METHODS

The study was performed in the A&E department of a district general hospital in South Lancashire, UK. A brief questionnaire survey, consisting of 15 questions using a 5 point Likert scale (strongly agree; agree; undecided; disagree; strongly disagree), was used to elicit patient attitudes. Age, sex, and marital status were also recorded. The questionnaire was developed and refined by a process of literature review,^10–12 discussion with A&E staff and an initial pilot study. The questionnaire was only available in English.

A predetermined number of questionnaires were distributed to consecutive patients attending at the start of every four hour time block over a seven day period. The questionnaires were given out by the triage nurse for self completion and placed anonymously by the patient, or in the case of “stretcher” patients by a family member or nurse, into a box at the A&E reception desk.

All adult patients (16 years plus), including “stretcher” patients, were eligible for inclusion. Patients requiring urgent medical attention and those with mental impairment or severe emotional distress were excluded. The survey was anonymous and not linked to the hospital records. The study was approved by the local research ethics committee.

The number of patients expected to attend A&E in each of 42 four hourly time blocks over seven days was estimated from average attendance over eight previous weeks. From these data the number of questionnaires to be distributed in each four hour time block was calculated to provide a representative sample of all patients attending A&E.

Sample size was based on an assumption that 68% of patients would consider routine questioning about violence acceptable.¹² This indicated that we would require 330 completed questionnaires to estimate such a proportion to a precision +/-5%. We allowed for a 20% non-completion rate and planned for the distribution of 432 questionnaires. Data were entered into SPSS (version 8) for analysis with the categories of strongly agree/agree combined for analysis.

RESULTS

The respondents

During the one week study period a total of 581 adult patients attended A&E. Of these, 457 (79%) consecutive attenders in the four hourly time blocks were considered for entry to the study. Of these 457 patients, 154 (34%) were excluded from the study (31 required immediate medical attention, 36 had severe emotional distress or mental impairment, 32 declined or had previously completed questionnaire, two were non-English speakers and some 53 patients were classified as “other”; for example, unable to complete questionnaire because of hand injury).

Altogether 303 questionnaires were distributed and 284 completed questionnaires were returned (94% response rate). Questionnaires were incorrectly completed by three patients aged under 16 years, leaving 281 for analysis (157 men, 116 women, eight unspecified). Fifty eight per cent of respondents were attending A&E for the first time in 12 months.

Some 70% (303 of 432) of the predetermined number of questionnaires were distributed. The majority of the questionnaires that were not distributed were accounted for during two consecutive four hour blocks on the Sunday afternoon when the A&E department struggled to cope with a heavy workload.

The median age of respondents was 33 years (IQR 25 to 49 years) and ranged from 16 to 84 years. Women were significantly older than men (36.5 and 30.0 years, Mann-Whitney U test p=0.003).

Attitude towards being asked about violence

Altogether 189 respondents (67%; 95%CI 60% to 74%) agreed that people attending A&E should routinely be asked whether they have been assaulted. There were no significant differences between the attitudes of men and women. The proportion of respondents who agreed with routine questioning increased with age (52% of 16–24 year olds, 65% of 25–44 year olds, and 85% of those over 45 years of age).

In relation to who should routinely ask about violence, 187 (67%) thought it would be appropriate for the doctors to ask, 174 (62%) for the nurses to ask and 73 (26%) for the reception staff to ask.

Only 89 (32 %; 95%CI 23% to 41%) agreed that adults attending A&E should routinely be asked about sexual abuse, although the proportion supporting this suggestion also increased with age (17% of 16–24 year olds, 28% of 25–44 year olds, and 51% of over 45 year olds).

A&E and the police

Altogether 251 (89%; 95%CI 85% to 93%) respondents agreed that health care staff should encourage victims of abuse or violence to inform the police and 209 (74%; 95%CI 68% to 80%) thought that health care staff should routinely inform the police. The proportion of respondents agreeing to these statements increased with age (82% of 16–24 year olds, 89% of 25–44 year olds, and 100% of over 45s supported encouragement to inform the police; while 61%, 77%, and 80% respectively, supported routinely informing the police).

Violence and the health service

A total of 189 (67%; 95%CI 60% to 74%) patients agreed that the health services should help in tackling the problem of violence in society, although support was higher among women (81%) than men (61%) (difference 20%, 95%CI 9% to 31%). In reality however, 71% (95%CI 65% to 77%) felt that health care staff could do little to stop violence or abuse re-occurring.

Telling A&E staff

Only 127 (45%; 95%CI 36% to 54%) of respondents thought that people who had experienced violence would be likely to tell staff if directly asked. However, when asked what they would do themselves if they were the victims of violence or abuse, 179 (64%; 95%CI 57% to 71%) indicated that they would tell without waiting to be asked by health care staff, and 227 (81%; 95%CI 76% to 86%) agreed that they would tell if asked directly. The proportion who would tell without waiting to be asked increased with age (51% of 16–24 year olds, 62% of 25–44 year olds, and 80% of over 45 year olds). Similarly, the proportion of those who would tell if directly asked also increased with age (67%, 84%, and 91% respectively). There were no significant differences between the responses of men and women.

Only 75 (27%; 95%CI 17% to 37%) felt that they would not be able to trust health care staff enough to tell them. Some 79 (28%; 95%CI 18% to 38%) were worried that the police may get involved if violence was reported to casualty staff. A significantly greater proportion of women than men (37% versus 24%) were worried about police involvement (difference 13%, 95%CI 1% to 25%). Indeed 36 (13%; 95%CI 2% to 24%) patients indicated that they would be too afraid to tell health care staff. Women (21%) were significantly more likely than men (8%) to indicate fear as a reason for not telling A&E staff (difference 13%, 95%CI 4% to 22%).

DISCUSSION

This study found high levels of acceptability among patients actually attending A&E for routine questioning about violence, by either doctors or nurses. Older patients particularly, supported a more active role for health professionals. The high proportion of patients favouring routine questioning in this study is comparable with the only other previous study from the USA.¹² There is also strong support for A&E staff to encourage the victims of violence to inform the police. Indeed, a large majority of patients supported the idea of health professionals routinely informing the police of violent assault. Mandatory reporting of domestic violence, as in parts of the USA,¹³ raises difficult ethical and moral issues.¹⁴ Staff may also have fears of provoking further violence by inappropriate questioning.

It is clear that many patients believe that the health service has a broader part to play in tackling violence in society, although there is recognition that preventing recurrent violence is more problematic. It was not possible in this study to perform a more in depth exploration of attitudes to working with the police, although the Data Protection Act provides challenges in sharing data from injury surveillance programmes.¹⁵

The returned questionnaires were representative (by time of attendance) of all adult A&E attenders. The high response rate reflects the use of a brief questionnaire that took only a couple of minutes to complete, and the involvement of the triage nurses in the planning and execution of the study. We suspect that similar views are likely to be held by patients attending other A&E departments in the UK. It should be noted however that the population of this area has a relatively small ethnic minority population and this study has not been able to explore how ethnic background and culture may affect attitudes. It is also possible that attitudes to confidentiality may depend on local population relationships with the police. We intentionally did not focus our attention on patients presenting with physical injuries. Targeted screening of selected injury groups will miss significant numbers of the victims of violence and abuse.⁷

We have demonstrated strong patient support for a more active approach towards identifying victims of violence in A&E. Further research into direct and routine questioning of patients attending A&E about their personal experience of violence is merited; in particular it will be important to determine the attitudes of patients who are actually victims of violence. As violence in the community becomes increasingly recognised as a serious and important threat to health, health care professionals will need to consider how best to protect patients under their care from violence.

Contributors

AH and MC designed the study, analysed the data and prepared the manuscript. RF participated in planning and supervising the study and revising the manuscript. AH is the guarantor of the paper.