Family history assessment for colorectal cancer susceptibility is a costly but potentially beneficial intervention that has feasibility for implementation in clinical practice. We built a decision model to evaluate the cost-effectiveness of implementing such a program. Under generally favorable assumptions towards family risk assessment and using the early-onset hypothesis of polyp behavior in persons with family history, family history assessment may be modestly cost-effective. The most important factors influencing the outcome are the life expectancy gain for persons who undergo regular colonoscopy screening and the cost of colonoscopy. If polyp behavior is substantially different than theorized, simply shifting screening to an earlier age is much less cost-effective. Additional research that better defines the benefits of colonoscopy in 40-year-olds with suggestive family histories and clearer delineation of polyp behavior are the most critical information that is needed to better define the cost-effectiveness of assessment.
Our model does not address several other issues that ultimately will determine the success and cost-effectiveness of family history assessment. First, it is not clear how many individuals visit primary care physicians for preventive care near age 40, particularly men. Second, individuals need to have an opportunity to receive family history assessment. Acute care visits for acute or chronic medical care needs are often poor opportunities for assessment. Reducing the proportion who present for assessment does not influence cost-effectiveness of assessment itself; yet, it greatly diminishes the benefits that are realized for the population. Ultimately, other approaches for risk assessment, such as mailed surveys, might be more effective means of obtaining family risk information. Third, it is not certain that providing individuals with information about their risk of colorectal cancer will influence adherence to screening, both for those with and without suggestive family histories. Colon cancer screening is currently underused in the United States (28
). Studies suggest that risk counseling can improve use of screening services both in those at increased and average risk for colorectal cancer (29
); yet, it is not known whether counseling at age 40 would translate into higher rates of screening 10 years hence (for those at average risk). This model does not quantify the costs and benefits of identifying those who carry rare mutations with high disease penetrance, primarily hereditary nonpolyposis colorectal cancer and familial adenomatous polyposis. Cost-effectiveness analyses of screening for these conditions have been reported previously (8
). This model does not include administrative costs associated with starting and maintaining family history assessment programs. We only consider the effect of adherence to screening during age 40 to 50, because this is the place where practice changes in response to the guidelines. Although we do not explicitly address the issue, persons with a family history might adhere more closely to screening over their lifetimes compared with average risk persons. In this case, the model may underestimate the cost-effectiveness of family history screening. For the sake of simplicity, and because of lack of agreement among experts about the ideal schedule and efficacy of repeat screening after polypectomy, the model does not address the issue of interval follow-up for those who have adenomatous polyps removed, other than assuming that repeat screening occurs at age 50. Increases in use of screening of persons with suggestive family histories may affect both demand for and supply of colonoscopy over time. These issues, combined with changes in screening efficiency will ultimately influence the future price of the procedure. The costs of colorectal cancer are based on estimates for persons predominantly over age 65 and thus may underestimate to true lifetime costs for a younger person with this disease. Finally, the opportunity for family history assessment affords an opportunity to determine risks for several conditions that may benefit from screening and prevention, such as other cancers, diabetes, and coronary artery disease. It is reasonable to consider the costs and benefits of “global” risk assessment and counseling based on family history, and there are efficiencies of gathering all relevant information together. In addition, risk assessment offers an opportunity to counsel patients on other ways to reduce risk, such as reducing dietary fat and increasing exercise. On the other hand, such practice runs the risk of losing effectiveness by overwhelming the recipient with information about risks for multiple diseases, particularly for those without interventions that are known to be effective. The list of family history information that is useful will probably grow over time, with more precise genomic knowledge and practical and effective interventions.
Public and private health insurance plans generally do not cover costs associated with family history assessment, except (in some cases) as part of a general medical evaluation. Certainly, direct payment for service (i.e., billing codes) would encourage providers to take and document a family history. Health plans might argue that such payment is not justified until we have direct evidence of benefit, as was obtained for other screening procedures like mammography and colorectal cancer screening after age 50. Certainly, a prospective controlled trial of family history assessment combined with early colonoscopy for those with a suggestive family history would be costly, time consuming, and challenging to implement on a population level. As costly as such trials are in prevention, our analysis suggests that such an investment may be worthwhile, given the substantial potential costs and benefits involved in such programs.
This analysis considers guidelines and screening practices that are widely agreed upon in the United States. Other countries with greater constraints on health resource expenditures might consider less costly screening modalities (e.g., fecal occult blood testing and flexible sigmoidoscopy) or screening schedules. Simulation models can address the tradeoffs between cost, sensitivity, specificity, and feasibility in these situations.