162–1167 with time as one would expect to occur because of inflation. This most likely occurred because cost data was extrapolated from hospital charges that tend to remain relatively constant with time and do not change at the same pace as medical inflation. We found that the US $706 IOC cost we estimated from population-based data is very similar to the approximately US $700 reported in four prior published series9–12. This confluence of findings at the US $700 range suggests that this is the most accurate estimate of the cost of IOC and should be the value used for cost–benefit analysis. Arguments in favor of routine IOC include reduced CBD injury when it is used. CBD injury is a rare event so that demonstrating reduced injury rates requires analysis of very large series. Studies examining administrative databases have shown that routine cholangiography is associated with reduced CBD injury rates13. Although these studies had power by virtue in the very large numbers of operations evaluated, administrative databases have limited patient information. This constrains the analysis because the overall outcomes of patients with these injuries is not known. Administrative databases are also limited by the accuracy of discharge coding that can be highly variable from one hospital to another. The simple occurrence of a CBD injury is undesirable but may have little clinical significance. Although injury in proximity to the hepatic confluence may be difficult to repair, some lower CD injuries recognized at the time of operation can be easily repaired with little patient morbidity. Alternatively, these injuries can substantially impact patients if they are unrecognized, resulting in postoperative sepsis and requiring complex major reconstructive operations. Thus, the range of adverse outcomes attributable to CDI is great. When determining the merits of routine IOC, the entire range of outcomes must be considered. Because some injuries are of little consequence, only CBD injuries that result in significant complications should be considered in the arguments in favor of routine IOC. To date, this has not been the case, and proponents of routine IOC have cited total, overall CBD injury rates, and their reduction by virtue of routine IOC in support of this practice. In conclusion, we have found that routine IOC is not widely practiced by the surgical community. Seventy-one percent of laparoscopic cholecystectomies are done with a 1167 selective IOC approach. Analysis of a very large database of nationally representative data revealed that IOC costs, on average, US $706 per case. This estimate appears reliable, as it is similar to those arrived at in four other studies reporting from single institution analysis. References 1. Livingston EH. Intraoperative cholangiography and risk of common bile duct injury. JAMA 2003;290:459–460. 2. Agency for Healthcare Research and Quality. HCUP nationwide inpatient sample: Design of the HCUP nationwide inpatient sample. Rockville, MD: Agency for Healthcare Research and Quality, 2003. 3. 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;40:373–383. 4. 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;46:1075–1079. 5. Flum DR, Flowers C, Veenstra DL. A cost-effectiveness analysis of intraoperative cholangiography in the prevention of bile duct injury during laparoscopic cholecystectomy. J Am Coll Surg 2003;196:385–393. 6. Fletcher DR. Laparoscopic cholecystectomy in Australia—Outcomes and costs. Surg Endosc 1995;9:1230–1235. 7. Traverso LW, Roush TS, Koo K. CBD stones-outcomes and costs. Laparoscopic transcystic techniques other than choledochoscopy. Surg Endosc 1995;9:1242–1244. 8. Flowers JL, Zucker KA, Graham SM, Scovill WA, Imbembo AL, Bailey RW. Laparoscopic cholangiography. Results and indications. Ann Surg 1992;215:209–216. 9. Podnos YD, Gelfand DV, Dulkanchainun TS, Wilson SE, Cao S, Ji P, Ortiz JA, Imagawa DK. Is intraoperative cholangiography during laparoscopic cholecystectomy cost effective? Am J Surg 2001;182:663–669. 10. Ladocsi LT, Benitez LD, Filippone DR, Nance FC. Intraoperative cholangiography in laparoscopic cholecystectomy: A review of 734 consecutive cases. Am Surg 1997;63:150–156. 11. Phillips EH. Routine versus selective intraoperative cholangiography. Am J Surg 1993;165:505–507. 12. Soper NJ, Dunnegan DL. Routine versus selective intra-operative cholangiography during laparoscopic cholecystectomy. World J Surg 1992;16:1133–1140. 13. Flum DR, Koepsell T, Heagerty P, Sinanan M, Dellinger EP. Common bile duct injury during laparoscopic cholecystectomy and the use of intraoperative cholangiography: Adverse outcome or preventable error? Arch Surg 2001;136:1287–1292. 14. Berci G. Static cholangiography vs digital fluoroscopy. Intraoperative cholangiography: Benefit and cost ratio. Surg Endosc 1995;9:1244–1248.