Estimating the Screening-Eligible Population Size, Ages 45–74, at Average Risk to Develop Colorectal Cancer in the United States

Colorectal cancer is the second deadliest malignancy in the United States, with an estimated 145,600 new cases and 51,020 deaths in 2019 (1). Although colorectal cancer is primarily diagnosed in older individuals, the incidence of colorectal cancer has been increasing in younger individuals. Since 1994, there has been a 51% increase in colorectal cancer incidence among those less than 50 years old (2). In addition, individuals less than 50 are more likely to have advanced-stage colorectal cancer at diagnosis than those over 50 (3), highlighting the potential benefit of colorectal cancer screening in this patient population.

The U.S. Preventive Services Task Force (USPSTF) published a recommendation statement in 2016 that colorectal cancer screening for eligible individuals should begin at age 50 (4). In 2018, based partly on a modeling study commissioned to expand the USPSTF analyses, the American Cancer Society (ACS) updated its colorectal cancer screening guideline and lowered the recommended screening age to 45 (2). This guideline modification has the potential to influence screening and healthcare decision-making for millions of eligible Americans. However, the number of individuals directly impacted by this potential decrease in the colorectal cancer screening start age, and a methodology to approximate that number, is not thoroughly described.

Using public databases and published literature, we established a methodology to estimate the current number of screening-eligible individuals ages 45–74 in the United States at average risk for developing colorectal cancer.

We identified the estimated number of men and women in 2019 in the United States ages 45–74 according to the U.S. Census Bureau (Table 1; ref. 5). We used the Surveillance Epidemiology and End Results (SEER) 2012–2016 age-adjusted annual incidence rate data from the November 2018 data submission to subtract the estimated number of individuals within each age group diagnosed with colorectal cancer prior to screening (Table 1; ref. 6). Notably, by subtracting all incidence colorectal cancer for individuals ages 45–49, we are conservatively assuming that the incidental cancers for all individuals would occur prior to any potential screening.

To identify the number of individuals with a family history of colorectal cancer, we reviewed the population-based studies included in the meta-analysis by Henrikson and colleagues, published in 2015 (7). We used two estimates for family history prevalence due to the heterogeneity of outcomes reported in the meta-analysis. For our smaller estimate, we used a conservative definition of family history (“one or more first-degree relatives with colorectal cancer”) and identified studies with prevalence values from populations that met our definition of family history (Table 2). We calculated an overall prevalence by weighing the prevalence values from each study based on population size (we removed one study from consideration, Taylor and colleagues; ref. 8, due to its outcomes being restricted to a single U.S. state and its disproportionate influence on a weighted prevalence calculation). For our larger estimate, we used the commonly cited value of 10%, which also corresponds to the upper bound of the prevalence estimate from the summary tables in Henrikson and colleagues (7). Using these two percentages, we subtracted the estimated number of people with colorectal cancer family history from our population.

Next, we estimated the total proportion of individuals with hereditary cancer syndromes associated with colorectal cancer: inflammatory bowel disease (IBD), hereditary nonpolyposis colorectal cancer (HNPCC or Lynch Syndrome), and familial adenomatous polyposis (FAP). We assumed the population-wide prevalence for these syndromes was representative of the 45- to 74-year-old population. For IBD, we used the combined estimated prevalence of ulcerative colitis and Crohn disease (0.5%) in the United States, made in 2016 (9). For HNPCC, we used 0.25%, which we extrapolated from the ACS's Cancer Facts and Figures 2017–2019 (800,000 individuals divided by an approximate U.S. population of 320,000,000; ref. 10). For FAP, we assumed that the population was negligible because most individuals with FAP would be diagnosed with colorectal cancer before age 45 (11) and therefore would not be included in this population.

Finally, we determined the screening-eligible population by subtracting the estimated number of people who already received colorectal cancer screening (both invasive and noninvasive modalities). Individuals ages 45–49 were considered separately because they are not currently recommended for colorectal cancer screening and therefore have different screening eligibility rates than individuals ages 50–74. For individuals ages 45–49, the estimated percentage of previously screened individuals is 7.2%, according to 2019 National Health Interview Survey data (12). For individuals ages 50–74, the estimated percentage of individuals up-to-date with colorectal cancer screening is 68.8%, according to 2018 data from the Behavioral Risk Factor Surveillance System (13).

The complete workflow and population estimate for each analysis subgroup is described in Fig. 1. We removed individuals diagnosed with colorectal cancer, without a family history of colorectal cancer, and with hereditary cancer syndromes associated with colorectal cancer from the total estimated population of the United States. We estimate the total average-risk population to be between 102.1–106.5 million people (18.5–19.3 million ages 45–49 and 83.6–87.2 million ages 50–74). After removing individuals who have undergone colorectal cancer screening, we estimate the total average-risk, screening-eligible population to be between 43.4 and 45.2 million people (17.2–18.0 million ages 45–49 and 26.1–27.3 million ages 50–74).

Lowering the recommended colorectal cancer screening start age to 45 will substantially increase the percentage of individuals who would be newly eligible for screening. By implementing the lowered screening age, we estimate that an additional 19 million people would be added to the average-risk population, or approximately 21.2%–23.1% of the current population. Among average-risk individuals, upon immediate adoption of a lowered colorectal cancer screening age, the screening-eligible population would increase by over 60%, from roughly 27 to 44 million people. This increase could have significant consequences for colorectal cancer prevention and early detection, given the increased incidence of colorectal cancer among younger individuals (2).

Overall, this analysis provides a comprehensive estimate of the average-risk population eligible for colorectal cancer screening in the United States, including whether the recommended start age is lowered to 45 years. One limitation of our analysis is overlapping populations probably exist among these groups, for example, individuals having received colorectal cancer screening because they have hereditary cancer syndromes. One study found that 20% of individuals under age 50 who had received colonoscopies reported that the reason was colorectal cancer screening (14). This percentage could be similar across screening modalities. In addition, our estimate does not incorporate subgroup analyses based on gender, race/ethnicity, or socioeconomic status, although notably, the ACS-recommended screening start age does not discriminate based on these characteristics. Finally, we anticipate that the percentage of individuals with hereditary colorectal cancer syndromes will increase over time as new acquired genetic abnormalities are discovered.

Ultimately, this estimate is important for policy makers and healthcare systems to better understand the population benefitting from changing colorectal cancer screening guidelines to improve cancer prevention and early detection.

A. Piscitello and D.K. Edwards V reports receiving personal fees from Exact Sciences during the conduct of the study. No potential conflicts of interest were disclosed.

Conception and design: A. Piscitello, D.K. Edward V

Development of methodology: A. Piscitello, D.K. Edwards V

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A. Piscitello, D.K. Edwards V

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): A. Piscitello, D.K. Edwards V

Writing, review, and/or revision of the manuscript: A. Piscitello, D.K. Edwards V

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): A. Piscitello, D.K. Edwards V

Study supervision: A. Piscitello, D.K. Edwards V

Special thanks to Patricia Thornewell, PharmD, Elle Grevstad, PhD, Kristin Lamont, PhD, and Amy Stern (all of Exact Sciences, Madison, WI) for their helpful comments and feedback. This study was funded by Exact Sciences.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.