- Email: [email protected]
Which colon cancer screening test? A comparison of costs, effectiveness, and compliance
CLINICAL STUDIES
Which Colon Cancer Screening Test? A Comparison of Costs, Effectiveness, and Compliance Sandeep Vijan, MD, MS, Erica W. Hwang, MD, Timothy P. Hofer, MD, MS, Rodney A. Hayward, MD PURPOSE: Recent media reports have advocated the use of colonoscopy for colorectal cancer screening. However, colonoscopy is expensive compared with other screening modalities, such as fecal occult blood testing and flexible sigmoidoscopy. We sought to determine the cost effectiveness of different screening strategies for colorectal cancer at levels of compliance likely to be achieved in clinical practice. METHODS: A Markov decision model was used to examine screening strategies, including fecal occult blood testing alone, fecal occult blood testing combined with flexible sigmoidoscopy, flexible sigmoidoscopy alone, and colonoscopy. The timing and frequency of screening was varied to assess optimal screening intervals. Sensitivity analyses were conducted to assess the factors that have the greatest effect on the cost effectiveness of screening. RESULTS: All strategies are cost effective versus no screening, at less than $20,000 per life-year saved. Direct comparison suggests that the most effective strategies are twice-lifetime colonoscopy and flexible sigmoidoscopy combined with fecal
occult blood testing. Assuming perfect compliance, flexible sigmoidoscopy combined with fecal occult blood testing is slightly more effective than twice-lifetime colonoscopy (at ages 50 and 60 years) but is substantially more expensive, with an incremental cost effectiveness of $390,000 per additional life-year saved. However, compliance with primary screening tests and colonoscopic follow-up for polyps affect screening decisions. Colonoscopy at ages 50 and 60 years is the preferred test regardless of compliance with the primary screening test. However, if follow-up colonoscopy for polyps is less than 75%, then even once-lifetime colonoscopy is preferred over most combinations of flexible sigmoidoscopy and fecal occult blood testing. Costs of colonoscopy and proportion of cancer arising from polyps also affect cost effectiveness. CONCLUSIONS: Colonoscopic screening for colorectal cancer appears preferable to current screening recommendations. Screening recommendations should be tailored to the compliance levels achievable in different practice settings. Am J Med. 2001;111:593– 601. 䉷2001 by Excerpta Medica, Inc.
C
In 1994, the Canadian Task Force on the Periodic Health Examination suggested that screening with flexible sigmoidoscopy was the best option for reducing colorectal cancer mortality. The panel concluded that colonoscopic screening was not feasible because of high cost and low compliance, and that fecal occult blood testing was not ideal because of low sensitivity and specificity (12). More recent guidelines from the American Gastroenterological Association recommend that any one of several screening procedures should be used in patients with an average risk of colorectal cancer. These procedures include annual fecal occult blood testing, flexible sigmoidoscopy every fifth year, combined annual fecal occult blood testing with every fifth-year flexible sigmoidoscopy, double-contrast barium enema every 5 to 10 years, or colonoscopy every 10 years (13). Previous decision analyses suggest that strategies that use combinations of sigmoidoscopy and fecal occult blood testing are cost effective (7,9,10,14 –17). More recently, there has been an increased emphasis on the possibility of colonoscopy as a primary screening test.
olorectal cancer is one of the leading causes of cancer death in the United States (1). Because colorectal cancer has a long asymptomatic phase but is often curable when found early (1), screening can substantially reduce colorectal cancer mortality. While experimental evidence of benefit from screening is thus far available only for fecal occult blood testing (2– 4), evidence from observational studies and simulation models suggests that flexible sigmoidoscopy (5–7) and colonoscopy may also reduce colorectal cancer mortality (7–11).
From The Veterans Affairs Center for Practice Management and Outcomes Research (SV, TPH, RAH), Ann Arbor, Michigan; the Department of Internal Medicine (SV, TPH, RAH), University of Michigan, Ann Arbor, Michigan and the Department of Internal Medicine, Georgetown University Medical Center (EWH), Washington, DC Timothy P. Hofer, MD, MS, and Sandeep Vijan, MD, MS, are recipients of the Veterans Affairs Health Services Research and Development Career Development Award, Ann Arbor, Michigan. Requests for reprints should be addressed to Sandeep Vijan, MD, MS, Veterans Health Services Research and Development, PO Box 130170, Ann Arbor, Michigan 48113-0170. Manuscript submitted February 14, 2001, and accepted in revised form August 23, 2001. 䉷2001 by Excerpta Medica, Inc. All rights reserved.
0002-9343/01/$–see front matter 593 PII S0002-9343(01)00977-9
Which Colon Cancer Screening Test?/Vijan et al
Figure 1. Structure of the Markov model. The branch for “Local, year 1” represents the first year of localized colon cancer; “Local, year 2” represents the second year of localized colon cancer (we assumed that colon cancer takes 2 years to progress from localized to regional cancer). The plus sign indicates additional branches that are not displayed; two branches exist beyond each plus sign, one for risk of progression to more advanced stages of premalignant polyps or colon cancer, and one for intervening mortality.
Although there are no trials comparing screening modalities, the recent debate has been clouded because of differing results from cost-effectiveness models (10,11,17). Moreover, these models have differing assumptions, which makes comparisons difficult. For example, the model by Frazier et al (10) used a higher cost of colonoscopy and assumed that polyps were removed during sigmoidoscopy, both of which favor noncolonoscopic screening. Indeed, they concluded that combined sigmoidoscopy and fecal occult blood testing is the test of choice (10). Another analysis did not compare strategies with each other, but rather with no screening, making the choice of the optimal test unclear (17). Sonnenberg et al. (11) did not use a natural history simulation, but assumed a risk reduction from different tests. They also did not test strategies such as combined flexible sigmoidoscopy and fecal occult blood testing (11). Interpretation of these analyses is also limited by the lack of multiway sensitivity analyses to estimate the effects of changing several factors simultaneously on choice of test. Furthermore, analyses of the effects of compliance were limited in these articles. In randomized trials, the highest compliance with screening was 75% (2), when compliance is much lower in actual practice (18,19). A preferred screening strategy for colon cancer should either have high rates of compliance or maintain cost effectiveness regardless of compliance. Colonoscopy may be a desirable screening option because, in addition to requir594
December 1, 2001
THE AMERICAN JOURNAL OF MEDICINE威
ing only a single test for screening and intervention, persons may prefer to have procedures once or twice in a lifetime to more frequent options (20,21). In addition, the possibility of performing office-based colonoscopy as a low-cost procedure makes colonoscopic screening intriguing (22). Thus, it is imperative to define how compliance affects cost effectiveness, the optimal timing and frequency of colonoscopy, and the effects of pricing on the relative cost effectiveness of different screening procedures. To help answer these questions, we constructed a Markov cost-effectiveness model to examine screening for colorectal cancer. Our model tests variations in the timing and frequency of tests and also estimates the effects of compliance, costs, test characteristics, and the natural history of colorectal cancer. We selected oncelifetime and twice-lifetime colonoscopy as strategies that might optimize compliance and compared their costs and effectiveness with those of fecal occult blood testing, flexible sigmoidoscopy, and sigmoidoscopy combined with fecal occult blood testing.
MATERIAL AND METHODS We created a Markov model to simulate the natural history of colorectal cancer (Figure 1), beginning at age 50
Volume 111
Which Colon Cancer Screening Test?/Vijan et al
Table 1. Model Assumptions Base Case
Range Used in Sensitivity Analysis
75%
50%–100%
20% 40% 50% 55%
10%–30% 30%–50% 40%–60% 45%–65%
15% 35%
5%–25% 25%–45%
0.05% 0.09% 0.14% 0.20% 0.27% 0.35% 0.43% 0.45%
— — — — — — — —
10.5% 35.1% 91.7%
— — —
5% 97.5%
2%–10% 90%–100%
30% 50% 55% 85% 95% 0.1% 7.5%
20%–40% 40%–60% 40%–75% 80%–95% 90%–100% 0.0%–0.3% 5%–10%
References
Natural history Proportion of cancers arising from polyps Prevalence of adenomatous polyps Age 50 years Age 60 years Age 70 years Age 80 years In patients with polyps Proportion of polyps ⬎1 cm Proportion with multiple polyps Annual incidence of colorectal cancer Age 50 years Age 55 years Age 60 years Age 65 years Age 70 years Age 75 years Age 80 years Age 85 years 5-Year colorectal cancer mortality Localized Regional Disseminated Test characteristics Sensitivity of fecal occult blood testing for polyps Specificity of fecal occult blood testing Sensitivity of fecal occult blood testing for cancer Localized Regional Polyps or cancer reachable by flexible sigmoidoscopy Sensitivity of colonoscopy or flexible sigmoidoscopy for polyps Sensitivity of colonoscopy or flexible sigmoidoscopy for cancer Perforation rate, colonoscopy Mortality rate, perforation Costs Fecal occult blood testing Flexible sigmoidoscopy Flexible sigmoidoscopy with biopsy Colonoscopy Polypectomy (including pathology) Cancer care Localized Regional Disseminated Cost of treating colon perforation
$17 $225 $240 $550 $215
$5–$30 $100–$500 $100–$550 $150–1000 $100–300
$60,000 $82,800 $73,000 $20,000
$40,000–$80,000 $60,000–$100,000 $50,000–$90,000 $10,000–$30,000
years. The key assumptions of the model are outlined in Table 1. The main sources of the estimates were colonoscopic screening studies and autopsy studies for the prevalence of polyps (27–31) and the Surveillance, Epidemiology, and End Results (SEER) registry data for the incidence and mortality rates of colorectal cancer (1,32). Mortality rates for the general population were derived from the National Center for Health Statistics publications (57).
We calculated the age-specific incidence of adenomas using the age-specific prevalence of polyps observed in screening and autopsy studies. The incidence rate was calculated so that at each 5-year interval, the prevalence of polyps predicted by the model matched that in published studies (Table 1). The incidence of hyperplastic polyps was defined in the same manner; prevalence ranged from 20% at age 50 years to 15% at age 80 years (not shown in Table 1) (27,28).
(23–26) (27–31)
(27–31) (27–31) (32)
(33)
December 1, 2001
THE AMERICAN JOURNAL OF MEDICINE威
(34–40) (34–36,39,40) (34,36–40)
(15,41–43) (41,44–49) (41,44–49) (50–52) (15,53) (54) (54) (54) (22,54) (54) (7,15,55,56)
(53)
Volume 111 595
Which Colon Cancer Screening Test?/Vijan et al
Figure 2. Observed cancer incidence versus predicted cancer incidence (model output). SEER ⫽ Surveillance, Epidemiology, and End Results [registry data].
We also calculated the rate of conversion of polyps to cancer to match cancer incidence rates. The proportion of incident cancers arising from polyps was assumed to be 75% in the base-case scenario and was varied from 50% to 100% in the sensitivity analyses (23–26). We assumed that it takes 10 years for a polyp to transform from benign to malignant. This 10-year polyp “dwell time” is supported by consensus opinion (13) but is based on indirect evidence, including studies showing that screening sigmoidoscopy provides a 10-year window of protection from colorectal cancer (5). To calculate the conversion rate from polyps to cancer, the prevalence of polyps at time “x” was matched with the incidence at time “x ⫹ 10,” and conversion rates were calculated, including intervening mortality. Thus, a proportion of polyps at each time stage was determined to be progressive, and entered the dwell state from which they gradually became malignant. The model predictions of incidence were nearly identical to the observed incidence in the SEER registry (Figure 2). If a colorectal malignancy developed, we assumed that it took 2 years to progress through localized cancer and an additional year to progress through regional cancer. Patients who developed disseminated cancer were diagnosed within 1 year, whether or not screening was employed. Mortality was varied by stage of colorectal cancer based on published results from the SEER registry (1,32).
Test Characteristics To simulate the effects of screening, strategies were interposed at different intervals. Screening affected mortality in two ways: prevention of cancer through removal of adenomatous polyps and detection of cancer in earlier 596
December 1, 2001
THE AMERICAN JOURNAL OF MEDICINE威
stages (Table 1). The sensitivity, specificity, and complication rates of the screening tests were derived from observational studies (34 – 41,44 – 49,58). For flexible sigmoidoscopy, we assumed that a 70-cm sigmoidoscope could detect 55% of polyps and cancers (eg, flexible sigmoidoscopy “reached” 55% of all lesions); this was varied from 40% to 75% in sensitivity analyses (15,41– 43). Prior compliance with a test did not predict future compliance. However, in the base case, compliance was consistent across tests. For example, in the flexible sigmoidoscopy combined with fecal occult blood testing strategy, if the compliance for fecal occult blood testing was 75%, it was also 75% for flexible sigmoidoscopy, which was performed only if fecal occult blood testing was negative. The effect of compliance with follow-up colonoscopy for positive initial screening tests (eg, positive fecal occult blood testing or flexible sigmoidoscopy) on the noncolonoscopic screening strategies was also evaluated. We assumed that a negative screening test did not reduce the risk of subsequent polyp growth (59 – 61). If an adenomatous polyp was detected, full colonoscopy was performed and the polyp was removed. If the polyp was greater than 1 cm or if there were multiple polyps, a follow-up colonoscopy was performed at 3 years. If this first follow-up was negative, further surveillance colonoscopy was performed at 5-year intervals (62– 64). No routine surveillance was performed for patients who had hyperplastic polyps or polyps less than 1 cm removed; these patients returned to usual screening. Adenomatous polyp recurrence rates at surveillance colonoscopy were taken from the National Polyp Study (64). After a negative follow-up colonoscopy, the rate of subsequent polyp devel-
Volume 111
Which Colon Cancer Screening Test?/Vijan et al
opment was assumed to be the same as that in the general population. Compliance with surveillance testing was assumed to be 100% in the base-case analysis; this was varied from 50% to 100% in sensitivity analyses.
Costs Costs were approached from the perspective of a thirdparty payer; thus, only direct medical costs were included. The costs of screening tests and interventions were taken from the 2000 Medicare reimbursement schedule (54). Because Medicare does not currently reimburse fecal occult blood testing, we based costs on past levels of reimbursement (0.5 relative value units). We included physician fees and facility expenses in procedural costs (colonoscopy, flexible sigmoidoscopy). Polypectomy costs included reimbursement for tissue pathology. Costs of caring for procedural complications (7,15,53) were inflated to 1999 US dollars using the medical care consumer price index. Costs of cancer care were taken from a 1987 Medicare claims analysis (56) and a 1997 Kaiser Permanente analysis (55), and were also inflated to 1999 dollars. All costs and years of life were discounted at 3% in the base case. The discount rate was varied between 0% and 6% in sensitivity analyses.
Sensitivity Analyses Sensitivity analyses were conducted to determine the effects of different variables on the cost effectiveness of the screening strategies. We tested each variable in the model in one-way sensitivity analyses (Table 1). We also conducted three-way sensitivity analyses on the assumptions that had the greatest effects on cost effectiveness in the one-way analyses.
RESULTS Using the base-case estimates, the model predicted that the lifetime risk of colorectal cancer is 5869 per 100,000 (5.9%). Model predictions closely matched the expected incidence from the SEER registry data (Figure 2). The average direct cost of caring for colorectal cancer in an unscreened population is $1,303 per person.
Cost Effectiveness and Incremental Cost Effectiveness The cost-effectiveness ratio of all screening strategies versus no screening is under $20,000 per life-year gained, across all levels of compliance. Once-lifetime colonoscopy at age 60 years is almost cost neutral (cost-effectiveness ratio of $130 per life-year gained), whereas once-lifetime colonoscopy at age 65 years is cost saving (savings of $2280 per person). Differences in life expectancy are small among strategies, so cost is the factor of primary importance (Table 2). Once-lifetime or twicelifetime colonoscopic screenings are less expensive than
combined flexible sigmoidoscopy and fecal occult blood testing. Regardless of compliance level, flexible sigmoidoscopy alone and fecal occult blood testing alone are dominated by colonoscopic screening; that is, colonoscopy is both more effective and less costly (Table 2). At 100% compliance, combined sigmoidoscopy and fecal occult blood testing is more effective than twice-lifetime colonoscopy but costs substantially more, with an incremental costeffectiveness ratio of more than $106,000 per life-year gained. However, as compliance decreases, the noncolonoscopic strategies become less effective and more costly. For example, at 75% compliance, the incremental cost effectiveness of combined sigmoidoscopy and fecal occult blood testing, compared with twice-lifetime colonoscopy, increases to over $330,000 per life-year gained. With further decreases in rates of compliance, combined sigmoidoscopy and fecal occult blood testing is dominated by the colonoscopic strategies. Because compliance with follow-up testing has been poorly described outside of randomized studies, we also evaluated a scenario in which patients had a constant 80% rate of compliance with follow-up colonoscopy. Using this assumption, the sigmoidoscopy alone and fecal occult blood testing alone strategies are dominated by the colonoscopic strategies. However, changing compliance with follow-up does attenuate some of the differences between combined sigmoidoscopy and fecal occult blood testing and the twice-lifetime colonoscopic strategies. For example, at a compliance level of 75%, the incremental cost effectiveness of combined sigmoidoscopy and fecal occult blood testing versus colonoscopy at ages 50 and 60 years is $116,900 (compared with over $330,000 in the base case).
Sensitivity Analyses The discount rate has the largest effect on costs and benefits. However, because the discount rate affects both costs and effectiveness, these differences are minimized when cost-effectiveness ratios are calculated. The major factors that affect cost-effectiveness ratios, in addition to compliance, are the cost of colonoscopy and the proportion of cancer arising from polyps. These three parameters were explored in multivariate sensitivity analyses. Variation in other factors (across the ranges defined in Table 1) and different methods of simulating compliance have small effects on the choice of strategies. We also tested the effect of including all polyps, regardless of size, in the surveillance program. This increases costs for all screening choices. The incremental cost effectiveness of colonoscopic screening is most affected (because of higher polyp sensitivity), but the differences in incremental cost effectiveness are less than $5000 per life-year from the base case.
December 1, 2001
THE AMERICAN JOURNAL OF MEDICINE威
Volume 111 597
Which Colon Cancer Screening Test?/Vijan et al
Table 2. Costs and Effectiveness of Screening
Strategy
Average Gain in Life Expectancy (days)*
Average Cost ($)*
Relative Reduction in Colorectal Cancer Mortality (%)
Incremental Cost per Added Life-year ($)†
— 100%
No screening Flexible sigmoidoscopy Colonoscopy at age 60 years Colonoscopy at age 55 years Fecal occult blood test Colonoscopy at ages 55 and 65 years Colonoscopy at ages 50 and 60 years Flexible sigmoidoscopy and fecal occult blood test
— 12.5 12.6 14.6 16.5 17.6 18.8 20.7
1300 1930 1310 1420 1550 1540 1750 2280
— 44.0 52.7 46.3 52.4 68.0 63.6 69.6
— D 130 20,770 D 14,870 62,140 106,860
75%
Fecal occult blood test Colonoscopy at age 60 years Flexible sigmoidoscopy Colonoscopy at age 55 years Colonoscopy at ages 55 and 65 years Colonoscopy at ages 50 and 60 years Flexible sigmoidoscopy and fecal occult blood test Fecal occult blood test Colonoscopy at age 60 years Flexible sigmoidoscopy Colonoscopy at age 55 years Flexible sigmoidoscopy and fecal occult blood test Colonoscopy at ages 55 and 65 years Colonoscopy at ages 50 and 60 years
8.0 9.5 10.7 11.0 14.3 15.6 15.8 5.8 6.3 7.1 7.3 8.5 10.2 11.4
1470 1310 1730 1390 1450 1600 1840 1420 1310 1590 1360 1570 1380 1480
43.4 39.5 41.0 34.7 56.3 52.7 61.0 32.1 26.3 35.0 23.1 45.6 41.1 38.4
D 130 D 20,940 6,500 42,670 332,630 D 130 D 21,150 D 2,130 32,990
Fecal occult blood test Flexible sigmoidoscopy and fecal occult blood test Flexible sigmoidoscopy Colonoscopy at age 60 years Colonoscopy at age 55 years Colonoscopy at ages 55 and 65 years Colonoscopy at ages 50 and 60 years
1.6 2.2 2.8 3.1 3.7 5.4 6.1
1370 1410 1470 1300 1330 1330 1380
17.9 23.1 23.2 13.2 11.6 22.3 20.9
D D D 130 D 4,140 26,880
Compliance
50%
25%
* Costs and life expectancy discounted at 3% per annum; values rounded to nearest $10 and 0.1 days. † Incremental cost effectiveness versus prior strategy. “D” represents a strategy that is dominated by one of the ensuing strategies, meaning that the ensuing strategies are both less costly and more effective. As a result, dominated strategies are not included in the incremental cost-effectiveness calculations.
Two-way sensitivity analyses varying the cost of colonoscopy and compliance level demonstrate that at the lower bounds of cost ($150 per colonoscopy), twicelifetime colonoscopy is favored at all levels of compliance and is cost saving in many circumstances. Low-cost office-based screening colonoscopy has been performed for a total charge of $150 (22), although reimbursement for colonoscopy is substantially higher than this in most circumstances. To compare incremental cost effectiveness, we conducted three-way analyses varying costs of colonoscopy, percentage of cancers arising from polyps, and compliance (Table 3). If only 50% of cancer arise from polyps, then compliance and cost of colonoscopy have pronounced effects on the choice of therapy. For example, if compliance is 75% or greater, then combined flexible sigmoidoscopy and fecal occult blood testing is a dominant 598
December 1, 2001
THE AMERICAN JOURNAL OF MEDICINE威
strategy if the cost of colonoscopy is $1000; this strategy costs about $39,000 per life-year gained if the cost is $575 and about $113,000 per life-year gained if the cost is $150. At more likely levels of compliance, or if more than 50% of cancers arise from polyps, colonoscopy is the screening strategy of choice unless its costs are very high (Table 3).
Model Validity We tested the model validity by altering the assumptions to fit the Minnesota fecal occult blood testing screening trial (2). We changed the compliance with fecal occult blood testing to 75% and the compliance of follow-up colonoscopy to 81%. We also adjusted the sensitivity and specificity of fecal occult blood testing, because fecal occult blood testing slides were rehydrated in that study. As in the trial, screening was limited to 10 years, with a 13year follow-up. Our model predicts a 39% reduction in
Volume 111
Which Colon Cancer Screening Test?/Vijan et al
Table 3. Three-way Sensitivity Analyses Varying Percentage of Cancers Arising from Polyps, Compliance, and Cost of Colonoscopy*: Comparison of Combined Flexible Sigmoidoscopy and Fecal Occult Blood Testing with Colonoscopy at Ages 55 and 65 Years Cost of Colonoscopy ($)† Compliance
Best Strategy
$150
Incremental cost effectiveness if 100% of cancers arise from polyps 75% Colonoscopy at ages 55 and 65 years Dominant 50% Colonoscopy at ages 55 and 65 years Dominant 25% Colonoscopy at ages 55 and 65 years Dominant Incremental cost effectiveness if 50% of cancers arise from polyps 75% Sigmoidoscopy and fecal occult 113,550 blood testing 50% Colonoscopy at ages 55 and 65 years Dominant 25% Colonoscopy at ages 55 and 65 years Dominant
$575
$1000
Dominant Dominant Dominant
40,120 20,347 11,915
39,106
Dominant
5674 3329
62,093 26,102
* Current evidence suggests that the majority (75% or more) of cancers arises from polyps. Compliance with screening ranges from 25% in casual programs to 50% to 75% in randomized trials. Colonoscopy is reimbursed by Medicare at $550. † Dollar figures represent the incremental costs per additional life-year saved versus the alternative, less effective strategy. A dominant strategy is one that is both more effective and less costly than the alternative. All costs and life-years are discounted at 3% per annum.
colorectal cancer mortality, slightly higher than the 33% seen in the trial, but within its 95% confidence interval. The additional mortality reduction predicted by the model is the result of the prevention of colorectal cancer through polyp detection and removal.
DISCUSSION Despite evidence that screening can reduce mortality from colorectal cancer, rates of screening have been disappointingly low. Colonoscopy has gained favor in the press and among many experts as a procedure that may optimize screening because of greater effectiveness and perhaps increased compliance (22,65– 68). We compared screening with focused colonoscopy with more traditional screening with combinations of flexible sigmoidoscopy and fecal occult blood testing, and evaluated these strategies at compliance levels that are feasible in clinical practice. Our results confirm earlier studies that have suggested that all of the screening strategies for colorectal cancer fall within the accepted range of cost effectiveness (7,9 – 11,15–17). However, the assertion that colonoscopy would be prohibitively expensive may not be correct. Colonoscopy may be the most cost-effective strategy, particularly at levels of compliance that are seen in clinical practice. Indeed, unless compliance with the noncolonoscopic screening strategies is much higher than generally reported (18,19), colonoscopy, particularly twice-lifetime colonoscopy, is probably the preferred strategy for colorectal cancer screening. Our sensitivity analyses demonstrate that, in addition to compliance, the cost of colonoscopy and the propor-
tion of cancers arising from polyps are the key factors that determine the cost effectiveness of screening for colorectal cancer. Combined flexible sigmoidoscopy and fecal occult blood testing would be the strategy of choice only when either 50% of cancers arise from polyps (which represents a lower bound on the actual rate) (25), compliance is very high (75% or greater), and costs are moderate; or 50% of cancers arise from polyps, compliance is 50%, and the cost of colonoscopy is $1000 or more. Meeting all these conditions seems unlikely. There are, however, limitations to this analysis. In particular, the effectiveness of endoscopic screening has not been demonstrated in randomized controlled trials, and the combination of flexible sigmoidoscopy and fecal occult blood testing, although often recommended, has not been evaluated in any clinical studies. Thus, our estimates of effectiveness are based on the natural history of colorectal cancer, particularly the evolution of adenoma to carcinoma. This rate of progression is based on indirect evidence, and the specific timing remains uncertain. It is possible that some polyps grow and degenerate rapidly into malignancy. If this is common, then infrequent screening procedures, such as twice-lifetime colonoscopy, may not be as effective in preventing cancer mortality. We also limited our perspective to that of a thirdparty payer. As a result, such issues as loss of productivity were not included in the analysis. However, we would expect the comparisons of strategies to be similar, because differences in screening effectiveness are fairly small. We also did not include barium enema as a screening test because of the debate about its effectiveness (13,69,70) and because recent studies have shown that it has a low sensitivity for polyps (71). Finally, it is not clear
December 1, 2001
THE AMERICAN JOURNAL OF MEDICINE威
Volume 111 599
Which Colon Cancer Screening Test?/Vijan et al
whether compliance is greater with colonoscopy, owing to decreased frequency and the use of sedation, or with a less invasive strategy, such as fecal occult blood testing. The value of screening to reduce colorectal cancer mortality is widely accepted. The estimated gains in life expectancy, a maximum of 21 days, compare well with other screening modalities, such as mammography for women aged 50 to 69 years, which results in a gain of about 12 days (72). Unfortunately, screening is still not performed in most people. Thus, we need to focus on developing strategies that optimize compliance or on adopting strategies that maintain a reasonable degree of cost effectiveness regardless of compliance. Our study demonstrates that colonoscopic screening is reasonable from a cost-effectiveness viewpoint. However, because colonoscopic screening maintains its cost effectiveness at lower compliance, it appears to be the preferred strategy under most modeled conditions.
REFERENCES 1. Parker SL, Tong T, Bolden S, Wingo PA. Cancer statistics, 1996. CA Cancer J Clin. 1996;46:5–27. 2. Mandel JS, Bond JH, Church TR, et al. Reducing mortality from colorectal cancer by screening for fecal occult blood. N Engl J Med. 1993;328:1365–1371. 3. Kronborg O, Fenger C, Olsen J, et al. Randomised study of screening for colorectal cancer with faecal-occult blood test. Lancet. 1996; 348:1467–1471. 4. Hardcastle JD, Chamberlain JO, Robinson MHE, et al. Randomised controlled trial of faecal-occult blood screening for colorectal cancer. Lancet. 1996;348:1472–1477. 5. Selby JV, Friedman GD, Quesenberry CP, Weiss NS. A case-control study of screening sigmoidoscopy and mortality from colorectal cancer. N Engl J Med. 1992;326:653–657. 6. Newcomb PA, Norfleet RG, Storer BE, et al. Screening sigmoidoscopy and colorectal cancer mortality. J Natl Cancer Inst. 1992;84: 1572–1575. 7. Wagner JL, Tunis S, Brown M, et al. Cost effectiveness of colorectal cancer screening in average-risk adults. In: Young GP, Rozen P, Levin B, eds. Prevention and Early Detection of Colorectal Cancer. Philadelphia: WB Saunders; 1996:321–341. 8. Lieberman D. Cost effectiveness of colon cancer screening. Am J Gastroenterol. 1991;86:1789 –1794. 9. Lieberman D. Cost effectiveness model for colon cancer screening. Gastroenterology. 1995;109:1781–1790. 10. Frazier AL, Colditz GA, Fuchs CS, Kuntz KM. Cost effectiveness of screening for colorectal cancer in the general population. JAMA. 2000;284:1954 –1961. 11. Sonnenberg A, Delco F, Inadomi JM. Cost effectiveness of colonoscopy in screening for colorectal cancer. Ann Intern Med. 2000;133: 573–584. 12. Solomon MJ, McLeod RS. Periodic health examination, 1994 update: 2. Screening strategies for colorectal cancer. Canadian Task Force on the Periodic Health Examination. Can Med Assoc J. 1994; 150:1961–1970. 13. Winawer SJ, Fletcher RH, Miller L, et al. Colorectal cancer screening: clinical guidelines and rationale. Gastroenterology. 1997; 112:594 –642. 600
December 1, 2001
THE AMERICAN JOURNAL OF MEDICINE威
14. Wagner JL, Herdman RC, Wadhwa S. Cost effectiveness of colorectal cancer screening in the elderly. Ann Intern Med. 1991;115:807– 817. 15. Eddy DM. Screening for colorectal cancer. Ann Intern Med. 1990; 113:373–384. 16. Loeve F, Brown ML, Boer R, et al. Endoscopic colorectal cancer screening: a cost-saving analysis. J Natl Cancer Inst. 2000;92:557– 563. 17. Khandker RK, Dulski JD, Kilpatrick JB, et al. A decision model and cost effectiveness analysis of colorectal cancer screening and surveillance guidelines for average-risk adults. Int J Technol Assess Health Care. 2000;16:799 –810. 18. Anderson LM, May DS. Has the use of cervical, breast, and colorectal cancer screening increased in the United States? Am J Public Health. 1995;85:840 –842. 19. Brown ML, Potosky AL, Thompson GB, Kessler LG. The knowledge and use of screening tests for colorectal and prostate cancer: data from the 1987 National Health Interview Survey. Prev Med. 1990;19:562–574. 20. Elwood JM, Ali G, Schlup MM, et al. Flexible sigmoidoscopy or colonoscopy for colorectal screening: a randomized trial of performance and acceptability. Cancer Detect Prev. 1995;19:337–347. 21. Lieberman D. Endoscopic colon screening: is less more? Gastroenterology. 1996;111:1385–1389. 22. Rogge JD, Elmore MF, Mahoney SJ, et al. Low-cost, office-based, screening colonoscopy. Am J Gastroenterol. 1994;89:1775–1780. 23. Winawer SJ, Zauber AG, Ho MN, et al. Prevention of colorectal cancer by colonoscopic polypectomy. N Engl J Med. 1993;329: 1977–1981. 24. Muto T, Bussey HJR, Morson BC. The evolution of cancer of the colon and rectum. Cancer. 1975;36:2251–2270. 25. Eide TJ. Remnants of adenomas in colorectal carcinomas. Cancer. 1983;51:1866 –1872. 26. Morson BC. The evolution of colorectal carcinoma. Clin Radiol. 1984;35:425–431. 27. Johnson DA, Gurney MS, Volpe RJ, et al. A prospective study of the prevalence of colonic neoplasms in asymptomatic patients with an age-related risk. Am J Gastroenterol. 1990;85:969 –974. 28. DiSario JA, Foutch PG, Mai HD, et al. Prevalence and malignant potential of colorectal polyps in asymptomatic, average-risk men. Am J Gastroenterol. 1991;86:941–945. 29. Rickert RR, Auerback O, Garfinkel L, et al. Adenomatous lesions of the large bowel: an autopsy survey. Cancer. 1979;43:1847–1857. 30. Blatt LJ. Polyps of the colon and rectum: incidence and distribution. Dis Colon Rectum. 1961;4:277–282. 31. Arminski TC, McLean W. Incidence and distribution of adenomatous polyps of the colon and rectum based on 1,000 autopsy examinations. Dis Colon Rectum. 1964;7:249 –261. 32. Ries LAG, Kosary CL, Hankey BF, et al. SEER Cancer Statistics Review—1973–1996. Bethesda: National Cancer Institute; 1999. 33. Myers M, Ries LAG. Cancer patient survival rates: SEER program results for 10 years of follow-up. CA Cancer J Clin. 1989;39:21–32. 34. Ahlquist DA, Wieand HS, Moertell CG, et al. Accuracy of fecal occult blood screening for colorectal neoplasia. JAMA. 1993;269: 1262–1267. 35. Mandel JS, Bond JH, Bradley M, et al. Sensitivity, specificity, and positive predictivity of the hemoccult test in screening for colorectal cancers. Gastroenterology. 1989;97:597–600. 36. Macrae FA, St John DJB. Relationship between patterns of bleeding with hemoccult sensitivity in patients with colorectal cancers or adenomas. Gastroenterology. 1982;82:891–898. 37. Crowley ML, Freeman LD, Mottet MD, et al. Sensitivity of guaiacimpregnated cards for the detection of colorectal neoplasia. J Clin Gastroenterol. 1983;5:127–130.
Volume 111
Which Colon Cancer Screening Test?/Vijan et al 38. Rex DK, Lehman GA, Hawes RH, et al. Screening colonoscopy in asymptomatic average-risk persons with negative fecal occult blood tests. Gastroenterology. 1991;100:64 –67. 39. Allison JE, Feldman R, Tekawa IS. Hemoccult screening in detecting colorectal neoplasm: sensitivity, specificity, and predictive value. Ann Intern Med. 1990;12:328 –333. 40. Niv Y, Sperber AD. Sensitivity, specificity, and predictive value of fecal occult blood testing (hemoccult II) for colorectal neoplasia in symptomatic patients: a prospective study with total colonoscopy. Am J Gastroenterol. 1995;90:1974 –1977. 41. Foutch PG, Mai H, Pardy K, et al. Flexible sigmoidoscopy may be ineffective for secondary prevention of colorectal cancer in asymptomatic, average-risk men. Dig Dis Sci. 1991;36:924 –928. 42. Lieberman DA, Smith FW. Screening for colon malignancy with colonoscopy. Am J Gastroenterol. 1991;86:946 –951. 43. Atkin WS, Morson BC, Cuzick J. Long-term risk of colorectal cancer after excision of rectosigmoid adenomas. N Engl J Med. 1992; 326:658 –662. 44. Rex DK, Cutler CS, Lemmel GT, et al. Colonoscopic miss rates of adenomas determined by back-to-back colonoscopies. Gastroenterology. 1997;112:24 –28. 45. Hixson LJ, Fennerty MB, Sampliner RE, Garewal HS. Prospective blinded trial of the colonoscopic miss-rate of large colorectal polyps. Gastrointest Endosc. 1991;37:125–127. 46. Rex DK. Colonoscopy: a review of its yield for cancers and adenomas by indication. Am J Gastroenterol. 1995;90:353–365. 47. Castiglione G, Mazzotta A, Grazzini G. Sensitivity of screening sigmoidoscopy for proximal colorectal tumours. Lancet. 1995;345: 726 –727. 48. Zarchy TM, Ershoff D. Do characteristics of adenomas on flexible sigmoidoscopy predict advanced lesions on baseline colonoscopy? Gastroenterology. 1994;106:1501–1504. 49. Achkar E, Carey W. Small polyps found during fiberoptic sigmoidoscopy in asymptomatic patients. Ann Intern Med. 1988;109:880 – 883. 50. Habr-Gama A, Waye JD. Complications and hazards of gastrointestinal endoscopy. World J Surg. 1989;13:193–201. 51. Macrae FA, Tan KG, Williams CB. Towards safer colonoscopy: a report on the complications of 5000 diagnostic or therapeutic colonoscopies. Gut. 1983;24:376 –383. 52. Waye JD, Lewis BS, Yessayan S. Colonoscopy: a prospective report of complications. J Clin Gastroenterol. 1992;15:347–351. 53. Eddy DM, Nugent FW, Eddy JF, et al. Screening for colorectal cancer in a high-risk population. Gastroenterology. 1987;92:682–692. 54. Department of Health, and Human Services HCFA. Medicare program; revisions to payment policies and five-year review of and adjustments to the relative value units under the physician fee schedule for calendar year 2000 and physician volume performance standard rates of increase for federal fiscal year 1999, notice. Fed Regist. 1999;64:59379 –59428.
55. Fireman BH, Quesenberry CP, Somkin CP, et al. Cost of care for cancer in a health maintenance organization. Health Care Financing Rev. 1997;18:51–76. 56. Baker MS, Kessler LG, Smucker RC. Analysis of the continuous Medicare history sample file: the cost of treating cancer. In: Proceedings of the Annual Meeting of the American Cancer Society Scientific Session; May 1987, San Diego, California. New York: American Cancer Society; 1987. 57. National Center for Health Statistics. Vital Statistics of the United States, 1991. Mortality. Vol. 2 (part B). Washington, DC: Public Health Service; 1995. 58. Cauffman JG, Hara JH, Rasgon IM, Clark VA. Flexible sigmoidoscopy in asymptomatic patients with negative fecal occult blood tests. J Fam Pract. 1992;34:281–286. 59. Squillace S, Berggreen P, Jaffe P, et al. A normal initial colonoscopy after age 50 does not predict a polyp-free status for life. Am J Gastroenterol. 1994;89:1156 –1159. 60. Rex DK, Cummings OW, Helper DJ, et al. 5-Year incidence of adenomas after negative colonoscopy in asymptomatic averagerisk persons. Gastroenterology. 1996;111:1178 –1181. 61. Rex DK, Lehman GA, Ulbright TM, et al. The yield of a second screening flexible sigmoidoscopy in average-risk persons after one negative examination. Gastroenterology. 1994;106:593–595. 62. Bond JH. Polyp guideline: diagnosis, treatment, and surveillance for patients with nonfamilial colorectal polyps. Ann Intern Med. 1993;119:836 –843. 63. Bond JH. Follow-up after polypectomy: consensus? Eur J Cancer. 1995;31A:1141–1144. 64. Winawer SJ, Zauber AG, O’Brien MJ, et al. Randomized comparison of surveillance intervals after colonoscopic removal of newly diagnosed adenomatous polyps. N Engl J Med. 1993;328:901–905. 65. Bhattacharya I, Sack EM. Screening colonoscopy: the cost of common sense. Lancet. 1996;347:1744 –1745. 66. Neugut AI, Forde KA. Screening colonoscopy: has the time come? Am J Gastroenterol. 1988;83:295–297. 67. Lieberman D. Screening colonoscopy: has the time come? Prim Care. 1995;22:501–511. 68. Ransohoff DF, Lang CA. Using colonoscopy to screen for colorectal cancer. Am J Gastroenterol. 1994;89:1765–1766. 69. Byers T, Levin B, Rothenberger D, et al. American Cancer Society guidelines for screening and surveillance for early detection of colorectal polyps and cancer: update 1997. CA Cancer J Clin. 1997;47: 154 –160. 70. Fletcher RH. The end of barium enemas? N Engl J Med. 2000;342: 1823–1824. 71. Winawer SJ, Stewart ET, Zauber AG, et al. A comparison of colonoscopy and double-contrast barium enema for surveillance after polypectomy. National Polyp Study Work Group. N Engl J Med. 2000;342:1766 –1772. 72. Salzmann P, Kerlikowske K, Phillips K. Cost effectiveness of extending screening mammography guidelines to include women 40 to 49 years of age. Ann Intern Med. 1997;127:955–965.
December 1, 2001
THE AMERICAN JOURNAL OF MEDICINE威
Volume 111 601
Which Colon Cancer Screening Test? A Comparison of Costs, Effectiveness, and Compliance Sandeep Vijan, MD, MS, Erica W. Hwang, MD, Timothy P. Hofer, MD, MS, Rodney A. Hayward, MD PURPOSE: Recent media reports have advocated the use of colonoscopy for colorectal cancer screening. However, colonoscopy is expensive compared with other screening modalities, such as fecal occult blood testing and flexible sigmoidoscopy. We sought to determine the cost effectiveness of different screening strategies for colorectal cancer at levels of compliance likely to be achieved in clinical practice. METHODS: A Markov decision model was used to examine screening strategies, including fecal occult blood testing alone, fecal occult blood testing combined with flexible sigmoidoscopy, flexible sigmoidoscopy alone, and colonoscopy. The timing and frequency of screening was varied to assess optimal screening intervals. Sensitivity analyses were conducted to assess the factors that have the greatest effect on the cost effectiveness of screening. RESULTS: All strategies are cost effective versus no screening, at less than $20,000 per life-year saved. Direct comparison suggests that the most effective strategies are twice-lifetime colonoscopy and flexible sigmoidoscopy combined with fecal
occult blood testing. Assuming perfect compliance, flexible sigmoidoscopy combined with fecal occult blood testing is slightly more effective than twice-lifetime colonoscopy (at ages 50 and 60 years) but is substantially more expensive, with an incremental cost effectiveness of $390,000 per additional life-year saved. However, compliance with primary screening tests and colonoscopic follow-up for polyps affect screening decisions. Colonoscopy at ages 50 and 60 years is the preferred test regardless of compliance with the primary screening test. However, if follow-up colonoscopy for polyps is less than 75%, then even once-lifetime colonoscopy is preferred over most combinations of flexible sigmoidoscopy and fecal occult blood testing. Costs of colonoscopy and proportion of cancer arising from polyps also affect cost effectiveness. CONCLUSIONS: Colonoscopic screening for colorectal cancer appears preferable to current screening recommendations. Screening recommendations should be tailored to the compliance levels achievable in different practice settings. Am J Med. 2001;111:593– 601. 䉷2001 by Excerpta Medica, Inc.
C
In 1994, the Canadian Task Force on the Periodic Health Examination suggested that screening with flexible sigmoidoscopy was the best option for reducing colorectal cancer mortality. The panel concluded that colonoscopic screening was not feasible because of high cost and low compliance, and that fecal occult blood testing was not ideal because of low sensitivity and specificity (12). More recent guidelines from the American Gastroenterological Association recommend that any one of several screening procedures should be used in patients with an average risk of colorectal cancer. These procedures include annual fecal occult blood testing, flexible sigmoidoscopy every fifth year, combined annual fecal occult blood testing with every fifth-year flexible sigmoidoscopy, double-contrast barium enema every 5 to 10 years, or colonoscopy every 10 years (13). Previous decision analyses suggest that strategies that use combinations of sigmoidoscopy and fecal occult blood testing are cost effective (7,9,10,14 –17). More recently, there has been an increased emphasis on the possibility of colonoscopy as a primary screening test.
olorectal cancer is one of the leading causes of cancer death in the United States (1). Because colorectal cancer has a long asymptomatic phase but is often curable when found early (1), screening can substantially reduce colorectal cancer mortality. While experimental evidence of benefit from screening is thus far available only for fecal occult blood testing (2– 4), evidence from observational studies and simulation models suggests that flexible sigmoidoscopy (5–7) and colonoscopy may also reduce colorectal cancer mortality (7–11).
From The Veterans Affairs Center for Practice Management and Outcomes Research (SV, TPH, RAH), Ann Arbor, Michigan; the Department of Internal Medicine (SV, TPH, RAH), University of Michigan, Ann Arbor, Michigan and the Department of Internal Medicine, Georgetown University Medical Center (EWH), Washington, DC Timothy P. Hofer, MD, MS, and Sandeep Vijan, MD, MS, are recipients of the Veterans Affairs Health Services Research and Development Career Development Award, Ann Arbor, Michigan. Requests for reprints should be addressed to Sandeep Vijan, MD, MS, Veterans Health Services Research and Development, PO Box 130170, Ann Arbor, Michigan 48113-0170. Manuscript submitted February 14, 2001, and accepted in revised form August 23, 2001. 䉷2001 by Excerpta Medica, Inc. All rights reserved.
0002-9343/01/$–see front matter 593 PII S0002-9343(01)00977-9
Which Colon Cancer Screening Test?/Vijan et al
Figure 1. Structure of the Markov model. The branch for “Local, year 1” represents the first year of localized colon cancer; “Local, year 2” represents the second year of localized colon cancer (we assumed that colon cancer takes 2 years to progress from localized to regional cancer). The plus sign indicates additional branches that are not displayed; two branches exist beyond each plus sign, one for risk of progression to more advanced stages of premalignant polyps or colon cancer, and one for intervening mortality.
Although there are no trials comparing screening modalities, the recent debate has been clouded because of differing results from cost-effectiveness models (10,11,17). Moreover, these models have differing assumptions, which makes comparisons difficult. For example, the model by Frazier et al (10) used a higher cost of colonoscopy and assumed that polyps were removed during sigmoidoscopy, both of which favor noncolonoscopic screening. Indeed, they concluded that combined sigmoidoscopy and fecal occult blood testing is the test of choice (10). Another analysis did not compare strategies with each other, but rather with no screening, making the choice of the optimal test unclear (17). Sonnenberg et al. (11) did not use a natural history simulation, but assumed a risk reduction from different tests. They also did not test strategies such as combined flexible sigmoidoscopy and fecal occult blood testing (11). Interpretation of these analyses is also limited by the lack of multiway sensitivity analyses to estimate the effects of changing several factors simultaneously on choice of test. Furthermore, analyses of the effects of compliance were limited in these articles. In randomized trials, the highest compliance with screening was 75% (2), when compliance is much lower in actual practice (18,19). A preferred screening strategy for colon cancer should either have high rates of compliance or maintain cost effectiveness regardless of compliance. Colonoscopy may be a desirable screening option because, in addition to requir594
December 1, 2001
THE AMERICAN JOURNAL OF MEDICINE威
ing only a single test for screening and intervention, persons may prefer to have procedures once or twice in a lifetime to more frequent options (20,21). In addition, the possibility of performing office-based colonoscopy as a low-cost procedure makes colonoscopic screening intriguing (22). Thus, it is imperative to define how compliance affects cost effectiveness, the optimal timing and frequency of colonoscopy, and the effects of pricing on the relative cost effectiveness of different screening procedures. To help answer these questions, we constructed a Markov cost-effectiveness model to examine screening for colorectal cancer. Our model tests variations in the timing and frequency of tests and also estimates the effects of compliance, costs, test characteristics, and the natural history of colorectal cancer. We selected oncelifetime and twice-lifetime colonoscopy as strategies that might optimize compliance and compared their costs and effectiveness with those of fecal occult blood testing, flexible sigmoidoscopy, and sigmoidoscopy combined with fecal occult blood testing.
MATERIAL AND METHODS We created a Markov model to simulate the natural history of colorectal cancer (Figure 1), beginning at age 50
Volume 111
Which Colon Cancer Screening Test?/Vijan et al
Table 1. Model Assumptions Base Case
Range Used in Sensitivity Analysis
75%
50%–100%
20% 40% 50% 55%
10%–30% 30%–50% 40%–60% 45%–65%
15% 35%
5%–25% 25%–45%
0.05% 0.09% 0.14% 0.20% 0.27% 0.35% 0.43% 0.45%
— — — — — — — —
10.5% 35.1% 91.7%
— — —
5% 97.5%
2%–10% 90%–100%
30% 50% 55% 85% 95% 0.1% 7.5%
20%–40% 40%–60% 40%–75% 80%–95% 90%–100% 0.0%–0.3% 5%–10%
References
Natural history Proportion of cancers arising from polyps Prevalence of adenomatous polyps Age 50 years Age 60 years Age 70 years Age 80 years In patients with polyps Proportion of polyps ⬎1 cm Proportion with multiple polyps Annual incidence of colorectal cancer Age 50 years Age 55 years Age 60 years Age 65 years Age 70 years Age 75 years Age 80 years Age 85 years 5-Year colorectal cancer mortality Localized Regional Disseminated Test characteristics Sensitivity of fecal occult blood testing for polyps Specificity of fecal occult blood testing Sensitivity of fecal occult blood testing for cancer Localized Regional Polyps or cancer reachable by flexible sigmoidoscopy Sensitivity of colonoscopy or flexible sigmoidoscopy for polyps Sensitivity of colonoscopy or flexible sigmoidoscopy for cancer Perforation rate, colonoscopy Mortality rate, perforation Costs Fecal occult blood testing Flexible sigmoidoscopy Flexible sigmoidoscopy with biopsy Colonoscopy Polypectomy (including pathology) Cancer care Localized Regional Disseminated Cost of treating colon perforation
$17 $225 $240 $550 $215
$5–$30 $100–$500 $100–$550 $150–1000 $100–300
$60,000 $82,800 $73,000 $20,000
$40,000–$80,000 $60,000–$100,000 $50,000–$90,000 $10,000–$30,000
years. The key assumptions of the model are outlined in Table 1. The main sources of the estimates were colonoscopic screening studies and autopsy studies for the prevalence of polyps (27–31) and the Surveillance, Epidemiology, and End Results (SEER) registry data for the incidence and mortality rates of colorectal cancer (1,32). Mortality rates for the general population were derived from the National Center for Health Statistics publications (57).
We calculated the age-specific incidence of adenomas using the age-specific prevalence of polyps observed in screening and autopsy studies. The incidence rate was calculated so that at each 5-year interval, the prevalence of polyps predicted by the model matched that in published studies (Table 1). The incidence of hyperplastic polyps was defined in the same manner; prevalence ranged from 20% at age 50 years to 15% at age 80 years (not shown in Table 1) (27,28).
(23–26) (27–31)
(27–31) (27–31) (32)
(33)
December 1, 2001
THE AMERICAN JOURNAL OF MEDICINE威
(34–40) (34–36,39,40) (34,36–40)
(15,41–43) (41,44–49) (41,44–49) (50–52) (15,53) (54) (54) (54) (22,54) (54) (7,15,55,56)
(53)
Volume 111 595
Which Colon Cancer Screening Test?/Vijan et al
Figure 2. Observed cancer incidence versus predicted cancer incidence (model output). SEER ⫽ Surveillance, Epidemiology, and End Results [registry data].
We also calculated the rate of conversion of polyps to cancer to match cancer incidence rates. The proportion of incident cancers arising from polyps was assumed to be 75% in the base-case scenario and was varied from 50% to 100% in the sensitivity analyses (23–26). We assumed that it takes 10 years for a polyp to transform from benign to malignant. This 10-year polyp “dwell time” is supported by consensus opinion (13) but is based on indirect evidence, including studies showing that screening sigmoidoscopy provides a 10-year window of protection from colorectal cancer (5). To calculate the conversion rate from polyps to cancer, the prevalence of polyps at time “x” was matched with the incidence at time “x ⫹ 10,” and conversion rates were calculated, including intervening mortality. Thus, a proportion of polyps at each time stage was determined to be progressive, and entered the dwell state from which they gradually became malignant. The model predictions of incidence were nearly identical to the observed incidence in the SEER registry (Figure 2). If a colorectal malignancy developed, we assumed that it took 2 years to progress through localized cancer and an additional year to progress through regional cancer. Patients who developed disseminated cancer were diagnosed within 1 year, whether or not screening was employed. Mortality was varied by stage of colorectal cancer based on published results from the SEER registry (1,32).
Test Characteristics To simulate the effects of screening, strategies were interposed at different intervals. Screening affected mortality in two ways: prevention of cancer through removal of adenomatous polyps and detection of cancer in earlier 596
December 1, 2001
THE AMERICAN JOURNAL OF MEDICINE威
stages (Table 1). The sensitivity, specificity, and complication rates of the screening tests were derived from observational studies (34 – 41,44 – 49,58). For flexible sigmoidoscopy, we assumed that a 70-cm sigmoidoscope could detect 55% of polyps and cancers (eg, flexible sigmoidoscopy “reached” 55% of all lesions); this was varied from 40% to 75% in sensitivity analyses (15,41– 43). Prior compliance with a test did not predict future compliance. However, in the base case, compliance was consistent across tests. For example, in the flexible sigmoidoscopy combined with fecal occult blood testing strategy, if the compliance for fecal occult blood testing was 75%, it was also 75% for flexible sigmoidoscopy, which was performed only if fecal occult blood testing was negative. The effect of compliance with follow-up colonoscopy for positive initial screening tests (eg, positive fecal occult blood testing or flexible sigmoidoscopy) on the noncolonoscopic screening strategies was also evaluated. We assumed that a negative screening test did not reduce the risk of subsequent polyp growth (59 – 61). If an adenomatous polyp was detected, full colonoscopy was performed and the polyp was removed. If the polyp was greater than 1 cm or if there were multiple polyps, a follow-up colonoscopy was performed at 3 years. If this first follow-up was negative, further surveillance colonoscopy was performed at 5-year intervals (62– 64). No routine surveillance was performed for patients who had hyperplastic polyps or polyps less than 1 cm removed; these patients returned to usual screening. Adenomatous polyp recurrence rates at surveillance colonoscopy were taken from the National Polyp Study (64). After a negative follow-up colonoscopy, the rate of subsequent polyp devel-
Volume 111
Which Colon Cancer Screening Test?/Vijan et al
opment was assumed to be the same as that in the general population. Compliance with surveillance testing was assumed to be 100% in the base-case analysis; this was varied from 50% to 100% in sensitivity analyses.
Costs Costs were approached from the perspective of a thirdparty payer; thus, only direct medical costs were included. The costs of screening tests and interventions were taken from the 2000 Medicare reimbursement schedule (54). Because Medicare does not currently reimburse fecal occult blood testing, we based costs on past levels of reimbursement (0.5 relative value units). We included physician fees and facility expenses in procedural costs (colonoscopy, flexible sigmoidoscopy). Polypectomy costs included reimbursement for tissue pathology. Costs of caring for procedural complications (7,15,53) were inflated to 1999 US dollars using the medical care consumer price index. Costs of cancer care were taken from a 1987 Medicare claims analysis (56) and a 1997 Kaiser Permanente analysis (55), and were also inflated to 1999 dollars. All costs and years of life were discounted at 3% in the base case. The discount rate was varied between 0% and 6% in sensitivity analyses.
Sensitivity Analyses Sensitivity analyses were conducted to determine the effects of different variables on the cost effectiveness of the screening strategies. We tested each variable in the model in one-way sensitivity analyses (Table 1). We also conducted three-way sensitivity analyses on the assumptions that had the greatest effects on cost effectiveness in the one-way analyses.
RESULTS Using the base-case estimates, the model predicted that the lifetime risk of colorectal cancer is 5869 per 100,000 (5.9%). Model predictions closely matched the expected incidence from the SEER registry data (Figure 2). The average direct cost of caring for colorectal cancer in an unscreened population is $1,303 per person.
Cost Effectiveness and Incremental Cost Effectiveness The cost-effectiveness ratio of all screening strategies versus no screening is under $20,000 per life-year gained, across all levels of compliance. Once-lifetime colonoscopy at age 60 years is almost cost neutral (cost-effectiveness ratio of $130 per life-year gained), whereas once-lifetime colonoscopy at age 65 years is cost saving (savings of $2280 per person). Differences in life expectancy are small among strategies, so cost is the factor of primary importance (Table 2). Once-lifetime or twicelifetime colonoscopic screenings are less expensive than
combined flexible sigmoidoscopy and fecal occult blood testing. Regardless of compliance level, flexible sigmoidoscopy alone and fecal occult blood testing alone are dominated by colonoscopic screening; that is, colonoscopy is both more effective and less costly (Table 2). At 100% compliance, combined sigmoidoscopy and fecal occult blood testing is more effective than twice-lifetime colonoscopy but costs substantially more, with an incremental costeffectiveness ratio of more than $106,000 per life-year gained. However, as compliance decreases, the noncolonoscopic strategies become less effective and more costly. For example, at 75% compliance, the incremental cost effectiveness of combined sigmoidoscopy and fecal occult blood testing, compared with twice-lifetime colonoscopy, increases to over $330,000 per life-year gained. With further decreases in rates of compliance, combined sigmoidoscopy and fecal occult blood testing is dominated by the colonoscopic strategies. Because compliance with follow-up testing has been poorly described outside of randomized studies, we also evaluated a scenario in which patients had a constant 80% rate of compliance with follow-up colonoscopy. Using this assumption, the sigmoidoscopy alone and fecal occult blood testing alone strategies are dominated by the colonoscopic strategies. However, changing compliance with follow-up does attenuate some of the differences between combined sigmoidoscopy and fecal occult blood testing and the twice-lifetime colonoscopic strategies. For example, at a compliance level of 75%, the incremental cost effectiveness of combined sigmoidoscopy and fecal occult blood testing versus colonoscopy at ages 50 and 60 years is $116,900 (compared with over $330,000 in the base case).
Sensitivity Analyses The discount rate has the largest effect on costs and benefits. However, because the discount rate affects both costs and effectiveness, these differences are minimized when cost-effectiveness ratios are calculated. The major factors that affect cost-effectiveness ratios, in addition to compliance, are the cost of colonoscopy and the proportion of cancer arising from polyps. These three parameters were explored in multivariate sensitivity analyses. Variation in other factors (across the ranges defined in Table 1) and different methods of simulating compliance have small effects on the choice of strategies. We also tested the effect of including all polyps, regardless of size, in the surveillance program. This increases costs for all screening choices. The incremental cost effectiveness of colonoscopic screening is most affected (because of higher polyp sensitivity), but the differences in incremental cost effectiveness are less than $5000 per life-year from the base case.
December 1, 2001
THE AMERICAN JOURNAL OF MEDICINE威
Volume 111 597
Which Colon Cancer Screening Test?/Vijan et al
Table 2. Costs and Effectiveness of Screening
Strategy
Average Gain in Life Expectancy (days)*
Average Cost ($)*
Relative Reduction in Colorectal Cancer Mortality (%)
Incremental Cost per Added Life-year ($)†
— 100%
No screening Flexible sigmoidoscopy Colonoscopy at age 60 years Colonoscopy at age 55 years Fecal occult blood test Colonoscopy at ages 55 and 65 years Colonoscopy at ages 50 and 60 years Flexible sigmoidoscopy and fecal occult blood test
— 12.5 12.6 14.6 16.5 17.6 18.8 20.7
1300 1930 1310 1420 1550 1540 1750 2280
— 44.0 52.7 46.3 52.4 68.0 63.6 69.6
— D 130 20,770 D 14,870 62,140 106,860
75%
Fecal occult blood test Colonoscopy at age 60 years Flexible sigmoidoscopy Colonoscopy at age 55 years Colonoscopy at ages 55 and 65 years Colonoscopy at ages 50 and 60 years Flexible sigmoidoscopy and fecal occult blood test Fecal occult blood test Colonoscopy at age 60 years Flexible sigmoidoscopy Colonoscopy at age 55 years Flexible sigmoidoscopy and fecal occult blood test Colonoscopy at ages 55 and 65 years Colonoscopy at ages 50 and 60 years
8.0 9.5 10.7 11.0 14.3 15.6 15.8 5.8 6.3 7.1 7.3 8.5 10.2 11.4
1470 1310 1730 1390 1450 1600 1840 1420 1310 1590 1360 1570 1380 1480
43.4 39.5 41.0 34.7 56.3 52.7 61.0 32.1 26.3 35.0 23.1 45.6 41.1 38.4
D 130 D 20,940 6,500 42,670 332,630 D 130 D 21,150 D 2,130 32,990
Fecal occult blood test Flexible sigmoidoscopy and fecal occult blood test Flexible sigmoidoscopy Colonoscopy at age 60 years Colonoscopy at age 55 years Colonoscopy at ages 55 and 65 years Colonoscopy at ages 50 and 60 years
1.6 2.2 2.8 3.1 3.7 5.4 6.1
1370 1410 1470 1300 1330 1330 1380
17.9 23.1 23.2 13.2 11.6 22.3 20.9
D D D 130 D 4,140 26,880
Compliance
50%
25%
* Costs and life expectancy discounted at 3% per annum; values rounded to nearest $10 and 0.1 days. † Incremental cost effectiveness versus prior strategy. “D” represents a strategy that is dominated by one of the ensuing strategies, meaning that the ensuing strategies are both less costly and more effective. As a result, dominated strategies are not included in the incremental cost-effectiveness calculations.
Two-way sensitivity analyses varying the cost of colonoscopy and compliance level demonstrate that at the lower bounds of cost ($150 per colonoscopy), twicelifetime colonoscopy is favored at all levels of compliance and is cost saving in many circumstances. Low-cost office-based screening colonoscopy has been performed for a total charge of $150 (22), although reimbursement for colonoscopy is substantially higher than this in most circumstances. To compare incremental cost effectiveness, we conducted three-way analyses varying costs of colonoscopy, percentage of cancers arising from polyps, and compliance (Table 3). If only 50% of cancer arise from polyps, then compliance and cost of colonoscopy have pronounced effects on the choice of therapy. For example, if compliance is 75% or greater, then combined flexible sigmoidoscopy and fecal occult blood testing is a dominant 598
December 1, 2001
THE AMERICAN JOURNAL OF MEDICINE威
strategy if the cost of colonoscopy is $1000; this strategy costs about $39,000 per life-year gained if the cost is $575 and about $113,000 per life-year gained if the cost is $150. At more likely levels of compliance, or if more than 50% of cancers arise from polyps, colonoscopy is the screening strategy of choice unless its costs are very high (Table 3).
Model Validity We tested the model validity by altering the assumptions to fit the Minnesota fecal occult blood testing screening trial (2). We changed the compliance with fecal occult blood testing to 75% and the compliance of follow-up colonoscopy to 81%. We also adjusted the sensitivity and specificity of fecal occult blood testing, because fecal occult blood testing slides were rehydrated in that study. As in the trial, screening was limited to 10 years, with a 13year follow-up. Our model predicts a 39% reduction in
Volume 111
Which Colon Cancer Screening Test?/Vijan et al
Table 3. Three-way Sensitivity Analyses Varying Percentage of Cancers Arising from Polyps, Compliance, and Cost of Colonoscopy*: Comparison of Combined Flexible Sigmoidoscopy and Fecal Occult Blood Testing with Colonoscopy at Ages 55 and 65 Years Cost of Colonoscopy ($)† Compliance
Best Strategy
$150
Incremental cost effectiveness if 100% of cancers arise from polyps 75% Colonoscopy at ages 55 and 65 years Dominant 50% Colonoscopy at ages 55 and 65 years Dominant 25% Colonoscopy at ages 55 and 65 years Dominant Incremental cost effectiveness if 50% of cancers arise from polyps 75% Sigmoidoscopy and fecal occult 113,550 blood testing 50% Colonoscopy at ages 55 and 65 years Dominant 25% Colonoscopy at ages 55 and 65 years Dominant
$575
$1000
Dominant Dominant Dominant
40,120 20,347 11,915
39,106
Dominant
5674 3329
62,093 26,102
* Current evidence suggests that the majority (75% or more) of cancers arises from polyps. Compliance with screening ranges from 25% in casual programs to 50% to 75% in randomized trials. Colonoscopy is reimbursed by Medicare at $550. † Dollar figures represent the incremental costs per additional life-year saved versus the alternative, less effective strategy. A dominant strategy is one that is both more effective and less costly than the alternative. All costs and life-years are discounted at 3% per annum.
colorectal cancer mortality, slightly higher than the 33% seen in the trial, but within its 95% confidence interval. The additional mortality reduction predicted by the model is the result of the prevention of colorectal cancer through polyp detection and removal.
DISCUSSION Despite evidence that screening can reduce mortality from colorectal cancer, rates of screening have been disappointingly low. Colonoscopy has gained favor in the press and among many experts as a procedure that may optimize screening because of greater effectiveness and perhaps increased compliance (22,65– 68). We compared screening with focused colonoscopy with more traditional screening with combinations of flexible sigmoidoscopy and fecal occult blood testing, and evaluated these strategies at compliance levels that are feasible in clinical practice. Our results confirm earlier studies that have suggested that all of the screening strategies for colorectal cancer fall within the accepted range of cost effectiveness (7,9 – 11,15–17). However, the assertion that colonoscopy would be prohibitively expensive may not be correct. Colonoscopy may be the most cost-effective strategy, particularly at levels of compliance that are seen in clinical practice. Indeed, unless compliance with the noncolonoscopic screening strategies is much higher than generally reported (18,19), colonoscopy, particularly twice-lifetime colonoscopy, is probably the preferred strategy for colorectal cancer screening. Our sensitivity analyses demonstrate that, in addition to compliance, the cost of colonoscopy and the propor-
tion of cancers arising from polyps are the key factors that determine the cost effectiveness of screening for colorectal cancer. Combined flexible sigmoidoscopy and fecal occult blood testing would be the strategy of choice only when either 50% of cancers arise from polyps (which represents a lower bound on the actual rate) (25), compliance is very high (75% or greater), and costs are moderate; or 50% of cancers arise from polyps, compliance is 50%, and the cost of colonoscopy is $1000 or more. Meeting all these conditions seems unlikely. There are, however, limitations to this analysis. In particular, the effectiveness of endoscopic screening has not been demonstrated in randomized controlled trials, and the combination of flexible sigmoidoscopy and fecal occult blood testing, although often recommended, has not been evaluated in any clinical studies. Thus, our estimates of effectiveness are based on the natural history of colorectal cancer, particularly the evolution of adenoma to carcinoma. This rate of progression is based on indirect evidence, and the specific timing remains uncertain. It is possible that some polyps grow and degenerate rapidly into malignancy. If this is common, then infrequent screening procedures, such as twice-lifetime colonoscopy, may not be as effective in preventing cancer mortality. We also limited our perspective to that of a thirdparty payer. As a result, such issues as loss of productivity were not included in the analysis. However, we would expect the comparisons of strategies to be similar, because differences in screening effectiveness are fairly small. We also did not include barium enema as a screening test because of the debate about its effectiveness (13,69,70) and because recent studies have shown that it has a low sensitivity for polyps (71). Finally, it is not clear
December 1, 2001
THE AMERICAN JOURNAL OF MEDICINE威
Volume 111 599
Which Colon Cancer Screening Test?/Vijan et al
whether compliance is greater with colonoscopy, owing to decreased frequency and the use of sedation, or with a less invasive strategy, such as fecal occult blood testing. The value of screening to reduce colorectal cancer mortality is widely accepted. The estimated gains in life expectancy, a maximum of 21 days, compare well with other screening modalities, such as mammography for women aged 50 to 69 years, which results in a gain of about 12 days (72). Unfortunately, screening is still not performed in most people. Thus, we need to focus on developing strategies that optimize compliance or on adopting strategies that maintain a reasonable degree of cost effectiveness regardless of compliance. Our study demonstrates that colonoscopic screening is reasonable from a cost-effectiveness viewpoint. However, because colonoscopic screening maintains its cost effectiveness at lower compliance, it appears to be the preferred strategy under most modeled conditions.
REFERENCES 1. Parker SL, Tong T, Bolden S, Wingo PA. Cancer statistics, 1996. CA Cancer J Clin. 1996;46:5–27. 2. Mandel JS, Bond JH, Church TR, et al. Reducing mortality from colorectal cancer by screening for fecal occult blood. N Engl J Med. 1993;328:1365–1371. 3. Kronborg O, Fenger C, Olsen J, et al. Randomised study of screening for colorectal cancer with faecal-occult blood test. Lancet. 1996; 348:1467–1471. 4. Hardcastle JD, Chamberlain JO, Robinson MHE, et al. Randomised controlled trial of faecal-occult blood screening for colorectal cancer. Lancet. 1996;348:1472–1477. 5. Selby JV, Friedman GD, Quesenberry CP, Weiss NS. A case-control study of screening sigmoidoscopy and mortality from colorectal cancer. N Engl J Med. 1992;326:653–657. 6. Newcomb PA, Norfleet RG, Storer BE, et al. Screening sigmoidoscopy and colorectal cancer mortality. J Natl Cancer Inst. 1992;84: 1572–1575. 7. Wagner JL, Tunis S, Brown M, et al. Cost effectiveness of colorectal cancer screening in average-risk adults. In: Young GP, Rozen P, Levin B, eds. Prevention and Early Detection of Colorectal Cancer. Philadelphia: WB Saunders; 1996:321–341. 8. Lieberman D. Cost effectiveness of colon cancer screening. Am J Gastroenterol. 1991;86:1789 –1794. 9. Lieberman D. Cost effectiveness model for colon cancer screening. Gastroenterology. 1995;109:1781–1790. 10. Frazier AL, Colditz GA, Fuchs CS, Kuntz KM. Cost effectiveness of screening for colorectal cancer in the general population. JAMA. 2000;284:1954 –1961. 11. Sonnenberg A, Delco F, Inadomi JM. Cost effectiveness of colonoscopy in screening for colorectal cancer. Ann Intern Med. 2000;133: 573–584. 12. Solomon MJ, McLeod RS. Periodic health examination, 1994 update: 2. Screening strategies for colorectal cancer. Canadian Task Force on the Periodic Health Examination. Can Med Assoc J. 1994; 150:1961–1970. 13. Winawer SJ, Fletcher RH, Miller L, et al. Colorectal cancer screening: clinical guidelines and rationale. Gastroenterology. 1997; 112:594 –642. 600
December 1, 2001
THE AMERICAN JOURNAL OF MEDICINE威
14. Wagner JL, Herdman RC, Wadhwa S. Cost effectiveness of colorectal cancer screening in the elderly. Ann Intern Med. 1991;115:807– 817. 15. Eddy DM. Screening for colorectal cancer. Ann Intern Med. 1990; 113:373–384. 16. Loeve F, Brown ML, Boer R, et al. Endoscopic colorectal cancer screening: a cost-saving analysis. J Natl Cancer Inst. 2000;92:557– 563. 17. Khandker RK, Dulski JD, Kilpatrick JB, et al. A decision model and cost effectiveness analysis of colorectal cancer screening and surveillance guidelines for average-risk adults. Int J Technol Assess Health Care. 2000;16:799 –810. 18. Anderson LM, May DS. Has the use of cervical, breast, and colorectal cancer screening increased in the United States? Am J Public Health. 1995;85:840 –842. 19. Brown ML, Potosky AL, Thompson GB, Kessler LG. The knowledge and use of screening tests for colorectal and prostate cancer: data from the 1987 National Health Interview Survey. Prev Med. 1990;19:562–574. 20. Elwood JM, Ali G, Schlup MM, et al. Flexible sigmoidoscopy or colonoscopy for colorectal screening: a randomized trial of performance and acceptability. Cancer Detect Prev. 1995;19:337–347. 21. Lieberman D. Endoscopic colon screening: is less more? Gastroenterology. 1996;111:1385–1389. 22. Rogge JD, Elmore MF, Mahoney SJ, et al. Low-cost, office-based, screening colonoscopy. Am J Gastroenterol. 1994;89:1775–1780. 23. Winawer SJ, Zauber AG, Ho MN, et al. Prevention of colorectal cancer by colonoscopic polypectomy. N Engl J Med. 1993;329: 1977–1981. 24. Muto T, Bussey HJR, Morson BC. The evolution of cancer of the colon and rectum. Cancer. 1975;36:2251–2270. 25. Eide TJ. Remnants of adenomas in colorectal carcinomas. Cancer. 1983;51:1866 –1872. 26. Morson BC. The evolution of colorectal carcinoma. Clin Radiol. 1984;35:425–431. 27. Johnson DA, Gurney MS, Volpe RJ, et al. A prospective study of the prevalence of colonic neoplasms in asymptomatic patients with an age-related risk. Am J Gastroenterol. 1990;85:969 –974. 28. DiSario JA, Foutch PG, Mai HD, et al. Prevalence and malignant potential of colorectal polyps in asymptomatic, average-risk men. Am J Gastroenterol. 1991;86:941–945. 29. Rickert RR, Auerback O, Garfinkel L, et al. Adenomatous lesions of the large bowel: an autopsy survey. Cancer. 1979;43:1847–1857. 30. Blatt LJ. Polyps of the colon and rectum: incidence and distribution. Dis Colon Rectum. 1961;4:277–282. 31. Arminski TC, McLean W. Incidence and distribution of adenomatous polyps of the colon and rectum based on 1,000 autopsy examinations. Dis Colon Rectum. 1964;7:249 –261. 32. Ries LAG, Kosary CL, Hankey BF, et al. SEER Cancer Statistics Review—1973–1996. Bethesda: National Cancer Institute; 1999. 33. Myers M, Ries LAG. Cancer patient survival rates: SEER program results for 10 years of follow-up. CA Cancer J Clin. 1989;39:21–32. 34. Ahlquist DA, Wieand HS, Moertell CG, et al. Accuracy of fecal occult blood screening for colorectal neoplasia. JAMA. 1993;269: 1262–1267. 35. Mandel JS, Bond JH, Bradley M, et al. Sensitivity, specificity, and positive predictivity of the hemoccult test in screening for colorectal cancers. Gastroenterology. 1989;97:597–600. 36. Macrae FA, St John DJB. Relationship between patterns of bleeding with hemoccult sensitivity in patients with colorectal cancers or adenomas. Gastroenterology. 1982;82:891–898. 37. Crowley ML, Freeman LD, Mottet MD, et al. Sensitivity of guaiacimpregnated cards for the detection of colorectal neoplasia. J Clin Gastroenterol. 1983;5:127–130.
Volume 111
Which Colon Cancer Screening Test?/Vijan et al 38. Rex DK, Lehman GA, Hawes RH, et al. Screening colonoscopy in asymptomatic average-risk persons with negative fecal occult blood tests. Gastroenterology. 1991;100:64 –67. 39. Allison JE, Feldman R, Tekawa IS. Hemoccult screening in detecting colorectal neoplasm: sensitivity, specificity, and predictive value. Ann Intern Med. 1990;12:328 –333. 40. Niv Y, Sperber AD. Sensitivity, specificity, and predictive value of fecal occult blood testing (hemoccult II) for colorectal neoplasia in symptomatic patients: a prospective study with total colonoscopy. Am J Gastroenterol. 1995;90:1974 –1977. 41. Foutch PG, Mai H, Pardy K, et al. Flexible sigmoidoscopy may be ineffective for secondary prevention of colorectal cancer in asymptomatic, average-risk men. Dig Dis Sci. 1991;36:924 –928. 42. Lieberman DA, Smith FW. Screening for colon malignancy with colonoscopy. Am J Gastroenterol. 1991;86:946 –951. 43. Atkin WS, Morson BC, Cuzick J. Long-term risk of colorectal cancer after excision of rectosigmoid adenomas. N Engl J Med. 1992; 326:658 –662. 44. Rex DK, Cutler CS, Lemmel GT, et al. Colonoscopic miss rates of adenomas determined by back-to-back colonoscopies. Gastroenterology. 1997;112:24 –28. 45. Hixson LJ, Fennerty MB, Sampliner RE, Garewal HS. Prospective blinded trial of the colonoscopic miss-rate of large colorectal polyps. Gastrointest Endosc. 1991;37:125–127. 46. Rex DK. Colonoscopy: a review of its yield for cancers and adenomas by indication. Am J Gastroenterol. 1995;90:353–365. 47. Castiglione G, Mazzotta A, Grazzini G. Sensitivity of screening sigmoidoscopy for proximal colorectal tumours. Lancet. 1995;345: 726 –727. 48. Zarchy TM, Ershoff D. Do characteristics of adenomas on flexible sigmoidoscopy predict advanced lesions on baseline colonoscopy? Gastroenterology. 1994;106:1501–1504. 49. Achkar E, Carey W. Small polyps found during fiberoptic sigmoidoscopy in asymptomatic patients. Ann Intern Med. 1988;109:880 – 883. 50. Habr-Gama A, Waye JD. Complications and hazards of gastrointestinal endoscopy. World J Surg. 1989;13:193–201. 51. Macrae FA, Tan KG, Williams CB. Towards safer colonoscopy: a report on the complications of 5000 diagnostic or therapeutic colonoscopies. Gut. 1983;24:376 –383. 52. Waye JD, Lewis BS, Yessayan S. Colonoscopy: a prospective report of complications. J Clin Gastroenterol. 1992;15:347–351. 53. Eddy DM, Nugent FW, Eddy JF, et al. Screening for colorectal cancer in a high-risk population. Gastroenterology. 1987;92:682–692. 54. Department of Health, and Human Services HCFA. Medicare program; revisions to payment policies and five-year review of and adjustments to the relative value units under the physician fee schedule for calendar year 2000 and physician volume performance standard rates of increase for federal fiscal year 1999, notice. Fed Regist. 1999;64:59379 –59428.
55. Fireman BH, Quesenberry CP, Somkin CP, et al. Cost of care for cancer in a health maintenance organization. Health Care Financing Rev. 1997;18:51–76. 56. Baker MS, Kessler LG, Smucker RC. Analysis of the continuous Medicare history sample file: the cost of treating cancer. In: Proceedings of the Annual Meeting of the American Cancer Society Scientific Session; May 1987, San Diego, California. New York: American Cancer Society; 1987. 57. National Center for Health Statistics. Vital Statistics of the United States, 1991. Mortality. Vol. 2 (part B). Washington, DC: Public Health Service; 1995. 58. Cauffman JG, Hara JH, Rasgon IM, Clark VA. Flexible sigmoidoscopy in asymptomatic patients with negative fecal occult blood tests. J Fam Pract. 1992;34:281–286. 59. Squillace S, Berggreen P, Jaffe P, et al. A normal initial colonoscopy after age 50 does not predict a polyp-free status for life. Am J Gastroenterol. 1994;89:1156 –1159. 60. Rex DK, Cummings OW, Helper DJ, et al. 5-Year incidence of adenomas after negative colonoscopy in asymptomatic averagerisk persons. Gastroenterology. 1996;111:1178 –1181. 61. Rex DK, Lehman GA, Ulbright TM, et al. The yield of a second screening flexible sigmoidoscopy in average-risk persons after one negative examination. Gastroenterology. 1994;106:593–595. 62. Bond JH. Polyp guideline: diagnosis, treatment, and surveillance for patients with nonfamilial colorectal polyps. Ann Intern Med. 1993;119:836 –843. 63. Bond JH. Follow-up after polypectomy: consensus? Eur J Cancer. 1995;31A:1141–1144. 64. Winawer SJ, Zauber AG, O’Brien MJ, et al. Randomized comparison of surveillance intervals after colonoscopic removal of newly diagnosed adenomatous polyps. N Engl J Med. 1993;328:901–905. 65. Bhattacharya I, Sack EM. Screening colonoscopy: the cost of common sense. Lancet. 1996;347:1744 –1745. 66. Neugut AI, Forde KA. Screening colonoscopy: has the time come? Am J Gastroenterol. 1988;83:295–297. 67. Lieberman D. Screening colonoscopy: has the time come? Prim Care. 1995;22:501–511. 68. Ransohoff DF, Lang CA. Using colonoscopy to screen for colorectal cancer. Am J Gastroenterol. 1994;89:1765–1766. 69. Byers T, Levin B, Rothenberger D, et al. American Cancer Society guidelines for screening and surveillance for early detection of colorectal polyps and cancer: update 1997. CA Cancer J Clin. 1997;47: 154 –160. 70. Fletcher RH. The end of barium enemas? N Engl J Med. 2000;342: 1823–1824. 71. Winawer SJ, Stewart ET, Zauber AG, et al. A comparison of colonoscopy and double-contrast barium enema for surveillance after polypectomy. National Polyp Study Work Group. N Engl J Med. 2000;342:1766 –1772. 72. Salzmann P, Kerlikowske K, Phillips K. Cost effectiveness of extending screening mammography guidelines to include women 40 to 49 years of age. Ann Intern Med. 1997;127:955–965.
December 1, 2001
THE AMERICAN JOURNAL OF MEDICINE威
Volume 111 601