Logistic regression identifying independent predictors of diabetes (KDM and NDM)* in a combined data set of AusDiab (1999–2000) and Busselton (1981)legend
| Variable as diabetes predictor . | β . | 95% CI for β . | Odds ratio (95% CI) . | P value . |
|---|---|---|---|---|
| Sex (male versus female) | 0.273 | 0.14–0.41 | 1.31 (1.15–1.50) | <0.001 |
| Age (per year) | 0.063 | 0.058–0.069 | 1.07 (1.06–1.07) | <0.001 |
| Survey (AusDiab versus Busselton) | 0.851 | 0.64–1.06 | 2.34 (1.90–2.90) | <0.001 |
| BMI (per unit increase) | 0.117 | 0.10–0.13 | 1.12 (1.10–1.14) | <0.001 |
| Constant | −8.90 | −9.45 to −8.35 |
| Variable as diabetes predictor . | β . | 95% CI for β . | Odds ratio (95% CI) . | P value . |
|---|---|---|---|---|
| Sex (male versus female) | 0.273 | 0.14–0.41 | 1.31 (1.15–1.50) | <0.001 |
| Age (per year) | 0.063 | 0.058–0.069 | 1.07 (1.06–1.07) | <0.001 |
| Survey (AusDiab versus Busselton) | 0.851 | 0.64–1.06 | 2.34 (1.90–2.90) | <0.001 |
| BMI (per unit increase) | 0.117 | 0.10–0.13 | 1.12 (1.10–1.14) | <0.001 |
| Constant | −8.90 | −9.45 to −8.35 |
β is the regression coefficient for the exponential. Sex, age, survey site, and BMI were the only variables entered into the regression. NDM was identified by 2hPG only in both surveys.