A Cost-Effectiveness Analysis of Early vs LateReconstruction of Iatrogenic Bile Duct InjuriesLeigh Anne Dageforde, MD, Matthew P Landman, MD, MPH, Irene D Feurer, PhD,Benjamin Poulose, MD, MPH, C Wright Pinson, MBA, MD, FACS, Derek E Moore, MD, MPH

BACKGROUND: Controversy exists regarding the optimal timing of repair after iatrogenic bile duct injuries(BDI). Several studies advocate late repair (�6 weeks after injury) with mandatory drainage andresolution of inflammation. Others indicate that early repair (�6 weeks after injury) producescomparable or superior clinical outcomes. Additionally, although most studies have reportedinferior outcomes with primary surgeon repair, this practice continues. With disparate pub-lished recommendations and rising health care costs, decision analysis was used to examine thecost-effectiveness of BDI repair.

STUDY DESIGN: A Markov model was developed to evaluate primary surgeon repair (PSR), late repair by ahepatobiliary surgeon (LHBS), and early repair by a hepatobiliary surgeon (EHBS). Baselinevalues and ranges were collected from the literature. Sensitivity analsyses were conducted to testthe strength of the model and variability of parameters.

RESULTS: The model demonstrated that EHBS was associated with lower costs, earlier return to normalactivity, and better quality of life. Specifically, 1 year after repair, PSR yielded 0.53 qualityadjusted life years (QALYs) ($120,000/QALY) and LHBS yielded 0.74 QALYs ($74,000/QALY); EHBS yielded 0.82 QALYs ($48,000/QALY). Sensitivity analyses supported thesefindings at clinically meaningful probabilities.

CONCLUSIONS: This cost-effectiveness model demonstrates that early repair by a hepatobiliary surgeon is thesuperior strategy for the treatment of BDI in properly selected patients. Although there is littleclinical difference between early and late repair, there is a great difference in cost and quality oflife. Ideally, costs and quality of life should be considered in decisions regarding strategiesof repair of injured bile ducts. ( J Am Coll Surg 2012;214:919–927. © 2012 by the AmericanCollege of Surgeons)

iqa

CME questions for this article available athttp://jacscme.facs.orgDisclosure Information: Authors have nothing to disclose. Timothy JEberlein, Editor-in-Chief, has nothing to disclose.Supported in part by the Institutional National Research Service Award T32HS 013833 (Dr Landman) from the Agency for Healthcare Research andQuality, US Department of Health and Human Services, and by an educa-tional grant from the Novartis Corporation.Presented at the American College of Surgeons 97th Clinical Congress, SanFrancisco, CA, October 2011.

Received November 23, 2011; Revised January 23, 2012; Accepted January23, 2012.From the Departments of Surgery (Dageforde, Landman, Feurer, Poulose,Pinson, Moore) and Biostatistics (Feurer), and the Vanderbilt Center forSurgical Quality and Outcomes Research (Feurer, Poulose, Moore), Vander-bilt University Medical Center, Nashville, TN.Correspondence address: Derek E Moore, MD, MPH, Division of Kidney/Pancreas Transplantation, Division of Hepatobiliary Surgery and Liver Trans-plantation, Vanderbilt University Medical Center, 912 Oxford House, Nash-

ville, TN 37232-4753. email: derek.moore@vanderbilt.edu

919© 2012 by the American College of SurgeonsPublished by Elsevier Inc.

Calculous biliary disease is one of the most common gas-trointestinal health problems in the United States today.1

The treatment of choice for this condition is laparoscopiccholecystectomy, and nearly 800,000 procedures are per-formed annually.2 While laparoscopic cholecystectomy isassociated with less overall morbidity and decreased lengthof stay compared with open cholecystectomy, the rate ofiatrogenic bile duct injury (BDI) is nearly double thatwith the open procedure. BDI in the era of laparoscopiccholecystectomy is reported to be between 0.4% and0.6% compared with between 0.1% and 0.2% for opencholecystectomy.3

Bile duct injury is associated with morbidity and costs tothe patient, the medical community, and society if notaddressed in an effective and timely manner.4-7 Some stud-es have shown a long-term decrement in health-relateduality of life (HRQOL) in patients who suffer BDI, evenfter repair.6,8,9 Additionally, these injuries are often associ-

ated with high rates of litigation and jury awards ranging6,10

from $250,000 to $500,000.

ISSN 1072-7515/12/$36.00http://dx.doi.org/10.1016/j.jamcollsurg.2012.01.054

920 Dageforde et al Costs of Bile Duct Injury Repair J Am Coll Surg

There is considerable debate in the literature regardingthe optimum setting for and timing of the repair of theseinjuries. Although it is clear that these injuries are bettercared for by trained hepatobiliary surgeons, there is noconsensus on the optimal timing of BDI repair.5,11,12 Someprograms advocate early repair by hepatobiliary surgeons(less than 6 weeks after injury); others advocate late repairby hepatobiliary surgeons (�6 weeks after the injury).7,13,14

Additionally, several studies have demonstrated poor out-comes with primary surgeon repair of BDI, but the practicecontinues.6

The objective of this study was to evaluate the cost-effectiveness of repair by a primary nonhepatobiliary surgeon,late repair by a hepatobiliary surgeon, and early repair by ahepatobiliary surgeon. By modeling the outcomes of these 3treatment strategies, we can better understand the determi-nants of costs and health-related quality of life associated withiatrogenic bile duct injury and subsequent repair.

METHODSMarkov decision modelThe Markov decision analytic technique is used to modeloutcomes for groups of hypothetical patients and analyzetime, value, and costs of patients in each state of health.These hypothetical patients are assigned to various healthstates, and outcomes for each patient group are simulatedover prespecified time intervals or cycles. Hypothetical pa-tients can change health states when the model is cycled.An “absorbing state” is a state the patient cannot leave onceit is entered. The most common absorbing state is death.The model is run either until all hypothetical patients havereached an “absorbing state” or over a dictated time horizon(such as 1 year in our model). When the time horizon islimited, the model does not run until all patients reach anabsorbing state, but instead stops when the predeterminedtime is complete. A comprehensive literature review is usedto determine the likelihood that in a cycle of the model, apatient will either remain in his or her current state ortransition to a new health state. A value, most commonly in

Abbreviations and Acronyms

BDI � bile duct injuryEHBS � early repair (� 6 weeks after injury) by

hepatobiliary surgeonHBS � hepatobiliary surgeonLHBS � late repair (� 6 weeks after injury) by hepatobiliary

surgeonPSR � primary surgeon repairQALY � quality adjusted life year

units of quality-adjusted life years (QALYs), is assigned to a

patient within a health state. These values accumulate overeach cycle. At the completion of all the cycles, the cost foreach patient is calculated. The costs per value (QALY) ofeach health state can be computed and compared with thevalue of the other treatment strategies.15,16

We constructed a Markov-based decision analytic modelto simulate outcomes for patients undergoing 3 differentstrategies of repair after BDI: primary repair by nonhepa-tobiliary surgeon (PSR), late repair by hepatobiliary sur-geon (LHBS), or early repair by hepatobiliary surgeon(EHBS). To construct and run the model we used TreeAgePro 2009 (TreeAge Software, Inc), a software specificallydesigned to create and evaluate decision trees and models.This model was created with methods similar to those usedpreviously by our group.17 The cost-effectiveness analysiswas performed and reported according to the Panel onCost- Effectiveness in Health and Medicine guidelines.11,18

Health statesThe Markov decision model is shown in Figure 1. As de-scribed earlier, patients initially underwent primary sur-geon repair, late repair by a hepatobiliary surgeon, or earlyrepair by a hepatobiliary surgeon. In each arm of themodel, after operative repair the hypothetical patient wassubjected to one of the following scenarios based on prede-termined probabilities from the literature: death, livertransplantation, uneventful recovery, perioperative compli-cations requiring reoperation, or perioperative complica-tions requiring radiologic or endoscopic intervention. Themodeled time horizon was 1 year.

Model assumptionsSeveral assumptions were made in designing the model inorder to create a clinically meaningful yet cost-consciousmodel. The base case patient was defined to be a 42-year-old woman employed outside of the home, which describesthe typical patient undergoing laparoscopic cholecystec-tomy. The time period considered for return to workstarted after the first operative repair in PSR and EHBSand after referral and onset of treatment in LHBS. Inthese models, injury severity was assumed to be at least aStrasberg E1 type injury. Strasberg’s classification of bileduct injuries can be separated into types amenable toendoscopic vs surgical repair. Types A to D include in-juries to tangential ducts that may be amenable to inter-ventional radiology or endoscopic repair and stricturesthat often do not present until months after injury. TypeE bile duct injuries involve complete transaction orstricture of the common bile duct, and therefore requireoperative repair.14 We assumed that patients did nothave septic physiology and therefore were candidates for

any of the 3 approaches of surgical repair. All operative

921Vol. 214, No. 6, June 2012 Dageforde et al Costs of Bile Duct Injury Repair

repairs were Roux-en-Y hepaticojejunostomies. The inter-ventional or endoscopic procedures were percutaneoustranshepatic cholangiocatheter placement, percutaneousdrain placement, and endoscopic retrograde cholangiogra-phy with stent placement. All modeled outcomes occurredduring the first year after the operative repair.

Probability and cost dataBoth probabilities and rates for the baseline analysis andthe ranges of these values for all sensitivity analyses arereported in Table 1. To determine these values we per-formed a systematic review of the MEDLINE/PubMeddatabase for all reports on BDI from 1995 to 2011,especially reviews and meta-analyses. Table 1 also pres-ents all cost estimates and ranges. Published data on spe-

Figure 1. Markov decision tree representing tstrategies to be chosen from are representedestimates of their probabilities are listed in(decision node); � the occurrence of recurin the simulation (terminal node). BDI, bile dutional radiology.

cific institutional costs, the Medicare database, or similar

databases were used for the cost analysis. Cost data werealso obtained from published studies identified by our sys-tematic review of the literature. All monetary values wereadjusted for inflation to 2010 US dollars using the Con-sumer Price Index for medical care (US Bureau of LaborStatistics). To account for the cost of spending money nowvs in the future, health benefits and future costs were dis-counted at a constant rate of 3%.12

All costs were approached from a societal perspective.19

Modeling from the societal or patient perspective allows forcomparisons with other similar studies often focused on pa-tient outcomes and costs to society. Both positive and negativecost changes resulting from an intervention into the systemwere considered. Furthermore, instead of interpreting thefindings for a particular patient population, our finding can be

oice of strategies for BDI repair. The 3 clinicalsquare decision node. The probabilities and

le 1. □ � the choice between strategiesvents (Markov node); and, � � a logic checkury; HBS, hepatobiliary surgeon; IR, interven-

he chat the

Tabsive ect inj

interpreted for the public interest.

P

, prim

922 Dageforde et al Costs of Bile Duct Injury Repair J Am Coll Surg

UtilitiesThe effectiveness of different repair strategies was measuredin terms of QALYs. This measure of health value incorpo-rates both quality of life and time into a composite statisticthat allows for comparison between health interventions.Quality of life is determined by health utilities reported inthe literature, which usually range from 0 (utility of death)to 1 (utility of perfect health). Utilities represent the re-

Table 2. Costs and Cost-Effectiveness of Strategies of Bile

Strategy Cost,$Incremental

cost, $Effec

Early HBS 39,000 —Late HBS 55,000 16,000

rimary surgeon 64,000 9,000

Early HBS, early repair (�6 weeks after injury) by hepatobiliary surgeon; late

Table 1. Literature-Based Probabilities and CostsBaseline parameters

Reoperation after first repair, %PSRLHBSEHBS

Stricture/leak requiring IR/endoscopic procedure, %PSRLHBSEHBS

Death, %PSRLHBSEHBS

Liver transplantation, %PSRLHBSEHBS

Days until return to workPSRLHBSEHBS

Cost, $Roux-en-Y HJ 1IR/endoscopic interventionReoperation 1Indirect/d 3

Utility, QALYBDI repairedIR/endoscopic interventionReoperationOLT

EHBS, early repair (�6 weeks after injury) by hepatobiliary surgeon; HJ, hepinjury) by hepatobiliary surgeon; OLT, orthotopic liver transplantation; PSR

adjusted life year.

ported health preferences of groups of patients who areeither presently ill or may be ill in the future20(Table 1).

Sensitivity analysisOne- and 2-way sensitivity analyses were performed to testthe model conclusions based on variations in the range ofvalues and costs reported in the literature. The ranges usedfor these analyses are described in Table 1. Multi-way prob-

t Injury Repairess, Incremental

effectiveness, QALYCost/

effectiveness, $/QALY

— 47,600�0.08 74,300�0.21 119,600

, late repair (� 6 weeks after injury) by hepatobiliary surgeon; QALY, quality

e Range References

30–90 4,5,7,24-302–255–30

5–20 4,5,7,24-303–152–15

0–8 3,11,24,30,310–50–8

0–8 25,320–50–5

15–180 3,13,24,30,3330–27015–180

10,000–35000 22,34,352,500–15,000

10,000–50,00050–500 18

.9 0.7–1.0 36-46

.6 0.4–0.8

.5 0.3–0.8

.7 0.5–0.9

ejunostomy; IR, interventional radiology; LHBS, late repair (�6 weeks afterary surgeon repair; QALY, quality adjusted life year.

DuctivenQALY

0.820.740.53

HBS

Valu

825

10

1085

213

322

309045

5,0002,5005,00085/d

0000

aticoj

923Vol. 214, No. 6, June 2012 Dageforde et al Costs of Bile Duct Injury Repair

abilistic sensitivity analyses using Monte Carlo methods,which change all probabilities and costs within the modelsimultaneously,21 provided additional tests of model sensi-tivity to changes in model parameters.

RESULTSBase case analysisThe previously described model assumptions and the basecase probabilities and costs from Table 1 were used in thebase-case analysis. This analysis evaluated the approach toBDI repair in a typical patient undergoing laparoscopiccholecystectomy complicated by an iatrogenic BDI. Whenthe model is run, the program simulates the transition ofhypothetical patients through the model. The results of thereference case analysis in the Markov model are essentiallythe averages associated with the different outcomes of thesehypothetical patients, and are listed in Table 2 and depictedgraphically in Figure 2. When thousands of hypotheticalpatients are run through the model using a 1-year timehorizon and the base case probabilities and costs describedin Table 1, patients treated with EHBS accrued an averageof 0.82 QALYs over the year. LHBS patients accrued 0.74QALYs and PSR patients accrued 0.53 QALYs. PSR re-sulted in costs of $64,000 to achieve 0.53 QALYs, or ap-proximately $119,600/QALY. Late hepatobiliary surgeon

Figure 2. Results of the base case analysis in the Markov modelcomparing cost-effectiveness. Early hepatobiliary surgeon repair isthe most cost-effective strategy, with costs of $39,000 to achieve0.82 quality-adjusted life years (QALYs) or $47,600/QALY. Greencircle, early hepatobiliary surgeon repair; blue diamond, late hepa-tobiliary surgeon repair; red triangle, primary surgeon repair.

repair resulted in costs of $55,000 to achieve 0.74 QALYs,

or approximately $74,300/QALY. EHBS repair resulted incosts of $39,000 to achieve 0.82 QALYs, or approximately$47,600/QALY. Therefore, the EHBS treatment strategyarm was superior to, and dominated, both PSR and LHBS.

Sensitivity analysisRealizing that the baseline probabilities and costs used inthis model vary between centers performing these proce-dures, we performed 1-way sensitivity analyses to test thevalidity of the conclusions over a range of probabilities andcosts. First we varied the probability of the need for reop-eration after EHBS. Late hepatobiliary surgeon repairnever became dominant (the better choice) over a clinicallymeaningful range of reoperative rates for EHBS (Fig. 3).Additionally, LHBS never became the dominant strategyover a clinically meaningful range of probabilities of inter-ventional radiology or endoscopic reinterventions afterEHBS (Fig. 4). Next, the probability of reoperation afterPSR was varied. Repair by the primary surgeon would be-come the least costly treatment approach only if the rate ofreoperation after PSR was below a threshold value of 14%.In the 2-way sensitivity analysis (Fig. 5), the number ofdays until returning to work after LHBS and the probabil-ity of reoperation after EHBS was varied simultaneously. Indoing so, EHBS remains the dominant strategy for a widerange of values and LHBS only becomes dominant at lownumbers of days till return to work for LHBS or very high

Figure 3. One-way sensitivity analysis demonstrating the effect ofincreasing the number of reoperations after early hepatobiliary sur-geon (EHBS) repair on the expected cost utility of early hepatobiliarysurgeon repair. Green circle, early hepatobiliary surgeon repair; bluediamond, late hepatobiliary surgeon repair; red triangle, primarysurgeon repair. Cost/Eff, cost-effectiveness.

rates of reoperation for EHBS.

stotd

dcdneoW

laa

924 Dageforde et al Costs of Bile Duct Injury Repair J Am Coll Surg

DISCUSSIONIatrogenic BDI during laparoscopic cholecystectomy is acatastrophic event associated with a great deal of cost to thepatient, the hospital, and society. Even after successful re-pair, BDI patients continue to have diminished health-related quality of life and increased morbidity and mortal-ity.9,11,19 Although there are no randomized controlled trialstudying early vs late repair (and are unlikely in the future),here is an abundance of literature detailing the outcomesf BDI and the different strategies of repair, yet the optimaliming of repair after BDI has not been conclusivelyetermined.In a series of 175 patients treated with surgical biliary

rainage after BDI between 1990 and 2003, Sicklick andolleagues3 noted that the timing of an operation for bileuct repair (early repair less than 1 month after injury) wasot associated with postoperative complications or postop-rative length of stay. In another series of 307 initial repairsf BDIs after laparoscopic cholecystectomy, Stewart anday13 noted that timing of the repair was not associated

with the success of the first repair. Resolution of intra-abdominal infection, use of correct surgical technique,complete cholangiography, and repair by an experiencedhepatobiliary surgeon were associated with success.7,13,14

Conversely, Sahajpal and colleagues21 recommend eitherimmediate repair (0 to 72 hours) or late repair (greater than6 weeks after injury) due to significantly higher morbidity

Figure 4. One-way sensitivity analysis demonstrating the effect ofincreasing the number of interventional radiology (IR)/endoscopicinterventions on the expected cost utility of early hepatobiliary sur-geon repair. Green circle, early hepatobiliary surgeon repair; bluediamond, late hepatobiliary surgeon repair; red triangle, primarysurgeon repair. Cost/Eff, cost-effectiveness.

rates in patients repaired after 72 hours but before 6 weeks.

Although the timing of BDI repair varies on a case-by-casebasis, general acceptance of late repair for all patientsshould be reconsidered. Additionally, costs and patientquality of life should be considered when determining tim-ing of BDI repair.

We used a cost-effectiveness analysis to determine themost prudent financial strategy for BDI repair. We dem-onstrated that EHBS repair is the most cost-effective strat-egy, costing $47,600 per QALY gained. In our model,EHBS repair costs substantially less than the other repairstrategies ($26,700/QALY less than LHBS and $72,000/QALY less than PSR). When the values were ranged overthose reported in the published literature for the sensitivityanalyses, the model remained robust. From these analyses,we concluded that EHBS is the most cost-effective strategyfor BDI repair from the societal and patient’s perspective.

In our model, lost wages from missed days of work werethe primary determinants of differences in cost. Our resultsare consistent with the cost analysis of Andersson and as-sociates,22 in which the loss of production and subsequentost wages was the primary cause of increased costs. In annalysis of hospital charges for 49 BDI patients, Savadernd colleagues23 reported that the overall costs associated

with the repair of iatrogenic BDIs were 4.5 to 26 times

Figure 5. Two-way sensitivity analysis varying both 1) days to returnto work after late hepatobiliary surgeon repair (LHBS) and 2) prob-ability of reoperation after early hepatobiliary surgeon repair (EHBS).EHBS repair dominates the LHBS strategy at all points except whenthe reoperation rate after EHBS is high or the number of days out ofwork after LHBS is low. At no point in these clinically importantranges does primary surgeon repair become a preferred strategyfrom a cost perspective. Blue diagonal diamond hatched area, earlyhepatobiliary surgeon repair; green square hatched area, late hepa-tobiliary surgeon repair.

higher than costs of uncomplicated cholecystectomies. Fac-

irl

crrm

clc(riofa

925Vol. 214, No. 6, June 2012 Dageforde et al Costs of Bile Duct Injury Repair

tors associated with increased costs were inpatient hospital-ization days and outpatient care days, with significantlylower costs for BDI recognized intraoperatively. It is notsurprising that intraoperative BDI recognition decreasescosts, especially given our findings of decreased costs forEHBP repair, but recognition of BDI during laparoscopiccholecystectomy only occurs approximately 30% of thetime. Furthermore, encouraging identification of BDI inthe operating room without increasing PSR rates may bedifficult because primary surgeons who identify the injuryat the time of operation may be inclined to attempt repair.

Repair by the surgeon responsible for the iatrogenic bil-iary injury is often associated with poor outcomes. Successby the primary surgeon has been found in numerous seriesto be less than 20%. Flum and colleagues11 found an 11%ncrease in the risk of mortality after a biliary injury if theepair was done by the primary surgeon. Fischer and col-eagues5 noted that a delayed referral and an increasednumber of pre-referral procedures significantly increasedthe number of postoperative complications after definitiverepair. Similar results were also noted by de Reuver andassociates.24 In 1995, Stewart and Way12 noted a statisti-ally increased length of hospital stay after primary surgeonepair compared with experienced hepatobiliary surgeonepair. Despite documentation of inferior outcomes, pri-ary surgeon repair continues.As noted in our model, there is significantly increased

ost associated with primary surgeon repair, driven by theow success and high complication rates. Only until theomplication rate after this strategy decreases below 20%markedly lower than the nearly 80% reported in moderneports) will this strategy be considered cost-effective. Thessues of familiarity with repair procedures and availabilityf skilled endoscopists and interventional radiologists mayurther limit this procedure to tertiary care centers. Silvand coworkers,20 from Queen Elizabeth Hospital in the

United Kingdom, have implemented a travel consultationservice to be available to surgeons regionally in an effort tobypass the issues of primary surgeon repair. Although thisgroup demonstrated positive outcomes, costs of such a ser-vice remain in question. Cost-effectiveness models, such asthe one presented here, would provide an excellent frame-work through which the generalizability of such programsmay be evaluated.

Ultimately, patients with these complex injuries requireindividualized approaches and multidisciplinary manage-ment. This model identifies the significant determinants ofcosts within several management strategies. The drivingforce behind the decreased costs of the EHBS was the in-creased length of time required for the patient to return to

work after the other strategies of repair. Furthermore, the

costs of LHBS were determined by endoscopic and inter-ventional radiologic procedures. Costs for PSR were alsodriven by reoperation rates. The most important factorcontributing to the cost-effectiveness of EHBS was theability to return the patient to a relatively normal life withas few interventions as possible. Given the rarity of thiscomplication and relatively small cost when compared withnational health care spending, surgeons should not lookprimarily to cost to determine an appropriate operativeplan for BDI patients. Instead, this model can be used as atool to guide individual case decision-making, where ac-counting for both costs and quality of life would benefit thepatient.

A limitation of this study is that BDI repair is distilledinto a very basic model. In this model we used publisheddata and meta-analyses to approximate the most likely clin-ical scenarios in these procedures. Any publication bias thatexists in the peer-reviewed literature on this topic wouldtherefore be reflected in our results. Also, while some of thestudies from which we derived our probabilities evaluateddifferences in injury characteristics, such as Strasberg levelor concomitant arterial injury, we chose to make a moregeneral model that included only Strasberg E1 or greaterinjuries that would require surgical repair. Other types ofbile duct injuries have been shown to alter procedural out-comes and require an alternative approach than the injuriesdiscussed in our model. Also, a major assumption of thismodel was that all patients had good source control andwere not in a proinflammatory state, which would allow achoice of early vs late repair. Obviously, patients with septicphysiology would require a delay in repair.

CONCLUSIONSThe strength of the decision analytic technique is thatthe cost-effectiveness models can be adapted to a widerange of clinical scenarios and patient characteristics.The models can be constructed to account for as manyvariables and outcomes of interest as there are data tosupport. Our cost-effectiveness model helps to elucidatethe factors that contribute to the determination of theoptimal economic strategy of BDI repair. In doing so,we have provided data that can allow centers to furthertailor their practice to perform the most cost-consciousprocedure for these injuries. Additionally, in an era ofincreasing health care costs, this model allows policymakers to further understand the costs associated withbile duct injuries and helps to identify policies and pro-cedures that are most likely to both optimize patient care

and minimize costs.

1

1

1

1

1

1

1

926 Dageforde et al Costs of Bile Duct Injury Repair J Am Coll Surg

Author ContributionsStudy conception and design: Dageforde, Landman,

Feurer, Poulose, Pinson, MooreAcquisition of data: Dageforde, Landman, MooreAnalysis and interpretation of data: Dageforde, Landman,

Feurer, Poulose, Pinson, MooreDrafting of manuscript: Dageforde, Landman, MooreCritical revision: Dageforde, Landman, Feurer, Poulose,

Pinson, Moore

REFERENCES

1. Nakeeb A, Comuzzie AG, Martin L, et al. Gallstones: geneticsversus environment. Ann Surg 2002;235:842–849.

2. Csikesz NG, Tseng JF, Shah SA. Trends in surgical managementfor acute cholecystitis. Surgery 2008;144:283–289.

3. Sicklick JK, Camp MS, Lillemoe KD, et al. Surgical manage-ment of bile duct injuries sustained during laparoscopic chole-cystectomy. Ann Surg 2005;241:786–795.

4. Schmidt SC, Langrehr JM, Hintze RE, Neuhaus P. Long-termresults and risk factors influencing outcome of major bile ductinjuries following cholecystectomy. Br J Surg 2005;92:76–82.

5. Fischer CP, Fahy BN, Aloia TA, et al. Timing of referral impactssurgical outcomes in patients undergoing repair of bile ductinjuries. HPB (Oxford) 2009;11:32–37.

6. Connor S, Garden OJ. Bile duct injury in the era of laparoscopiccholecystectomy. Br J Surg 2006;93:158–168.

7. Kapoor VK. Bile duct injury repair: when? what? who? J Hepa-tobiliary Pancreat Surg 2007;14:476–479.

8. de Reuver PR, Sprangers MA, Gouma DJ. Quality of life in bileduct injury patients. Ann Surg 2007;246:161–163.

9. Moore DE, Feurer ID, Holzman MD, et al. Long-term detri-mental effect of bile duct injury on health-related quality of life.Arch Surg 2004;139:476–481; discussion 481–482.

10. Roy PG, Soonawalla ZF, Grant HW. Medicolegal costs of bileduct injuries incurred during laparoscopic cholecystectomy.HPB (Oxford) 2009;11:130–134.

1. Flum DR, Cheadle A, Prela C, et al. Bile duct injury duringcholecystectomy and survival in medicare beneficiaries. JAMA2003;290:2168–2173.

2. Stewart L, Way LW. Bile duct injuries during laparoscopic cho-lecystectomy. Factors that influence the results of treatment.Arch Surg 1995;130:1123–1128; discussion 1129.

3. Stewart L, Way LW. Laparoscopic bile duct injuries: timing ofsurgical repair does not influence success rate. A multivariateanalysis of factors influencing surgical outcomes. HPB (Oxford)2009;11:516–522.

4. Strasberg SM, Hertl M, Soper NJ. An analysis of the problem ofbiliary injury during laparoscopic cholecystectomy. J Am CollSurg 1995;180:101–125.

5. Inadomi JM. Decision analysis and economic modelling: aprimer. Eur J Gastroenterol Hepatol 2004;16:535–542.

6. Birkmeyer JD, Liu JY. Decision analysis models: opening theblack box. Surgery 2003;133:1–4.

7. Landman MP, Feurer ID, Pinson CW, Moore DE. Which ismore cost-effective under the MELD system: primary livertransplantation, or salvage transplantation after hepatic resec-tion or after loco-regional therapy for hepatocellular carcinoma

within Milan criteria? HPB 2011;13:783–791.

18. U.S. Department of Labor. Bureau of Labor Statistics. Wages byArea and Occupation. www.bls.gov/bls/blswage.htm. AccessedMarch 11, 2011.

19. de Reuver PR, Sprangers MA, Rauws EA, et al. Impact of bileduct injury after laparoscopic cholecystectomy on quality of life:a longitudinal study after multidisciplinary treatment. Endos-copy 2008;40:637–643.

20. Silva MA, Coldham C, Mayer AD, et al. Specialist outreachservice for on-table repair of iatrogenic bile duct injuries—a newkind of ’travelling surgeon’. Ann R Coll Surg Engl 2008;90:243–246.

21. Sahajpal AK, Chow SC, Dixon E, et al. Bile duct injuries asso-ciated with laparoscopic cholecystectomy: timing of repair andlong-term outcomes. Arch Surg 2010;145:757–763.

22. Andersson R, Eriksson K, Blind PJ, Tingstedt B. Iatrogenic bileduct injury—a cost analysis. HPB (Oxford) 2008;10:416–419.

23. Savader SJ, Lillemoe KD, Prescott CA, et al. Laparoscopiccholecystectomy-related bile duct injuries: a health and financialdisaster. Ann Surg 1997;225:268–273.

24. de Reuver PR, Grossmann I, Busch OR, et al. Referral pattern andtiming of repair are risk factors for complications after reconstruc-tive surgery for bile duct injury. Ann Surg 2007;245:763–770.

25. de Santibanes E, Palavecino M, Ardiles V, Pekolj J. Bile ductinjuries: management of late complications. Surg Endosc 2006;20:1648–1653.

26. Johnson SR, Koehler A, Pennington LK, Hanto DW. Long-term results of surgical repair of bile duct injuries followinglaparoscopic cholecystectomy. Surgery 2000;128:668–677.

27. Nordin A, Halme L, Makisalo H, et al. Management and out-come of major bile duct injuries after laparoscopic cholecystec-tomy: from therapeutic endoscopy to liver transplantation.Liver Transpl 2002;8:1036–1043.

28. Ozturk E, Can MF, Yagci G, et al. Management and mid- to long-term results of early referred bile duct injuries during laparoscopiccholecystectomy. Hepatogastroenterology 2009;56:17–25.

29. Schmidt SC, Settmacher U, Langrehr JM, Neuhaus P. Manage-ment and outcome of patients with combined bile duct andhepatic arterial injuries after laparoscopic cholecystectomy. Sur-gery 2004;135:613–618.

30. Thomson BN, Parks RW, Madhavan KK, et al. Early specialistrepair of biliary injury. Br J Surg 2006;93:216–220.

31. Lillemoe KD, Martin SA, Cameron JL, et al. Major bile ductinjuries during laparoscopic cholecystectomy. Follow-up aftercombined surgical and radiologic management. Ann Surg 1997;225:459–468; discussion 468–471.

32. Walsh RM, Henderson JM, Vogt DP, Brown N. Long-termoutcome of biliary reconstruction for bile duct injuries fromlaparoscopic cholecystectomies. Surgery 2007;142:450–456;discussion 456–457.

33. Pitt HA, Murray KP, Bowman HM, et al. Clinical pathwayimplementation improves outcomes for complex biliary surgery.Surgery 1999;126:751–756; discussion 756–758.

34. Woods MS. Estimated costs of biliary tract complications inlaparoscopic cholecystectomy based upon Medicare cost/chargeratios. A case-control study. Surg Endosc 1996;10:1004–1007.

35. Flum DR, Flowers C, Veenstra DL. A cost-effectiveness analysisof intraoperative cholangiography in the prevention of bile ductinjury during laparoscopic cholecystectomy. J Am Coll Surg2003;196:385–393.

36. Eid S, Stromberg AJ, Ames S, et al. Assessment of symptomexperience in patients undergoing hepatic resection or ablation.

Cancer 2006;107:2715–2722.

3

3

4

4

927Vol. 214, No. 6, June 2012 Dageforde et al Costs of Bile Duct Injury Repair

37. Nahrwold DL, Demuth WE. Diverticulitis with perforationinto the peritoneal cavity. Ann Surg 1977;185:80–83.

8. Weinstein MC, Stason WB. Foundations of cost-effectivenessanalysis for health and medical practices. N Engl J Med 1977;296:716–721.

9. Martin RC, Eid S, Scoggins CR, McMasters KM. Health-related quality of life: return to baseline after major and minorliver resection. Surgery 2007;142:676–684.

0. Jablonska B, Lampe P, Olakowski M, et al. Hepaticojejunos-tomy vs end-to-end biliary reconstructions in the treatment ofiatrogenic bile duct injuries. J Gastrointest Surg 2009;13:1084–1093.

1. George DC, Krag MH, Johnson CC, et al. Hole preparationtechniques for transpedicle screws. Effect on pull-out strengthfrom human cadaveric vertebrae. Spine 1991;16:181–184.

42. Kamphues C, Rather M, Engel S, et al. Laparoscopic fenestra-tion of non-parasitic liver cysts and health-related quality of lifeassessment. Updates Surg 2011;63:243–247.

43. Torrance GW. Preferences for health outcomes and cost-utilityanalysis. Am J Managed Care 1997;(3 Suppl):S8–S20.

44. Fennessy FM, Kong CY, Tempany CM, Swan JS. Quality-of-lifeassessment of fibroid treatment options and outcomes. Radiol-ogy 2011;259:785–792.

45. Kamphues C, Engel S, Denecke T, et al. Safety of liver resectionand effect on quality of life in patients with benign hepaticdisease: single center experience. BMC Surg 2011;11:16.

46. Naugler WE, Sonnenberg A. Survival and cost-effectivenessanalysis of competing strategies in the management of smallhepatocellular carcinoma. Liver Transpl 2010;16:1186–1194.