Increasing colorectal cancer screening adherence: comment on “A randomized comparison of print and web communication on colorectal cancer screening”

Colorectal cancer (CRC) is the second leading cause of cancer death in the United States.¹ Furthermore, it is eminently preventable with routine screening that reducing CRC mortality as well as incidence via identification and removal of pre-cancerous adenomatous polyps. Nearly all authoritative groups endorse CRC screening, yet over 40% of the population remains unscreened.² While much effort has been invested in trying to improve rates of CRC screening, the overall effect of individual strategies at the patient-, provider-, or system-level has been relatively modest to date. In particular, web-based approaches have been viewed as having tremendous promise as a modality to address participation, but to date there has been little supporting evidence.

To this end, Weinberg et al. report in this issue of the Archives a trial of 904 average risk women ≥ 50 years old not previously screened for CRC, who were randomized to receive education materials via web-based communication, mailed print communication, or usual care. Disappointingly, even after factoring in attentional style to better target individual learning styles, the primary endpoint (completion of any CRC screening test including stool based, flexible sigmoidoscopy, colonoscopy, barium enema) was no different among the groups. One of the most alarming findings was that the rate of uptake across the board was appalling low at 4 months (12.3%) and 12 months (21%) follow-up.

While the study has a few limitations, particularly in its absence of description of how participants interested in CRC screening would pursue testing (e.g. contact PCP for referral, wait for next appointment, register online?) and its lack of generalizability beyond predominantly white women, it provides an important caveat to ongoing discussion regarding methods for improving rates of CRC screening participation. Over the past several decades, a multitude of studies have evaluated the potential impact of patient-centered interventions by education (print, multi-media, telephone, in person),^{3, 4} reminder systems (mailed, telephone, and electronic),^{3, 5} and engagement at other points of health care maintenance (at mammography, during flu shot clinic)⁶ with minimal to modest improvements in participation rates. Nationally, the 58.6% rate of CRC screening participation lags well behind that of breast cancer screening at 72.4% and cervical cancer screening at 83.0%.⁷ In Weinberg’s study, 70% of participants were up-to-date with mammography suggesting a predilection toward preventative health behavior, yet only 12.3% participated in CRC screening despite the interventions. This begs the question of how and why CRC screening is different from other types of cancer screening.

One important difference is that for breast and cervical cancer screening, a single test is recommended whereas there is a menu of accepted options for CRC screening. These choices can be time-consuming and cumbersome for providers to discuss as well as overwhelming and confusing for some patients to interpret. While colonoscopy has more recently emerged as the preferred test by both patients and providers alike because of its accuracy,⁸ the logistic barriers to colonoscopy are plenty (e.g. securing time off work, transportation, navigating a fragmented medical system). Furthermore, concerns about the purgative regimen, discomfort, embarrassment, complication rates and cost can foster patient reticence. The importance of participation for CRC outcomes is underscored by a recent report that when compared to colonoscopy, fecal immunohistochemical test (FIT) had an equivalent rate of CRC diagnosis with the significantly lower accuracy of FIT being offset superior screening participation rate..⁹

While these challenges are not insurmountable, they do highlight that a web-based communication without a “human” connection is unlikely to improve substantially on current CRC participation rates. Web-based interventions are appealing because they are relatively low cost with the ability to be widely disseminated, but as demonstrated by Weinberg and colleagues there is a great disconnect between having internet access in daily life and using it for preventative care and medical decision making. While internet access at home or at work was an inclusion criterion for the study, only 24.4% of women logged into the site and self-reported use was even lower, suggesting that the site made no imprint on their memory. Furthermore, there was no significant difference in screening participation rates among those who logged into the site and those who did not. Perhaps web-based communication has the potential for information transfer for the particularly savvy, but this subset would are probably more likely to be compliant with screening regardless of prompting. For those who interventions may make a difference, web-based or mailed contacts are unlikely to substitute for face-to-face interactions with a health-care provider. Patient-provider shared decision making allows for exchange of information, exploration of patient preferences, and clarification of values.

Unfortunately, time constraints during the physician visit may represent a considerable hurdle for effective counseling. Clearly, the web is not a panacea and unable to harness the power of interpersonal communications. One potential alternative is the use patient navigators, lay personnel trained to guide patients through the health care system to receive appropriate services and in doing so, provide a functionally integrated medical system for each individual patient. The use of patient navigation in improving screening colonoscopy rates, particularly among underserved populations is appealing because it is a system-level intervention that is simultaneously patient-centered and adaptable to individual patient needs. Patient navigation for CRC screening participation has shown great promise^{10, 11} and perhaps when coupled with other forms of education has the potential for even greater improvement in screening rates.

Another vexing issue for CRC population screening is individualizing screening. Current practice considers the majority of patients over age 50 as “average risk,” yet the yield of during colonoscopy is generally low (~6–7% rate of advanced adenomas), meaning the vast majority of screening procedures have no significant CRC prevention connotations. Selective use of colonoscopy among subjects identified a priori as high risk could reduce the volume of normal colonoscopies performed and allow transfer of limited screening resources to the subset of the population more likely to derive a benefit. Patients identified as having a higher risk of CRC may then be more likely to adhere to recommendation for colonoscopy. Such risk stratification techniques would need to be inexpensive, minimally intrusive, and above all accurate. Unfortunately, currently used non-invasive tests (e.g. stool-based test) do not completely fulfill these criteria. Newer approaches including novel blood tests, prep-less CT colography, next generation fecal DNA analysis and biophotonic detection of field carcinogenesis (implemented via a simple rectal exams) appear to have potential, but obstacles include expense, accuracy or the relatively early stage of development. ¹²,

Until sound approaches for accurate risk stratification are developed, the focus of Weinberg’s study is right on. Evaluating methods for increasing participation in already accepted screening tests, whether through clinical trials or clinical effectiveness research should be our focus of attention. The abysmal rate of CRC screening noted in this study, especially when compared to other malignancies, is truly sobering. There is clearly much more work to be done. With the abundance of accepted modalities for CRC screening, adherence and personalization of screening appears to be paramount. Future research should continue to explore patient beliefs and values with an eye toward developing multi-modal interventions in conjunction with personalizing screening strategies. This will bring us closer to achieving the goal of 70.5% of the US population screened as set out by Health People 2020 and consequently reduce the toll of CRC.¹³