U.S National Institutes fo Health | cancer.gov NCI Logo

Cancer Intervention and Surveillance Modeling Network

Modeling to guide public health research and priorities

Other Achievements: Highlights

How do Bayesian methods facilitate calibration of microsimulation models?

Microsimulation models that describe disease processes synthesize information from multiple sources and can be used to estimate the effects of screening and treatment on cancer incidence and mortality at a population level.  Microsimulation models are complex and invariably include parameters that are not well informed by existing data.  Calibration of the parameters of the model to reflect the observed data for incidence and mortality can be achieved using a Bayesian method with Markov chain Monte Carlo methods. This approach is computationally intensive but uses statistical criteria to selected parameter values, allows simultaneous calibration of multiple parameters to multiple data sources, incorporates information via prior distributions, describes parameter identifiability, and provides interval estimates of model parameters.  Rutter, Miglorietti and Savarino developed a microsimulation model for colorectal cancer (CRC-SPIN) using this proposed method.  The authors suggest that multiple sources of variability are needed when calibrating and applying microsimulation models (Rutter 2009).

Return to Top

Does individualizing colonoscopy screening guidelines by gender and race reduce health disparities?

The Memorial Sloan-Kettering and ErasmusMC group (MISCAN colorectal model) showed only marginal improvements in the costs and effects of individualized screening recommendations on a population level. Individualized recommendations, however, could contribute to decreasing disparities between African-American and white populations. This study informs the discussion on individualization of colorectal cancer screening recommendations by quantifying the potential benefit (Lansdorp-Vogelaar 2009).

Return to Top

What estimates of fecal occult blood test screening sensitivity and sojourn time best explain trial data?

Estimates of the fecal occult blood test (FOBT) (Hemoccult II) sensitivity differ widely between screening trials and could lead to divergent conclusions on the effects of FOBT screening. Three hypotheses for FOBT sensitivity were tested:

The latter hypothesis gave the best fit to results of several randomized controlled trials (those conducted in Minnesota, Nottingham, and Funen). (Lansdorp-Vogelaar 2009).

Return to Top

At what age do people undergo their first colorectal cancer test? How has this changed over time?

Miglioretti (2008) estimated the age at first fecal occult blood test, first lower endoscopy examination, and first colorectal cancer screening test of either type by using two series of cross-sectional surveys: the National Health Interview Survey (NHIS) and the Behavioral Risk Factor Surveillance System (BRFSS). The model for the national data (NHIS) incorporated birth cohort effects and was linked to state-specific models for BRFSS data. These results will be used to estimate the contribution of stool and endoscopy examinations to observed changes in colorectal cancer incidence and mortality.

Return to Top

Do individuals with a positive fecal occult blood test receive appropriate follow-up?

Screening for fecal occult blood can be effective in reducing the rate of colorectal cancer mortality only if positive tests are appropriately followed up with complete diagnostic evaluation (i.e., colonoscopy or flexible sigmoidoscopy with double contrast barium enema) and treatment. Changes in the rates of complete diagnostic evaluation were examined within 1 year following a positive fecal occult blood test (FOBT) performed from 1993 to 2005 by Group Health (GH), a large integrated health care organization. In 1998, GH started an electronic registry to track enrollees who had not received a complete diagnostic evaluation within 4 months after a positive FOBT, and in 2003, GH implemented a centralized manual audit system to supplement the electronic system. Rates of complete follow-up increased from 57-64% between 1993 and 1996 to 82-86% between 2000 and 2005. This study demonstrates great improvements in adherence to recommendations for complete diagnostic evaluation after a positive FOBT, and it also showed that it may be possible to reach levels of follow-up that are comparable to those observed in randomized trials through the use of tracking systems and screening recommendations (Miglioretti 2008).

Return to Top

How prevalent are adenomas, and how does prevalence vary with age?

Adenoma prevalence rates vary widely in the literature and depends on the age and sex of the individuals included in the sample and the methods used to measure adenomas. A study was conducted to estimate adenoma prevalence by age and sex by combining information from 14 autopsy studies. All studies reported adenoma prevalence by age. Most also reported adenoma prevalence by sex, and several also reported adenoma counts. A meta-analytic model developed by the Group Health Research Institute was used to combine the results across studies, which were consistent with results from recent colonoscopy studies. Adenomas were more prevalent in men than women, and adenoma prevalence and multiplicity both increased with age. For example, the prevalence of adenomas in men was estimated to increase from 32% at age 50 to 40% at age 60; the prevalence of adenomas in women was estimated to increase from 23% at age 50 to 29% at age 60. In addition, 9% of 50-year-old men and 5% of 50-year-old women have three or more adenomas (Rutter 2007). These results are important in understanding the natural history of adenoma initiation and progression.    

Return to Top

How has microsimulation modeling been used to inform health policy on colorectal cancer screening in the Netherlands?

The MISCAN colorectal group at ErasmusMC in Rotterdam works with the Dutch government to plan programs for population-based colorectal cancer screening.  This work has included microsimulation modeling to assess key screening programs issues and components such as limited endoscopy capacity, the optimal cut point for fecal immunochemical tests, and screening for higher risk individuals because of positive colorectal cancer family history.

Wilschut JA, Habbema JD, van Leerdam ME, Hol L, Lansdorp-Vogelaar I, Kuipers EJ, van Ballegooijen M. Fecal occult blood testing when colonoscopy capacity is limited. J Natl Cancer Inst 2011 Dec 7;103(23):1741-51. [Abstract]

Wilschut JA, Hol L, Dekker E, Jansen JB, Van Leerdam ME, Lansdorp-Vogelaar I, Kuipers EJ, Habbema JD, Van Ballegooijen M. Cost-effectiveness analysis of a quantitative immunochemical test for colorectal cancer screening. Gastroenterology 2011 Nov;141(5):1648-55.e1. [Abstract]

Wilschut JA, Steyerberg EW, van Leerdam ME, Lansdorp-Vogelaar I, Habbema JD, van Ballegooijen M. How much colonoscopy screening should be recommended to individuals with various degrees of family history of colorectal cancer? Cancer 2011 Sep 15;117(18):4166-74. doi: 10.1002/cncr.26009. [Abstract]