Domestic violence and mental health: a cross-sectional survey of women seeking help from domestic violence support services

Figures & data

Table 1 Recruitment

	Cardiff					Bristol				Total
	Women's centre	Community outreach	Residential	Total	% of entered	Community outreach	Residential	Total	% of entered	N	% of entered
Entered service	444	534	317	1295		519	126	645		1940
Screened	162	408	209	779	60	372	121	493	76	1272	66
Ineligible 31		73	20	124	10	49	30	79		203	10
Drugs and alcohol 6		9	1	16		7	0	7		23	1
Language barrier 14		14	7	35		20	27	47		82	4
Male 1		0	0	1		0	0	0		1	0
Psychotic 1		5	1	7		7	0	7		14	1
Psychological therapy 9		46	11	66		15	3	18		84	4
Eligible but not approached	115	51	52	218	17	67	19	86	13	304	16
One-off contact	29	0	0	29						29	1
SPA capacity	60	53	42	155		50	17	67		222	11
Researcher capacity	22	8	6	36		14	1	15		51	3
Other	4	19	4	27		3	1	4		31	2
Unable to contact/declined	4	59	58	121	9	65	10	75	12	196	10
Approached	16	284	137	437	34	274	81	355	55	792	41
Did not consent to contact	6	117	15	138	11	115	26	141	22	279	14
Consented to contact	10	167	122	299	23	159	55	214	33	513	26
Met with researcher	6	108	64	178	14	92	45	137	21	315	16
Recruited	4	95	47	146	11	86	31	117	18	263	14
Not recruited	2	13	17	32	2	6	6	20	3	52	3
Wanted counselling	1	2	3	6		0	–	0		6	0
Time commitment	1	6	5	12		2	2	4		16	1
Other	0	5	9	14		4	4	16		30	2
Withdrawal	–	–	–	0		2	1	3		3	0
Total	4	95	47	146	11	84	30	114	18	260	13

SPA, specialist psychological advocates.

Table 2 Sociodemographic profile of the sample

	Mean	Median	%	Minimum	Max	Standard deviation (IQR)	N
Age	33	31		18	63	17	248
Maximum education level		City & Guilds and similar		None	Bachelor's degree or higher	(GCSE to A-level)	233
Income bracket		Up to £10,999		Up to £10,999	More than £60,000	(Up to £10,999, to £11,000–£20,999)	156
White			87			34%	253
Currently in a relationship			20			40%	250
Perpetrator is current partner			23			42%	236
Is a parent			81			39%	254
Has children under 4 years of age			37			48%	260
Works in the household			38			49%	237
Not in formal employment (excluding retirees and students)			78			42%	236
Hazardous drinking (AUDIT-C ≥ 3)			54			50%	251
Smoked cannabis in past 12 months			26			44%	245
Witnessed DVA as a child			52			50%	257
Was abused as a child			50			50%	257
Had a mental health problem in the past			82			38%	251

IQR, interquartile range; DVA, domestic violence and abuse.

Table 3 Exposure to abuse

	CAS measure
	Mean	Median	%	SD	Minimum	Maximum	Interquartile range	N
Severe abuse	6	3		8	0	33		248
Emotional abuse	31	31		16	0	55		248
Physical abuse	13	11		10	0	35		248
Harassment	8	7		6	0	20		247
Total abuse	57	49		34	0	136		245
Severe abuse >1			69	46%				248
Emotional abuse >3			96	20%				248
Physical abuse >1			92	28%				248
Harassment >2			86	35%				247
Total abuse >3			97	18%				245
Type of abuse, ordinal measure		SCA			None	SCA	(Physical and other – SCA)	251
Recency		In the past 3 months			More than 1 year ago	Past month	Between 6 months and less than 1 month ago	243
Length of exposure		Up to 3 years			Never	More than 5 years	Between (up to) 1 to more than 5 years	244

CAS, Composite Abuse Scale; SCA, severe combined abuse.

Table 4 Mental health, health utility, and quality of life measures

	Mean	SD	Median	Minimum	Maximum	N
CORE-OM
Subjective well-being	24	8	25	3	40
Percentage with mean ≥1.77	74%					259
Problems	22	10	23	0	40
Percentage with mean ≥1.62	70%					259
Functioning	20	8	20	2	36
Percentage with mean ≥1.3	80%					259
Risk	4	7	0	0	30
Percentage with mean ≥0.31	41%					259
CORE-OM	18	7	19	2	35
Percentage with mean ≥1.29	76%					259
Depression, anxiety, stress
Depression (PHQ-9)	14	7	14	0	27
PHQ-9 score >9	72%					258
Anxiety (GAD-7)	13	6	14	0	21
GAD-7 score >9	70%					255
Post-traumatic stress (PTSD test for civilians)	26	12	27	0	50
PTSD score ≥17	77%					256
Utility
EQ-5D-5L	0.6	0.3	0.7	−0.2	1.0	249
Quality of life
SF-12 Aggregate physical health	48	12	51	19	68	236
SF-12 Aggregate mental health	31	14	30	6	62	236

CORE-OM, Clinical Outcomes in Routine Evaluation – Outcome Measure; PHQ-9, nine-item Patient Health Questionnaire; GAD-7, seven-item Generalised Anxiety Disorder questionnaire; PTSD, posttraumatic stress disorder.

Table 5 Associations between mental health and health state utility and severity of exposure to violence

Variable	Coefficient	Adjusted coefficient
Measures of mental health
CORE-OM	0.081	0.1
95% CI	(0.050, 0.10)	(0.043, 0.2)
p value	0.004	0.013
N	245	174
PTSD	0.2	0.2
95% CI	(0.1, 0.2)	(0.1, 0.2)
p value	0.004	0.002
N	243	172
Measures of health state utility
EQ-5D	−0.0028	−0.0037
95% CI	(−0.0038, −0.0018)	(−0.0052, −0.0023)
p value	0.003	0.003
N	238	170
Quality of life
Aggregate physical health (T score)	−0.080	−0.093
95% CI	(−0.12, −0.040)	(−0.17, −0.012)
p value	0.008	0.035
N	228	165
Aggregate mental health (T score)	−0.10	−0.12
95% CI	(−0.18, −0.026)	(−0.23, 0.015)
p value	0.023	0.036
N	228	165

The first column of results reports coefficients from a normal univariable regression of the mental health or utility variable (COREOM, PTSD, SF-6D, EQ-5D) on exposure to abuse as captured by a continuous measure of the Composite Abuse Scale (CAS); the second column reports coefficients from a regression of CORE-OM, EQ-5D, SF-6D, and PTSD on CAS, and sociodemographic confounders (age, number of live-in children under 4, maximum level of education, use of drugs and alcohol, and work status) as well as measures of recency and length of exposure, previous mental health issues, exposure to non-IPV domestic abuse, and exposure to child abuse.

Table 6 Associations between binary mental health states and severity of exposure to violence

Variable	Odds ratios	Adjusted odds ratios
PHQ-9 ≥ 10	1.02	1.03
95% CI	(1.01, 1.03)	(0.99, 1.05)
p value	0.002	0.113
N	244	174
GAD-7 ≥ 10	1.02	1.03
95% CI	(1.01, 1.02)	(1.01, 1.05)
p value	<0.0001	<0.0001
N	241	174
PTSD ≥ 17	1.03	1.03
95% CI	(1.02, 1.03)	(1.03, 1.04)
p value	<0.0001	<0.0001
N	243	172

The first column of results reports odds ratios from a univariable logistic regression of the mental health variable (PHQ-9, GAD-7, PTSD) on exposure to abuse as captured by a continuous measure of the Composite Abuse Scale (CAS); the second column reports adjusted odds ratios from a logistic regression of PHQ-9, GAD-7, and PTSD on CAS, and sociodemographic confounders (age, number of live-in children under 4, maximum level of education, use of drugs and alcohol, and work status) as well as measures of recency and length of exposure, previous mental health issues, exposure to non-IPV domestic abuse, and exposure to child abuse.