|Multivariate analysis#of predictors of intent to participate in CRC screening with FOBT|
|Model 1 (excluding variable: bowel cancer knowledge score from the model) (n = 89)|
|Predictor||Adjusted OR (95% CI)||p-value|
|Age (>44 years vs ≤44 years^)||4.5 (1.2-16.5)||0.026*|
|Participated in screening in the past (yes vs no^)||6.8 (2.0-23.3)||0.002*|
|Perceived self-efficacy (confident-very confident vs not confident^)||19.8 (5.5-71.8)||<0.0001*|
|Model 2 (excluding the variables: perceived self-efficacy, perceived susceptibility and previous participation in cancer screening from the model) (n = 86)|
|Marital status (married vs unmarried^)||6.91 (2.1-22.5)||0.001*|
|Bowel cancer knowledge score (overall)||<0.0001*|
|Medium vs Low^||9.97 (2.4-41.3)||0.002*|
|High vs Low^||13.6 (3.4-54.0)||<0.0001*|
OR- odds ratio; CI- confidence interval; FOBT- faecal occult blood test.
#Multivariate analysis using binary logistic regression with a backwards step-wise approach. Outcome variable refers to the item, ≤Would you consider screening using FOBT in the next 6-12 months if a kit was given or sent to you?’
^reference category.*significant at 0.05.
The following variables were excluded from multivariate analyses due to:
1. High number of missing responses: Income, Shame, Perceived benefits, and Fatalistic attitude.
2. p > 0.30 in bivariate analysis: Gender, Self-rated current health status, Perceived severity (The thought of bowel cancer scares you and If you had bowel cancer your whole life would change).
3. Taken into account in the total knowledge score variable, Aware bowel cancer/polyps can be asymptomatic.
4. Bowel cancer knowledge score was excluded from Model 1 due to strong association with perceived self-efficacy, perceived susceptibility and participated in screening in the past.
5. Perceived self-efficacy, perceived susceptibility and participated in screening in the past excluded from Model 2 due to correlation with bowel cancer knowledge score.
6. Self-rated English reading ability and Employment status excluded due to strong association with Education.
7. Frequency of doctor visits excluded due to high correlation with Last time saw doctor.
8. Education excluded due to high correlation with Marital status.
9. Marital status excluded from Model 1 due to unstable OR and wide CI.
Christou and Thompson
Christou and Thompson BMC Public Health 2012 12:528 doi:10.1186/1471-2458-12-528