Abstract
Summary
Domestic violence (DV) occurs frequently within the UK, with one in four women and one in six men experiencing DV during their life. DV is the leading cause of morbidity for women aged 19–44 years and is associated with sexual health problems. However, few data exist on the prevalence of DV in UK genitourinary (GU) medicine settings. An anonymous questionnaire was self-completed by patients (n = 476/500) attending Bournemouth GU medicine clinic during July 2009 to explore associations between lifetime DV and sociodemographic, sexual and behavioural factors. Overall, 21% (98/472) reported that they had ever been abused by a partner (12% [27/229] of men and 29% [71/243] of women). Logistic regression highlighted that being female, having children/dependants and use of illicit drugs were the most important factors associated with lifetime DV. Regular staff training on DV is recommended to increase awareness and signposting to relevant services.
Keywords
INTRODUCTION
The Home Office (page 7) defines domestic violence (DV) as: ‘Any incident of threatening behaviour, violence or abuse (psychological, physical, sexual, financial or emotional) between adults who are or have been intimate partners or family members, regardless of gender or sexuality’. 1 DV has a widespread occurrence in the UK, with one in four women and one in six men experiencing DV during their lifetime. 2 One incident of DV is reported to the police every minute 3 and two women are killed every week by a current or former partner. 4 Based on the 2009/2010 British Crime Survey, 7% of women aged 16–59 years were victims of domestic abuse in the previous year compared with 4% of men. 5 Partner abuse was the most common type of domestic abuse, experienced by 5% of women and 3% of men. 5
DV is the leading cause of morbidity for women aged 19–44 years 2 and has an association with sexual health problems.5,6 A World Health Organisation (WHO) multicentre study revealed that factors influencing DV risk were similar across areas, despite wide variations in DV prevalence. 7 Alcohol abuse, co-habitation, pregnancy, younger age, lower socioeconomic status and growing up with DV were common risk factors.7,8 A recent study from a genitourinary (GU) medicine clinic in Australia found an association between intimate partner violence, sex work, prior sexually transmitted infections (STIs) and lower condom usage, suggesting that sexual health clinics might be appropriate venues to assess women for intimate partner violence. 9
However, to the authors’ knowledge, few data exist on the prevalence and factors associated with DV in UK GU medicine settings. 10 Loke et al. 10 highlighted high lifetime prevalence of DV among female patients within an inner city GU medicine service. Furthermore, there are no published UK guidelines on how best to manage DV if identified in GU medicine. This is despite GU medicine being an optimal environment to assess for DV due to staff being highly skilled in asking sensitive questions about sexual abuse.
Currently, Bournemouth GU medicine clinic does not routinely assess patients for DV, despite having referral mechanisms in place if DV is identified or reported. In order to evaluate service provision, the first step was to determine local prevalence of DV. The main aims of this study were to establish both the lifetime and the preceding 12 months’ prevalence of DV in both male and female patients and to investigate the associations between demographic, socioeconomic, sexual and behavioural characteristics and the self-reporting of lifetime DV.
METHODS
An anonymous, cross-sectional questionnaire was self-completed by clinic attendees from the coastal town of Bournemouth during July 2009. This study was registered at the local Clinical Audit Department since it was considered a service development activity by the Ethics committee.
The clinic is located within the District General Hospital and offers both Sexual Health and HIV services. The service offers a mix of walk-in and appointment clinics. During the month of July 2009, a total of 2259 patients (55% male) were seen.
A self-complete questionnaire was offered after patients’ consent to the first 500 patients (n = 250 females, n = 250 males). Questionnaire completion took place in a separate room to ensure privacy. DV leaflets were made available for all patients and those who requested further help were signposted to health advisers. DV was assessed using a modified version of the Abuse Assessment Screening tool. 11
The following data were collected:
Demographics Socioeconomics Sexual behaviour Drug/alcohol behaviour Reporting and the type of DV (similar to the Home Office definition on different types of DV) – physical (hit/punched/kicked/other injury), emotional (yelling insults/threats/abusive language), sexual (unwanted sex), financial (control of finances) and spiritual (not allowing you to practice your religion/forcing you to practice a different religion), whether carried out in the previous 12 months, who by, abuse because of STIs or fear to disclose DV support-seeking intentions.
Frequencies were produced to provide descriptive information for the whole sample (median and interquartile range for age at first sex and number of sexual partners in the preceding year), as well as for men and women separately given gender is such an important component of DV affecting its prevalence, nature and severity. Logistic regression was used to investigate the associations between demographics (gender, age, ethnicity, marital status and children/dependants to care for), socioeconomics (occupation and area of residence), sexual behavioural factors (age at first sex, sexual orientation, number of sexual partners, condom use, other contraceptive methods and diagnosis of STIs in the preceding year) and drug/alcohol behavioural factors (illicit drug use and consumption of alcohol over the recommended weekly levels) and the reporting of lifetime DV (dependent variable) using Predictive Analytics Software (PASW) version 18.0 (Education IBM Software Group, Surrey, UK).
The model-building process was in stages, adding each set of variables at a time, but is not presented here for brevity. To ensure there were enough cases in each category, ethnicity was divided into ‘White’ versus ‘Other’ and the two patients under 16 years old and the four patients who identified themselves as ‘Retired’ were excluded as the coefficients created unusually large standard errors. The models were therefore based on a sample of 390 male and female respondents with complete data for the variables of interest. Although it is noted that gender is such an important issue, with the risk factors for DV being very different by gender, separate models for men/women would not be able to identify associations by gender. Indeed, separate models by men and women, as well as a combined model with interactions had standard error/confidence interval (CI) issues because of too few cases in some cells. Furthermore, separate models by gender would have less statistical power with smaller sample sizes and issues of multiplicity.
RESULTS
Percentage and frequencies of characteristics of domestic violence and support points of access as reported by 476 respondents, as well as separately for men and women, to an anonymous self-completed questionnaire, Bournemouth Genitourinary Medicine Clinic, July 2009
N/A = not applicable
Valid percentage and n varies because of item non-response
Descriptive characteristics of demographic, socioeconomic, sexual behaviour and drug/alcohol behavioural characteristics as reported by 476 respondents, as well as separately for men and women, to an anonymous self-completed questionnaire, Bournemouth Genitourinary Medicine Clinic, July 2009
Valid percentage and n varies because of item non-response. Median and interquartile range presented where data were non-normally distributed
Number and odds ratios of demographic, socioeconomic and drug/alcohol behavioural characteristics from logistic regression of lifetime domestic violence as reported by 390 male and female respondents to an anonymous self-completed questionnaire, Bournemouth Genitourinary Medicine Clinic, July 2009
Analyses based on n = 390 with complete data for the variables of interest and under 16-year-olds and retired excluded. Illicit drug use includes cocaine, cannabis, etc.; recommended units per week for alcohol are up to 14 units for women and up to 21 units for men. Nagelkerke R2 = 0.295; −2 log-likelihood = 321.414. RC = reference category
In contrast, those who were currently married were less likely to report abuse than those who were single (OR: 0.1; 95% CI: 0.0–0.6), although marital status was non-significant overall. With regard to socioeconomic status, the unemployed were more likely to report DV during their life compared with full-time workers (OR = 2.9; 95% CI: 1.2–7.2), although occupation was non-significant overall.
Finally, with regard to drug/alcohol behaviour, those who reported the use of illicit drugs in the last 12 months were more likely to report abuse than those who did not (OR = 2.0; 95% CI: 1.0–3.7). In summary, being female, having children/dependants and illicit drug use were the variables significantly associated with lifetime reporting of DV in this study.
DISCUSSION
To our knowledge, this is the first UK study of DV prevalence among both male and female GU medicine clinic attendees. This study demonstrated that lifetime prevalence of DV was higher in our female clinic population compared with males but found lower prevalence of DV among women compared with one published study from an inner city London GU medicine clinic. 10 However, the previous study was in women from a very different sociodemographic area. 10 Lifetime DV prevalence among female attendees within our service was also lower compared with other UK health-care settings – 29% compared with 34.9% in a family planning clinic 12 – but higher compared with the prevalence found in general practice (23%), 13 gynaecology clinic (21%) 14 and antenatal clinic (17%). 15 With regard to abuse during the preceding 12 months, nearly 10% of the whole sample were abused (8% for men and 11% for women) and an ex-partner was the perpetrator in most cases. To the authors’ knowledge, there are no published data to compare the DV prevalence in both men and women. Therefore, further GU medicine clinic studies would be warranted within variable settings to establish prevalence.
Being female, having dependants and illicit drug use were found to be the most important factors associated with lifetime DV in this current study. Not unexpectedly, female gender was significantly associated with DV.1–3 The negative impact of DV on women's health has been well recognised, e.g. unwanted pregnancy, mental health problems, STIs, increased risky sexual behaviours and negative behavioural impact on children.
In this current study, 12% of men reported a life experience of DV. Data from Home Office statistical bulletins and the British Crime Survey show that men constituted about 40% of DV victims within the years 2004–2005 and 2008–2009.1–3 There is possibly an underestimation factor as men are more likely find it culturally unacceptable to report to authorities. This factor may have impacted on our results. Health-care professionals need to be alerted to the risk factors, such as illicit drug use, highlighted by the study and be prepared to ask all patients, including men, about DV. 16
In this study, patients with children/dependants to care for were more likely to report lifetime DV. This could be related to the added stress to a relationship or people may be more likely to report DV for the sake of children. Although our study did not find ethnicity to be significantly associated with ever reporting of DV, in contrast, one study from an inner city London GU medicine clinic showed significant associations with ethnicity. 10 However, the majority of the respondents in our study were White. Both studies found associations with illicit drug use which has a known association with unsafe sexual behaviour.
Surprisingly, alcohol abuse and younger age were not significantly associated with lifetime DV in our study (35–45-year-olds had increased odds compared with 16–24-year-olds but overall age was non-significant) in contrast to the most recent WHO multicentre study on intimate partner violence. 7 Respondents could have underestimated their alcohol consumption leading to self-reporting bias. Perhaps the contradictory finding by age could be explained by local demographics and those of older age having more time to experience lifetime history of DV, that is, time related rather than age related per se.
Although the occupation variable was non-significant overall in this current study, lower socioeconomic status (i.e. being unemployed) was significantly associated with reporting of lifetime DV, as in the WHO multicentre study. 7 This association was independent of ethnicity. In contrast to other studies, having an STI in the preceding 12 months was not significantly associated with DV.9,10,16 These findings raise important issues and potential subgroups to be aware of when patients are booked into clinics.
In this study, the DV support options were similar among male and female respondents, apart from ‘going nowhere’, which was the least frequent option stated by women. This may reflect the fact that a lot of services are not perceived as ‘DV friendly’. National DV services should be publicised more widely within National Health Service (NHS) services as, in this study, friends and family were the commonest preferred point of contact for DV support, which may delay or prevent access to suitable DV services. Not surprisingly, accident and emergency was the least preferred option possibly due to perceived lack of continuity of care, privacy and time pressure. Only 5% of the respondents from our study would go to ‘another source’ for help, which reiterates the fact that increasing NHS staff awareness and making services appear ‘DV friendly’ would allow patients a wider choice in where to go for that initial contact. Social media/networks will be an invaluable source of information and support for many patients in the future and was highlighted as a source of information used by patients within this study.
Although it usefully provides an idea of prevalence among both men and women in the GU medicine setting and identifies associations with lifetime DV, this study has a number of limitations. Firstly, only 500 patients were asked to complete the questionnaire during the study period and it is not known how representative these were from within the clinic population and from Bournemouth as a whole. Indeed, the timing of the study period may have had an impact on the findings. The finding that those with dependants were more likely to report lifetime DV could be related to the fact that the study period was during July, when, due to the summer vacation, it may have been easier for those with dependants to visit. Also, the questionnaire on different types of abuse may have measurement error as it lists suggestions for the inclusions. Hence the findings may be skewed dependent on how the respondents interpreted physical, emotional, sexual, financial and spiritual abuse definitions included in the questionnaire. The questionnaire used in this current study also did not completely differentiate sexual behaviour from risk-taking behaviour. For example, the questionnaire design did not account for the use of condoms not being mutually exclusive to use of other methods of contraception. With regard to lifestyle factors, the study only examined the drug, alcohol and mental health behaviour for current but not ex-partners who were likely to be the abusers. Finally, and importantly, findings are from a self-completed questionnaire and this might introduce a socio-desirability bias. Privacy and anonymity may remove the fear of disclosure and a further study is planned to assess reporting of DV from direct patient questioning compared with an anonymous self-completed questionnaire such as that used within this study.
A study from a London teaching hospital looking into routine enquiry of DV showed that maternity and sexual health professionals reported positive attitudes towards women affected by abuse, but had limited domestic abuse training. 17 It has been suggested that institutional guidelines need to be formulated to enhance staff training. 17 Regular training of GU medicine staff should be in place so they can address DV in a sensitive manner, are able to identify patients and refer to the appropriate services.
Depending on the individual clinic prevalence, the routine assessment of DV could be included in well-designed clinic proforma. Robust referral mechanisms should be in place to signpost DV victims. Services that combine routine enquiry, support after disclosure and attention to harm reduction have been found to be beneficial in previous studies.18,19 Furthermore, feasibility of routine enquiry of DV has been demonstrated in Australian sexual health clinics. 20 All clinics should have patient-visible literature on DV.
Consideration should be given by the British Association for Sexual Health and HIV to recommend routine assessment of DV along with sexual history depending on the local population prevalence, very similar to UK National HIV testing guidelines.21–23 Further prevalence studies are warranted to inform the ‘cut-off level’ where DV assessment would be mandatory. Having ‘DV Champions’ within each service (i.e. a DV specialist with up-to-date training) might overcome resource pressures for training all members within the service, meaning potentially more cost-efficient and specialist support services for this group of vulnerable patients.
Footnotes
ACKNOWLEDGEMENTS
The authors would like to thank the respondents and Peter Thomas for his statistical advice.