Cover Image

PAPERBACK
$63.25



View/Hide Left Panel

Appendix E
Description of the Laudabaum Colorectal Cancer Screening Model Uri Ladabaum, M.D., M.S.

SLIDE 1

SLIDE 1 NOTES: This summary describes key elements of the current version of a model used in a cost-effectiveness analysis published by my colleagues and me (Song et al., 2004)



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 61
Economic Models of Colorectal Cancer Screening in Average-Risk Adults: Workshop Summary Appendix E Description of the Laudabaum Colorectal Cancer Screening Model Uri Ladabaum, M.D., M.S. SLIDE 1 SLIDE 1 NOTES: This summary describes key elements of the current version of a model used in a cost-effectiveness analysis published by my colleagues and me (Song et al., 2004)

OCR for page 61
Economic Models of Colorectal Cancer Screening in Average-Risk Adults: Workshop Summary SLIDE 2 SLIDE 2 NOTES: Our model is a cohort model. The main outputs are average years of life lived and accrued costs per person. It can be an converted to an aggregate annual model by combining the estimates for every age group in each year and projecting to the national population. Year-2000 U.S. Census data can be combined with age-specific model outputs to make predictions for the U.S. population (aggregate annual model), as opposed to a hypothetical cohort of a given size starting at age 50 years. The model is a Markov model. For the purposes of the pre-workshop modeling exercise, it followed people from age 50 to 85 years of age, or until death if that came before age 85. The model can incorporate stopping ages up to age 100. It is also possible to treat each sex separately, though most of our work has been with average values for the entire population. The current version runs on DataPro™, a commercial software package.

OCR for page 61
Economic Models of Colorectal Cancer Screening in Average-Risk Adults: Workshop Summary SLIDE 3 SLIDE 3 NOTES: This is a schematic of the natural history model. At the left, we start with people in the normal state. They can progress to a small polyp (less than 10 mm), which is our nomenclature for a low-risk polyp. Over time small polyps can progress to large polyps (greater than or equal to 10 mm) and then to localized, regional, or disseminated colorectal cancer. Some patients can progress directly from the normal state to localized colorectal cancer. In our model, approximately 85 percent of colorectal cancers arise from the adenoma-to-carcinoma sequence. The remaining 15 percent of cancers arise de-novo. Our model posits that in the absence of screening the only way a cancer can be detected is through the emergence of symptoms. If patients present with symptoms, they are assumed to get the appropriate diagnostic workup and the appropriate therapy. If they survive, they enter a new state of “history of cancer.” From each state of cancer patients can die. The probability of death was estimated from epidemiological data on stage-specific survival for the average population. Definitions: CRC-L=colorectal cancer, localized. CRC-R=colorectal cancer, regional. CRC-D=colorectal cancer, distant. Sx=symptoms. Rx=treatment s/p=status-post.

OCR for page 61
Economic Models of Colorectal Cancer Screening in Average-Risk Adults: Workshop Summary SLIDE 4 SLIDE NOTES 4: Our model starts with the adenoma-carcinoma sequence. We model adenomatous polyps, which are generally agreed to be the precursor lesions for most colorectal cancers. We started with polyp prevalence by age from onset (50 years old). We assumed total polyp prevalence at age 50 (15%, average of men and women; 5% of polyps are “large”). We derived transition probabilities from small to large polyps from data on prevalence at various ages. We used age- and stage- specific cancer incidence rates reported in SEER (1990–1994). We assumed that 85% of cancers arise from polyps. We derived age-specific transition probabilities for normal to localized cancer (same for small polyp to localized cancer) and a fixed annual transition probability from large polyp to localized cancer. (It turns out that a fixed transition rate fits the data well, which is biologically plausible.) We assumed a dwelling time of 2 years each in localized and regional cancer, rates of symptomatic presentation derived to match SEER stage distribution (22%/yr for localized and 40%/yr for regional; 100% for distant). Stage-specific yearly mortality rates were derived for first five years after diagnosis (1.74%/yr for localized and 8.6%/yr for regional). For patients who survive more than 5 years, we applied age-specific mortality from all causes. We assumed distant cancer has average survival of 1.9 years.

OCR for page 61
Economic Models of Colorectal Cancer Screening in Average-Risk Adults: Workshop Summary SLIDE 5 SLIDE 5 NOTES: To estimate the transition probabilities we used various data sources. For transitions from normal mucosa to small polyp, and from small polyps to large polyps we used age-specific data from autopsy studies (Rickert et al., 1979; Arminski and McLean, 1964; Williams et al., 1982; Vatn and Stalsberg, 1982; Clark et al., 1985). The graphs above show how the predictions from our model so calibrated compare to the average of the published data on polyp prevalence for men and women.

OCR for page 61
Economic Models of Colorectal Cancer Screening in Average-Risk Adults: Workshop Summary SLIDE 6 SLIDE 6 NOTES: The graphs in this slide show how the model, as calibrated by SEER data (Ries et al., 1997), compares to the age- and stage-specific incidence rates.

OCR for page 61
Economic Models of Colorectal Cancer Screening in Average-Risk Adults: Workshop Summary SLIDE 7 SLIDE 7 NOTES: Incremental effectiveness of screening compared with no screening is affected by compliance, but incremental cost-effectiveness is not. Note that when we assume less than 100 percent compliance, the incremental cost-effectiveness ratios and the rankings across screening strategies remain the same, provided that compliance is equal across all strategies. However, the incremental cost-effectiveness ratios and rankings across strategies would be affected if compliance rates vary across strategies. Definitions: ICE=incremental cost-effectiveness.

OCR for page 61
Economic Models of Colorectal Cancer Screening in Average-Risk Adults: Workshop Summary SLIDE 8 SLIDE 8 NOTES: Until now, the health care costs associated with dying of other causes has been set at 0. The model could accommodate other assumptions, however. In addition, those costs could be age-dependent, if reasonable data were available.

OCR for page 61
Economic Models of Colorectal Cancer Screening in Average-Risk Adults: Workshop Summary SLIDE 9 SLIDE 9 NOTES: To validate the model, we examined the age-specific outcomes of the model and compared them with national data for the year 2000. We were gratified that the estimated number of cancer cases in our model is consistent with published data (Jemal et al., 2003; Sandler et al., 2002).

OCR for page 61
Economic Models of Colorectal Cancer Screening in Average-Risk Adults: Workshop Summary SLIDE 10 SLIDE 10 NOTES: In the Minnesota study, screening occurred in only some of the years over the course of the trial, and imperfect adherence with screening was attained (Mandel et al., 2000; Mandel et al., 1993). Overall, in that study, about 50 percent of all the potential yearly screenings were actually completed. With full adherence with yearly FOBT, our model predicts approximately double the reductions in cancer incidence and mortality that were observed in the Minnesota trial. For sigmoidoscopy every 5 years, our model predicts a 56 percent reduction in colorectal cancer incidence. That result is consistent with the Kaiser data for left-side lesions (Selby et al., 1992). The model’s estimate is a bit higher, however, because in the model we account for a reduction in CRC incidence due to removal of polyps in the proximal colon at the time of follow-up colonoscopy (subsequent to a positive sigmoidoscopy In the Kaiser study, the benefit of sigmoidoscopy was limited to the part of the colon examined during screening. The model predicted a reduction of 71 percent in cancer incidence for colonoscopy screening every 10 years. That result is consistent with findings of the National Polyp Study (Winawer et al., 1993)

OCR for page 61
Economic Models of Colorectal Cancer Screening in Average-Risk Adults: Workshop Summary SLIDE 11 SLIDE 11 NOTES: It is possible for the model to evaluate the costs and effectiveness of screening in populations with different levels of risk for polyps and cancer. We are currently working on making predictions at the level of the entire population. We have evaluated adjunct interventions, namely chemoprevention withy aspirin in average-risk patients and Cox-2 inhibitors in average-risk and high-risk patients (Ladabaum et al., 2001; Ladabaum et al., 2003). The model does have certain limitations and potential biases. In addition to those listed, we are currently not satisfied with our ability to model complex patterns of compliance with screening, follow-up and surveillance. We are working to improve the model in that regard. The question of independence between sequential tests (e.g. FOBT and FS) or repeated tests (e.g. annual FOBT in a cancer that is “dwelling”) is probably important to predictions from the model (e.g. independent annual FOBT at 40% sensitivity for cancer has only a 0.6×0.6×0.6×0.6=13% chance of not picking up a cancer at some point before it is distant if cancer dwells in localized×2 years and regional×2 years) As the model is currently structured, transitions among most patient states are probabilistic. However, for individuals with cancer, we assumed a fixed dwelling time in each stage. And the percent of cancers that arise from polyps has been fixed at 85 percent. We are not currently modeling the details of polyp location (distal vs. proximal) or histology (high-risk vs. low-risk). However, we do assume that sigmoidoscopy can

OCR for page 61
Economic Models of Colorectal Cancer Screening in Average-Risk Adults: Workshop Summary reach 50 percent of all lesions. So, that is one strategy which might be evaluated differently if our model were to identify polyps by location. REFERENCES Arminski TC, McLean MD. 1964. Incidence and distribution of adenomatous polyps of the colon and rectum based on 1,000 autopsy examinations. Dis Colon Rectum. 7:249–261. Clark JC, Collan Y, Eide TJ, Esteve J, Ewen S, Gibbs NM, et al. 1985. Prevalence of polyps in an autopsy series from areas with varying incidence of large-bowel cancer. Int J Cancer. 36(2):179–186. Jemal A, Murray T, Samuels A, Ghafoor A, Ward E, Thun MJ. 2003. Cancer statistics, 2003. CA Cancer J Clin. 53(1):5–26. Ladabaum U, Chopra CL, Huang G, Scheiman JM, Chernew ME, Fendrick AM. 2001. Aspirin as an adjunct to screening for prevention of sporadic colorectal cancer: A cost-effectiveness analysis. Ann Intern Med. 135(9):769–781. Ladabaum U, Scheiman JM, Fendrick AM. 2003. Potential effect of cyclooxygenase-2—Specific inhibitors on the prevention of colorectal cancer: A cost-effectiveness analysis. Am J Med. 114(7): 546–554. Mandel JS, Bond JH, Church TR, Snover DC, Bradley GM, Schuman LM, Ederer F. 1993. Reducing mortality from colorectal cancer by screening for fecal occult blood. N Engl J Med. 328(19): 1365–1371. Mandel JS, Church TR, Bond JH, Ederer F, Geisser MS, Mongin SJ, Snover DC, Schuman LM. 2000. The effect of fecal occult-blood screening on the incidence of colorectal cancer. N Engl J Med. 343(22):1603–1607. Rickert RR, Auerbach O, Garfinkel L, Hammond EC, Frasca JM. 1979. Adenomatous lesions of the large bowel: an autopsy survey. Cancer. 43(5):1847–1857. Ries LAG, Kosary CL, Hankey BF, Miller BA, Harras A, Edwards BK. 1997. SEER Cancer Statistics Review, 1973–1994. Bethesda, MD: National Cancer Institute. Sandier RS, Everhart JE, Donowitz M, Adams E, Cronin K, Goodman C, Gemmen E, Shah S, Avdic A, Rubin R. 2002. The burden of selected digestive diseases in the United States. Gastroenterol. 122(5):1500–1511. Selby JV, Friedman GD, Quesenberry Jr CP, Weiss NS. 1992. A case-control study of screening sigmoidoscopy and mortality from colorectal cancer. N Engl J Med. 326(10):653–657. Song K, Fendrick AM, Ladabaum U. 2004. Fecal DNA Testing Compared with Conventional Colorectal Cancer Screening Methods: A Decision Analysis. Gastroenterol. 126(5):1270–1279. Vatn MH, Stalsberg H. 1982. The prevalence of polyps of the large intestine in Oslo: An autopsy study. Cancer. 49(4):819–825. Williams AR, Balasooriya BA, Day DW. 1982. Polyps and cancer of the large bowel: a necropsy study in Liverpool. Gut 23(10):835–842 Winawer SJ, Zauber AG, May Nah Ho, O’Brien MJ, Gottlieb LS, Sternberg SS, Waye JD, Schapiro M, Bond JH, Panish JF, Ackroyd F, Shike M, Kurtz RC, Hornsby-Lewis L, Gerdes H, Stewart ET, Lightdale CJ, Edelman M, Fleisher M. 1993. Prevention of colorectal cancer by colonoscopic polypectomy. N Engl J Med 329(27):1977–1981.