This study investigated the prevalence of elder abuse in a hospitalized population, and we found that one in six older adults had been subjected to elder abuse. Of those, a majority had also been subjected to abuse before the age of 65, and earlier life abuse was the only background factor significantly associated with subsequent elder abuse. Reporting victimization both before and after the age of 65 was associated with poor mental-health outcomes, while no association was found between mental ill-health and reporting only elder abuse or only earlier in life abuse. Altogether, our findings underline the importance of a life-course perspective in research on elder abuse.
Prevalence of elder abuse
The prevalence of elder abuse in this hospitalized population was 18%. Hospital wards are a relatively unexplored setting for investigating elder abuse prevalence. There is one example of a study from India, which used a self-report measure and estimated a prevalence of 16%, which is similar to our findings [21]. Our reported prevalence is slightly higher compared to the 10%–15% often reported in community samples [5, 48, 49]. Hospitalized older adults often have risk factors for elder abuse such as high dependence on others for their daily living, and hence one could have expected an even higher prevalence compared to community samples. Why this was not the case could possibly be explained by a Neyman bias, whereby those who were most affected by disease were not included. Elder abuse has previously been associated with functional dependency as well as cognitive impairment. In this study, however, older adults who did not have the physical, cognitive, or linguistic capacity to fill out the REAGERA-S and participate in the interview were excluded. During the study period, 224 older adults (34%) admitted to the clinic were excluded for these reasons. It is plausible that the prevalence rate would have been higher if those older adults could have been included in the study. However, to include older adults with, e.g., cognitive impairment would require another study methodology.
It should also be noted that a prevalence of 10–15% in community samples often refers to victimization in the past year. In this study, we defined elder abuse as one or more abusive experiences after the age of 65. This definition is likely to lead to a higher prevalence compared to only past-year prevalence. Also, in meta-analyses and reviews on elder abuse, there is a great variation in estimated prevalence, even if data is limited to the past 12 months [5]. This could be a symptom of the methodological challenges in violence research, including elder abuse, where differences in definitions, operationalizations, and settings have a strong impact on the reported prevalence [10, 11, 50]. To be able to make comparisons between different populations, studies using the same methodology need to be carried out in different settings. However, our findings confirm that experiences of elder abuse are common among older adult hospitalized patients in Sweden.
The fact that victimization data was collected using a semi-structured qualitative interview rather than a questionnaire or structured interview could have affected the results in several ways. There was a dropout, whereby one in four who completed the instrument declined to participate in the interview. When validating the REAGERA-S, the non-interviewed group reported somewhat less abuse than the interviewed group [38]. This corresponds with previous research on non-response bias, where individuals who are not exposed are less inclined to participate in a study regarding the subject [51]. With this rationale, there might be a slight overestimation of prevalence in our results, due to the dropout rate. However, there should be a greater degree of certainty in the data, as the answers were validated against the participants in the interview, and misconceptions regarding questions could be clarified directly. It was also possible to address the challenge of defining abuse in the interviews. What constitutes an abusive experience must be defined partly by the victim, and the cultural norms and circumstances surrounding the abusive experience must be considered. As presented in the methods, this issue was handled systematically via discussions in the research group prior to classification.
Background factors associated with elder abuse
Regarding predictors of elder abuse, previous studies have found divergent results concerning the associations between elder abuse and sex, age, and marital status, although a consistent association between elder abuse and functional dependency is reported [15, 16, 52]. It was hence unexpected that we did not find a significant association between reporting elder abuse and ADL status. This could potentially be explained by the rather low sample size, increasing the risk of a false negative. Hence, the lack of association with all background factors should be interpreted with some caution. However, we found that earlier life abuse increases the odds of reporting elder abuse, which is in line with previous research [15, 24, 26], and further underlines the importance of a life-course perspective in research on elder abuse. From a cumulative inequality perspective, it would have been interesting to differentiate the effect on subsequent elder abuse between earlier childhood abuse and other types of abuse, for example intimate partner violence and community violence. Cumulative inequality theory specifically states that childhood experiences play a special role in the development of a person’s life trajectory, and hence it would have been valuable to know whether it was uniquely childhood abuse that increased the odds of reporting elder abuse or whether previous adult victimization was also relevant. One Canadian study managed to make this differentiation, and found that only childhood abuse retained its effects after adjusting for other background variables [24]. This suggests that childhood abuse could, as cumulative inequality theory states, have a specific effect on vulnerability to elder abuse.
Life-course victimization and mental ill-health
When exploring associations between abuse and mental ill-health, we found that only those reporting abusive experiences both before and after the age of 65 had increased odds of poor mental health. The associations were similar for all four outcomes, but were not significant in terms of having an anxiety or depression diagnosis. This indicates that a negative life trajectory with experiences of abuse earlier in life as well as in later life has the strongest impact on mental health. The difference in mental ill-health may also be attributed to the fact that all cases in that category were polyvictims, i.e., were victimized on several occasions or by several perpetrators. Polyvictimization has previously been reported to be associated with greater ill-health than any single victimization among older adults [37], and in this particular sample previous experiences of abuse were found to influence both the experience of elder abuse and how it was managed [39, 40]. It is therefore unfortunate that many studies on elder abuse focus on past-year exposure, and that polyvictimization is not further explored. Our finding that only the combination of victimization earlier in life and elder abuse was associated with mental ill-health indicates that there is a risk of disregarding substantial information when taking a narrow time perspective in the field of elder abuse. If only past-year victimization is considered, the association between abuse and mental ill-health may be misinterpreted, i.e., overestimated or underestimated due to disregarding previous life-course experiences of abuse.
As this was a cross-sectional study, causality cannot be assessed. The association between victimization and ill-health is complex, and the relationships may be bidirectional. Previously, childhood victimization has been found to be associated with ill-health in middle age, which in turn was associated with elder abuse [32]. There is also a well-established association between childhood abuse and adult intimate partner violence, and childhood victimization has been linked to later life mental ill-health [28, 53]. A few longitudinal studies have also been conducted in which abusive experiences in later life are concluded to be a risk factor for later mental ill-health, such as anxiety symptoms and major depression [4, 54]. Altogether, this supports the cumulative inequality theory, stipulating that abuse and ill-health aggravate each other throughout the life course.
Clinical implications
Our finding that one in six patients had experienced elder abuse and the association found between life-course victimization and mental ill-health indicate that abusive experiences are important to consider in health care encounters. Previous studies have found that abusive experiences often go unnoticed in Swedish health care, and that many health care providers have never spoken to older patients about abuse [20, 55]. Hence, there is a need to increase the detection of elder abuse and earlier life experiences of abuse to improve the care given to older adults. One potential way forward could be to increase the use of screening instruments for detecting elder abuse, such as the REAGERA-S which also uses a life-course perspective [38].
Limitations
One major limitation of this study was the rather small study sample, which was a consequence of using data previously collected for another purpose. Hence, the lack of significant associations between background characteristics and elder abuse (Table 2), as well as between some of the background characteristics and reporting the different forms of mental ill-health (Table 3), should be interpreted with caution. Also, because of the rather small sample size, all estimates come with wide confidence intervals and the odds ratios are less precise. However, despite the rather low number of participants, previous victimization was associated with elder abuse, and experiences of victimization both before and after 65 years were associated with different forms of ill-health. This finding indicates that a life-course perspective in research on elder abuse is important.
Data was only collected from internal medicine and geriatric wards, which limits the generalizability of the results. The prevalence rate might have been different had we also included patients from other settings, such as orthopedics or surgery. Also, as shown in Fig. 1, 38% (n = 115) of those asked to participate in the study (n = 306) declined and 24% (n = 46) of those who completed the first step by answering REAGERA-S (n = 191) did not participate in the interview. The total participation rate was 44% (n = 135/306), which is a limitation in terms of generalizability but is in line with previous research, e.g., a survey about elder abuse conducted in seven European countries, including Sweden, where the reported response rate was 45% [9]. Likewise, an overview of studies about intimate partner violence conducted in Sweden reported response rates in included studies ranging from 35 to 64% [50].
As previously mentioned, this was a cross-sectional study, and hence causality between background characteristics, abusive experiences, and mental health cannot be assessed. It should also be noted that the used measures of mental ill-health have some limitations. Diagnoses and medications do not necessarily mirror either the respondents’ level of suffering or their quality of life, and it would have been interesting to include measures of such aspects. Also, we did not consider protective factors for abuse or ill-health. Many previous studies on the relationship between victimization and mental ill-health highlight the impact and mediating effect of social support [4, 37], which we did not examine. Nor did we consider potential risk factors for elder abuse at societal level, such as family norms or organization of health care. For example, ageism is a factor at societal level which has been suggested as a risk factor for elder abuse that is rarely considered in studies [56, 57]. Future studies could consider this factor by measuring the effect of individual experiences of ageism, or ageist patterns in the older adult’s environment.