Predictors of colorectal cancer screening-related decisional conflict in a general population

Background: Decisional conflict about colorectal cancer screening (CRCS) options may act as a barrier, limiting screening adherence. Little is known about factors related to decisional conflict in the general population.

Methods: In December 2012, a self-administered questionnaire was mailed to randomly selected adults, 50-75 years, in six rural Minnesota communities using a probabilistic sampling scheme (N=2,150). Questions ascertained CRCS decisional conflict (validated scale/subscales); CRCS-related behavior, intentions, test preference, and options discussed with clinician; other screening behaviors; and demographics. Descriptive statistics were calculated and multivariate logistic regression performed to examine respondent characteristics associated with decisional conflict and subscale scores among average-risk respondents (N=1,292).

Results: Overall, 74% of respondents were 50-64 years, 53% female, and 76% reported CRCS per guidelines. Mean decisional conflict was 28.3 (standard error: 0.53; range: 0-100, higher score denotes higher decisional conflict). Mean subscale scores (i.e., informed, uncertainty, support, values clarity, and effective decision) were 22.2-34.9. Predictors of decisional conflict included: CRCS-related behavior, intentions, test preference, and options discussed as well as insurance, age, and other screening behaviors. Respondents without a CRCS test preference were 4.81 times more likely to have decisional conflict than respondents who preferred endoscopy (95% CI: 3.20, 7.23). Compared to those who strongly intended to have CRCS, respondents who somewhat intended or did not intend to have CRCS were respectively 4.22 (95% CI: 3.05, 5.82) and 3.62 (95% CI: 2.16, 6.07) times more likely to have decisional conflict. In subscale models, CRCS intention, test preference, and age were significant predictors.

Conclusions: While mean decisional conflict was moderate in a general population of average-risk adults, decisional conflict varied significantly by respondent characteristics. Knowing respondent characteristics associated with decisional conflict has implications for future CRCS interventions.