Phenotyping Chronic Pelvic Pain Based on Latent Class Modeling of Physical Examination

Table 3

(a) Statistical fit criterion for multiple latent class models of abdominal wall testing site pain thresholds. (b) Statistical fit criterion for multiple latent class models of vulvar testing site pain thresholds. (c) Statistical fit criterion for multiple latent class models of pelvic floor muscle testing site pain thresholds. (d) Statistical fit criterion for multiple latent class models of internal pelvic organ testing site pain thresholds.

(a)  

Classes	Parameters	Log-likelihood	BIC	Entropy	Size of smallest class, 𝑛 (%)
1	28	−6503	13177	N/A	N/A
2	43	−4771	9806	0.952	129 (27.9%)
3	58	−4310	8977	0.908	95 (20.6%)
4	73	−4142	8732	0.906	43 (9.4%)
5	88	−4007	8554	0.924	40 (8.6%)
6	103	−3898	8429	0.931	11 (2.5%)

(b)   

Classes	Parameters	Log-likelihood	BIC	Entropy	Size of smallest class, 𝑛 (%)
1	12	−6170	12413	N/A	N/A
2	19	−5366	10849	0.989	55 (11.6%)
3	26	−5134	10428	0.984	20 (4.2%)
4	33	−5005	10213	0.980	13 (3.0%)
5	40	−4827	9900	0.949	13 (3.0%)
6	47	−4754	9798	0.933	13 (3.0%)

(c)  

Classes	Parameters	Log-likelihood	BIC	Entropy	Size of smallest class, 𝑛 (%)
1	12	−7941	15968	N/A	N/A
2	19	−7208	14552	0.984	122 (26.0%)
3	26	−6985	14154	0.962	23 (4.9%)
4	33	−6867	13968	0.959	10 (2.1%)
5	40	−6762	13807	0.933	5 (1.1%)
6	47	−6678	13689	0.930	5 (1.1%)

(d)  

Classes	Parameters	Log-likelihood	BIC	Entropy	Size of smallest class, 𝑛 (%)
1	8	−4657	9364	N/A	N/A
2	13	−4188	8456	0.902	101 (22.1%)
3	18	−4073	8257	0.899	66 (14.2%)
4	23	−3968	8078	0.851	56 (12.1%)
5	28	−3872	7916	0.872	31 (6.8%)
6	33	−3789	7780	0.846	19 (4.2%)

BIC: Bayesian Information Criterion.