Fig. 4

Nomogram for the prediction of postoperative hypoxemia. The logistic regression model were described as a series of straight lines with a common linear scale in the nomogram, with the scale factors of the individual lines given by the coefficients (beta) of the covariates in the model.The distribution of each variable is superimposed on each scale. A representative patient was shown to illustrate how to use the nomogram. Given values of the eight predictors, the patient can be mapped onto the nomogram. Note there is a red dot at each scale, representing the value of each of the 8 predictors for the patient. The total point of the patient was 370, corresponding to a probability of 0.7 for developing postoperative hypoxemia