Preparedness to provide nursing care to women exposed to IPV
Several shortcomings were found regarding support at the level of the organisation. Our results highlighted lack of relevant mandate to deal with the issue which was reflected in a lack of guidelines on proper nursing care and cooperation with other authorities. Several studies have shown that organisational support is important to improve care of women exposed to abuse. According to Minsky-Kelly et al. , support in the form of continuing education along with guidelines are necessary steps to be taken by a health provider to improve care of women exposed to IPV. Waalen et al.  showed that education increased preparedness in identifying women exposed to IPV and that strategic interventions coupled with training improved screening rates. In addition, another study showed that guidelines had an impact on the nurses' willingness and ability to ask women about violence and on to properly manage their care .
Although guidelines are meant to facilitate IPV detection and implementation of the appropriate intervention methods, results from several studies are so far inconclusive [17, 32, 33, 35]. At the same time, it is well known that only when guidelines are implemented in the organisation can nurses effectively support women exposed to IPV . When guidelines are, therefore, lacking nurses may have to improvise with uncertain outcomes . In this study, only 5% of the participants stated they were aware of written guidelines. It was not known whether this was due to complete lack of guidelines or lack of knowledge about existing guidelines. Either way, special attention should be paid on the impact lack of guidelines or lack of awareness of existing guidelines has on the nurses' self-rated preparedness. In this study, nurses considered the lack of guidelines as inhibiting in dealing with women exposed to IPV.
Conditions at the organisation and personal attitudes towards IPV equally affect nurses' preparedness. When nurses were asked to explain the reasons for not asking women about violence many stated that they felt uncomfortable about addressing the issue. Awareness of one's own attitudes plays an important role in one's preparedness to deal with abused women . Participants' attitudes towards women exposed to IPV matched general societal views and was consistent with findings by others . Training programmes must deal with this problem, since having the 'prejudicial' attitudes is known to negatively impact the nurses' interaction with abused women as well as their ability to identify them and properly care for them . As an example, almost every fourth respondent in this study agreed with the general notion that "Victims of IPV can always leave the perpetrator if they want to". This view could have a significant impact on the nurses' decision of whether to ask women about violence or not. One study reported that in-service training gave the nurses a better understanding of the difficulties women have in leaving their abusive partners. Knowing the difficulties for women to leave the perpetrator may have a positive impact on the nurses' encounters with women exposed to IPV and the quality nursing care they could offer them. Many nurses in the present study stated a need for training which was found to be closely associated with 'feeling sufficiently prepared' to identify abused women.
In previous studies, women exposed to IPV preferred to be asked directly about IPV [39–41]. In this study, approximately half of the nurses stated that they always did so when they suspected IPV. This rate seems to be higher than that reported in earlier studies, except that of nurses working in obstetrics and gynaecology . Overall, studies have shown that very few women are asked directly about IPV by health care professionals [25, 43–46]. This might also be the case for the population in this study, since, even though half of the respondents stated that they always asked women about IPV when they suspected it, most of them met women exposed to IPV less than once a month. This could result in lack of knowledge about health problems caused by IPV, which, in turn, may lead to cases of IPV remaining undetected. This is supported by the finding that nurses suspected IPV only when women had visible injuries, which is a common misunderstanding about health problems caused by IPV [47, 48]. The most common health effects are psychological and psychosomatic problems such as depression and chronic pain without an obvious cause. In the present study, these signs did not cause suspicion as often as bruises and injuries did. Furthermore, this is probably closely associated with lack of knowledge. Interestingly, few of the nurses in this study had received training about IPV during their professional education or during their employment. It is, thus, not surprising that those who could identify IPV victims are those who had obtained knowledge on their own initiative.
This study showed that 13% of the nurses had personal experiences with IPV and one out of three had a relative or near relation who had been exposed to IPV. Even though there was no indication of an association between having experienced IPV and higher odds of successfully identifying IPV victims, the personal experience may influence the quality of nursing care given to IPV victims .
The word 'caring' in nursing care, means 'providing for', and there are many nursing interventions that should be considered when caring for women exposed to IPV . The questionnaire suggested 18 possible interventions the nurses could choose from when encountering a woman they knew had been exposed to IPV. However, only two nurses reported that they had used all 18 interventions. Nurses' most common intervention was to refer the woman to a doctor. A doctor's appointment is a necessary intervention but it might also, be a way of 'passing the buck' when one is not aware of other nursing interventions, does not have written guidelines, and/or feels uncomfortable encountering someone who has been exposed to IPV. It could also mean that nurses believed that doctors were more prepared to intervene in cases of IPV which may also be the case sometimes but not always.
Caring for an abused woman may also include caring for her children who may also been exposed to violence. It is commonly known that IPV affects children's health just as much as it affects the women's. Nurses are obligated to report children living in an abusive home to social services. However, very few nurses stated that they reported these families to social services. This is a potentially serious shortcoming, as the families where IPV occurs need considerable support. Nurses also need to respond appropriately to women's disclosure, as it seems that the effect of distressful disclosure experiences may lead to a gradual reduction in health care seeking among women exposed to IPV .
Analysis of estimated sample size was performed and yielded a power of 90%. Totally 125 participants were needed and in fact, 192 nurses participated, why the power was judged sufficient. The confidence intervals in the regression models were rather wide, which could indicate a low statistical power. However, the response rate was 69.3% which must be considered large in this kind of a study, and the calculated power to detect a significant difference in the study sample was 90%, which is satisfactory. Among the 83 nurses who did not answer the questionnaire reasons for this are known for more than half of them since they returned the questionnaire with written comments regarding this. However, the nurses, in total 35, who did not return their questionnaire or returned it but gave no written reasons for not answering, were not further contacted. It is therefore not known if they introduce a selected bias. This was, however, a decision taken from an ethical point of view and out of respect for the nurses' privacy, since IPV is considered sensitive. The internal dropout was between 0% and 5% except for two questions where it was 9%, even though it must be considered as low. The decision was made to not exclude the questionnaires or the variables with missing values since they occurred randomly and did not affect the outcome of the study.