Table 2.

Fit statistics of latent class analysis results using clinical and protein biomarker data at study enrollment

Number of subtypes	BIC	Entropy	Number of participants per subtype				P value*
			N1	N2	N3	N4
1	29,780		400
2	29,454	0.83	292	108			0.01
3†	29,370	0.82	189	170	41		0.10
4†	29,328	0.83	152	132	75	41	0.11

BIC Bayesian Information Criterion

*Calculated using Vuong–Lo–Mendell–Rubin method to test whether a model with an additional class fits better than a model with one fewer class

^†Model failed to replicate the maximum likelihood with up to 400 random starts