Table 3.

Summary of regression analyses, logistic regression and logistic regression using GEE to account for potential clustering effect.

Logistic regressionLogistic regression using GEE
ParameterEstimatePOR (95% CL)EstimatePOR (95% CL)
Intercept−0.110.56−0.110.56
Study arm - intervention with social support (social pledge) (unadjusted)0.780.00252.17 (1.31–3.59)0.720.00592.06 (1.23–3.44)
Intercept−2.390.0038−2.500.0036
Study arm - intervention with social support (social pledge)0.980.00152.67 (1.46–4.89)0.960.00172.62 (1.44–4.79)
Study site 1 and 20.990.00253.03 (1.38–6.61)1.080.00892.96 (1.31–6.67)
Study site 31.280.00834.04 (1.27–12.88)1.380.013.98 (1.34–11.83)
Study site 4−1.240.010.33 (0.10–1.06)−1.150.040.32 (0.11–0.93)
Study site 5 and 6−1.950.00500.16 (0.03–0.85)−1.800.030.17 (0.03–0.87)
Study site 71.050.00233.24 (1.41–7.43)1.150.00773.15 (1.35–7.32)
Study site 8 (Reference)ReferenceReference
Preeducation: “A colonoscopy is one of the tests to screen for CRC”1.570.044.80 (1.11–20.74)1.590.044.88 (1.05–22.72)
Preeducation: “Having a family history of CRC decreases your risk of CRC”−0.630.060.53 (0.27–1.03)−0.630.060.53 (0.28–1.02)
Preeducation: “Would you make lifestyle changes?”0.260.041.30 (1.02–1.67)0.260.041.29 (1.01–1.65)
  • Abbreviations: CL, confidence limits; CRC, colorectal cancer.