Inpatient Costs of Routine Endovascular Repair ofAbdominal Aortic Aneurysm1
Jessica S. Lester, MM, Johanna L. Bosch, PhD, John A. Kaufman, MD, Elkan F. Halpern, PhDG. Scott Gazelle, MD, MPH, PhD
Rationale and Objectives. The purpose of this study was to determine the inpatient cost of routine (ie, without emergentconversion to open repair during the hospital stay) endovascular stent-graft placement in a consecutive series of patientsundergoing elective endovascular repair of abdominal aortic aneurysm (AAA) at a single institution.
Materials and Methods. Inpatient hospital costs of 91 patients who underwent initial elective endovascular repair ofAAA were analyzed retrospectively. All patients had participated in clinical trials at the authors institution during theprevious 6 years. Financial data were derived from the hospitals cost-accounting system; additional procedural data werecollected from a departmental database and with chart review. Stent-graft and professional costs were excluded.
Results. The mean total cost for endovascular repair was $11,842 (standard deviation [SD], $5,127), mean procedure timewas 149 minutes (SD, 79 minutes), and mean length of stay was 3.5 days (SD, 2.3 days). Total cost depended on stent-graft type (means, $12,428 [bifurcated] vs $9,622 [tube]; P .0002) and strongly correlated with procedure time andlength of hospital stay (r 0.78 and 0.66, respectively; P .0001). Ninety-six percent of total costs for all patients wereattributable to the following departments: operating theater (31%), radiology (31%), nursing (22%), and anesthesia (12%).
Conclusion. Overall costs are greater with bifurcated than with tube stent-grafts. Total procedure-related costs are dividedrelatively equally between the operating theater, the radiology department, and the combination of the nursing and anes-thesia departments.
Key Words. Aortic aneurysm, abdominal; hospital costs; stents
The U.S. Food and Drug Administration first approvedstent-grafts for elective endovascular repair of abdominalaortic aneurysms (AAA) in September 1999. Clinical tri-als of stent-grafts are essential to receiving approval fromthe U.S. Food and Drug Administration, and such trialsare ongoing, with new approvals likely to be forthcoming.
Preliminary clinical evaluation of the new procedure sug-gests that it may be as effective asand less invasivethanopen surgical repair. To evaluate this new technol-ogy more fully, however, and to understand the full im-pact of its use on the health care system in particular, it isimportant to understand not only the associated healthbenefits but also the economic costs.
Cost studies of the traditional, open surgical techniqueare abundant in the literature (118). To our knowledge,nine published studies have analyzed the costs of the al-ternative stent-graft procedure (1220). Additional dataare needed, however, to build on this previous work andto address the potential limitations of these earlier studies.For example, four studies (1215) based their results onsmall sample sizes (n 22). Three studies (16,17,20)compared the costs of endovascular repair with those of
Acad Radiol 2001; 8:639646
1 From the Decision Analysis and Technology Assessment (DATA) Group,Department of Radiology, Massachusetts General Hospital and HarvardMedical School, Zero Emerson Place, Ste 2H, Boston, MA 02114 (J.S.L.,J.L.B., J.A.K., E.F.H., G.S.G.); the Department of Health Policy and Man-agement, Harvard School of Public Health, Boston, Mass (G.S.G.); and theDepartment of Epidemiology and Biostatistics, Erasmus University MedicalCenter, Rotterdam, the Netherlands (J.L.B.). Received January 15, 2001;revision requested January 31; revision received March 5; accepted March7. Address correspondence to G.S.G.
open surgery, but absolute cost figures were not pre-sented. The remaining two studies (18,19) did not useprimary cost data. One of these two studies (18) calcu-lated cost data on the basis of average resource use asreported in the literature, and the other (19) estimatedcosts by converting charges found in the Health Care Fi-nancing Administrations Medicare Provider Analysis andReview (MEDPAR) database with the use of cost-to-charge ratios (21).
The purpose of this study was to determine the inpa-tient cost of routine (ie, without emergent conversion toopen repair during the hospital stay) endovascular stent-graft placement in a consecutive series of patients under-going elective endovascular repair of AAA at a singleinstitution. The analysis focused on the cost of the princi-pal hospital stay and did not include preadmission orpostdischarge costs, device costs, or professional fees.Our analysis was conducted with primary cost data, pro-vided estimates in terms of real-dollar figures, and used arelatively large patient cohort as the study group.
MATERIALS AND METHODS
Patient Population and Procedure
Our study group included consecutive patients enrolledin all clinical trials involving AAA repair with bifurcatedor tube stent-grafts, such as the Ancure (Guidant/Endo-Vascular Technology, Menlo Park, Calif), Vanguard(Boston Scientific/Medi-tech, Wayne, NJ), and Excluder(W.L. Gore & Associates, Flagstaff, Ariz) devices, at asingle, large, urban teaching hospital between October1994 and October 1999. Internal Review Board approvaland informed consent were obtained for all patients inthese trials and for the retrospective use of patient infor-mation in this study. None of the trials was consideredhigh risk or compassionate use during this time period;thus, all patients were also surgical candidates. The twomost important inclusion criteria of the clinical trials wereappropriate anatomy for the dimensions of the device andabsence of exclusionary, comorbid medical conditions,such as renal failure, metastatic carcinoma, or uncorrect-able coagulopathy. Selection criteria for endovascular ver-sus traditional repair were at the discretion of the refer-ring surgeon, provided that the patient met the inclusioncriteria of the clinical trial. Once a patient was acceptedinto a trial, the device configuration was chosen solely onthe basis of the patients anatomy and on the availabilityof device itself within the trial (22).
Ninety-one patients were included in our study andselected from a total of 99 consecutive patients. Six ofthe 99 patients were excluded from our analysis due tomissing data or unresolvable inconsistencies in the data(eg, dates of hospital stay not matching date of the proce-dure performed, cost fields left blank), which caused us toquestion the validity of some (or all) of the data for thatpatient. Two patients who required immediate conversionduring the procedure to open surgical repair also wereexcluded. Of the final 91 patients included in our study,19 (21%) underwent tube and 72 (79%) underwent bifur-cated stent-graft placement. Sixty-eight (75%) patientswere men, 85 (93%) were white, and the mean patientage was 75 years (standard deviation [SD], 7.5 years; agerange, 5191 years). Twenty-two patients (24%) under-went stent-graft placement with regional anesthesia; theother 69 patients (76%) were treated under general anes-thesia.
The procedure was performed on the day of or the dayafter hospital admission. Bilateral femoral artery cut-downs were performed in all patients scheduled to un-dergo bifurcated stent-graft placement, and ipsilateralfemoral artery cutdowns with percutaneous contralateralaccess were performed in patients scheduled to undergotube stent-graft placement. Flush aortography was per-formed before and after stent-graft deployment in allcases. Pelvic and additional, selective angiograms wereobtained as needed during the procedure. Angioplasty ofstent-graft elements was performed as required by theprotocol being followed (eg, in the contralateral limb formodular devices) or as needed (with or without additionalbare-metal stents) to resolve attachment endoleaks, kinks,or dissections. Computed tomography (CT), angiography,radiography, ultrasonography, laboratory tests, and physi-cal examinations were performed according to the proto-col being followed. Choice of anesthesia was determinedon the basis of patient risk and hospital practice. Initially,patients received general anesthesia, as was the customfor patients with AAA undergoing open repair proce-dures. More recently, however, epidural anesthesia hasbeen preferred and used whenever possible unless contra-indicated (eg, spine problems, failed epidural anesthesia,or medical conditions requiring intubation for airway con-trol during sedation).
The stent-graft procedure generally involved one sur-gery staff physician, one surgery fellow, and two radiol-ogy staff physicians. The senior members of the team didnot change during the study period. Thus, the same sur-gery and radiology attending physicians were present for
LESTER ET AL Academic Radiology, Vol 8, No 7, July 2001
almost all cases, so the experience level of the team in-creased over time. Vascular surgery and vascular radiol-ogy fellows changed each July; however, they did notperform the critical aspects of these procedures. With oneexception in 1994 (ie, an attending physician with 3 yearsof experience), the attending physicians each had morethan 10 years of clinical experience. In this study, weincluded all actions that were performed during the hospi-tal stay in which the procedure was performed. Preadmis-sion tests or those performed during patient follow-upwere not included in this analysis.
Data SourcesWe obtained data regarding each patient from three
sources: (a) the hospital accounting system (TSI software;Eclipsys, Delray Beach, Fla), (b) a database kept by vas-cular radiologists regarding all endovascular proceduresperformed during the past 6 years, and (c) hospital medi-cal records. We cross-linked these three separate data setswith the aim of increasing the reliability of the data usedin this study.
Our hospitals accounting system provided all cost,demographic, and resource utilization data. The costs de-scribed here are the sum of both direct and indirect costs.Our hospital uses standard accounting methods, which aredescribed in detail elsewhere (2327). The accountingsystem defines direct costs in terms of specific interme-diate products, and the set of intermediate products for adepartment represents a comprehensive breakdown of ac-tual departmental costs into units, including such diverseitems as labor, disposable supplies, and equipment. Indi-rect expenses (eg, building depreciation, information sys-tems) are allocated to departments with an accountingmethod that is based, in part, on standard stepdown ac-counting, which spreads these costs evenly to the individ-ual clinical departments (26,27). Thus, the total cost foreach patient is the sum of costs across all departments inwhich resources were used during that patients hospitalstay. The departmental cost for each patient stay is thesum of all intermediate products in that department thatwere consumed by the patient during his or her stay. Phy-sicians fees are not included in the accounting databaseand, therefore, were excluded from our analysis. The pur-chase price of the stent-grafts also is not included in theaccounting system (because the cost is in the context of aclinical trial).
A second database, which is kept by clinicians in theradiology department, was used solely to determine whichtype of stent-graft (ie, bifurcated or tube) was used for
each case. A third data source was patient medicalrecords, which provided data regarding procedure time.Medical records also were used to corroborate the ac-counting systems data regarding mortality, diagnoses,and procedures.
Data AnalysisData were analyzed with both Excel (Microsoft Corpo-
ration, Redmond, Wash) and SAS version 6.1 (SAS Insti-tute, Cary, NC) software. All hospital costs incurred dur-ing the principal stay for the procedure, excluding devicecosts and professional fees, were normalized to 1999 dol-lars by using the Medical Care component of the Con-sumer Price Index (28). Descriptive statistics (ie, mean,SD, median, range, frequencies) were calculated for theprocedure time (ie, initial incision to closure), length ofstay, and actual total cost.
In addition, the mean actual total costs of tube and ofbifurcated stent-graft placements were compared using theStudent t test (two-sided, .05). Pearson correlationcoefficients (r) of the actual total cost with the proceduretime and length of stay were calculated for all patientsand for each stent-graft group. Costs were modeled sepa-rately with a one-way analysis of covariance by usinglength of stay (in days) and procedure time (in minutes)as the covariates. For each model, uniformity of the rela-tionship across stent-graft types was tested by inclusion ofthe interaction of the covariate and the stent-graft type( .05). Changes in cost from year to year also wereexamined using a general linear model.
Furthermore, we identified the mean costs attributed toeach department that was involved in endovascular repairof AAA. All intermediate products were identified andsummarized by department.
The distribution of total costs is shown in Figure 1.The mean cost per patient was $11,842 (SD, $5,127; me-dian, $10,471; range, $6,372$44,086). One patient hadan exceptionally high cost associated with mortality($44,086). Three other patients who experienced compli-cations also had total costs of more than $24,000. Exclud-ing these four patients, the mean total cost was $10,992(SD, $2,815). Including outliers, the total cost of the pro-cedure was statistically significantly higher for patientswho required bifurcated stent-grafts than for patients whoreceived tube stent-grafts (means, $12,428 [SD, $5,591]vs $9,622 [SD, $1,247], respectively; P .0002).
Academic Radiology, Vol 8, No 7, July 2001 INPATIENT COSTS OF ANEURYSM REPAIR
The mean procedure time for all 91 patients was 149minutes (SD, 79 minutes; median, 135 minutes; range,65780 minutes), and the mean length of stay was 3.5days (SD, 2.3 days; median, 3.0 days; range, 119 days).The mean procedure time was 155 minutes for the bifur-cated stent-graft group (SD, 86 minutes; n 72) and 129minutes for the tube stent-graft group (SD, 45; n 19;P .20). The mean length of stay was 3.6 days (SD, 3.0days) and 3.1 days (SD, 0.9 days), respectively (P .40).
Procedure time and length of stay each correlated wellwith cost (r 0.78 and 0.66, respectively; P .0001).The relationship of procedure time to cost depended onthe stent-graft type (P .01). Actual total cost increasedby an average of $52 (standard error [SE], $5; P .0001) for each minute of procedure time for the bifur-cated grafts, but the estimated increase in cost for thetube grafts was only $9 per minute (SE, $6; P .16).Actual total cost increased by an average of $1,429 perday across all groups (SE, $176); however, the added costper additional day of hospital stay did not depend in astatistically significant manner on the stent-graft type(P .22). Using both procedure time and length of stayin a single multiple-regression equation, we achieved sim-ilar estimates (increase in tube graft cost of $6 perminute, increase in bifurcated graft cost of $43 perminute, and cost increase of $1,002 per additional day ofstay for both; r2 .84).
Overall, costs did not change statistically significantlyin a consistent direction from year to year. In the bifur-
cated stent-graft cases, however, a decrease was found incost over time (r 0.24, P .04). Although our studyspanned 5 years, nearly all patients (85 of 91 patients,93%) were examined during the last 3 years.
Figure 2 shows the costs distributed among the indi-vidual departments. Of these departments, anesthesia,nursing, operating theater, and radiology made the great-est contributions (in aggregate, 96%). The other 4% ofcosts was distributed among 10 additional departments,each accounting for 2% or less of the total costs: intrave-nous services, emergency room, cardiac, neurology/psy-chology, respiratory/pulmonary, blood, laboratory, phar-macy, supplies, and physical therapy.
Within departments, direct costs are assigned to inter-mediate products in a variety of ways. The anesthesiadepartment, for example, uses only one intermediate prod-uct (ie, time) with which to calculate cost, and each pa-tient consumed different quantities of this product. Costsin the nursing department are categorized by the ward inwhich each patient stayed; room and board, labor, andsupplies are calculated on the basis of an estimate ofnursing use, which, in turn, is calculated on the basis ofthe specific ward and level of patient acuity.
The operating theater costs are detailed in Figure 3a.This department distinguishes major resource componentson the basis of salary and room use (as distinct from sup-plies). Operating theater labor and supplies, which incor-porate the fixed costs of maintaining the operating theateritself, account for 85% of the mean total departmental
Figure 1. Distribution of actual total costs presented as histogram with half stan-dard-error bars.
LESTER ET AL Academic Radiology, Vol 8, No 7, July 2001
costs. Postoperative costs accounted for 13% of depart-mental costs.
The radiology department incorporates labor and sup-plies into product categories on the basis of each test andprocedure that is performed. Figure 3b shows the detailedcosts in this department. The catheter-based stent-graftplacement, which actually is performed in the operatingtheater, accounts for 70% of the mean total departmentalcosts. Only 11% of the radiology department costs are notrelated to direct performance of the intervention (eg, CTand chest radiography).
Until recently, the evaluation of new medical technolo-gies focused only (or principally) on effectiveness (29).The combination of more expensive technologies with theincreasing attention now being paid to limiting the ratesof increases in medical spending, however, has resulted inthe need to more carefully evaluate costs in addition tobenefits.
In this study, we estimated the total inpatient hospitalcosts (excluding professional fees and device costs), pro-cedure time, and length of stay for routine endovascularrepair of AAA. In general, the total hospital costsstrongly related to the procedure time and the total lengthof stay. We also quantified and compared resource use forboth bifurcated and tube stent-graft placements. Mean
total hospital costs were higher for procedures that in-volved placement of bifurcated stent-grafts than for proce-dures that involved placement of tube stent-grafts, as wasthe cost per minute of procedure time. The cost of anadditional day spent in the hospital, however, was notsignificantly affected by the type of stent-graft used. Onepossible explanation for this is that bifurcated stent-graftswere routinely placed with bilateral femoral artery cut-downs (involving a surgical cutdown on one side and per-cutaneous access on the other); this is a more labor-inten-sive procedure than is typically performed for tube stent-graft placement. Additionally, we determined that roughlytwo-thirds of the mean cost per patient could be ac-counted for by two departments (ie, operating theater andradiology), with nearly all of the one-third remaining be-ing accounted for by the anesthesiology and nursing de-partments. Within the operating theater department, non-nursing labor and supplies shared the largest cost burden.In the radiology department, the catheter-based interven-tion itself accounted for the greatest cost.
Despite differences in methods of data collection andanalysis and in hospital setting between our study andothers published in the literature, the mean total hospitalcost for patients in our study ($11,842) was within therange of costs reported previously for endovascular repairof AAA (1215,18,19). Three studies performed in theUnited States (14,18,19) also presented absolute total hos-pital costs, excluding stent-graft costs and professionalfees. Costs reported in those studies ranged from $8,444to $10,239. Three additional studies performed in Europe(12,13,15) reported mean total hospital costs that rangedfrom $11,645 and $26,933, but these studies includedstent-graft costs and professional fees. (For comparison,we converted the nondollar numbers to U.S. dollars byusing the Organization for Economic Co-operation andDevelopments Purchasing Power Parities index and thennormalizing the result to 1999 U.S. dollars by using theU.S. Bureau of Labor Statistics Medical Care componentof the Consumer Price Index.) In comparison to some (orall) of the previous studies, however, our analysis in-cluded a larger study population, calculated costs on thebasis of actual accounting data, reported measures of vari-ation, and identified major hospital cost components bydepartment. Furthermore, actual accounting costs wereused as opposed to budgeted or standard costs, charges,or reimbursements. Finally, we also used full costs in-stead of only direct or variable costs to incorporate over-head expenses (30).
Figure 2. Distribution of costs by hospital department.
Academic Radiology, Vol 8, No 7, July 2001 INPATIENT COSTS OF ANEURYSM REPAIR
The principal limitations of this analysis relate to theaccuracy of the hospital accounting system. The numberassigned to each intermediate product represents the hos-pitals best estimate of cost. Such estimates more closelyapproximate true costs than do proxies based on eitherreimbursement rates or charges (even if deflated by cost-to-charge ratios). True microcosting, however, may repre-sent a better estimate of cost but is considerably morelabor-intensive and can be done only prospectively. Thus,it is less practical. For this reason, and supported by theconclusions of the 1996 U.S. Panel on Cost-Effectivenessin Health and Medicine, the present method is both apractical and a viable alternative (30,31).
As previously stated, this study was performed at alarge, urban teaching hospital. We believe that our re-ported costs likely are similar to those incurred at similarinstitutions, but they may differ substantially from thosein other settings. All procedures included in this studyalso were performed in the setting of a clinical trial,which may not be representative of general clinical prac-tice. In our sample, no additional procedures wereplanned, and patients with severe comorbidities (eg, renalfailure, metastatic carcinoma) were excluded. We furtherexcluded two patients whose procedures were emergentlyconverted to open surgery so that we could focus theanalysis on routine endovascular repair. With these exclu-
sions, our costs may be an underestimate of the actualcosts of the procedure if applied to the much larger groupof potentially eligible patients (ie, due to longer hospitalstays, higher complication rates, or both). We also ex-cluded six patients because of questionable validity of thedata; however, these exclusions are unlikely to have intro-duced any systematic bias into our results.
Our study also was limited to only two types of AAAstent-grafts. A third type, the aorto-unilateral iliac stent-graft, requires a femoral-to-femoral surgical bypass proce-dure and, therefore, may be more expensive to use. Dur-ing the past 6 years at our institution, approximately 20%of all stent-graft placements involved the aorto-unilateraliliac type.
A priori considerations suggest that the costs of thisprocedure may decrease over time at a given institutiondue to learning-curve effects (32). In our study, however,only procedures involving bifurcated stent-grafts showedany decrease in cost over time, although the importanceof this finding is mitigated by the great majority of pa-tients being treated only during the last 3 years of thestudy period. Furthermore, the literature reports that tubestent-graft placements are occurring less frequently overtime (22,33). Future average costs may be higher due tothe change in the case mix, reflecting higher proportionsof bifurcated stent-graft placements. It also is possible
Figure 3. Distribution of (a) operating theater and (b) radiology department costs.
LESTER ET AL Academic Radiology, Vol 8, No 7, July 2001
that routine stent-graft placements may be performed inless-expensive angiography suites as opposed to the oper-ating theater (as in our series). Alternatively, however,costs may increase over time as the procedure is per-formed on more varied patients with more complicatedconditions. The magnitude and direction of any changesin cost, therefore, remain to be determined.
Finally, our study focused on the costs incurred bypatients during their principal hospital stay. When consid-ered from a broader perspective, many other costs areassociated with the endovascular procedure, such aspreadmission work-ups, follow-up costs, professional fees,the cost of the device itself, and patient time and out-of-pocket expenses. The costs reported in this article repre-sent only part of all the costs associated with endovascu-lar repair of AAA, and, if taken out of context, theywould reflect an underestimation. On the basis of stent-graft costs as reported in the literature (ranging from$5,000 to $11,100) and the manufacturers list prices forthe two devices already approved (approximately $10,000for a simple stent-graft, but substantially more when mul-tiple graft elements are needed), inclusion of the cost ofthe stent-grafts themselves clearly may double the totalprocedure cost (1315,1820,34,35). Further research isneeded to compare the total costs of endovascular repairof AAA with the costs of elective, open surgery.
In summary, we have presented a detailed cost analysisregarding endovascular repair of AAA by using tube orbifurcated stent-grafts. This analysis demonstrates thatcosts increase significantly as operating theater time andlength of stay increase; that total procedure-related costsare divided relatively equally between the operating the-ater, the radiology department, and the combination of thenursing and anesthesia departments; and that the overallcosts are greater when bifurcated rather than tube stent-grafts are used.
The authors thank Stuart Geller, MD, for his supportand help with data acquisition.
1. Breckwoldt WL, Mackey WC, ODonnell TF. The economic implica-tions of high-risk abdominal aortic aneurysms. J Vasc Surg 1991; 13:798804.
2. Katz DA, Cronenwett JL. The cost-effectiveness of early surgery ver-sus watchful waiting in the management of small abdominal aortic an-eurysms. J Vasc Surg 1994; 19:980991.
3. Frame PS, Fryback DG, Patterson C. Screening for abdominal aorticaneurysm in men ages 60 to 80 years: a cost-effectiveness analysis.Ann Intern Med 1993; 119:411416.
4. Muluk SC, Painter L, Saba S, et al. Utility of clinical pathway and pro-spective case management to achieve cost and hospital stay reduc-tion for aortic aneurysm surgery at a tertiary care hospital. J VascSurg 1997; 25:8493.
5. Benzaquen BS, Eisenberg MJ, Challapalli R, Nguyen T, Brown KJ,Topol EJ. Correlates of in-hospital cost among patients undergoingabdominal aortic aneurysm repair. Am Heart J 1998; 136:696702.
6. Dardik A, Lin JW, Gordon TA, Williams M, Perler BA. Results of elec-tive abdominal aortic aneurysm repair in the 1990s: a population-based analysis of 2335 cases. J Vasc Surg 1999; 30:985995.
7. Bagia JS, Robinson D, Kennedy RM, Englund R, Hanel K. The cost ofelective and emergency repair of AAA in patients under and over theage of 80. Aust N Z J Surg 1999; 69:651654.
8. Morehouse DL, Elmore JR, Franklin DP, Youkey JR. Physician workeffort and reimbursement for ruptured abdominal aortic aneurysms.Am J Surg 1997; 174:136139.
9. Schermerhorn ML, Birkmeyer HD, Gould DA, Cronenwett JL. Cost-effectiveness of surgery for small abdominal aortic aneurysms on thebasis of data from the United Kingdom Small Aneurysm Trial. J VascSurg 2000; 31:217226.
10. Jepson RG, Forbes JF, Fowkes FGR. Resource use and costs of elec-tive surgery for asymptomatic abdominal aortic aneurysm. Eur J VascEndovasc Surg 1997; 14:143148.
11. Seiwert AJ, Elmore JR, Youkey JR, Franklin DP. Ruptured abdominalaortic aneurysm repair: the financial analysis. Am J Surg 1995; 170:9196.
12. Forbes JF, Brady AR, Brown LC. Health service costs and quality oflife for early elective surgery or ultrasonographic surveillance for smallabdominal aortic aneurysms. Lancet 1998; 352:16561660.
13. Ceelen W, Sonneville T, Randon C, DeRoose J, Vermassen F. Cost-benefit analysis of endovascular versus open abdominal aortic aneu-rysm treatment. Acta Chir Belg 1999; 99:6467.
14. Seiwert AJ, Wolfe J, Whalen RC, Pigott JP, Kritpracha B, Beebe HG.Cost comparison of aortic aneurysm endograft exclusion versus opensurgical repair. Am J Surg 1999; 178:117120.
15. Holzenbein TJ, Kretschmer G, Glanzl R, et al. Endovascular AAAtreatment: expensive prestige or economic alternative? Eur J Vasc En-dovasc Surg 1997; 14:265272.
16. Moore WS, Kashyap VS, Vescera CL, Quinones-Baldrich WJ. Abdomi-nal aortic aneurysm: a 6 year comparison of endovascular versustransabdominal repair. Ann Surg 1999; 230:298306.
17. Quinones-Baldrich WJ, Garner C, Caswell D, et al. Endovascular,transperitoneal, and retroperitoneal abdominal aortic aneurysm repair:results and costs. J Vasc Surg 1999; 30:5967.
18. Patel ST, Haser PB, Bush HL, Kent CK. The cost-effectiveness of en-dovascular repair versus open surgical repair of abdominal aorticaneurysms: a decision analysis model. J Vasc Surg 1999; 29:958972.
19. Sternbergh WC, Money SR. Hospital cost of endovascular versusopen repair of abdominal aortic aneurysms: a multicenter study. JVasc Surg 2000; 31:237244.
20. Clair DG, Gray B, OHara PJ, Ouriel K. An evaluation of the costs tohealth care institutions of endovascular aortic aneurysm repair. J VascSurg 2000; 32:148152.
21. Medicare Provider Analysis and Review (MEDPAR). U.S. Federal Gov-ernment Health Care Financing Administration Web site. Available at:http://www.hcfa.gov/stats/medpar/medpar.htm. Accessed February2000.
22. Brewster DC, Geller SC, Kaufman JA, et al. Initial experience with en-dovascular aneurysm repair: comparison of early results with outcomeof conventional open repair. J Vasc Surg 1998; 27:9921005.
23. Roberts RR, Frutos PW, Ciavarella GG, et al. Distribution of variablevs fixed costs of hospital care. JAMA 1999; 281:644649.
24. Woolhandler S, Himmelstein DU, Lewontin JP. Administrative costs inU.S. hospitals. N Engl J Med 1993; 329:400403.
25. Woolhandler S, Himmelstein DU. Costs of care and administration atfor-profit and other hospitals in the United States. N Engl J Med 1997;336:769774.
Academic Radiology, Vol 8, No 7, July 2001 INPATIENT COSTS OF ANEURYSM REPAIR
26. Drummond MF, OBrien B, Stoddart GL, Torrance GW. Cost analysis.In: Methods for the economic evaluation of health care programmes.2nd ed. Oxford, England: Oxford University Press, 1997; 6465.
27. Anthony RN, Young DW. The allocation method. In: Management con-trol in nonprofit organizations. 5th ed. Homewood, Ill: Irwin, 1994;185186.
28. U.S. Federal Government Bureau of Labor Statistics, Consumer PriceIndex, All Urban Consumers, Medical Care, 2000. Bureau of LaborStatistics Web page. Available at: http://188.8.131.52/cgi-bin/surveymost. Accessed March 10, 2000.
29. Kaufman JA, Geller SC, Brewster DC, et al. Endovascular repair ofabdominal aortic aneurysms: current status and future directions. AJRAm J Roentgenol 2000; 175:289302.
30. Luce BR, Manning WG, Siegel JE, Lipscomb J. Estimating costs incost-effectiveness analysis. In: Gold MR, Siegel JE, Russel LB, Wein-
stein MC, eds. Cost-effectiveness in health and medicine. New York,NY: Oxford University Press, 1996; 176213.
31. Shwartz M, Young DW, Siegrist R. The ratio of costs to charges: howgood a basis for estimating costs? Inquiry 1995; 32:476481.
32. Krajewski LJ, Ritzman LP. Learning curves. In: Payne M, ed. Opera-tions management: strategy and analysis. 4th ed. Reading, Mass: Ad-dison-Wesley, 1996; 263269.
33. Makaroun M, Zajko A, Orons P, et al. The experience of an academicmedical center with endovascular treatment of abdominal aortic aneu-rysms. Am J Surg 1998; 176:198202.
34. Ancure endograft system, U.S. price list. Menlo Park, Calif: GuidantCardiac and Vascular Surgery, 2000.
35. AneuRx stent graft product price list. Santa Rosa, Calif: MedtronicAVE, Peripheral Technologies, 1999.
LESTER ET AL Academic Radiology, Vol 8, No 7, July 2001
Inpatient Costs of Routine Endovascular Repair of Abdominal Aortic AneurysmMATERIALS AND METHODSPatient Population and ProcedureData SourcesData Analysis