Skip to main content

Table 2 Fit statistics for latent class models

From: Latent class analysis to predict intensive care outcomes in Acute Respiratory Distress Syndrome: a proposal of two pulmonary phenotypes

No. of classes

Bayesian information criteria°

Entropy*

Number of individual per class

p value†

1

2

3

4

5

1

10,227

 

238

     

2

10,202

0.84

132

106

   

0.001

3

10,317

0.90

65

107

66

  

0.205

4

10,472

0.92

71

75

47

45

 

0.266

5

10,639

0.95

37

59

53

44

45

0.676

  1. °Bayesian information criterion (BIC) is a likelihood function derived criterion for model selection among a set of models; lower BICs indicate better model fit
  2. *Entropy is a measure to assess the degree of association between an individual and a class based on the posterior class membership probabilities; values above 0.8 define good class distinction
  3. †The p value is calculated by means of the bootstrap likelihood ratio test; it addresses if a model with k classes provides increased fit compared to a model with k − 1 classes