Table 3.

Unadjusted and adjusted associations of IPV measures with depressive symptoms

N = 743 observations (using pooled 12/24 data in GEE models; N = 436 men)		Clinically significant depressive symptoms (PHQ-9 ≥ 10) 14.4% (107/743)
		% (n/N)	Unadjusted PR [95% CI] Overall p valueb	Adjusteda PR [95% CI] Overall p valueb
Lifetime IPV victimization	No/missing	8.9% (38/425)	1	1
	Yes	21.7% (69/318)	2.45 [1.63, 3.67]	2.57 [1.71, 3.86]
			< 0.001	< 0.001
IPV victimization in last year	No/missing	11.1% (70/630)	1	1
	Yes	32.7% (37/113)	2.82 [1.88, 4.22]	2.93 [1.96, 4.40]
			< 0.001	< 0.001
Lifetime IPV perpetration	No/missing	10.8% (65/603)	1	1
	Yes	30.0% (42/140)	2.83 [1.89, 4.22]	2.87 [1.91, 4.32]
			< 0.001	< 0.001
IPV perpetration in last year	No/missing	12.1% (83/688)	1	1
	Yes	43.6% (24/55)	3.40 [2.13, 5.41]	3.47 [2.13, 5.64]
			< 0.001	< 0.001
Combined lifetime IPV victimization/ perpetrationc	Vict. & perp.	31.2% (39/125)	3.69 [2.33, 5.86]	3.87 [2.43, 6.16]
	Undirectional vict.	15.5% (30/193)	1.74 [1.07, 2.82]	1.83 [1.13, 2.98]
	Neither /missing	8.9% (38/425)	1	1
			< 0.001	< 0.001

^aAge (included as four categories: < 25, 25–29, 30–39, 40+), born in the UK, sexual identity (gay or bisexual/straight), university education, and London clinic site

^bp value by Wald test using GEEs. p values< 0.1 are indicated in bold

^cMen who reported unidirectional IPV perpetration were excluded since the number of men reporting this measure (n = 11 at month-12 and n = 4 at month-24) was too small to allow for meaningful analysis, and these men did not fit into the ‘neither/missing’ category