Elsevier

Preventive Medicine

Volume 45, Issue 4, October 2007, Pages 267-273
Preventive Medicine

Preference-based electronic decision aid to promote colorectal cancer screening: Results of a randomized controlled trial

https://doi.org/10.1016/j.ypmed.2007.07.003Get rights and content

Abstract

Objective

Despite the burden of colorectal cancer and improved health care outcomes with early detection and treatment, screening rates among eligible adults are low. We previously developed through a series of studies an interactive electronic tool, Colorectal Web, to promote colorectal cancer screening.

Method

From May 2002 to December 2003, we conducted a randomized controlled trial of Colorectal Web compared to a standard Web site on colorectal cancer screening in urban, suburban, and rural communities in Michigan with high colorectal cancer burden. Study participants were age 50 years and older, with no previous colorectal cancer screening. Major outcome was screened for colorectal cancer by 24 weeks post-intervention.

Results

174 eligible adults were randomized and participated. Immediately post-intervention, Colorectal Web participants were significantly more likely to have a preferred colorectal cancer screening method, but this difference did not persist at subsequent follow-up. Eighty-nine participants had been screened for colorectal cancer by 24 weeks post-intervention. The probability of being screened for the Colorectal Web intervention study arm compared to the control is OR = 3.23 (2.73–3.50 95% Confidence Interval).

Conclusion

Colorectal Web is more effective than a standard colorectal cancer Web site at prompting previously unscreened individuals to choose a preferred colorectal cancer screening test and to be screened for colorectal cancer.

Introduction

Colorectal cancer (CRC) is the third leading cause of new cancer cases and deaths in men and women each year (Jemal et al., 2005). It has a lengthy, detectable, premalignant phase (Donald and Burhenne, 1993) during which a simple intervention (polypectomy) can prevent progression to more advanced stages of cancer (Burt et al., 1985, Decosse et al., 1994, Fuchs et al., 1994). When detected early, more than 90% of persons with the disease live at least 5 years beyond the diagnosis; however, only 37% of CRCs are diagnosed before metastasizing (Ries et al., 1998).

CRC is unique among preventable cancers. It has four screening options that are similarly endorsed by professional groups: fecal occult blood testing (FOBT), flexible sigmoidoscopy, double-contrast barium enema, and colonoscopy. No strong scientific evidence supports favoring one test over another for reducing mortality from CRC. CRC screening thus represents a form of “preference sensitive care,” defined by Wennberg and colleagues as “a care situation in which two or more medically justifiable options exist” (Wennberg, 2002). The US Preventive Services Task Force (Sheridan et al., 2004) recommends CRC screening be based on patient preferences among the screening options.

Patient preferences among CRC screening options have been examined in several published studies (Van Ness et al., 1987, Dominitz and Provenzale, 1997, Leard et al., 1997, Pignone et al., 1999, Wolf and Schorling, 2000, Sheikh et al., 2004, Nelson and Schwartz, 2004, Frew et al., 2001a, Frew et al., 2001b, Frew et al., 2004, Frew et al., 2005) finding that patient preferences for CRC screening were highly individual and variable. In the most recent published study, 55% preferred sigmoidoscopy and FOBT, 29% chose colonoscopy, and 16% wanted no screening (Sheikh et al., 2004). A recent synthesis of all previous studies and new data concluded that FOBT was the most preferred CRC screening choice by asymptomatic patients (2004).

Patient preferences for colon cancer screening are modestly sensitive to information about test performance and strongly sensitive to out-of-pocket costs (Pignone et al., 1999). In patients for whom accuracy was the most important test feature, colonoscopy (62%) was the preferred screening method. Patients for whom invasive test features were more important preferred fecal occult blood testing compared to patients for whom the invasive test features were not important (76%; p < 001) (Ling et al., 2001). Patients have distinct preferences for CRC screening tests that are associated with the importance placed on certain test features. Physicians incorrectly perceive those factors that are important to patients (Ling et al., 2001). Some investigators suggest that physicians incorporate patient values in regard to certain test features when discussing CRC screening with their patients and when eliciting their screening preferences (Ling et al., 2001). Patient's general lack of knowledge regarding CRC screening and its modalities and the heterogeneity of patients' preferences would make this recommendation impractical (Shokar et al., 2005). However, this insight is extremely critical toward developing decision tools to promote CRC screening.

We have previously reported that adults needing CRC screening can establish a preference among the options. In 10 focus groups of black and white men and women aged 50 to 70 years not screened for CRC, participants were able to articulate their preferences for a particular CRC screening test after receiving appropriate information. In addition, study participants suggested that an electronic tool, such as a Web site to help them choose among CRC screening options, would be useful (Fetters et al., 2004). Using information obtained from the focus groups, we developed Colorectal Web, an interactive electronic tool promoting CRC screening. The tool was further refined with 30 intensive individual interviews with adults using the program (Power, 2001, Power et al., 2002, Fetters et al., 2004). We report here the results of a randomized controlled trial (RCT) to test the efficacy of Colorectal Web compared to an existing Web site promoting CRC screening.

As a guide for the data collection and study design, the Elaboration Likelihood Model by Petty and Cacioppo (1986) was chosen. The Elaboration Likelihood Model suggests two routes by which health messages may be processed: central and peripheral. Central processing of information occurs when individuals are both motivated and have the ability to process and think about the information. The resultant behavior change is long-term, in contrast to peripheral processing, which occurs when a person cannot elaborate on a message extensively but may still be persuaded temporarily by factors unrelated to the actual message content. We believe that the interactive feature unique to Colorectal Web increases elaboration and central processing of the health message leading to the behavior (Brandstatter et al., 2001). Hence the study hypotheses were that the study participants who used Colorectal Web would be more likely to have a clear preference in CRC screening modality and to undergo CRC screening.

Section snippets

Design

We compared Colorectal Web and an existing stand-alone Web site in a RCT. The University of Michigan Institutional Review Board approved this investigation.

Setting

We conducted the RCT in urban (Detroit), suburban (Flint, Saginaw), and rural (St. Joseph, Benton Harbor) communities selected in Michigan. These communities were chosen: (1) because of a high burden of CRC, namely, high incidence, mortality, and percent of cases presenting at more advanced stage than state of Michigan average; and (2) for

Results

From the list of 3000 residential telephone numbers, 1789 unique numbers were called (see Fig. 1). Up to five telephone calls were made for each number if no answer was obtained. No contact was made with 992 numbers for the following reasons: answering machine 521 (52.5%), no answer 254 (25.6%), disconnected number 77 (7.8%), busy 71 (7.2%), not a residence 21 (2.1%), and other reasons for no contact 48 (4.8%). Telephone contacts were made with 797 potentially eligible participants. Of these

Discussion

Exposure to the interactive decision aid Colorectal Web significantly increased the percent of adults screened for CRC compared to the control exposure. The absolute difference was relatively small—only 23 people. However, this difference is significant given the one-time exposure and relatively short-term follow-up of 24 weeks. Some study participants did complete flexible sigmoidoscopy or colonoscopy. So the difference was not solely due to simple, easy to complete procedures like FOBT.

CRC

Conclusions

This type of interactive decision aid may be extremely useful in promoting CRC. Unlike breast, cervical, and prostate cancer screening, CRC offers multiple screening options to the average-risk adult. Adults unscreened for CRC exposed to a new interactive decision aid were significantly more likely to get screened by 24 weeks than a non-interactive electronic program. Future research needs to clarify what specific features of Colorectal Web affect patients' decision making on CRC screening and

Acknowledgments

The authors wish to thank all the participants for their precious time and participation. Kathleen Crawford provided invaluable assistance with recruitment and follow-up of study participants. The collaboration of the Center for Health Communications Research under the leadership of Dr. Victor Strecher and Ed Saunders was critical to the development of the Colorectal Web program.

References (29)

  • E. Frew et al.

    Mass population screening for colorectal cancer: factors influencing subjects' choice of screening test

    J. Health Serv. Res. Policy

    (2001)
  • E.J. Frew et al.

    Eliciting relative preferences for two methods of colorectal cancer screening

    Eur. J. Cancer Care (Engl.)

    (2005)
  • S.J. Fuchs et al.

    A prospective study of family history and the risk of colorectal cancer

    N. Engl. J. Med.

    (1994)
  • A. Jemal et al.

    Cancer statistics, 2005

    CA Cancer J. Clin.

    (2005)
  • Cited by (88)

    • Interactivity in a Decision Aid: Findings From a Decision Aid to Technologically Enhance Shared Decision Making RCT

      2019, American Journal of Preventive Medicine
      Citation Excerpt :

      Although there is a possibility that the usual care patients would have behaved differently in other practices, it is notable that the difference in CRCS rate between the two arms and the usual care group was even more significant when the comparison was limited to the three university-based practices. In the pilot RCT, the interactive DA was compared with a generic website not designed as a DA and was publicly available (http://preventcancer.org/colorectal/; no longer available online).35 The graphics and reading levels were different from the interactive DA.

    • Lay Beliefs About the Accuracy and Value of Cancer Screening

      2018, American Journal of Preventive Medicine
    • Video-based patient decision aids: A scoping review

      2018, Patient Education and Counseling
    View all citing articles on Scopus

    Support: Michigan Department of Community Health and the National Cancer Institute provided funding for this research. Dr. Ruffin's participation was also made possible by support from the National Cancer Institute (K24-CA80846-010). Dr. Fetters' participation was also made possible in part by the generous support of the Robert Wood Johnson Foundation Generalist Physician Faculty Scholars Program.

    View full text