Residents (N = 345) were recruited in 8 rural Kentucky counties to enroll in the study and receive an FIT kit, yielding a participation rate of 90.7%. Recruitment occurred in 8 economically distressed counties of rural Appalachia. Based on rural-urban commuting area codes,¹¹ 7 of these counties were ranked extremely rural (code of 10), with 1 being ranked less extreme in its rural classification (ie, code of 8). Eligibility criteria were: (1) being between 50 and 75 years of age (however, those 30 and older with a first-degree relative previously diagnosed with CRC were also eligible) and (2) reporting a previous CRC screening history that is not compliant with recommendations.

Recruitment methods included flyers posted in local health departments. Flyers described the FIT kit as a test to screen for CRC and noted that people enrolling in this study would be compensated for their time. Staff were trained to describe the FIT kit as a simple at-home test that can determine the presence of precancerous polyps. Participants were also recruited at community outreach events. Direct referrals were provided by local health departments. Additionally, project staff conducted outreach to senior citizen centers throughout the 8-county area and at health and wellness events sponsored by local employers.

Participants completed a paper-and-pencil survey instrument prior to receiving instructions for specimen collection. Survey questions collected demographic and health information. All study procedures were approved by the Office of Research Integrity at the University of Kentucky.

A 4-item scale assessed perceptions of fatalism regarding CRC. Items were: (1) “I am likely to develop colorectal cancer in my lifetime,” (2) “I am worried that I will develop colorectal cancer in my lifetime,” (3) “If it was meant for me to develop colorectal cancer there is nothing that I can do about it,” and (4) “There is nothing I can do to reduce my risk of developing colorectal cancer.” The scale produced an excellent interitem reliability coefficient (Cronbach’s alpha = .93). This measure has been tested and used previously with populations of rural Appalachians.¹²

Associations between dichotomous correlates and return of FIT kits were tested using chi-square tests. Associations between continuous correlates and return were tested by independent group t tests. Correlates obtaining significance at the bivariate level were entered into a stepwise logistic regression model to calculate adjusted OR and their 95% CI.

Table 1 displays the bivariate associations between the dichotomous correlates and the return of FIT kits. As shown, 4 of the 8 correlates obtained significance. Those indicating an annual income of less than $15,000 per year were more likely to return the kits (89.8%) than those earning more than this amount (73.8%). Also, those indicating they did not have a regular health care provider were more likely to return their kits (92.3%) compared with those who indicated having a regular provider (79.9%). Persons rating themselves as overweight or obese were less likely to return kits (76.4%) than those not having this perception of their weight (87.3%). Finally, people who did not graduate from high school were more likely to return their kits (88.6%) than those with at least a high school education (79.2%).

Table 2 displays the bivariate associations between the 2 continuous correlates and the return of FIT kits. As shown, neither age nor the scale measure of CRC fatalism was associated with return of FIT kits.

The multivariate model achieved significance (χ² = 191.4, 2 df, P < .001) and had adequate Goodness of Fit (χ² = 1.43, 2 df, P = .49). The final model included only 2 correlates. Those indicating an annual income of less than $15,000 were 2.85 times more likely to return their kits (95% CI: 1.56–5.24; P < .001). Also, those not perceiving themselves to be overweight/obese were 1.95 times more likely to return their kits (95% CI: 1.07–3.55; P = .029). Not having a regular health care provider (P = .06) and not graduating from high school (P = .49) were not retained in the final iteration of the model.

Our findings are limited by the use of a convenience sample. Furthermore, generalization of the study findings to other underserved, rural populations is not possible. Also, our selection of covariates was limited; clearly, unmeasured confounding may have occurred.

The prevailing strategy for promoting CRC screening is embedded in the medical paradigm of serving patients in clinics. A community-based approach to delivering FIT kits challenges the assumptions underlying current CRC screening methods, and an in-person, community-based screening program may have important implications for isolated populations with limited access to health care. This low-intensity, low-resource service delivery intervention transcends geography and the limitations of rural health care systems. A key challenge will be boosting return rates among those who are not classified as low-income and among those who perceive themselves to be overweight or obese.