Screening high-risk clusters for developing birth defects in mothers in Shanxi Province, China: application of latent class cluster analysis

Table 3 Standardized characteristics of latent classes among respondents^a

Indicator variables	Cluster 1^a	Cluster 2^a	Cluster 3^a	F ^b	Multiple comparisons^c
1 Maternal delivery age	0.27 (1.07)	−0.09 (0.97)	0.00 (1.00)	113.39^*	1 > 3 > 2
2 Genetic factors	1.73 (3.61)	−0.11 (0.00)	−0.11 (0.00)	4498.98^*	1 > 2, 1 > 3
3 Medical history	3.37 (2.03)	−0.22 (0.00)	−0.22 (0.00)	53915.62^*	1 > 2, 1 > 3
4 Nutrition and folic acid	0.05 (1.02)	−0.14 (0.99)	0.04 (1.00)	100.06^*	1 > 2, 3 > 2
5 Maternal illness	0.45 (1.46)	1.33 (1.18)	−0.46 (0.00)	23225.03^*	2 > 1 > 3
6 Drug use	0.31 (1.40)	0.91 (1.66)	−0.31 (0.00)	6468.21^*	2 > 1 > 3
7 Environmental^e	0.13 (1.11)	0.27 (1.15)	−0.10 (0.92)	445.78^*	2 > 1 > 3
8 Unhealthy lifestyle	0.05 (1.00)	0.17 (0.98)	−0.06 (1.00)	160.82^*	2 > 1 > 3

^aEntries in column = Mean (Standard deviation)
^bOne-way ANOVA was conducted to compare means of each indicator across three clusters, ^* P <0.0001
^cStudent–Newman–Keuls test was used for multiple comparisons among different clusters. Means were significantly different at P < 0.05
^dBold text indicated highest mean of the indicator in three different latent classes
^eEnvironmental = Environmental risk factors

ISSN: 1471-2393