The Impact of Colorectal Cancer Screening on the United States Population: is it time to celebrate?

The goal of cancer screening is to reduce disease-related death through the detection and effective treatment of precursor or early invasive lesions without undue harm. The effectiveness of colorectal cancer (CRC) screening is well-established, but less is known about its impact at the population level.¹ In the United States, CRC screening has been practiced since the 1960s or earlier and began to be tracked by the National Health Interview Survey (NHIS) since 1987 (Figure).^{2, 3} Although national surveys overestimate the true prevalence of screening,⁴ they in combination with other data systems demonstrate a steady increase in the use of CRC testing over the years in tandem with declines in both disease incidence and mortality rates since the mid-1980s.³ In men, CRC incidence dropped by 2.6% during 1992-1995, and by 4.2% during 2008-2010, with similar changes observed in women.⁵ Remarkably, there has also been a decline in the absolute number of CRC cases over the years. For instance, there will be about 136,830 new CRC cases in 2014 compared to 145,290 in 2005.⁶ The reasons for this decline have been speculated and naturally, many are apt to attribute it to the increased uptake of screening. However, little empirical data exist to support that view.

Figure.

Colorectal Cancer Test Use Rates for Adults 50 years and older by Race/Ethnicity: NHIS 1987-2010

In this issue of the Journal, Yang et al. published an ecological study of the public health impact of CRC screening in the United States.³ The study analyzed over three decades of Surveillance, Epidemiology, and End Result data on CRC diagnosis and NHIS data on CRC screening use during 1987-2010. The incidence data from 1976-1978 served as the baseline for disease rate in the absence of screening to derive the number of prevented cancers. The authors report that approximately one-quarter to one-half of a million CRC cases were averted during the 1979-2009 period. These estimates were based on two scenarios. First, the authors assumed that the CRC incidence rate did not change between 1979 and 2009. The second more realistic estimates incorporated a 0.4% annual percentage incidence increase over the base rate. In contrast to screening for other cancers, there was no evidence for over-diagnosis bias.⁷ There was a decrease in incidence for both early and late-stage disease. Not surprisingly, the greatest number of averted cancers was in the left colon/rectum. In the right colon, which is less effectively screened by fecal occult blood testing (FOBT) or sigmoidoscopy, there was a slight increase in early stage cancers but a decrease in the rate of incident advanced cancers. Inclusion of in situ lesions, which are precursor lesions, in the analysis may have underestimated the number of averted cancers. They also excluded unstaged cancer, which may have underestimated the total number of observed incident cancers.

Although no direct causal conclusions can be drawn from an ecological analysis, increased use of screening seems to be the most plausible explanation for the decreased incidence. This hypothesis is strongly supported by results of randomized trials showing that screening with sigmoidoscopy reduced the incidence of CRC within the reach of the sigmoidoscope by 50%.⁸ Case-control studies have also shown that screening colonoscopy substantially reduces the risk of late-stage disease.⁹ Using micro-simulation modeling, Edwards et al. reported that a substantial portion of the decrease in CRC death in the United States was due to increased uptake of screening.¹⁰ Further, the favorable trends in CRC disease burden have occurred during a period of continued increases in risk factors in the United States. Approximately 70% CRC cases in the United States are believed to be attributable to unhealthful lifestyles.¹¹ In the United States, these risk factors, particularly obesity, are highly prevalent with high calorie intake and only modest improvements in physical activity levels.¹² The prevalence of obesity among individuals 20-74 years old increased from 15.1% in 1976-80 to 35.3% in 2007-2010. In some countries, increased westernization has been accompanied by an increase in the incidence of CRC.¹³ However, CRC rates are lower in Hispanics than in non-Hispanic whites, but so are their screening rates. Thus, alternative explanations are possible. Increasing use of drugs with chemopreventive properties such as non-steroidal anti-inflammatory agents may contribute to the observed CRC incidence trends.¹⁴

Before we celebrate, the report also underscores that a substantial portion of the current cases of CRC is attributable to non-use of screening. In particular, some groups have not realized the public health benefits of screening equally. Substantial uptake of screening did not occur in the United States until the start of the 21^st century in tandem with the ascendancy of interest in colonoscopy. Unfortunately, even with overestimation in national surveys,⁴ screening rates remain below the public health goal of 70.5%, and progress has been particularly slow in some minority populations.¹⁵ For instance, the screening rates of African-Americans lag about two years behind those of non-Hispanic whites and their incidence and mortality rates are higher. For the 2001-2010 period, the CRC incidence rate was 50.5 per 100,000 for white men and 62.5 per 100,000 among black men and the percentage decreases in incidence were 4.0% and 2.0%, respectively, with similar mortality trends.⁵ In addition to race/ethnicity, low-income, geography, or lack of insurance coverage, regular place of health care, or physicians recommendation for screening are significant barriers to use of CRC screening.¹⁶ Increasing the use of screening in these underserved populations, who also have disproportionally high disease rates, will have great impact on progress towards the HealthyPeople 2020 screening goal.

In a recent paper, Gupta and colleagues provided four key multilevel recommendations to boost use of screening in underserved populations.¹⁶ The first was to avoid a colonoscopy-only screening policy in clinical settings and actively promote the message that “the best test is the one that gets done well.” Currently, the United States Preventive Services Task Force recommends highly-sensitive FOBT annually, flexible sigmoidoscopy every 5 years with mid-interval FOBT, or optical colonoscopy every 10 years as equally acceptable screening strategies.¹ Since 2001, colonoscopy has rapidly become the most commonly used screening test and is considered the preferred test by some national groups, even as we wait for studies to determine if it is superior to other strategies. Studies show that one size does not fit all and providing choice in screening can boost screening rates. Second, it is critical to develop and implement strategies to efficiently identify screen-eligible people in both clinical settings and outreach programs to maximize use and minimize overuse or misuse. As adoption of health IT including electronic medical record (EMR) systems gains momentum, greater collaboration across existing healthcare delivery silos through health information exchanges can make it possible to document screening use across delivery systems and health plans. Third, it is critical to assure provision of the entire CRC screening continuum including timely diagnostic testing for abnormal screening and treatment when cancer is diagnosed. Quality metrics for CRC screening such as Healthcare Effectiveness Data and Information Set measures should be updated to quantify not just the receipt of a CRC test but separately document proportion of tests used for screening purposes and the timely (within 90 days) receipt of diagnostic testing for positive screens. This will allow us to identify failures in the screening process to target for appropriate interventions.

Fourth, there is now strong evidence that organized screening strategies are effective in increasing use.^{16, 17} Newer FOBT technologies do not require dietary restrictions and can be sent in the mail and tested using high-throughput systems. It is feasible to implement organized programs in many health care systems or practices through local or regional collaborations to gain economies of scale and reduce implementation cost while freeing up healthcare providers for other tasks. Therefore, visit-based approaches should no longer be used as the sole approach to provide screening.

These recommendations provide a framework for hastening the public health benefits of CRC screening in the United States. Programs in New York City and the state of Delaware are examples of successful public health efforts.¹⁶ They dovetail into provisions in the Patient Protection and Affordable Care Act of 2010 for broadening access to preventive care services. The recommendations are “back to the future.” Screening for CRC dates back to the 1800s when proctoscopy was first used to visualize the rectum and sigmoid.¹⁸ Guaiac-based gFOBT was developed in the 1950s, and immunochemical FOBT (FIT) in the 1970s.¹⁹ The introduction of colonoscopy in 1969 made it possible to visualize the colon beyond the reach of the sigmoidoscope.²⁰ These three tests remain central to efforts to prevent premature death from CRC even as technologies and testing patterns continue to evolve. The concept of home-based screening was first introduced by Greegor in 1967.¹⁹ Improvements in FIT technology now make it possible to realize the ideals of increasing access to screening with large-scale population-based screening program, particularly for those who do not seek care or for whom screening was not provided during the course of a medical care visit.¹⁷ However, all screening programs need adequate colonoscopy capacity and it is critical to identify and address variations in performance characteristics for all CRC tests.

Finally, the study by Yang and colleagues suggests that we can reduce the public health burden of CRC through increased CRC screening. However, current practices of opportunistic screening delivery in most healthcare systems deprive many individuals of the benefits of screening either because they do not receive regular healthcare or because their preventative care needs are crowded out by competing economic, social, medical and administrative needs of both the patient and/or provider. Fortunately, technologies have advanced to the point that opportunistic screening during a visit with a healthcare provider need not be relied on as the sole approach to deliver CRC screening. United States healthcare systems should adopt programs of population-based delivery through screening outreach and facilitated in-reach by leveraging population health capabilities in health IT systems. It is also important that public health programs and the medical community develop more effective ways to communicate both the importance of and the options for CRC screening to the public. Capitalizing on opportunities for provider recommendation of appropriate CRC screening can boost screening rates. All said, our success in maintaining sustained decreases in CRC incidence may be defined by our ability to improve delivery to underserved populations.

Précis.

Increasing use of screening has contributed to a substantial decline in the incidence of colorectal cancer in the United States. However, more attention needs to be focused on improving access to screening for underserved populations to further advance the public health goal of reducing the burden of colorectal cancer on the United States population.