Factors associated with health care providers speaking with older patients about being subjected to abuse

ABSTRACT Health care providers have difficulties responding to elder abuse. This study aimed to investigate factors associated with health care providers speaking with older patients about being subjected to abuse, and what facilitating measures staff preferred to help them achieve this. A cross-sectional questionnaire survey was conducted among hospital health care providers (n = 154) in Sweden. Half of the respondents had experience of speaking about elder abuse. A high sense of professional responsibility (OR 3.23) and being less concerned about inflicting damage to the therapeutic relationship (OR 3.97) were associated with having spoken with older patients about being subjected to abuse. Written guidelines about elder abuse and a patient information sheet were the most preferred facilitating measures. Our findings indicate that increasing care providers’ sense of responsibility and addressing concerns about damaging the therapeutic relationship might be important factors to target in future interventions to improve health care response to elder abuse.


Introduction
The global prevalence rate of elder abuse in community settings is estimated to be around 16% in a recent meta-analysis study (Yon et al., 2017). Elder abuse is strongly associated with different kinds of ill-health, increased mortality risk, and increased hospitalization (Dong, 2005;Dong et al., 2009;Dong & Simon, Dominguez et al., 2019). Health care organizations, therefore, play an important role in identifying and reporting elder abuse cases in order to coordinate care or refer patients to suitable resources, e.g., social services or the police (Dong, 2005;Lachs & Pillemer, 2015;Touza Garma, 2017). Despite this, every fourth older adult in Sweden has reported life-time experience of abuse, but only 2% have been asked questions in health care encounters about abusive experiences (Simmons & Swahnberg, 2020). In line with this, care providers have been found to have difficulty both in recognizing and reporting suspected cases of elder abuse (Dong, 2015;Rosenblatt et al., 1996). Studies have found a lack of awareness and low levels of knowledge of elder abuse and the related legislation, among care providers in Sweden and internationally (Ahmed et al., 2016;Almogue et al., 2010;Corbi et al., 2019;Erlingsson et al., 2006;Saveman & Sandvide, 2001;Touza Garma, 2017;Yi & Hohashi, 2018).
Previous research on care providers' knowledge of elder abuse is somewhat inconsistent; some studies indicate that physicians are more knowledgeable than nurses (Ahmed et al., 2016), and others indicate the opposite (Yi & Hohashi, 2018). Reporting of suspected elder abuse cases have also been found to vary depending on occupational category (Ahmed et al., 2016;Almogue et al., 2010;Rosenblatt et al., 1996). Some studies have found that nurses more readily manage suspected abuse cases than physicians, and possible explanations could be that nurses spend more time with their patients, or that education for physicians does not cover the topic of elder abuse (Mandiracioglu et al., 2006;Yi & Hohashi, 2018).
A broad variety of other factors at both personal and organizational levels have been identified as barriers to health care providers detecting or reporting suspected cases of elder abuse. In addition to a lack of knowledge, studies show that health care providers report a lack of training, that they are insecure about how to identify or report elder abuse, and that they are uncomfortable with asking questions about abuse (Corbi et al., 2019;Daly & Coffey, 2010;Schmeidel et al., 2012;Touza Garma, 2017). Research shows that nurses and physicians consider assessing older patients for potential abuse is outside their role, or that they rely on other professionals or their supervisors to approach this issue with patients (Schmeidel et al., 2012). However, to the best of our knowledge, no previous studies have specifically examined how a sense of responsibility influences care providers' response to elder abuse.
Reported internal barriers among care providers include causes of concern about asking abuse-related questions, e.g., worries that the abused patient may deny mistreatment, a fear of offending the patient, or that asking questions about abuse will have a negative effect on the patient-provider relationship (Almogue et al., 2010;Schmeidel et al., 2012;Touza Garma, 2017). Health care providers also worry about being wrong or about not having enough proof or sufficiently strong suspicions for reporting cases (Schmeidel et al., 2012;Touza Garma, 2017). In addition, organizational barriers to detecting abuse have frequently been reported, e.g., time constraints, difficulties obtaining privacy during visits, and a lack of protocols on how to manage cases (Schmeidel et al., 2012;Touza Garma, 2017).
Feeling confident about one's ability to manage elder abuse cases might facilitate identifying and managing cases. Self-efficacy pertains to a person's self-perceived capability to perform a certain task (Bandura, 2006) and has been associated with a higher screening rate for intimate partner violence in health care (Lawoko et al., 2011).

Aims and hypotheses
Previous studies have investigated factors that might constitute barriers to or facilitators for speaking with older patients about abuse, but the findings have partly been inconsistent or yet too limited to give a comprehensive explanation of care providers' actions. More knowledge about the factors characterizing health care providers with experience of raising the topic of elder abuse could provide better support to victims of elder abuse, through educational interventions or organizational changes targeting such factors. Therefore, the aims of this study were to: (1) Investigate factors associated with health care providers speaking with older patients about being subjected to abuse. More specifically, we hypothesized that the following factors would be associated with higher odds for care providers to have such experience: a) Working as a nurse. b) Longer work experience. c) Having received education in violence in close relationships or in elder abuse. d) Having a high sense of professional responsibility to ask older patients questions about abuse. e) Reporting low levels of concern for asking older patients questions about abuse.
f) Reporting high levels of self-efficacy in asking abuse-related questions and in managing abuse cases. g) Reporting low levels of organizational barriers, i.e., fewer perceived time constraints, considering workplace preparedness to manage abuse cases as good, and knowing where to turn for help in managing elder abuse cases.
(1) Our second aim was to investigate which facilitating measures in the workplace health care providers would prefer to help them respond to elder abuse.

Study sample
This survey was conducted among health care professionals working at one acute internal medicine ward and one acute geriatric ward at a university hospital clinic in Sweden. All employees participating in a continuing educational programme for staff, with a theme unrelated to elder abuse, were asked to participate in the study. Of 166 eligible participants, 165 (99.4%) agreed to fill out the self-administered questionnaire. Participants who responded that they work strictly administratively (n = 11; 7%) were excluded from the study, leaving a total sample of 154 respondents. All data was collected in 2018.

Measurement
Since an existing instrument measuring preparedness to detect and manage elder abuse was not found, we constructed a new questionnaire, the REAGERA-P (Responding to Elder Abuse in GERiAtric Care -Provider questionnaire). The development and validation of this questionnaire is presented separately, and this study is based on the same data collection that was used in phase I of the validation study (Johanna Simmons et al., 2021). The variables relevant to this study are presented below, and the exact wording of the questions can be found in supplementary file 1. In short, questions were examined by a review committee to ensure content validity, whereafter face validity and cognitive functioning were ensured by conducting cognitive interviews. The construct validity of scales was tested by using factor analysis and a test of internal consistency (Cronbach's alpha = 0.75 and 0.87). Convergent validity was assessed by investigating if different items within the questionnaire correlated in predictable and logical ways, a method recommended when developing new self-efficacy scales (Bandura, 2006). The items were found to correlate in the anticipated way, e.g., respondents with high self-efficacy for asking questions about abuse reported lower level of concern for negative reactions from the patients (Spearman correlation −0.3, p < .01) as well as less concern about asking questions having a negative impact on the patient-provider relationship (Spearman correlation −0.33, p < .01). Also, the self-efficacy scales showed responsiveness, i.e., there was a significant increase in selfefficacy for asking questions about abuse (p < .01) and managing the response (p < .01) comparing before and after an educational intervention about elder abuse given to medical interns. Altogether the validity of the instrument was satisfactory . Completing the questionnaire required approximately 10-15 minutes.

Aim 1
To evaluate participants' previous history of speaking with older patients about being subjected to abuse, we asked: In your work, have you ever spoken with an older patient about abuse he or she might have been subjected to? Response alternatives were: (a) No; (b) Yes, a few times; (c) Yes, many times; and (d) Do not remember. Because a small number of respondents selected alternatives (c) (N = 1; 0.7%) and (d) (N = 8; 5.2%), the categories were dichotomized in the following manner: (a) No or do not remember; (b) Yes, a few or many times. The demographic background characteristics of respondents were established using questions pertaining to, e.g., age, occupation, education in elder abuse or in violence in close relationships, and years of work experience.
Respondents were asked to evaluate how much responsibility they thought each professional category (i.e., nurse, assistant nurse, physiotherapist, occupational therapist, physician) at their workplace has for asking older patients questions about abuse. A 4-point Likert-like scale ranging from "None" to "A lot" was used. Respondents' answers to this part of the questionnaire were also categorized to show how much responsibility they attributed to their own professional category, which was interpreted as their own sense of professional responsibility. Respondents were either categorized as having: (a) a high sense of professional responsibility, i.e., they attributed "a lot" of responsibility to their own professional category or (b) a low to medium sense of professional responsibility, i.e., they attributed "none," "fairly little," or "quite a lot" of responsibility to their own profession.
Cause for concern was measured using three items pertaining to respondents' concern that asking abuse-related questions would have a negative impact on the patient-provider relationship, concerns that questions would result in negative reactions from the patient, and concerns about not being able to offer good follow-up to the patient. Respondents were asked to rate how concerned they were on a 4-point Likert-like scale ranging from "Not at all concerned" to "Very concerned." The categories were dichotomized into (a) Very or somewhat concerned and (b) A little or not at all concerned, for all three items.
Self-efficacy was measured using two scales. One scale measured preparedness to ask questions related to elder abuse, derived from 3 items asking participants to rate their preparedness from very bad (1) to very good (10). Ratings were then summed up to determine the respondent's level of selfefficacy (Cronbach's alpha = 0.75). The other scale measured preparedness to manage cases where elder abuse emerges, using 5 items asking participants to rate their preparedness from very bad (1) to very good (10). Ratings were summed up in the same manner to determine the respondent's level of selfefficacy (Cronbach's alpha = 0.87). Both scales were subdivided into three equally large parts, and each part represented low, medium, or high levels of self-efficacy.
Organizational barriers to responding to elder abuse were evaluated using items pertaining to organizational conditions. Respondents were asked to evaluate the preparedness at their workplace to take care of older patients who had been subjected to abuse using a 4-point Likert-like scale ranging from "Very good" to "Very inadequate," or the alternative "Don't know what preparedness there is." The categories of the Likert-like scale were merged into (a) "Very or fairly good" and (b) "Very or somewhat inadequate," while the alternative of (c) "Don't know what preparedness there is" was maintained.
Respondents were asked to evaluate how often they had time to bring up the issue of abuse with their older patients, and response alternatives were dichotomized into: (a) "Never or rarely" and (b) "Often or always." We also asked whether respondents knew where to turn for help in handling elder abuse cases with response alternatives: (a) "Yes" and (b) "No."

Aim 2
We suggested five measures intending to facilitate health care providers' handling of encounters with older patients who have been subjected to abuse. We asked respondents to evaluate how helpful they would find these different measures on a 4-point Likert-like scale ranging from "A lot" to "None."

Statistical analysis
Statistical analysis was performed using SPSS software (version 26), and the significance level was set to 95% for all analyses.

Aim 1
To investigate aim 1, we explored factors associated with speaking with older patients about being subjected to abuse. First, Pearson's chi-square test was used to test for associations between the dependent variable and the following independent variables: demographic background characteristics, sense of professional responsibility, self-efficacy, cause for concern, and organizational barriers. Factors found to be significantly associated with the dependent variable in the bivariate analyses were then included in a binary logistic regression analysis. The independent variables were placed in one block and analyzed using the forward conditional method in SPSS. Variables that were significant (p < .05) in the last step of the regression model were included in the final model. The model fit was good with Hosmer-Lemeshow p = .33 and Nagelkerke R 2 = 0.44.

Aim 2
To investigate aim 2, descriptive statistics were used to evaluate how helpful respondents found the suggested measures meant to facilitate the handling of encounters with older patients subjected to abuse.

Ethical approvals
Ethical approval was given by the regional ethical review board in Linköping Sweden (Registration no. 2017/181-31 and2017/564-32). Eligible participants were provided with information about the research project, and a returned filled-out questionnaire was considered as informed consent.

Aim 1: speaking with older patients about being subjected to abuse
About half (n = 76, 49.7%) of the respondents had experience of speaking with an older patient about being subjected to abuse (Table 1). A majority of physicians (n = 17, 81.0%) had this experience, while the corresponding rate was 51.7% for nurses (n = 30), 54.5% for paramedical personnel (n = 6), and 36.5% for assistant nurses (n = 23) ( Table 1).
Most respondents (79.7%, n = 122) thought physicians had a lot of responsibility for asking older patients questions about abuse, while the corresponding proportions were 51.3% (n = 78) for nurses, 34.0% (n = 52) for assistant nurses, and 29.6% (n = 45) for paramedical personnel (Figure 1). This data was also analyzed to visualize how each respondent assessed their own sense of professional responsibility. Slightly over half of the respondents (n = 81, 52.9%) had a low to medium sense of professional responsibility for asking older patients questions about abuse (Table 1).
Bivariate associations are presented in Table 1. In the binary logistic regression analysis (Table 2), the following variables were associated with speaking with an older patient about being subjected to abuse: A) working as a nurse (adj OR 0.21, CI 0.05-0.93), assistant nurse (adj OR 0.11, CI 0.02-0.46) or paramedical personnel (ajd OR 0.1, CI 0.01-0.88), B) having received education in both elder abuse and violence in close relationships (adj OR 11.51, CI 2.41-55.00), C) having over 5 years of work experience (adj OR 8.45, CI 1.84-38.86), D) having a high sense of professional responsibility for asking older patients questions about abuse (adj OR 3.23; CI 1.39-7.54), E) being not at all or a little concerned that asking abuserelated questions would have a negative impact on the patient-provider relationship (adj OR 3.97, CI 1.41-11.14), and F) considering the workplace preparedness for taking care of older patients subjected to abuse as very or fairly good (adj OR 3.27, CI 1.09-9.79) or very or somewhat inadequate (adj OR 2.88, CI 1.10-7.53). *Paramedical personnel are referred to as "Other" in the questionnaire, and include the following occupational groups: Occupational therapist, physiotherapist, pharmacologist, and dietician.

Aim 2: facilitating measures
Respondents' evaluation of different measures that might facilitate health care professionals in handling encounters with older patients who have been subjected to abuse are presented in Figure 2. Most respondents (n = 101, 66%) considered written guidelines for staff that included contact information to available support services to be a very helpful facilitating measure.

Discussion
Even though elder abuse is relatively common, half of our respondents had never spoken with an older patient about being subjected to abuse. We found that experiencing a higher sense of professional responsibility to ask older patients about abuse, having less concern for a negative impact on the patientprovider relationship, and having a formed opinion of workplace preparedness to handle elder abuse were factors associated with speaking with older patients about being subjected to abuse. These factors may also be reasonable targets for future interventions aiming to improve the response of health care providers to elder abuse. Respondents' assessment of how much responsibility they think each professional category has for asking older patients questions about being subjected to abuse. Amount of responsibility presented as percentage of answers for each profession (%) (n = 154 a,b ). a 1 missing case for categories assistant nurse and physician, (0.6%). b 2 missing cases for categories nurse and occupational therapist or physiotherapist, (1.3%).

Occupation and sense of responsibility
Contrary to our hypothesis, nurses were less likely than physicians to speak with older patients about being subjected abuse (Table 2). In fact, a much larger proportion of physicians reported having such experience (Table 1). This stands in contrast to previous findings that nurses might feel more comfortable and confident in managing and detecting elder abuse than physicians (Ahmed et al., 2016;Mandiracioglu et al., 2006). Several causes might contribute to these differences. Our results show that respondents attributed physicians as largely responsible for asking questions about elder abuse (Figure 1), and previous studies have hypothesized that expectations may influence how health care providers respond to suspected cases of elder abuse. For example, Japanese public health nurses and social workers, who are expected to assess and manage elder abuse cases, were found to have better knowledge of and more willingness to deal with these issues than other occupational categories (Yi & Hohashi, 2018). Other studies show that nurses and physicians rely on, e.g., social workers to approach or deal with elder abuse issues (Ahmed et al., 2016;Schmeidel et al., 2012), which might act as a barrier to dealing with the issue themselves. In line with this, our results showed that considering your own profession as being very responsible for asking questions about elder abuse was associated with speaking with older patients about being subjected to abuse (Table 2). This further indicates that a sense of professional responsibility may be a facilitator for detecting elder abuse in health care organizations.
Varying results from different studies might be due to different methodology and study samples. However, differences might also be attributed to cultural and legislative differences regarding the responsibilities of and expectations on different occupational categories. Clarifying and strengthening care providers' sense of responsibility could potentially be targeted in interventions to improve health care response to elder abuse.

Internal and organizational barriers
Less concern about abuse-related questions negatively impacting the patientprovider relationship was associated with higher odds of speaking with older patients about being subjected to abuse (Table 2). This is in accordance with previous studies that show how greater concern for damage to the therapeutic relationship impedes professionals from detecting and reporting abuse (Schmeidel et al., 2012;Touza Garma, 2017). It is also possible that this   association is bidirectional, i.e., participants who had spoken with older patients about being subjected to abuse may be less concerned after learning their experience did not impact the relationship in a negative way. If so, this might also be reassuring for providers who want to speak with older patients about abuse. Health care providers with a formed opinion of the workplace preparedness to manage cases of elder abuse were more likely to speak with older patients about being subjected to abuse, compared with those that did not know about the preparedness (Table 2). This association might be because those with experience of speaking with older patients about being subjected to abuse probably have stronger motivation to familiarize themselves with the organizational conditions at the workplace, whether good or bad. Studies have, however, shown that the organizational environment may positively influence health care providers' response to elder abuse through, e.g., clear procedures and protocols (Touza Garma, 2017). In line with this finding, a majority of respondents in this study (n = 101, 66.0%) responded that written guidelines to aid staff in managing elder abuse would be very helpful when meeting older patients who had been subjected to abuse (Figure 2). Such factors have also been reported as facilitators of detection and reporting of elder abuse in previous research (Touza Garma, 2017). In addition, respondents also advocated the usage of patient-directed documents entailing contact information to support services (Figure 2). This further emphasizes the importance of supportive organizational measures to aid and encourage health care providers to adequately respond to elder abuse. The absence of such facilitating measures may reflect a lack of awareness and knowledge of elder abuse on an organizational level, and not be limited to just health care providers as individuals. This would motivate the inclusion of e.g., managers in interventions aiming to raise awareness of elder abuse, to possibly facilitate organizational changes.
Respondents' self-efficacy, in terms of preparedness to ask questions related to elder abuse, was not significantly associated with speaking with older patients about being subjected to abuse (Table 2). This was unexpected because previous studies have shown that health care providers having strong self-efficacy is associated with increased screening rates for intimate partner violence (Lawoko et al., 2011). Self-efficacy is likely affected by variables pertaining to causes for concern when speaking with patients about abuse and by having received education in elder abuse, which might explain the difference between this and previous studies.

Education and work experience
As expected, having received education in both elder abuse and violence in close relationships was associated with speaking with older patients about being subjected to abuse (Table 2). In line with this, previous research has shown that knowledge and education about elder abuse improves health care providers' recognition and management of abuse cases. (Almogue et al., 2010;Daly & Coffey, 2010;McCreadie et al., 2000;Touza Garma, 2017;Wagenaar et al., 2010;Yi & Hohashi, 2018). Somewhat surprisingly our analysis showed that having received education only in elder abuse or violence in close relationships, one without the other, was not significantly associated with speaking with older patients about being subjected to abuse (Table 2). This outcome could be due to our relatively small study sample, or due to the fact that the format and extent of the participants' education may vary greatly. It is possible that having a greater interest for elder abuse characterizes participants with a more comprehensive education in both elder abuse and violence in close relationships. Even so, our results suggest that more extensive education in violence in close relationships and elder abuse is likely associated with raising this topic with patients.
Only a small minority of participants (n = 21; 13.7%) had received education in both elder abuse and violence in close relationships (Table 1). This is in accordance with previous studies that found limited knowledge and awareness of elder abuse amongst health care workers, and a need for more elder-abusefocused education is concluded by a number of authors and requested by health care workers themselves (Almogue et al., 2010;Corbi et al., 2019;Daly & Coffey, 2010;Erlingsson et al., 2006;Schmeidel et al., 2012;Wagenaar et al., 2010). In fact, a majority of participants in our survey responded that education in how to ask questions about abuse (n = 86, 56.2%) and education in what support services are available (n = 89, 58.2%) would be very helpful facilitating in handling of encounters with older patients who had been subjected to abuse.
We found an association between having over five years of work experience and speaking with older patients about being subjected to abuse (Table 2). Previous research shows inconsistent results regarding how years of work experience influence health workers' behavior in this respect. Some studies indicate that the factor fewer years of professional experience is associated with more reporting and greater recognition of abuse (Touza Garma, 2017), but a study in accordance with our results shows that long work experience is associated with addressing suspected elder abuse (Meeks-Sjostrom, 2013). More research is needed to explore how age or professional experience may influence health care providers' response to elder abuse.

Clinical implications
We found several factors that could potentially be targeted in future interventions to improve health care providers' response to elder abuse. A strong sense of professional responsibility was found to characterize health care providers with experience of speaking with older patients about being subjected to abuse, possibly reflecting a higher awareness of elder abuse or a greater willingness to live up to expectations. Health care providers in Sweden, regardless of their occupational category, are obliged to ask questions when adult patients show signs and symptoms of possible abuse (The National Board of Health and Welfare, 2014). However, we found that slightly over half of the respondents experienced a low to medium sense of professional responsibility for asking older patients questions about abuse. If the sense of responsibility can be strengthened amongst all professional categories, this might reduce care providers' reluctance to do tasks presently considered to be outside their professional role or tasks considered as entailing extra work. Hence, our results indicate that policy changes or interventions that increase care providers' sense of responsibility, e.g., by clarifying or further emphasizing their requirement to ask abuse-related questions, may be effective strategies to improve the response given to older patients who have been subjected to abuse.
We also found that less concern that abuse-related questions would damage the therapeutic relationship was associated with speaking with older patients about being subjected to abuse. Addressing such concerns actively through, e.g., educational interventions or providing tools that enable staff to ask abuserelated questions while maintaining their alliance with patients could possibly be important strategies to encourage more care providers to ask older patients about abuse.
Having a formed opinion of workplace preparedness was also associated with having spoken to older patients about being subjected to abuse, and written guidelines for staff was the most sought-after facilitating measure. This further indicates that organizational measures that facilitate workers in handling the encounter with older patients who have been subjected to abuse might encourage a better response, and this outcome could potentially mitigate the reluctance of staff to spend time and extra work finding ways to manage abuse cases. Research on interventions with such organizational measures would be useful to draw additional conclusions.

Limitations
The relatively small number of participants (n = 154) in this study was a limiting factor. The small sample size, together with some co-variating variables, is reflected in the relatively large confidence intervals in the logistic regression model. Also, the majority of participants (>75%) were nurses and assistant nurses, and most respondents were female (>90%). One could argue that a more even distribution of occupational categories and sex would have been preferable. Since we found that physicians were more likely to have spoken with older patients about being subjected to abuse than other professionals, a more even distribution of professional categories might have increased the overall estimate of how common speaking with older patients about abuse was. Since we only collected data from two wards at one clinic that was not randomly selected, it is possible that other important factors not considered, e.g., the organizational culture or other local preconditions, may have influenced our results. In addition, two of the authors of the present study are employed at the clinic, which may have influenced respondents' answers. However, anonymity was assured by not using questions that could reveal the identity of respondents, e.g., we asked about predetermined age categories and work experience categories instead of asking for specific age or years of work experience. On the other hand, requesting respondents' exact age and work experience would have enabled us to create a different variety of categories, possibly revealing other statistical relationships.
We had an exceptionally good response rate (99.4%), and our study sample well represents typical staff composition in Swedish inpatient wards in terms of sex and occupation. It is, therefore, likely that our results well represent a typical Swedish hospital ward and are reliable. However, further studies that include larger samples from a greater number of in-patient wards as well as primary health care centers are needed to lay a stronger foundation for interventions intended to improve health care response to elder abuse.
In our questionnaire the wording "speaking with older patients about being subjected to abuse" is used to examine health care providers' experience of raising the subject of elder abuse with patients through various aspects. However, the terminology doesn't allow us to distinguish whether respondents directly asked their patients about possible abuse, if the patients spontaneously disclosed abuse-related information, or if the subject was raised as part of an intervention.
Our questionnaire does not enquire specific details, such as format or duration, of the respondents' education in elder abuse and violence in close relationships. Even though education was associated with speaking to older patients about being subjected to abuse, we cannot distinguish what kind of education might have been most useful. This would be valuable to further investigate in future research.

Conclusions
In this study we investigated factors associated with health care providers speaking with older patients about being subjected to abuse. We found that only half of the care providers had ever spoken with an older patient about being subjected to abuse, even though the prevalence of elder abuse is relatively high. Health care providers with experience of speaking with older patients about being subjected to abuse were characterized by considering their own profession as highly responsible for asking questions about abuse and by having less concern about abuse-related questions damaging the therapeutic relationship. These findings indicate that strengthening health care providers' sense of responsibility and addressing concerns about damaging the therapeutic relationship are important to target in future interventions intended to improve the health care response to elder abuse.

Disclosure statement
No potential conflict of interest was reported by the author(s).