Table 2.
Univariate and multivariable logistic regression models associated with bladder cancer in the development cohort. N = 3539 (bladder cancer = 285)
| Predictor | Unit | Univariate | Multivariable | ||
|---|---|---|---|---|---|
| IQR‐ORa (95% CI) | LR χ² (d.f., P) | IQR‐ORa (95% CI) | Δχ² (d.f., P) | ||
| Age | years | 2.931 (2.377, 3.614) | 120.07 (1, <2.2e‐16) | 2.892 (2.319, 3.605) | 120.07 (1, <2.2e‐16) |
| Haematuria | Nonvisible | 1 (ref) | |||
| Visible | 4.526 (3.127, 6.551) | 89.007 (1, <2.2e‐16) | 3.850 (2.629, 5.638) | 84.119 (1, <2.2e‐16) | |
| Smokerb | Nonsmoker | 1 (ref) | |||
| Ex‐smoker | 1.917 (1.453, 2.531) | 1.512 (1.132, 2.018) | |||
| Current smoker | 1.619 (1.112, 2.357) | 2.568 (1.719, 3.836) | |||
| Missingb | 1.223 (0.621, 2.410) | 22.638 (3, 4.8e‐05) | 1.283 (0.636, 2.585) | 24.257 (3, 2.2e‐05) | |
| Gender | Female | 1 (ref) | |||
| Male | 2.960 (2.196, 3.990) | 60.044 (1, 9.3e‐15) | 1.779 (1.298, 2.438) | 13.812 (1, 2.0e‐04) | |
| Ethnicityc | White | 1 (ref) | |||
| Non‐White | 0.490 (0.248, 0.967) | NSS | |||
| Missingb | 0.496 (0.274, 0.896) | 11.097 (2, 0.00389) | |||
CI, confidence interval; IQR, interquartile range; LR, likelihood ratio; NSS, not statistically significant; OR, odds ratios; Δχ2, delta chi‐square (degrees of freedom, P‐value), terms added sequentially (first to last); χ2, chi‐square test (degrees of freedom, P‐value).
aInterquartile range odds ratios for continuous predictors and simple odds ratios for categorical predictors. Model LR χ² (d.f, P) = 242.257 (6, <2.2e‐16). bFourth category for smoking ‘missing’ was created and compared to nonsmoker category in the logistic regression analysis. cThird category for ethnicity and compared to White category in the logistic regression analysis. Harrell's c‐index = 0.768 (95%CI: 0.741, 0.795).