We found that an automated telephone contact intervention for FOBT screening was effective in increasing CRC screening when compared to usual care in a population aged 51–80 at average risk for CRC. At 6 months, FOBT screening rates were 22.5% in the intervention group and 16.0% for UC. The intervention increased the likelihood of FOBT screening by about 30%, markedly so for patients aged 71–80 compared to those aged 51–60. Although statistically significant, the effect size of the intervention was modest in that less than 25% of the intervention population completed FOBT.
Our study findings are consistent with previous research that reported positive effects of telephone contact interventions. Previous studies have found that telephone contact interventions using live callers were associated with increased screening for CRC (16
) and breast and cervical cancers (19
). These interventions are least effective when using telephone contact alone (OR = 1.7–1.9), (24
) and most effective when using targeted and tailored contacts by telephone and mail (OR = 1.9–4.4), (20
This study is the first randomized trial to examine the effect of an automated telephone contact intervention with CRC screening. Only one other study has evaluated the use of automated calls to improve cancer screening in a similar population overdue for screening: Feldstein (27
) found that automated telephone calls after a reminder mailing and followed by a live call for non-responders were effective in increasing use of mammography, compared to usual care (HR 1.51; 95% CI = 1.40–1.62). In this study, an informational postcard was sent prior to delivery of automated calls, and live-callers were used to schedule mammography for non-responders. The multi-modal approach of this intervention may explain the stronger effect size, compared to results found in our study (HR 1.31; 95% CI = 1.10–1.56). The automated call content regarding mammography was similar to the message regarding FOBT delivered in our study. In both studies, respondents were informed of the importance of screening and encouraged to complete screening.
The main finding—that simple automated telephone calls can increase use of FOBT screening—has large implications for real-world medical practices. These types of patient calls could be implemented in a variety of delivery systems. Moreover, the effect size of the intervention may be even stronger in delivery systems that do not have other active system interventions, such as the clinician electronic visit-based CRC screening prompts at this study site (KPNW).
Although this study did not include a formal cost-effectiveness analysis, automated telephone contact interventions are relatively inexpensive compared to live-person telephone-contact interventions. Live-person interventions are often labor-intensive and require hiring telephone staff for implementation. Conversely, automated telephone contact interventions do not require telephone-based staff and are much less resource-intensive. These less-costly interventions can be delivered through public delivery systems, which are often constrained for resources. Moreover, because the intervention’s maximum effect occurs by 12–14 weeks, automated reminder campaigns do not need to be lengthy, which further reduces delivery costs.
Of particular interest was the finding that individuals aged 71–80 were most likely to complete a FOBT in response to automated telephone calls, compared to younger individuals (aged 51–60). Supporting our findings, a recent study that evaluated telephone-based motivational interviewing followed by telephone reminder calls also found that older individuals were more responsive to mailed reminders for CRC screening (21
). While our study did not isolate the reasons for such higher responsiveness, we hypothesize that older individuals are more receptive to the intervention because they are more sensitive to their increased risks for CRC; furthermore, older people are generally home more often to receive calls.
For two reasons, the responsiveness of older participants (ages 71–80) remains important, despite the recent USPSTF guidelines recommending against CRC screening for populations between ages 76–85. First, the USPSTF recommendation was based on life expectancy concerns and time needed to benefit from screening but didn’t largely apply to this study population, since fewer than 10% of participants were 76–80. Second, the intervention was equally effective even after removing individuals age 76–80.
Interestingly, while both Rosen (28
) and Ferrante (29
) found that obese patients were less likely to complete FOBT screening, our research demonstrated that obesity was not associated with completion of FOBT. It is unclear why our findings differed; more research is needed to fully understand the effect of obesity status on completion of FOBT and CRC screening.
Finally, we found that baseline CRC screening status did not interact with the intervention group. This suggests that the automated messages may be effective for previously unscreened patients and in encouraging repeat CRC screenings. More research is needed to understand what best facilitates a pattern of routine screening.
Although we found increased use of FOBT after implementation of the intervention in KPNW, national trends suggest that colonoscopy is being used over FOBT (8
). Given this trend, we believe that aspects of this automated telephone intervention could be applied to facilitate scheduling of colonoscopy. Feldstein (27
) found that automated telephone reminder calls resulted in increased mammography use.
This study is not without limitations. First, the magnitude of the results may not be generalizable beyond a group-model HMO setting or populations without telephones. However, notwithstanding this limitation, study results are likely applicable to other delivery systems, given that automated reminder calls are a simple direct-to-patient intervention. Second, the study included few racial/ethnic minorities, which precluded our ability to assess differences in FOBT screening by race/ethnicity. However, the study population was representative of the KPNW population, which is predominantly white, and most members (99%) have telephones. Nevertheless, implementation of automated telephone interventions may have generalizability concerns, since certain sub-populations, such as African-Americans and Hispanics, are more likely to not have telephones compared to the overall U.S. population (30
Findings from this research suggest several additional areas for future research. First, work is needed to implement and evaluate automated CRC telephone contact interventions in a variety of public and private delivery systems. Second, more research is needed to better understand provider and patient characteristics associated with non-completion of FOBT or alternatively a pattern of routine screening. Moreover, given the modest effect size, further methods are needed to strengthen the intervention, perhaps by combining mailings and/or live calls, or switching to a fecal test that is easier to use, such as the fecal immunochemical test (FIT).
Last, given advancements in communication technologies, alternative methods for delivering screening reminders, such as e-mail and text messaging, should be tested, especially in younger populations.