Skip to main content

Table 2 Summary of logistic regression results for the six informative metabolites. The predictors used or selected by the logistic regression model are listed as Predictors selected. Other columns report the model fit results (Max Rescaled R-squared), the relative variance explained (−2LL), the calibration (HL p-value), and the classification ability (AUC and AUC (LOO CV)) of each model

From: A diagnostic biomarker profile for fibromyalgia syndrome based on an NMR metabolomics study of selected patients and controls

Selection method

Predictors selected

-2LL

HL p-value

Max rescaled R-squared

AUC

AUC (LOO CV)

Forward

Creatine & succinic acid

36.15

0.0273

0.47

0.8917

0.8583

Backward

Taurine

40.16

0.6336

0.37

0.8056

0.7556

Stepwise

Succinic acid

40.87

0.5496

0.35

0.8583

0.8306

Forced Entry

Creatine; succinic acid & taurine

29.66

0.0932

0.60

0.8972

0.875