Table 5.

Maternal childhood maltreatment type and associations with maternal perinatal IPV combined latent class membership (n = 832).

	Childhood maltreatment typea	ORb	AORc	95%CI	p-value
		High vs Low class
Combined IPV classes	Sexual Abuse	4.23	2.97	1.50; 5.91	0.002
	Emotional Neglect	2.45	1.99	1.11: 3.59	0.020
	Emotional Abuse	3.89	2.46	1.31; 4.61	0.004
	Physical Neglect	4.5	2.93	1.63; 5.31	<0.001
	Physical Abuse	4.71	3.61	1.85; 7.01	<0.001
		Moderate vs Low class
	Sexual Abuse	1.48	1.33	0.79; 2.27	0.296
	Emotional Neglect	1.01	0.96	0.55; 1.66	0.890
	Emotional Abuse	1.75	1.64	0.99; 2.75	0.061
	Physical Neglect	1.36	1.16	0.73; 1.86	0.540
	Physical Abuse	2.06	2.23	1.26; 3.93	0.006
		Decreasing vs Low class
	Sexual Abuse	2.37	1.93	0.94; 4.00	0.074
	Emotional Neglect	2.26	1.96	1.04; 3.66	0.037
	Emotional Abuse	2.33	1.75	0.92; 3.35	0.090
	Physical Neglect	1.36	1.16	0.73; 1.86	0.540
	Physical Abuse	2.02	1.46	0.66; 3.27	0.350

^aChildhood maltreatment sub-types included as dichotomous exposure (data collected antenatally at 28–32 weeks’ gestation).

^bUnadjusted multinomial logistic regression model. Odds ratios generated for each childhood maltreatment type and association with maternal membership in indicated IPV exposure class (high versus low, moderate versus low and decreasing versus low exposure over time). All childhood maltreatment types were included in separate models because of collinearity between maltreatment sub-types.

^cAdjusted multinomial logistic regression models were controlled for key sociodemographic variables (maternal age, maternal income, maternal employment, maternal education, site of enrolment and partnership status, collected antenatally at 28–32 weeks’ gestation). All childhood maltreatment types were included in separate models because of collinearity between maltreatment sub-type.