Table 1.
Odds ratios for variables in the Logistic thin plate regression spline model
| OR (95% CI) | |
|---|---|
| Sex | |
| Male | reference |
| Female | 1.00 (0.96, 1.05) |
| Age (per 10 years) | 0.84 (0.82, 0.85) |
| Race | |
| White | reference |
| Black | 1.78 (1.63, 1.94) |
| Asian | 0.66 (0.57, 0.77) |
| American Indian | 1.25 (1.00, 1.56) |
| Hawaiian or Pacific Islander | 1.29 (0.77, 2.18) |
| Unknown | 0.81 (0.68, 0.96) |
| Ethnicity | |
| Non-Hispanic | reference |
| Hispanic | 0.79 (0.69, 0.90) |
| Unknown | 0.81 (0.68, 0.96) |
| BMI (per 5 kg/m2) | 1.18 (1.16, 1.20) |
| Encounter days in EHR dataset (per 30 days) | 1.05 (1.04, 1.05) |
| Distance to clinic (per 10 mile) | 1.02 (1.00, 1.03) |
| Sparse Principal Component 1 (per 5 units) | 1.00 (0.96, 1.05) |
| Sparse Principal Component 2 (per 5 units) | 0.95 (0.89, 0.99) |
| Sparse Principal Component 3 (per 5 units) | 0.94 (0.86, 1.03) |
| Sparse Principal Component 4 (per 5 units) | 1.13 (1.01, 1.27) |
OR, odds ratio, CI, confidence interval, BMI, body mass index, EHR, electronic health record Age, BMI, encounter days in EHR, distance to clinic, and sparse principal component odds ratios are scaled by 10 years, 5 kg/m2, 30 days, 10 miles, and 5 units, respectively