Table 2 Multivariable fractional polynomial logistic regression model for diagnostic prediction of multiple myeloma. (imputed dataset with a total n = 586)
Predictor | Covariate transformation | ß | 95% CI | P-value | ||
|---|---|---|---|---|---|---|
Terms | df | Formula | ||||
Intercept | Â | Â | Â | -2.28 | -2.63, -1.93 | <0.001 |
Hemoglobin | Out | 0 | - | - | - | - |
Log serum creatinine | Linear | 1 | Log creatinine-0.0237 | 1.28 | 0.80, 1.75 | <0.001 |
Log serum globulin | Linear | 4 | Log globulin-0.5-0.8714 | -92.64 | -114.80, -70.49 | <0.001 |
| Â | FP2 | Â | Log globulin-0.5*Log (Log globulin)-0.2400 | -48.14 | -60.13, -36.15 | <0.001 |
Log alkaline phosphatase | Linear | 1 | Log ALP-4.9318 | -0.97 | -1.38, -0.56 | <0.001 |
Serum calcium | Out | 0 | - | - | - | - |