- Email: [email protected]
Endoscopic surveillance of barrett’s esophagus: a cost-effectiveness comparison with mammographic surveillance for breast cancer
THE AMERICAN JOURNAL OF GASTROENTEROLOGY Copyright © 1998 by Am. Coll. of Gastroenterology Published by Elsevier Science Inc.
Vol. 93, No. 6, 1998 ISSN 0002-9270/98/$19.00 PII S0002-9270(98)00156-7
Endoscopic Surveillance of Barrett’s Esophagus: A Cost-Effectiveness Comparison With Mammographic Surveillance for Breast Cancer John M. Streitz, Jr., M.D., F. Henry Ellis, Jr., M.D., Ph.D., Robert L. Tilden, DrPH., and Robert V. Erickson, M.D. Duluth Clinic, University of Minnesota School of Medicine, Duluth, Minnesota; Whiteside Institute for Clinical Research, Duluth, Minnesota; and Deaconess Hospital, Harvard Medical School, Boston, Massachusetts
Objective: Endoscopic surveillance of Barrett’s esophagus is commonly practiced to detect malignancy in an early and curable stage. However, the cost-effectiveness of this practice has been questioned. To clarify this issue, we undertook a cost analysis of endoscopic surveillance to detect adenocarcinoma in Barrett’s esophagus compared with mammography used to detect occult carcinoma of the breast, a widely accepted cancer surveillance technique. Methods: The rate of esophageal adenocarcinoma detected by endoscopic surveillance was calculated for Duluth Clinic patients with Barrett’s esophagus seen from 1980 to 1995 and compared with published rates. The rate of occult breast cancer detection was calculated for all women undergoing surveillance mammography at the Duluth Clinic for the year 1994 and compared with published rates. Costs for screening studies and therapy for cancer treatment for both cancers were calculated based on clinical results and assumptions regarding outcomes derived from published reports, and the costs were compared. Results: Endoscopic surveillance of 149 patients with benign Barrett’s esophagus was performed for a total of 510 patient-yr, during which time seven patients developed adenocarcinoma, an incidence of one case per 73 patient-yr of follow-up. Occult breast cancer was detected in 50 of 12,537 mammograms, a detection rate of 0.4%. The incidences in both cases were comparable to published figures. The costs of detecting a case of adenocarcinoma in Barrett’s esophagus and occult breast cancer were $37,928 and $54,513, respectively, and those for treatment resulting in cure were $83,340 and $83,292. Cost per life-yr saved was $4,151 for adenocarcinoma in Barrett’s esophagus and $57,926 for breast cancer. Conclusion: Endoscopic surveillance of patients with Barrett’s esophagus compares favorably with the common practice of surveillance mammography to detect early breast cancer, and should therefore be considered to be
as cost-effective as surveillance mammography. (Am J Gastroenterol 1998;93:911–915. © 1998 by Am. Coll. of Gastroenterology)
INTRODUCTION Although endoscopic surveillance of patients with Barrett’s esophagus is commonly practiced in an effort to discover malignant epithelial changes in an early and curable stage, little is known about the cost-effectiveness of such surveillance. Regular endoscopy with biopsy has been shown to be effective in detecting the early histological changes of malignancy and to result in significant improvement in postoperative survival in those undergoing resection. We (1) and others (2, 3) have shown that surveillance endoscopy increases long term survival after resection of esophageal adenocarcinoma from 20% to 62% or more in patients undergoing yearly endoscopy. Notwithstanding this demonstrated effectiveness, the cost of such a surveillance program has been considered by some to be prohibitively high. The cost-effectiveness of endoscopic surveillance of patients with Barrett’s esophagus has been estimated by others, but absolute dollar amounts are subject to marked variability depending on the assumptions used. A relative comparison with an established screening practice seemed more likely to clarify this issue. Accordingly, we compared the cost incurred, in our current clinical practice, to detect and cure a patient with adenocarcinoma arising in Barrett’s esophagus versus the cost of detecting and curing a patient with carcinoma of the breast using surveillance mammography, a technique widely regarded as cost-effective. MATERIALS AND METHODS Data This study was designed to be a cost-effective analysis comparing the direct surveillance costs for patients with adenocarcinoma arising in the breast and in Barrett’s esophagus, and the cost of curing them, based on our clinical
Presented in part at the 61st Annual Meeting of the American College of Gastroenterology, October 22, 1996, Seattle, Washington. Received Dec. 5, 1996; accepted Feb. 11, 1998. 911
912
STREITZ et al.
experience. To determine whether our experience could be extrapolated to larger patient populations, the cancer detection rates in this study were compared with published results from other studies. Indirect costs such as time lost from work or medical costs to palliate late-stage disease were not included in our calculations, nor were indirect benefits. Cancer detection rates for both groups were determined by retrospective record review of patients followed at the Duluth Clinic. In the case of breast cancer, a sufficient number of cancers were detected in a single year to yield a yearly cancer rate suitable for the calculations. In the case of Barrett’s esophagus, far fewer cancers are detected yearly. Therefore, a cancer rate was established by review of 15 yr of endoscopic patient follow-up. This detection rate was then used as the yearly cancer incidence in the cost calculations. Data regarding cancer detection rates and incidence in our patients was compared with other published data to ensure that our own experience was within established limits. Where reliably significant data regarding our own patients (such as long-term survival following therapy) was lacking, assumptions regarding cure rates were extracted from accepted published values. Data from the literature regarding yearly surveillance mammography were limited to women age 50 –74 yr, for whom there is unequivocal evidence of cancer death reduction (4). Cost analysis Patient charges for 1995 were used as cost estimates with the assumption that variation between them would be similar for both groups of patients and would provide a fair relative comparison, if not a precise absolute cost figure. Only direct costs related to detection and therapy aimed at cure were included. Indirect costs, such as time lost from work, and indirect benefits were not included in the calculations. Cost of palliating late stage disease was also excluded, because assumptions regarding such therapy would vary widely, and such care for both groups is assumed to be similar in cost. Breast cancer. Data from 1994 were used to avoid the changes in costs related to the introduction in 1995 of stereotactic breast biopsy. Costs for breast cancer detection included mammography charges for the entire group, and the cost of excisional biopsy for those found to have positive mammograms. Occult cancers were defined as nonpalpable tumors discovered as a result of surveillance mammography. The cost of curing a cancer included detection costs and the cost of subsequent cancer therapy. Biopsy costs for patients with positive mammograms but with benign biopsies were included. Cost of therapy was based on the assumption that half of those found to have cancer would undergo lumpectomy and radiation therapy and the other half mastectomy, and included adjuvant chemotherapy for both groups. Operative mortality was assumed to be 0%. Life-years saved were calculated using the median age of onset for breast cancer published in the National Cancer Institute’s Statistics, Epidemiology and End Results (SEER) (5) data and compared with the life expectancy of women in
AJG – Vol. 93, No. 6, 1998 TABLE 1 Surgery for Adenocarcinoma in Barrett’s Esophagus Detected by Surveillance Endoscopy Study, year (ref)
Median Age
Operative Mortality
Stage 0 (%)
5-Yr survival (%)
Altorki 1991 (2) Pera 1992 (3) Streitz 1993 (1) Present study Mean
47 66 62 64 60
0 0 0 0 0
55% 55% 38% 57% 51%
100% 66% 62% 76%
the United States (6). The increase in number of patients cured as a result of mammographic detection was estimated using published relative risk reduction values and applying it to our patient group. Adenocarcinoma in Barrett’s esophagus. Surveillance and therapy for adenocarcinoma in Barrett’s esophagus included endoscopy and biopsy charges for the entire screened group and the cost of esophagogastrectomy for those found to have cancer, which included the diagnosis of high grade dysplasia. Operative mortality was assumed to be 0% (Table 1). No adjuvant therapy was included in the charges. Lifeyears saved were calculated using the mean age at onset of adenocarcinoma in Barrett’s esophagus from published results compared with the life expectancy of white males in the United States. The expected increase in number of patients cured following resection was estimated using long term survival figures published for both surveillance and nonsurveillance patients. RESULTS Cancer detection rates and costs Breast cancer. During 1994, 12,537 mammograms were performed at the Duluth Clinic. During that year, 267 patients underwent needle-localization biopsies for clinically inapparent, mammographically suspicious lesions. Of these, 50 proved to be malignant. This yields a positive biopsy rate of 19% and a rate of occult malignancy detection of 0.4%, values within the range reported in other series (Table 2). Cost calculations are shown in Table 3. At $125 per mammogram and including the surgical biopsy cost of $4,339 each for 267 patients with positive mammograms, a total of $2,725,638 was expended to discover 50 occult cancers, a cost of $54,513 per case detected. Studies from the last decade regarding survival of patients with nodenegative (15), mammographically detected (16), or T1 N0 tumors (17) show an average long term survival of 86%. Assuming that therapy would be divided equally between lumpectomy and radiation ($21,437) and mastectomy ($12,800), a total cost of $3,581,563 would be incurred to discover and treat these 50 patients, 43 (86%) of whom would be cured at a cost of $83,292 per cure. Barrett’s esophagus and adenocarcinoma. From 1980 to 1995, 149 patients with Barrett’s esophagus were seen at the Duluth Clinic, of whom 13 had adenocarcinoma when first
AJG – June 1998
ENDOSCOPIC SURVEILLANCE OF BARRETT’S ESOPHAGUS
TABLE 2 Cancer Detection Rates for Screening Mammography Number of Patients
Study, Year (Ref) Ohuchi, 1993 (7) Paci, 1993 (8) Burhenne, 1992 (9) Present study Menon, 1994 (10) Daures, 1992 (11) Chamberlain, 1993 (12) Faulk, 1995 (13) Morimoto, 1994 (14) Mean
9,634 10,407 11,824 12,537 8,000 5,098 1,059,703 32,140 950
% Cancer Rate 0.31 0.37 0.40 0.40 0.51 0.59 0.62 0.74 0.84 0.53
TABLE 3 Cost of Detection and Treatment for Surveillance-Detected Breast Cancer Number Mammograms Biopsies
12,537 267
Unit Cost
Total Cost
$ 125 $ 4,339
$1,567,125 $1,158,513
913
TABLE 4 Incidence of Adenocarcinoma in Barrett’s Esophagus Study, Year (ref)
Case/Patient-year follow-up
Skinner, 1989 (18) Hameeteman, 1989 (19) Robertson, 1988 (20) Present study Williamson, 1991 (21) Van der Veen, 1989 (22) Spechler, 1984 (23) Mean
1/48 1/52 1/56 1/73 1/99 1/170 1/175 1/100
TABLE 5 Cost of Detection and Treatment of Surveillance Detected Adenocarcinoma in Barrett’s Esophagus Number Endoscopy with biopsy Cancers detected Cancers resected Cancers cured (76%) Cost per cure
354 7 7 (76%) 5
Unit Cost $ 750 $37,928 $21,600
Total Cost $265,500 $151,200 $416,700
$83,340
$2,725,638 Cancers detected Lumpectomy-RoRx Mastectomy Cures (86%) Cost per cure
50 (0.4%) 25 25 43
$54,513 $21,437 $12,800
$ 535,925 $ 320,000 $3,581,563
$83,292
seen, a prevalence of 9.4%. The 136 patients with benign Barrett’s esophagus were followed for 510 patient yr, and underwent 354 endoscopic examinations with biopsy, an average interval of one every 17 months. No endoscopic esophageal perforations occurred during the exams. During the follow-up period, seven adenocarcinomas were detected. This cancer incidence of one case per 73 yr of patient follow-up, or 1,370 cases per 100,000 population per year, is similar to other reported rates (Table 4). All seven patients underwent esophagogastrectomy. Five are alive without disease at 1, 2, 4, 6, and 7 yr postoperatively, and two have died, one 6 months after surgery of unrelated causes without evidence of disease, and one of metastatic cancer 28 months postoperatively. The pathological stages of the resected specimens were stage 0 in three patients, stage I in two, and stage IIA in two. This early stage distribution is similar to that which we have reported previously in a different group of patients (1), and is similar to other published results (Table 1). Cost calculations for detection and treatment of these cancers are shown in Table 5. Assuming that all seven patients reported here underwent resection at a cost of $21,600 per case, and that five patients, or 76% (Table 1) were cured, the total cost per cure would equal $83,340. Life-yr gained The cost per life-yr gained in both groups is based on the formula Cost S 2 Cost NS, 4 # life-years gained where
Cost S 5 costs of detection and treatment with surveillance and Cost NS is that for no surveillance. It is not known how long it takes high grade dysplasia to become a clinically detectable tumor, nor is the time separating a mammographically detected cancer from a clinically palpable one well defined. For the purposes of these calculations, it was assumed that all surveillance-detected cancers would have become clinically detectable within equal time periods, and that all would then have been biopsied and treated in the same way as surveillance-detected cancers. Thus, the difference, Cost/Cure S 2 Cost/Cure NS, is comprised of the surveillance costs for patients who do not develop cancer. Breast cancer. Recalculation of costs for the clinically detected breast cancers, shown in Table 6, exclude mammography and biopsy costs for all but the 50 patients whose cancers were mammographically detected, inasmuch as mammography would be performed when these tumors became clinically evident. Reduction of breast cancer mortality for women age 50 –74 yr undergoing surveillance mammography yearly averages 26% (4) or a relative risk of 0.74. This risk difference would mean 9.4 cancer deaths in 50 patients as compared with the seven deaths assumed for the surveillance group. According to the SEER (5) data, the mean age at onset for women with breast cancer is 64 yr. The life expectancy for women in the United States is 82 years of age (6). The 2.4 cancer deaths averted by a surveillance strategy would yield 43.2 life-yr gained. Therefore, $3,581,563 2 $1,079,125 4 43.2 5 $57,926 per life-yr gained. Barrett’s esophagus. Recalculation of costs for the clinically detected Barrett’s esophagus cancers, shown in Table 7, excludes surveillance endoscopy costs for all those who
914
STREITZ et al.
AJG – Vol. 93, No. 6, 1998 TABLE 8 Sensitivity Analysis Results
TABLE 6 Cost of Treatment for Clinically Detected Breast Cancer Number Mammograms Biopsies Cancers Lumpectomy-RoRx Mastectomy
50 50 50 25 25
Unit Cost
Total Cost
Cost
Mammography
Endoscopy
$ 125 $ 4,339
$ 6,250 $ 216,950
Cost per cancer detected Cost per cure Cost per life-year saved
$25,714 $50,446 $25,302
$ 88,500 $173,842 $ 16,650
$21,437 $12,800
$ 535,925 $ 320,000
TABLE 9 Discount Values
$1,079,125 Cures Cost per cure
40.6
Discount $26,579
TABLE 7 Cost of Treatment for Clinically Detected Adenocarcinoma in Barrett’s Esophagus Number
Unit Cost
Total Cost $
Endoscopy with biopsy Cancers detected Cancers resected
7 7 7
$ 750 $ 750 $ 21,600
Cures Cost per cure
1.4
$111,750 $ 79,821
5,250
$151,200 $156,450
did not develop cancer. The cost per cancer cured is $79,821. Long term survival after esophagectomy for surveillancedetected Barrett’s adenocarcinoma has been shown to be 76% on average for those undergoing yearly endoscopy, three times the survival of patients with clinically detected cancers. The mean age of those developing esophageal adenocarcinoma is 60 yr (Table 1). In the United States the life expectancy of white men, who comprise the vast majority of those developing cancer in Barrett’s esophagus, is 76 yr. Assuming that surveillance increased the cures from 1.4 of seven to 5.3 of seven patients (from 20% to 76%), a relative mortality risk of 0.26, the number of life-yr gained by surveillance is 62.7 (3.9 3 16). Cost per life-yr gained is therefore $416,700 2 $156,450 4 62.7 5 $4,151. Sensitivity analysis The greatest uncertainties regarding cost-effectiveness conclusions in this study arise from the wide range of reported cancer incidences and detection rates for both cancers. The sensitivity analysis substitutes for the average values used in the initial calculations, the extreme values from the tables listing incidence of adenocarcinoma in Barrett’s, mammographic cancer detection rates, and cure rates for esophageal resection. Using values that improve the cost-effectiveness results for breast cancer and worsen those for Barrett’s cancers, recalculations of cost per cancer detected, per cancer cured, and per life-yr gained were performed. Values were chosen from the extremes of Table 1 (62% cure rate), Table 2 (0.84 cancer rate), and Table 4 (1 case per 175 patient-yr). (Mortality risk reduction for breast cancer detected by mammo-
0% 3% 5% 7%
Mammography
Endoscopy
$ 56,360 $ 73,761 $ 86,784 $100,851
$4,310 $5,490 $6,363 $7,300
graphic surveillance is a well enough established value that it was not varied for this analysis.) Although cost per cancer detected and cured is substantially higher for Barrett’s surveillance, the cost per life-yr gained is still lower for the Barrett’s cancers than for breast cancer. Results are shown in Table 8. The only assumption that leads to a cost-effectiveness advantage for mammographic surveillance was that treatment of all 50 mammographically detected tumors resulted in cures in women who would otherwise have died within 1 yr of breast cancer (cost per life-yr saved of $3,233 and $5,490 for mammography and endoscopy respectively). Such a reduction of mortality risk is, of course, well in excess of the documented value of 0.74. Varying discount rates, from 0% to 7%, were applied to the cost per life-yr saved values and the results are shown in Table 9. Because both groups develop cancer at a similar age and the cancers display similar lead times from a surveillance-detected state to a clinically detectable tumor, a relative comparison of discounting shows little difference regardless of the rate chosen. DISCUSSION A cancer surveillance program has value if a population at risk for developing cancer is willing to undergo an examination that can reliably detect cancer in an early and curable stage, resulting in lives saved. Surveillance endoscopy for patients with Barrett’s esophagus has been shown by us (1) and by others (2, 25) to fulfill these requirements. However, the cost-effectiveness of this practice has been called into question (24, 26). The data presented in this study show that endoscopic surveillance of Barrett’s esophagus patients is more cost-effective than mammographic surveillance for breast cancer in terms of cost per cancer detected, cancer cured, and life-yr saved. The cost-effectiveness of an intervention requires that it be compared with something, usually the costs that result from failure to perform the intervention. Our simple, direct cost methodology was meant to produce values meaningful only in the context of this comparison. To be sure, absolute
AJG – June 1998
ENDOSCOPIC SURVEILLANCE OF BARRETT’S ESOPHAGUS
dollar amounts can vary widely depending on the assumptions made regarding screening technique and frequency, cancer rates, therapeutic efficacy, and the treatment of indirect and intangible costs (27). However, an analysis of this sort, based on established clinical practices and using incidence and efficacy assumptions that are widely accepted, allows a fair relative comparison of cost-effectiveness. Achkar and colleagues (28) have estimated that detecting a case of carcinoma arising in Barrett’s esophagus costs $62,000, a value somewhat higher than our estimate of $37,928. Skinner has estimated the cost to cure a case of carcinoma in Barrett’s esophagus to be $80,000 (1), virtually identical to our own estimate. Others have estimated the cost per quality-adjusted life-yr gained to be several times more than our estimate (25), as high as $118,000 (29). This difference is accounted for by our use of direct costs only, excluding indirect benefits such as palliative care for undetected cancers, and indirect costs such as lost wages and quality of life assessment. In terms of overall benefit to society, breast cancer screening will save more lives than will screening for carcinoma in Barrett’s esophagus, because of the larger number of women at risk for breast cancer than for adenocarcinoma in Barrett’s esophagus. The cost-effectiveness of Barrett’s esophagus surveillance results from the high cancer incidence, the ease of cancer detection at an early stage, and the marked difference in surgical outcome between surveillance-detected and clinically detected cancers. Mammographic surveillance decreases the relative mortality risk to 0.74 compared with endoscopic surveillance of Barrett’s esophagus, with a relative mortality risk of 0.26. According to the SEER (5) data, women $50 yr of age have a breast cancer incidence of 345 cases per 100,000 population per year. In our study, the cancer incidence of patients with Barrett’s esophagus was 1,370 cases per 100,000 population per year, amounting to a yearly cancer risk of 1.37%. This is substantial, especially for young patients, and warrants surveillance, the cost of which is adequately recovered in years of life saved by early intervention. Reprint requests and correspondence: John M. Streitz, Jr., M.D., Department of Thoracic and Cardiovascular Surgery, Duluth Clinic, 400 East Third Street, Duluth, MN 55805.
REFERENCES 1. Streitz JM Jr, Andrews CW Jr, Ellis FH Jr. Endoscopic surveillance of Barrett’s esophagus: Does it help? J Thor Cardiovasc Surg 1993;105: 383– 8. 2. Altorki NK, Sunagawa M, Little AG, et al. High-grade dysplasia in the columnar-lined esophagus. Am J Surg 1991;161:97–100. 3. Pera M, Tastek VF, Carpenter HA, et al. Barrett’s esophagus with high-grade dysplasia: An indication for esophagectomy? Ann Thorac Surg 1992;54:199 –204. 4. Kerlikowske K, Grady D, Ruben SM, et al. Efficacy of screening mammography; a meta analysis. JAMA 1995;273:149 –54.
915
5. National Cancer Institute’s “Statistics, epidemiology and end results” data, 1995 ed. 6. Statistical Abstract of the United States, 1995 ed., p. 86. 7. Ohuchi N, Yoshida K, Kimura M, et al. Improved detection rate of early breast cancer in mass screening combined with mammography. Jpn J Cancer Res 1993;84:807–12. 8. Paci L, Perfetti P, Zappa M, et al. First results of a mammographic screening program in two municipalities of Massa-Carrara Province. Tumori 1993;79:26 –9. 9. Burhenne LJ, Hislop TG, Burhenne HJ. The British Columbia Mammography Screening Program: Evaluation of the first 15 months. AJR 1992;158:45–9. 10. Menon A, Devlin NJ, Richardson AK. The costs of mammography screening in New Zealand: Evidence from the pilot programmes. N Z J 1994;107:501–3. 11. Daures JP, Peray P, Bousquet J, et al. Mass screening for breast cancer in Herault county. First results in women between 40 and 70 years old. Journal de Gynecologie, Obstetrique et Biologie de la Reproduction 1992;21:173– 8. 12. Chamberlain J, Moss SM, Kirkpatrick AE, et al. National Health Service breast screening programme results for 1991–2. Br Med J 1993;307:353– 6. 13. Faulk RM, Sickles EA, Sollitto RA, et al. Clinical efficacy of mammographic screening in the elderly. Radiology 1995;194:193–7. 14. Morimoto T, Sasa M, Yamaguchi T, et al. High detection rate of breast cancer by mass screening using mammography in Japan. Jpn J Cancer Res 1994;85:1193–5. 15. Quiet CA, Ferguson DJ, Weichselbaum RR, et al. Natural history of node-negative breast cancer: A study of 826 patients with long-term follow-up. J Clin Oncol 1995;13:1144 –51. 16. Letton AH, Mason EM, Ramshaw BJ. Twenty-year review of a breast cancer screening project. Cancer 1996;77:104 – 6. 17. Carter CL, Allen C, Henson DE. Relation of tumor size, lymph node status and survival in 24,740 breast cancer cases. Cancer 1989;63: 181–7. 18. Skinner DB. The incidence of cancer in Barrett’s esophagus varies according to series. In: Giuli R, McCallum RW, eds. Benign lesions of the esophagus and cancer. New York: Springer-Verlag, 1989:764 –5. 19. Hameeteman W, Tytgat GNJ, Houthoff J, et al. Barrett’s esophagus: Development of dysplasia and adenocarcinoma. Gastroenterology 1989;96:1249 –56. 20. Robertson CS, Mayberry JF, Nicholson DA, et al. Value of endoscopic surveillance in the detection of neoplastic change in Barrett’s oesophagus. Br J Surg 1988;75:760 –3. 21. Williamson WA, Ellis FH Jr, Gibb PS, et al. Barrett’s esophagus: Prevalence and incidence of adenocarcinoma. Arch Intern Med 1991; 151:2212– 6. 22. Van der Veen AH, Dees J, Blankensteijn JD, et al. Adenocarcinoma in Barrett’s oesophagus: An overrated risk. Gut 1989;30:14 – 8. 23. Spechler SJ, Robbins AH, Rubins HB, et al. Adenocarcinoma and Barrett’s esophagus: An overrated risk? Gastroenterology 1984;87: 927–33. 24. vander Burgh A, Dees J, Hop WCJ, et al. Oesophageal cancer is an uncommon cause of death in patients with Barrett’s esophagus. Gut 1996;39:5– 8. 25. Provenzale D, Kemp JA, Arora S, et al. A guide for surveillance of patients with Barrett’s esophagus. Am J Gastroenterol 1994;89:670 – 80. 26. Richter JE, Falk GW. Barrett’s esophagus and adenocarcinoma: The need for a consensus conference. J Clin Gastroenterol 1996;23:88 –90. 27. Eisenberg JM. Clinical economics: A guide to the economic analysis of clinical practices. JAMA 1989;262:2879 – 86. 28. Achkar E, Carey W. The cost of surveillance for adenocarcinoma complicating Barrett’s esophagus. Am J Gastroenterol 1988;83:291– 4. 29. Provenzale D, Wong JB, Kemp JA, et al. Endoscopic surveillance for Barrett’s esophagus; can we afford the cost? Am J Gastroenterol 1990;85:1224 (abstract).
Vol. 93, No. 6, 1998 ISSN 0002-9270/98/$19.00 PII S0002-9270(98)00156-7
Endoscopic Surveillance of Barrett’s Esophagus: A Cost-Effectiveness Comparison With Mammographic Surveillance for Breast Cancer John M. Streitz, Jr., M.D., F. Henry Ellis, Jr., M.D., Ph.D., Robert L. Tilden, DrPH., and Robert V. Erickson, M.D. Duluth Clinic, University of Minnesota School of Medicine, Duluth, Minnesota; Whiteside Institute for Clinical Research, Duluth, Minnesota; and Deaconess Hospital, Harvard Medical School, Boston, Massachusetts
Objective: Endoscopic surveillance of Barrett’s esophagus is commonly practiced to detect malignancy in an early and curable stage. However, the cost-effectiveness of this practice has been questioned. To clarify this issue, we undertook a cost analysis of endoscopic surveillance to detect adenocarcinoma in Barrett’s esophagus compared with mammography used to detect occult carcinoma of the breast, a widely accepted cancer surveillance technique. Methods: The rate of esophageal adenocarcinoma detected by endoscopic surveillance was calculated for Duluth Clinic patients with Barrett’s esophagus seen from 1980 to 1995 and compared with published rates. The rate of occult breast cancer detection was calculated for all women undergoing surveillance mammography at the Duluth Clinic for the year 1994 and compared with published rates. Costs for screening studies and therapy for cancer treatment for both cancers were calculated based on clinical results and assumptions regarding outcomes derived from published reports, and the costs were compared. Results: Endoscopic surveillance of 149 patients with benign Barrett’s esophagus was performed for a total of 510 patient-yr, during which time seven patients developed adenocarcinoma, an incidence of one case per 73 patient-yr of follow-up. Occult breast cancer was detected in 50 of 12,537 mammograms, a detection rate of 0.4%. The incidences in both cases were comparable to published figures. The costs of detecting a case of adenocarcinoma in Barrett’s esophagus and occult breast cancer were $37,928 and $54,513, respectively, and those for treatment resulting in cure were $83,340 and $83,292. Cost per life-yr saved was $4,151 for adenocarcinoma in Barrett’s esophagus and $57,926 for breast cancer. Conclusion: Endoscopic surveillance of patients with Barrett’s esophagus compares favorably with the common practice of surveillance mammography to detect early breast cancer, and should therefore be considered to be
as cost-effective as surveillance mammography. (Am J Gastroenterol 1998;93:911–915. © 1998 by Am. Coll. of Gastroenterology)
INTRODUCTION Although endoscopic surveillance of patients with Barrett’s esophagus is commonly practiced in an effort to discover malignant epithelial changes in an early and curable stage, little is known about the cost-effectiveness of such surveillance. Regular endoscopy with biopsy has been shown to be effective in detecting the early histological changes of malignancy and to result in significant improvement in postoperative survival in those undergoing resection. We (1) and others (2, 3) have shown that surveillance endoscopy increases long term survival after resection of esophageal adenocarcinoma from 20% to 62% or more in patients undergoing yearly endoscopy. Notwithstanding this demonstrated effectiveness, the cost of such a surveillance program has been considered by some to be prohibitively high. The cost-effectiveness of endoscopic surveillance of patients with Barrett’s esophagus has been estimated by others, but absolute dollar amounts are subject to marked variability depending on the assumptions used. A relative comparison with an established screening practice seemed more likely to clarify this issue. Accordingly, we compared the cost incurred, in our current clinical practice, to detect and cure a patient with adenocarcinoma arising in Barrett’s esophagus versus the cost of detecting and curing a patient with carcinoma of the breast using surveillance mammography, a technique widely regarded as cost-effective. MATERIALS AND METHODS Data This study was designed to be a cost-effective analysis comparing the direct surveillance costs for patients with adenocarcinoma arising in the breast and in Barrett’s esophagus, and the cost of curing them, based on our clinical
Presented in part at the 61st Annual Meeting of the American College of Gastroenterology, October 22, 1996, Seattle, Washington. Received Dec. 5, 1996; accepted Feb. 11, 1998. 911
912
STREITZ et al.
experience. To determine whether our experience could be extrapolated to larger patient populations, the cancer detection rates in this study were compared with published results from other studies. Indirect costs such as time lost from work or medical costs to palliate late-stage disease were not included in our calculations, nor were indirect benefits. Cancer detection rates for both groups were determined by retrospective record review of patients followed at the Duluth Clinic. In the case of breast cancer, a sufficient number of cancers were detected in a single year to yield a yearly cancer rate suitable for the calculations. In the case of Barrett’s esophagus, far fewer cancers are detected yearly. Therefore, a cancer rate was established by review of 15 yr of endoscopic patient follow-up. This detection rate was then used as the yearly cancer incidence in the cost calculations. Data regarding cancer detection rates and incidence in our patients was compared with other published data to ensure that our own experience was within established limits. Where reliably significant data regarding our own patients (such as long-term survival following therapy) was lacking, assumptions regarding cure rates were extracted from accepted published values. Data from the literature regarding yearly surveillance mammography were limited to women age 50 –74 yr, for whom there is unequivocal evidence of cancer death reduction (4). Cost analysis Patient charges for 1995 were used as cost estimates with the assumption that variation between them would be similar for both groups of patients and would provide a fair relative comparison, if not a precise absolute cost figure. Only direct costs related to detection and therapy aimed at cure were included. Indirect costs, such as time lost from work, and indirect benefits were not included in the calculations. Cost of palliating late stage disease was also excluded, because assumptions regarding such therapy would vary widely, and such care for both groups is assumed to be similar in cost. Breast cancer. Data from 1994 were used to avoid the changes in costs related to the introduction in 1995 of stereotactic breast biopsy. Costs for breast cancer detection included mammography charges for the entire group, and the cost of excisional biopsy for those found to have positive mammograms. Occult cancers were defined as nonpalpable tumors discovered as a result of surveillance mammography. The cost of curing a cancer included detection costs and the cost of subsequent cancer therapy. Biopsy costs for patients with positive mammograms but with benign biopsies were included. Cost of therapy was based on the assumption that half of those found to have cancer would undergo lumpectomy and radiation therapy and the other half mastectomy, and included adjuvant chemotherapy for both groups. Operative mortality was assumed to be 0%. Life-years saved were calculated using the median age of onset for breast cancer published in the National Cancer Institute’s Statistics, Epidemiology and End Results (SEER) (5) data and compared with the life expectancy of women in
AJG – Vol. 93, No. 6, 1998 TABLE 1 Surgery for Adenocarcinoma in Barrett’s Esophagus Detected by Surveillance Endoscopy Study, year (ref)
Median Age
Operative Mortality
Stage 0 (%)
5-Yr survival (%)
Altorki 1991 (2) Pera 1992 (3) Streitz 1993 (1) Present study Mean
47 66 62 64 60
0 0 0 0 0
55% 55% 38% 57% 51%
100% 66% 62% 76%
the United States (6). The increase in number of patients cured as a result of mammographic detection was estimated using published relative risk reduction values and applying it to our patient group. Adenocarcinoma in Barrett’s esophagus. Surveillance and therapy for adenocarcinoma in Barrett’s esophagus included endoscopy and biopsy charges for the entire screened group and the cost of esophagogastrectomy for those found to have cancer, which included the diagnosis of high grade dysplasia. Operative mortality was assumed to be 0% (Table 1). No adjuvant therapy was included in the charges. Lifeyears saved were calculated using the mean age at onset of adenocarcinoma in Barrett’s esophagus from published results compared with the life expectancy of white males in the United States. The expected increase in number of patients cured following resection was estimated using long term survival figures published for both surveillance and nonsurveillance patients. RESULTS Cancer detection rates and costs Breast cancer. During 1994, 12,537 mammograms were performed at the Duluth Clinic. During that year, 267 patients underwent needle-localization biopsies for clinically inapparent, mammographically suspicious lesions. Of these, 50 proved to be malignant. This yields a positive biopsy rate of 19% and a rate of occult malignancy detection of 0.4%, values within the range reported in other series (Table 2). Cost calculations are shown in Table 3. At $125 per mammogram and including the surgical biopsy cost of $4,339 each for 267 patients with positive mammograms, a total of $2,725,638 was expended to discover 50 occult cancers, a cost of $54,513 per case detected. Studies from the last decade regarding survival of patients with nodenegative (15), mammographically detected (16), or T1 N0 tumors (17) show an average long term survival of 86%. Assuming that therapy would be divided equally between lumpectomy and radiation ($21,437) and mastectomy ($12,800), a total cost of $3,581,563 would be incurred to discover and treat these 50 patients, 43 (86%) of whom would be cured at a cost of $83,292 per cure. Barrett’s esophagus and adenocarcinoma. From 1980 to 1995, 149 patients with Barrett’s esophagus were seen at the Duluth Clinic, of whom 13 had adenocarcinoma when first
AJG – June 1998
ENDOSCOPIC SURVEILLANCE OF BARRETT’S ESOPHAGUS
TABLE 2 Cancer Detection Rates for Screening Mammography Number of Patients
Study, Year (Ref) Ohuchi, 1993 (7) Paci, 1993 (8) Burhenne, 1992 (9) Present study Menon, 1994 (10) Daures, 1992 (11) Chamberlain, 1993 (12) Faulk, 1995 (13) Morimoto, 1994 (14) Mean
9,634 10,407 11,824 12,537 8,000 5,098 1,059,703 32,140 950
% Cancer Rate 0.31 0.37 0.40 0.40 0.51 0.59 0.62 0.74 0.84 0.53
TABLE 3 Cost of Detection and Treatment for Surveillance-Detected Breast Cancer Number Mammograms Biopsies
12,537 267
Unit Cost
Total Cost
$ 125 $ 4,339
$1,567,125 $1,158,513
913
TABLE 4 Incidence of Adenocarcinoma in Barrett’s Esophagus Study, Year (ref)
Case/Patient-year follow-up
Skinner, 1989 (18) Hameeteman, 1989 (19) Robertson, 1988 (20) Present study Williamson, 1991 (21) Van der Veen, 1989 (22) Spechler, 1984 (23) Mean
1/48 1/52 1/56 1/73 1/99 1/170 1/175 1/100
TABLE 5 Cost of Detection and Treatment of Surveillance Detected Adenocarcinoma in Barrett’s Esophagus Number Endoscopy with biopsy Cancers detected Cancers resected Cancers cured (76%) Cost per cure
354 7 7 (76%) 5
Unit Cost $ 750 $37,928 $21,600
Total Cost $265,500 $151,200 $416,700
$83,340
$2,725,638 Cancers detected Lumpectomy-RoRx Mastectomy Cures (86%) Cost per cure
50 (0.4%) 25 25 43
$54,513 $21,437 $12,800
$ 535,925 $ 320,000 $3,581,563
$83,292
seen, a prevalence of 9.4%. The 136 patients with benign Barrett’s esophagus were followed for 510 patient yr, and underwent 354 endoscopic examinations with biopsy, an average interval of one every 17 months. No endoscopic esophageal perforations occurred during the exams. During the follow-up period, seven adenocarcinomas were detected. This cancer incidence of one case per 73 yr of patient follow-up, or 1,370 cases per 100,000 population per year, is similar to other reported rates (Table 4). All seven patients underwent esophagogastrectomy. Five are alive without disease at 1, 2, 4, 6, and 7 yr postoperatively, and two have died, one 6 months after surgery of unrelated causes without evidence of disease, and one of metastatic cancer 28 months postoperatively. The pathological stages of the resected specimens were stage 0 in three patients, stage I in two, and stage IIA in two. This early stage distribution is similar to that which we have reported previously in a different group of patients (1), and is similar to other published results (Table 1). Cost calculations for detection and treatment of these cancers are shown in Table 5. Assuming that all seven patients reported here underwent resection at a cost of $21,600 per case, and that five patients, or 76% (Table 1) were cured, the total cost per cure would equal $83,340. Life-yr gained The cost per life-yr gained in both groups is based on the formula Cost S 2 Cost NS, 4 # life-years gained where
Cost S 5 costs of detection and treatment with surveillance and Cost NS is that for no surveillance. It is not known how long it takes high grade dysplasia to become a clinically detectable tumor, nor is the time separating a mammographically detected cancer from a clinically palpable one well defined. For the purposes of these calculations, it was assumed that all surveillance-detected cancers would have become clinically detectable within equal time periods, and that all would then have been biopsied and treated in the same way as surveillance-detected cancers. Thus, the difference, Cost/Cure S 2 Cost/Cure NS, is comprised of the surveillance costs for patients who do not develop cancer. Breast cancer. Recalculation of costs for the clinically detected breast cancers, shown in Table 6, exclude mammography and biopsy costs for all but the 50 patients whose cancers were mammographically detected, inasmuch as mammography would be performed when these tumors became clinically evident. Reduction of breast cancer mortality for women age 50 –74 yr undergoing surveillance mammography yearly averages 26% (4) or a relative risk of 0.74. This risk difference would mean 9.4 cancer deaths in 50 patients as compared with the seven deaths assumed for the surveillance group. According to the SEER (5) data, the mean age at onset for women with breast cancer is 64 yr. The life expectancy for women in the United States is 82 years of age (6). The 2.4 cancer deaths averted by a surveillance strategy would yield 43.2 life-yr gained. Therefore, $3,581,563 2 $1,079,125 4 43.2 5 $57,926 per life-yr gained. Barrett’s esophagus. Recalculation of costs for the clinically detected Barrett’s esophagus cancers, shown in Table 7, excludes surveillance endoscopy costs for all those who
914
STREITZ et al.
AJG – Vol. 93, No. 6, 1998 TABLE 8 Sensitivity Analysis Results
TABLE 6 Cost of Treatment for Clinically Detected Breast Cancer Number Mammograms Biopsies Cancers Lumpectomy-RoRx Mastectomy
50 50 50 25 25
Unit Cost
Total Cost
Cost
Mammography
Endoscopy
$ 125 $ 4,339
$ 6,250 $ 216,950
Cost per cancer detected Cost per cure Cost per life-year saved
$25,714 $50,446 $25,302
$ 88,500 $173,842 $ 16,650
$21,437 $12,800
$ 535,925 $ 320,000
TABLE 9 Discount Values
$1,079,125 Cures Cost per cure
40.6
Discount $26,579
TABLE 7 Cost of Treatment for Clinically Detected Adenocarcinoma in Barrett’s Esophagus Number
Unit Cost
Total Cost $
Endoscopy with biopsy Cancers detected Cancers resected
7 7 7
$ 750 $ 750 $ 21,600
Cures Cost per cure
1.4
$111,750 $ 79,821
5,250
$151,200 $156,450
did not develop cancer. The cost per cancer cured is $79,821. Long term survival after esophagectomy for surveillancedetected Barrett’s adenocarcinoma has been shown to be 76% on average for those undergoing yearly endoscopy, three times the survival of patients with clinically detected cancers. The mean age of those developing esophageal adenocarcinoma is 60 yr (Table 1). In the United States the life expectancy of white men, who comprise the vast majority of those developing cancer in Barrett’s esophagus, is 76 yr. Assuming that surveillance increased the cures from 1.4 of seven to 5.3 of seven patients (from 20% to 76%), a relative mortality risk of 0.26, the number of life-yr gained by surveillance is 62.7 (3.9 3 16). Cost per life-yr gained is therefore $416,700 2 $156,450 4 62.7 5 $4,151. Sensitivity analysis The greatest uncertainties regarding cost-effectiveness conclusions in this study arise from the wide range of reported cancer incidences and detection rates for both cancers. The sensitivity analysis substitutes for the average values used in the initial calculations, the extreme values from the tables listing incidence of adenocarcinoma in Barrett’s, mammographic cancer detection rates, and cure rates for esophageal resection. Using values that improve the cost-effectiveness results for breast cancer and worsen those for Barrett’s cancers, recalculations of cost per cancer detected, per cancer cured, and per life-yr gained were performed. Values were chosen from the extremes of Table 1 (62% cure rate), Table 2 (0.84 cancer rate), and Table 4 (1 case per 175 patient-yr). (Mortality risk reduction for breast cancer detected by mammo-
0% 3% 5% 7%
Mammography
Endoscopy
$ 56,360 $ 73,761 $ 86,784 $100,851
$4,310 $5,490 $6,363 $7,300
graphic surveillance is a well enough established value that it was not varied for this analysis.) Although cost per cancer detected and cured is substantially higher for Barrett’s surveillance, the cost per life-yr gained is still lower for the Barrett’s cancers than for breast cancer. Results are shown in Table 8. The only assumption that leads to a cost-effectiveness advantage for mammographic surveillance was that treatment of all 50 mammographically detected tumors resulted in cures in women who would otherwise have died within 1 yr of breast cancer (cost per life-yr saved of $3,233 and $5,490 for mammography and endoscopy respectively). Such a reduction of mortality risk is, of course, well in excess of the documented value of 0.74. Varying discount rates, from 0% to 7%, were applied to the cost per life-yr saved values and the results are shown in Table 9. Because both groups develop cancer at a similar age and the cancers display similar lead times from a surveillance-detected state to a clinically detectable tumor, a relative comparison of discounting shows little difference regardless of the rate chosen. DISCUSSION A cancer surveillance program has value if a population at risk for developing cancer is willing to undergo an examination that can reliably detect cancer in an early and curable stage, resulting in lives saved. Surveillance endoscopy for patients with Barrett’s esophagus has been shown by us (1) and by others (2, 25) to fulfill these requirements. However, the cost-effectiveness of this practice has been called into question (24, 26). The data presented in this study show that endoscopic surveillance of Barrett’s esophagus patients is more cost-effective than mammographic surveillance for breast cancer in terms of cost per cancer detected, cancer cured, and life-yr saved. The cost-effectiveness of an intervention requires that it be compared with something, usually the costs that result from failure to perform the intervention. Our simple, direct cost methodology was meant to produce values meaningful only in the context of this comparison. To be sure, absolute
AJG – June 1998
ENDOSCOPIC SURVEILLANCE OF BARRETT’S ESOPHAGUS
dollar amounts can vary widely depending on the assumptions made regarding screening technique and frequency, cancer rates, therapeutic efficacy, and the treatment of indirect and intangible costs (27). However, an analysis of this sort, based on established clinical practices and using incidence and efficacy assumptions that are widely accepted, allows a fair relative comparison of cost-effectiveness. Achkar and colleagues (28) have estimated that detecting a case of carcinoma arising in Barrett’s esophagus costs $62,000, a value somewhat higher than our estimate of $37,928. Skinner has estimated the cost to cure a case of carcinoma in Barrett’s esophagus to be $80,000 (1), virtually identical to our own estimate. Others have estimated the cost per quality-adjusted life-yr gained to be several times more than our estimate (25), as high as $118,000 (29). This difference is accounted for by our use of direct costs only, excluding indirect benefits such as palliative care for undetected cancers, and indirect costs such as lost wages and quality of life assessment. In terms of overall benefit to society, breast cancer screening will save more lives than will screening for carcinoma in Barrett’s esophagus, because of the larger number of women at risk for breast cancer than for adenocarcinoma in Barrett’s esophagus. The cost-effectiveness of Barrett’s esophagus surveillance results from the high cancer incidence, the ease of cancer detection at an early stage, and the marked difference in surgical outcome between surveillance-detected and clinically detected cancers. Mammographic surveillance decreases the relative mortality risk to 0.74 compared with endoscopic surveillance of Barrett’s esophagus, with a relative mortality risk of 0.26. According to the SEER (5) data, women $50 yr of age have a breast cancer incidence of 345 cases per 100,000 population per year. In our study, the cancer incidence of patients with Barrett’s esophagus was 1,370 cases per 100,000 population per year, amounting to a yearly cancer risk of 1.37%. This is substantial, especially for young patients, and warrants surveillance, the cost of which is adequately recovered in years of life saved by early intervention. Reprint requests and correspondence: John M. Streitz, Jr., M.D., Department of Thoracic and Cardiovascular Surgery, Duluth Clinic, 400 East Third Street, Duluth, MN 55805.
REFERENCES 1. Streitz JM Jr, Andrews CW Jr, Ellis FH Jr. Endoscopic surveillance of Barrett’s esophagus: Does it help? J Thor Cardiovasc Surg 1993;105: 383– 8. 2. Altorki NK, Sunagawa M, Little AG, et al. High-grade dysplasia in the columnar-lined esophagus. Am J Surg 1991;161:97–100. 3. Pera M, Tastek VF, Carpenter HA, et al. Barrett’s esophagus with high-grade dysplasia: An indication for esophagectomy? Ann Thorac Surg 1992;54:199 –204. 4. Kerlikowske K, Grady D, Ruben SM, et al. Efficacy of screening mammography; a meta analysis. JAMA 1995;273:149 –54.
915
5. National Cancer Institute’s “Statistics, epidemiology and end results” data, 1995 ed. 6. Statistical Abstract of the United States, 1995 ed., p. 86. 7. Ohuchi N, Yoshida K, Kimura M, et al. Improved detection rate of early breast cancer in mass screening combined with mammography. Jpn J Cancer Res 1993;84:807–12. 8. Paci L, Perfetti P, Zappa M, et al. First results of a mammographic screening program in two municipalities of Massa-Carrara Province. Tumori 1993;79:26 –9. 9. Burhenne LJ, Hislop TG, Burhenne HJ. The British Columbia Mammography Screening Program: Evaluation of the first 15 months. AJR 1992;158:45–9. 10. Menon A, Devlin NJ, Richardson AK. The costs of mammography screening in New Zealand: Evidence from the pilot programmes. N Z J 1994;107:501–3. 11. Daures JP, Peray P, Bousquet J, et al. Mass screening for breast cancer in Herault county. First results in women between 40 and 70 years old. Journal de Gynecologie, Obstetrique et Biologie de la Reproduction 1992;21:173– 8. 12. Chamberlain J, Moss SM, Kirkpatrick AE, et al. National Health Service breast screening programme results for 1991–2. Br Med J 1993;307:353– 6. 13. Faulk RM, Sickles EA, Sollitto RA, et al. Clinical efficacy of mammographic screening in the elderly. Radiology 1995;194:193–7. 14. Morimoto T, Sasa M, Yamaguchi T, et al. High detection rate of breast cancer by mass screening using mammography in Japan. Jpn J Cancer Res 1994;85:1193–5. 15. Quiet CA, Ferguson DJ, Weichselbaum RR, et al. Natural history of node-negative breast cancer: A study of 826 patients with long-term follow-up. J Clin Oncol 1995;13:1144 –51. 16. Letton AH, Mason EM, Ramshaw BJ. Twenty-year review of a breast cancer screening project. Cancer 1996;77:104 – 6. 17. Carter CL, Allen C, Henson DE. Relation of tumor size, lymph node status and survival in 24,740 breast cancer cases. Cancer 1989;63: 181–7. 18. Skinner DB. The incidence of cancer in Barrett’s esophagus varies according to series. In: Giuli R, McCallum RW, eds. Benign lesions of the esophagus and cancer. New York: Springer-Verlag, 1989:764 –5. 19. Hameeteman W, Tytgat GNJ, Houthoff J, et al. Barrett’s esophagus: Development of dysplasia and adenocarcinoma. Gastroenterology 1989;96:1249 –56. 20. Robertson CS, Mayberry JF, Nicholson DA, et al. Value of endoscopic surveillance in the detection of neoplastic change in Barrett’s oesophagus. Br J Surg 1988;75:760 –3. 21. Williamson WA, Ellis FH Jr, Gibb PS, et al. Barrett’s esophagus: Prevalence and incidence of adenocarcinoma. Arch Intern Med 1991; 151:2212– 6. 22. Van der Veen AH, Dees J, Blankensteijn JD, et al. Adenocarcinoma in Barrett’s oesophagus: An overrated risk. Gut 1989;30:14 – 8. 23. Spechler SJ, Robbins AH, Rubins HB, et al. Adenocarcinoma and Barrett’s esophagus: An overrated risk? Gastroenterology 1984;87: 927–33. 24. vander Burgh A, Dees J, Hop WCJ, et al. Oesophageal cancer is an uncommon cause of death in patients with Barrett’s esophagus. Gut 1996;39:5– 8. 25. Provenzale D, Kemp JA, Arora S, et al. A guide for surveillance of patients with Barrett’s esophagus. Am J Gastroenterol 1994;89:670 – 80. 26. Richter JE, Falk GW. Barrett’s esophagus and adenocarcinoma: The need for a consensus conference. J Clin Gastroenterol 1996;23:88 –90. 27. Eisenberg JM. Clinical economics: A guide to the economic analysis of clinical practices. JAMA 1989;262:2879 – 86. 28. Achkar E, Carey W. The cost of surveillance for adenocarcinoma complicating Barrett’s esophagus. Am J Gastroenterol 1988;83:291– 4. 29. Provenzale D, Wong JB, Kemp JA, et al. Endoscopic surveillance for Barrett’s esophagus; can we afford the cost? Am J Gastroenterol 1990;85:1224 (abstract).