Skip to main content

Table 1 Latent profile analysis for choosing the best number of profiles

From: Identification of subclasses of sepsis that showed different clinical outcomes and responses to amount of fluid resuscitation: a latent profile analysis

       Number of patients in each latent profile  
Number of classes AIC CAIC BIC SABIC Entropy 1 2 3 4 5 6 7 P value
2 2,254,774.736 2,255,429.502 2,255,353.502 2,255,111.98 0.975 12,865 (86) 2128 (14)       0.001
3 2,244,946.905 2,245,825.67 2,245,723.67 2,245,399.522 0.955 11,525 (77) 1580 (11) 1888 (13)      0.001
4 2,235,891.639 2,236,994.403 2,236,866.403 2,236,459.629 0.953 10,340 (69) 1305 (9) 1635 (11) 1713 (11)     0.001
5 2,232,244.359 2,233,571.121 2,233,417.121 2,232,927.722 0.887 8010 (53) 1188 (8) 2561 (17) 1552 (10) 1682 (11)    0.063
6 2,228,493.639 2,230,044.4 2,229,864.4 2,229,292.374 0.886 7586 (51) 1200 (8) 2367 (16) 1476 (10) 1599 (11) 765 (5)   0.075
7 2,223,495.867 2,225,270.627 2,225,064.627 2,224,409.976 0.887 7108 (47) 1080 (7) 2272 (15) 1293 (9) 1051 (7) 1505 (10) 684 (5) 0.081
  1. P value was reported for the bootstrap likelihood ratio test comparing the current model (k class) to the model with k-1 class.
  2. Abbreviations: AIC Akaike information criterion, BIC Bayesian information criteria, CAIC consistent Akaike information criterion, SABIC sample size adjusted Bayesian information criteria