Table 3.

Multinomial logistic regression analysis exploiting ORs-like to evaluate the predictive strength of the derived clusters on the overall outcomes of patients

	ORs-like	SE	95% CI	P value
Monocyclic pattern
Base outcome
Polycyclic pattern
Cluster 1	Reference
Cluster 2	1.42	0.43	0.78 to 2.57	0.252
Cluster 3	0.66	0.31	0.27 to 1.64	0.373
Cluster 4	1.99	0.72	0.98 to 4.04	0.058
Chronic pattern
Cluster 1	Reference
Cluster 2	1.31	0.44	0.68 to 2.53	0.424
Cluster 3	0.40	0.24	0.12 to 1.31	0.132
Cluster 4	1.45	0.60	0.65 to 3.26	0.366
Mortality
Cluster 1	Reference
Cluster 2	4.56	3.73	0.92 to 22.63	0.064
Cluster 3	Not estimable
Cluster 4	11.03	9.10	2.19 to 55.60	0.004
χ2 = 21.68; p=0.010
Monocyclic pattern
Base outcome
Polycyclic pattern
Cluster 1	Reference
Cluster 2	1.46	0.45	0.80 to 2.67	0.220
Cluster 3	0.65	0.30	0.26 to 1.62	0.358
Cluster 4	2.00	0.73	0.98 to 4.07	0.057
Male sex	0.96	0.25	0.58 to 1.60	0.884
MAS	1.25	0.47	0.60 to 2.61	0.552
Chronic pattern
Cluster 1	Reference
Cluster 2	1.28	0.44	0.66 to 2.51	0.468
Cluster 3	0.41	0.25	0.13 to 1.35	0.144
Cluster 4	1.36	0.57	0.60 to 3.07	0.461
Male sex	1.67	0.49	0.94 to 2.97	0.079
MAS	1.21	0.52	0.52 to 2.82	0.652
Mortality
Cluster 1	Reference
Cluster 2	5.99	5.07	0.94 to 31.48	0.064
Cluster 3	Not estimable
Cluster 4	9.55	8.07	1.82 to 49.99	0.008
Male sex	2.36	1.29	0.80 to 6.91	0.118
MAS	5.99	3.52	1.89 to 18.95	0.002
χ2 = 37.48; p=0.001

Bold values are statistically significant results.

MAS, macrophage activation syndrome.