Skip to main content
. 2019 Apr 16;8(6):3206–3215. doi: 10.1002/cam4.2142

Table 3.

AORs associated with HCC by race/ethnicity (aOR associated with metformin use was adjusted for covariates, inverse propensity scores weighting of metformin use for each race/ethnic group, and inverse propensity scores weighting of race/ethnicity group membership)

  NHW NHAA Hispanics P‐values for difference between race/ethnicity groups
aOR (95% CI) aOR (95% CI) aOR (95% CI) NHAA/NHW Hispanics/NHW Hispanics/NHAA
Metformin 0.570 (0.400, 0.810) 0.350 (0.250, 0.470) 0.310 (0.220, 0.430) 0.025 0.008 0.31
Statin use 0.442 (0.317, 0.616) 0.727 (0.298, 1.770) 0.192 (0.046, 0.794) 0.152 0.131 0.06
Beta blocker use 1.185 (0.857, 1.637) 0.466 (0.199, 1.089) 0.164 (0.037, 0.734) 0.022a 0.006a 0.118
Age (y) 1.019 (1.003, 1.034) 1.019 (0.984, 1.055) 0.884 (0.807, 0.969) 0.50 0.001 0.002
Charlson comorbidity 0.981 (0.967, 0.997) 0.981 (0.948, 1.016) 1.131 (1.032, 1.239) 0.50 0.001a 0.002
ΔA1C (%) 1.140 (0.989, 1.316) 1.277 (0.964, 1.692) 0.750 (0.446, 1.261) 0.24 0.064 0.039
ΔBMI (kg/m2) 0.977 (0.962, 0.991) 0.996 (0.984, 1.008) 1.220 (0.796, 1.869) 0.024a 0.154 0.176
ΔLDL (mg/L) 1.013 (1.007, 1.019) 0.997 (0.987, 1.007) 1.026 (0.995, 1.058) 0.003 0.219 0.041a
Neighborhood poverty (%) 1.005 (0.980, 1.031) 0.946 (0.886, 1.010) 0.938 (0.848, 1.037) 0.046 0.096b 0.445
Neighborhood High School completion (%) 1.062 (1.036, 1.088) 0.883 (0.780, 1.000) 0.933 (0.789, 1.103) 0.002 0.067 0.304
Neighborhood public Health insurance coverage (%) 0.992 (0.971, 1.013) 0.982 (0.933, 1.035) 1.016 (0.935, 1.105) 0.365 0.293 0.249
Neighborhood private Health insurance coverage (%) 1.016 (0.997, 1.036) 1.008 (0.951, 1.070) 1.025 (0.952, 1.101) 0.399 0.416 0.367

aOR, adjusted odds ratios; BMI, body mass index; HCC, hepatocellular carcinoma; LDL, low‐density lipoprotein; NHAA, non‐Hispanic African American; NHW, non‐Hispanic white.

a

Model with interactions involving race/ethnicity suggested nonsignificant interaction effect.

b

Model with interactions involving race/ethnicity suggested significant interaction effect.