Characteristics of Elder Abuse Perpetrators by Sexual Orientation and Gender Identity of the Abused: Findings from the Canadian Longitudinal Study on Aging

Gloria Gutman ^{1, *}, Mojgan Karbakhsh ², Heather Stewart ³

1. Simon Fraser University, Gerontology Research Centre & Gerontology Dept, British Columbia, Canada

2. Simon Fraser University, Gerontology Research Centre, British Columbia, Canada

3. Simon Fraser University, Gerontology Research Centre & University of British Columbia, Djavad Mowafaghian Centre for Brain Health, British Columbia, Canada

* Correspondence: Gloria Gutman

Academic Editor: Ines Testoni

Special Issue: Elder Abuse in the LGBT Community: A Hidden Problem

Received: December 20, 2023 | Accepted: April 08, 2024 | Published: April 16, 2024

OBM Geriatrics 2024, Volume 8, Issue 2, doi:10.21926/obm.geriatr.2402278

Recommended citation: Gutman G, Karbakhsh M, Stewart H. Characteristics of Elder Abuse Perpetrators by Sexual Orientation and Gender Identity of the Abused: Findings from the Canadian Longitudinal Study on Aging. OBM Geriatrics 2024; 8(2): 278; doi:10.21926/obm.geriatr.2402278.

© 2024 by the authors. This is an open access article distributed under the conditions of the Creative Commons by Attribution License, which permits unrestricted use, distribution, and reproduction in any medium or format, provided the original work is correctly cited.

1. Introduction

While a universally agreed upon definition of elder abuse (EA) remains elusive, the abuse of older people is generally defined as a single or repeated act, or lack of appropriate or required action, occurring within any relationship where there is an expectation of trust, and which causes harm or distress [1,2]. Common types of EA include psychological, physical, sexual, financial, and neglect [3]. Our prior research [4] and that of Burnes et al. [5] explored data from community-dwelling participants of the Canadian Longitudinal Study of Aging (CLSA) as does this study and found that psychological abuse is more common than financial and physical abuse respectively.

While characteristics of victims, perpetrators and their relationship to each other can contribute to the risk or occurrence of abuse, scant empirical studies exploring perpetrators by abuse type has stalled the development of targeted, effective preventive actions [6]. Jackson and Hafemeister observed that perpetrators of financial abuse were more likely to be non-relatives e.g., paid care providers or strangers. The duration of abuse was significantly shorter when the perpetrator was a non-relative [7]. Burnes et al. examined victim-perpetrator dyads and found that those characterized by a separate living arrangement had shorter duration abuse. They concluded that a perpetrator-oriented approach in elder mistreatment cases is a priority research area, in order to optimize protective services outcomes [8].

These historical observations demonstrate a need for better understanding of how the relationships between perpetrators and victims impacts the abuse process, and outcomes of abuse [9]. Individuals already marginalized, with heightened vulnerability, warrant additional study. The focus of this paper is on LGBT+ older adults. Compared to their heterosexual peers, LGBT+ older adults are more likely to live alone and less likely to be partnered [10]. Estrangement of biological family leads to reliance on friends and “chosen” family as caregivers [11]. The intersection of aging, diminishing power, HIV-related morbidity, lifetime discrimination and social exclusion, and unique care needs may make LGBT+ older adults particularly vulnerable to certain abuse types and to abuse by multiple perpetrators [12].

Our prior research showed weighted prevalences of psychological, physical, and financial abuse to be 8.8%, 1.3% and 1.3% respectively among heterosexuals (overall 10.0%), compared with 10.5%, 1.9% and 5.2% (overall 12.0%) among LGB participants [4]. Prevalence rates for psychological and financial abuse were highest among LB females. Differences in socio-demographic characteristics, health, and lifestyle variables may explain LB women’s heightened vulnerability, as might differences in perpetrator characteristics, the focus of the present study. The wide-reaching and cumulative effects of societal prejudices may have a significant impact on the quality of life for lesbian couples in later life. Due to the fear of discrimination and social isolation, intimate relationships are often viewed as a safe space for older lesbians, making the circumstances even more challenging when things go wrong [13]. As Cook-Daniels describes, same-sex couples may have all their assets under one name in order to avoid “questions” about having the names of two men or two women on a property or bank account. When financial abuse by a partner occurs, victims are left in a difficult position to reclaim their belongings and may feel that they have no choice, except to stay with their perpetrating partner [14]. Other studies show that LB women are at a significant financial disadvantage in general and face a more difficult task when planning for their retirement. While the statistics on lesbian retirement challenges are limited, LB women are less likely to transit into a financially secure retirement [15].

While our own studies [4,16] have shown higher rates of abuse among sexually diverse older adults compared to their heterosexual peers, studies comparing perpetrator characteristics of LGBT+ and heterosexual victims are scarce [17]. In this study, we examined perpetrator characteristics, including their sex, relationship to and living arrangement vis-à-vis the victim. Analyses explored characteristics of perpetrators of EA, by abuse type (psychological, physical, financial), and by sexual orientation-gender identity (SOGI) of victims (heterosexual men, heterosexual women, gay and bisexual men, lesbian and bisexual women). Prior research has shown that most victims of EA are abused by a lone perpetrator [18,19]. We also aim to examine the prevalence of poly-victimization, defined by elder abuse researchers [18,19] as where an older adult is harmed through multiple co-occurring or sequential types of abuse by one or more perpetrators or experiences one form of abuse perpetrated by multiple others with whom they have a personal, professional, or care recipient relationship in which there is a societal expectation of trust.

Our research questions explored perpetrator numbers and their characteristics both overall and by SOGI of victims within the Canadian Longitudinal Study on Aging (CLSA) cohort. These questions include:

1. Of the CLSA participants who reported they were victims of EA, what proportion were abused by one vs. multiple perpetrators, and what is the pattern according to SOGI of victims?
2. Which type of abuse (psychological, financial or physical) has the highest perpetrator-to-victim ratio?
3. What are the perpetrator profiles across abuse types? Characteristics of interest included perpetrator sex, their relationship with the victim, and whether they lived in the same household as the victim.

2. Methods

2.1 Data and Sample

Data derive from the Canadian Longitudinal Study on Aging (CLSA), a national cohort study of 51, 338 Canadians aged 45-85 at enrollment (2012-2015), with follow-up testing and interviews conducted every 3 years for at least 20 years or until death or loss to follow-up. Eligibility criteria included being physically and cognitively able to participate on their own at enrollment, able to communicate in English or French, not being full-time members of the Canadian Armed Forces, or residing in Canada’s three territories, First Nation reserves, or in long-term care facilities. Following baseline data collection, accommodations are made for participants to continue in CLSA future cycles if diminished physical or mental capacity occurs [20,21].

The CLSA is composed of two complementary cohorts. The Comprehensive cohort (n = 30,097) includes participants randomly selected from within 25-50 km of 11 data collection sites in 7 provinces. Comprehensive cohort participants are interviewed in person, assessed physically and cognitively, and provide blood and urine samples. The Tracking cohort (n = 21,241) includes persons randomly selected from 10 provinces, and all questionnaires are administered by computer-assisted telephone interviews. Tracking cohort participants were recruited from persons expressing interest in the CLSA following participation in the Canadian Community Health Survey on Healthy Aging or by receiving mail-outs from provincial health ministries or, by means of random-digit dialing. Upon enrollment, all participants were given the option of providing their health card number for future linkage to provincial administrative health databases [22].

Data from both CLSA cohorts includes psychological, medical, social, lifestyle, economic and biological variables. Overall, 48,893 participants completed the first follow-up (95% retention) in mid-2018, the second follow-up was completed in mid-2021, and follow-up 3 is in progress.

2.2 Ethical Approval

The CLSA is approved by McMaster University’s Health Integrated Research Ethics Board (HiREB) and by research ethics boards at all collaborating institutions. The current study is a secondary analysis of CLSA data, approved by the Research Ethics Board of Simon Fraser University (#30000518). CLSA design and methods are described in more detail elsewhere [20,22] and all data collection tools and survey questionnaires are accessible on the CLSA website [23].

2.3 Measures

For this study, we analyzed data from participants aged 65 and over at follow-up 1 and who completed the CLSA elder abuse module questions (n = 23,466). The CLSA elder abuse module is not administered to participants under age 65. The CLSA elder abuse module is adapted from work by the National Initiative for the Care of the Elderly [24]. The conceptual definitions were developed in a series of pilot studies [25] and estimates for Canadian older adults are provided in the Canadian National Survey on the Mistreatment of Older Canadians in 2015 report [3].

The CLSA elder abuse module focuses on 3 types of abuse: psychological, physical, and financial. To provide context for the participants, the following sentences were provided as part of the preamble to the elder abuse module: “next I would like to ask you about some things that might cause you emotional distress. Sometimes, people close to you such as a partner, spouse, family member, friend or someone who takes care of you can cause you emotional distress.”

EA module questions probe for 4 forms of psychological abuse (being criticized; insulted; threatened or intimidated; excluded or ignored), 6 forms of physical abuse (pushed, shoved, or grabbed; had something thrown at; hit or slapped; hit with something; someone tried to choke; threatened with a weapon) and 3 forms of financial abuse (made to handover money, possessions, or property; money, possessions or property were taken; access to money, possessions, or property were deliberately blocked).

A positive response to any abuse question was followed by additional questions about frequency of the abuse over the past year (once; a few times; many times; every day; almost every day). For this study, physical and financial abuse were deemed to have occurred if the participant reported at least one instance of abuse in the past year. Psychological abuse was deemed to have occurred if criticism, insulting, or exclusion/ignoring were experienced many times or every day in the past year, with an exception for threats/intimidation, which was deemed to have occurred if experienced at least once in the past year [5]. Dichotomous variables (yes or no) were constructed to capture the occurrence of each EA type and overall EA, based on these definitions.

Precautionary techniques were used by CLSA interviewers when administering the EA module, either in person or by telephone. The module includes a mandatory, preamble script advising participants that questions about mistreatment and abuse might trigger emotional distress, that participation was voluntary, that their responses were confidential, and that they could opt out of answering any question or discontinue the module at any time.

To ensure privacy, EA questions were specifically worded to elicit only yes/no responses about abuse experiences. Data on selected perpetrator characteristics were captured, but perpetrators were not named, and participants could decline providing any perpetrator information if they preferred. For the Tracking cohort telephone interviews, additional techniques were used to enhance participants’ safety and privacy, for example, if a perpetrator might become present during the interview. They were asked if there was anyone whom they would feel uncomfortable asking to leave if that person entered the space where the interview was taking place. If yes, a code word was established for the participant to use to indicate that this person was present. Tracking cohort participants were also asked if they anticipated anyone entering the room during the EA module administration, and if so options were provided to continue or schedule a callback to complete the interview when privacy was possible.

Every question in the EA module was followed by questions regarding the following perpetrator characteristics: 1. relationship with the victim (spouse or partner, sibling, child, grandchild, other family member, friend, paid caregiver, or other); 2. sex of the perpetrator (male or female) and 3. living arrangement of the perpetrator (residing in the same household or separately from the victim). Participants could choose more than one perpetrator relationship for each abuse form probed within the 3 EA types.

The explanatory variables in this study included victim sexual orientation (heterosexual, LGB) and victim current gender identity (man, woman). Current gender identity and sexual orientation were combined to derive the variable sexual orientation-gender identity (SOGI) with four categories- heterosexual men, gay-bisexual men, heterosexual women, and lesbian-bisexual women.

2.4 Analytic Approach

The complex survey design was accounted for by using inflation weights, which were proportional to the reciprocals of the individual inclusion probabilities, and were provided in the dataset released to us by the CLSA. Inflation weights were implemented to make the results generalizable to the Canadian population [26]. All analyses were performed with IBM SPSS 22, using descriptive statistics (relative frequencies) to report perpetrator relationships with participants, perpetrator sex and living arrangements with victims at the time of abuse, for all participants combined, and also by victim SOGI. Chi square test was implemented for comparing the association of SOGI and perpetrator relationship type. Number of perpetrator relationship types was considered as a proxy for the number of perpetrators.

3. Results

In reporting results, we first present data comprised of the whole sample which included 11,337 heterosexual men, 11,527 heterosexual women, 298 GB men and 184 LB women. This is followed by findings related to each of the abuse types examined (psychological, physical, financial). We then report findings by type and SOGI.

While most participants experiencing abuse reported only one perpetrator relationship type, 15.0% of psychological abuse victims, 2.2% of physical abuse, and 5.0% of financial abuse victims reported two or more relationship types (Table 1).

Table 1 Number of perpetrators according to relationship types reported and abuse type.

3.1 Psychological Abuse

Among older adults reporting psychological abuse within the preceding year (weighted sample n = 411,526), 67.5% said the perpetrator was a family member, was most likely to be male (55.5%), and did not live with the victim (54.5%) (Table 2). The most common perpetrator relationship was ‘spouse or partner’ (41.8%), followed by ‘others’ (28.1%). ‘Other’ perpetrators were more commonly male than female (17.3% vs 7.7%). For participants who experienced psychological abuse from a ‘spouse or partner’, the perpetrator most likely lived with them when the abuse occurred.

Table 2 Distribution of perpetrator characteristics according to abuse type.

3.1.1 Characteristics of Psychological Abuse Perpetrators by Victim SOGI

Of the older adults who had experienced psychological abuse within the preceding year (weighted sample n = 411,526), 52.2% were heterosexual women (n = 215,003), 45.4% were heterosexual men (n = 186,782), 1.3% were LB women (5,447) and 0.8% were GB men (n = 3,474) (0.2% missing).

While most victims experienced psychological abuse via only one perpetrator relationship, 31.1% of LB women reported experiencing abuse via two or more perpetrator relationships (Table 3). This proportion was 15.1% among heterosexual women, 14.3% of heterosexual men and 7.9% of GB men who had experienced psychological abuse within the preceding year by more than one perpetrator (P < 0.001).

Table 3 Number and percentage of perpetrator relationship types according to abuse type and sexual orientation-gender identity (SOGI) of victim.

The majority of psychologically abused heterosexual women said that the perpetrator was a family member (74.8%). This was less common among heterosexual men (59.9%), followed by LB women (56.1%), and GB men (37.9%) (P < 0.001) (Table 4).

Table 4 Distribution of perpetrator characteristics according to abuse type and SOGI.

Males were identified as perpetrators of psychological abuse by 62.5% of heterosexual women, 49% of GB men, 48.1% of heterosexual men, and 40.1.% of LB women. It is noteworthy that 10.2% of LB women were psychologically abused solely by males, 53% exclusively by females and 29.9% were abused by both males and females.

Most LB women (97.0%) and GB men (82.7%) did not live with their psychological abusers, whereas close to half of heterosexual women (46%) and heterosexual men (48.2%) did live with their abusers (P < 0.001) (Table 4).

While ‘spouse or partner’ was the most common perpetrator-victim relationship for psychologically abused heterosexual men (49.6%) and heterosexual women (39.2%), ‘other’ was the most common relationship for GB men (61.6%) and ‘other family members’ for LB women (41.6%). For GB men, after ‘other’ the next most common perpetrators of psychological abuse were siblings, friends or ‘other family members’ who were usually males not living with them. For LB women experiencing psychological abuse from a sibling, perpetrators were usually females not living with them (Figure 1 & Figure 2).

Figure 1 Psychological abuse perpetrators’ sex by sexual orientation-gender identity of the victim (weighted sample n = 411,526).

Figure 2 Living arrangement with the perpetrator(s) among victims of psychological abuse by sexual orientation-gender identity of victim (weighted sample n = 411,526).

3.2 Physical Abuse

Among older adults reporting physical abuse within the preceding year (weighted sample n = 59,482), 68.4% said the perpetrator was a family member, was male in 53.3% of cases, and lived with the victim in 52.3% of cases (Table 2).

The most common relationship type was a ‘spouse or partner’ (47.3%), followed by ‘other’ (19.5%), ‘child’ (11%) and ‘grandchild’ (7.6%). As expected, most victims experiencing physical abuse from a spouse/partner, lived with them at the time of abuse, while this was much less common regarding other relationship types.

3.2.1 Characteristics of Physical Abuse Perpetrators by Victim SOGI

Of the older adults who had experienced physical abuse within the preceding year (weighted sample n = 59,482), 50.0% were heterosexual women (n = 29,733), 47.2% were heterosexual men (n = 28,071), 0.6% were LB women (385) and 2.0% were GB men (n = 1,198) (0.2% missing).

While most of victims had experienced physical abuse from only one relationship type, 53.1% of physically abused LB women did not report the relation type with the perpetrator and responded to the perpetrator relationship question as “Don’t know/No answer” or “Refused (Table 3).

The majority of physically abused heterosexual women said that the perpetrator had been a family member (79.3%). This was less common among heterosexual men (59.9%), followed by LB women (18%) and GB men (8.6%) (P < 0.001) (Table 4).

Regarding the sex of the perpetrator, men were involved in 71.4% of cases involving physically abused heterosexual women, followed by 56.1% of GB men, 34.3% of heterosexual men and 18% of LB women (P < 0.001). It is noteworthy that 66% of physically abused LB women did not disclose the perpetrator’s sex.

The preponderance of physically abused GB men did not live with the perpetrator (86.2%), nor did 39.6% of heterosexual men, 32.4% of heterosexual women and 28.9% of LB women c It should be noted that 53.1% of physically abused LB women did not disclose whether the perpetrator did or did not live with them.

While the most common perpetrators of physical abuse of heterosexual men and women were their ‘spouse or partner’, GB men and LB women reported physical abuse perpetrated most often by ‘others’. For GB men, the ‘other’ abuser was male in 45.8% and female in 41.7% cases, and did not live with them in most cases. For LB women, ‘other’ abusers were exclusively females who did not live with them (Figure 3 & Figure 4).

Figure 3 Physical abuse perpetrators’ sex by sexual orientation-gender identity of victim (Weighted sample n = 59,482).

Figure 4 Living arrangement with the perpetrator(s) among victims of physical abuse by sexual orientation-gender identity of the victim (Weighted sample n = 59,482).

3.3 Financial Abuse

Among older adults reporting financial abuse within the preceding year (weighted sample n = 59,482), 46.1% said the perpetrator was a family member, was male in 68.0% of cases, and did not live with the victim in 67.6% of cases (Table 2).

The most common relationship type with the perpetrator was ‘other’ (40.3%), followed by ‘child’ (19%). ‘Other’ perpetrators were more commonly male than female (23.4% vs 7.1%) and did not live with the victim when abuse occurred.

3.3.1 Characteristics of Financial Abuse Perpetrators by Victim SOGI

Of the older adults who had experienced financial abuse within the preceding year (weighted sample n = 65,575), 48.9% were heterosexual women (n = 32,036), 43.2% were heterosexual men (n = 28,347), 3.7% were LB women (2,466) and 3.1% were GB men (n = 2,019) (1.1% missing).

While most victims had experienced financial abuse from only one relationship type, 66.5% of financially abused LB women and 24.3% of GB men reported financial abuse by more than one perpetrator relationship type. Also, 24.7% of GB men chose not to report the relationship type of their financial abuse perpetrators (Table 3).

The majority of financially abused LB women said that the perpetrator had been a family member (67.8%). This was less common among heterosexual women (60.4%), followed by heterosexual men (31.6%) and GB men (1.5%) (P < 0.001).

Regarding the sex of the perpetrator, males were involved in 75.2% of cases of financially abused heterosexual women, followed by 67.8% of LB women, 63.4% of heterosexual men, and 32.4% of GB men (P < 0.001). Approximately a quarter of heterosexual men and GB men experiencing financial abuse did not report the perpetrator’s sex.

The majority of LB women did not live with the perpetrator (98.2%), followed by 82.9% of heterosexual men, 75.3% of GB men and 50.8% of heterosexual women (P < 0.001) (Table 4).

The most common financial abuse perpetrator relationship type among heterosexual men and GB men was ‘other’ (56.5%). ‘Child’ was the next most common relationship type among heterosexual men (11.3%) and ‘friend’ was the second most common among GB men (42.6%). Heterosexual women were financially abused by a ‘child’ most commonly (28.2%), followed by ‘spouse/partner’ (25.5%) and ‘other’ (25.2%). ‘Sibling’ (66.5%) and ‘other family members’ (66.5%) were more commonly involved in financial abuse of LB women.

‘Other’ perpetrators of financial abuse among heterosexual men and women were often a male (65.2% and 57.1% of heterosexual men and heterosexual women experiencing financial abuse from ‘other’, respectively), but were commonly a female among GB men (77.3%) and LB women (100%). The ‘other’ perpetrators usually did not live with the victims, except in the case of heterosexual women (57.1% of heterosexual women who experienced financial abuse from ‘other’) (Figure 5 & Figure 6).

Figure 5 Financial abuse perpetrators’ sex by sexual orientation-gender identity of the victim (Weighted sample n = 65,575).

Figure 6 Living arrangement with the perpetrator(s) among victims of financial abuse by sexual orientation-gender identity of the victim (Weighted sample n = 65,575).

4. Discussion

This study is among the first to examine characteristics of elder abuse perpetrators by abuse types and across victim SOGI groups. It yielded novel answers to research questions about abuse experienced by Canadian older adults. We identified differences in the number of victims experiencing abuse perpetrated by multiple individuals, in different relationship roles, among the three abuse types studied. We found the distribution of perpetrator characteristics (relationships and living arrangements with victims, and sex of perpetrators) varied between abuse types, and notably, across SOGI groups. To date only a limited number of studies have examined the characteristics of perpetrators of EA, focusing on EA generally and not considering the SOGI of victims.

4.1 Perpetrator Characteristics

A study of 558 older adults (85% women) exposed to abuse in Israel showed approximately three-quarters of abusers were men, half were adult offspring, and a third were spouses [27]. In our study, perpetrators were most often male (68.0% for financial, 55.5% for psychological, and 53.3% for physical abuse) and most often family members (68.4% for physical, 67.5% for psychological and 46.1% for financial abuse). When stratifying for SOGI however, it was discovered that perpetrators of psychological abuse were more often female, except where victims were heterosexual women. Our findings highlight - as Kosberg elucidates that - “elder abuse should be seen to be perpetrated by, and to, both males and females” [28] and that the SOGI of the victim makes a difference as does the type of abuse when it comes to whether the perpetrator is more likely to be male or female.

A contextual gender-based analysis of EA in Canada showed that women were more likely than men to be victims of abuse by family members, with male spouses and adult children most common perpetrators. In contrast, male victims of EA were more likely to be victimized by their adult children, with only a small proportion experiencing abuse from their spouses [29]. In the US, analysis of 818 calls to the National Center on Elder Abuse (NCEA) resource line showed that family members were the most commonly alleged perpetrators of psychological, physical and financial abuse. This study also revealed that individuals who were ‘known’ to the victim (non-family, non-caregivers) were common perpetrators of psychological and financial abuse [30]. Another study of 528,641 cases of EA reported to Brazil’s Notifiable Disease Information System revealed that adult children were involved as perpetrators in approximately 30 percent of reported EA incidents [31].

In our study, key differences were observed across SOGI groups. Psychological abusers of heterosexual men and heterosexual women were most often spouses or partners. In contrast, psychological abusers of GB men were most often adult children, and for LB women were most often siblings. Physical abusers were most often spouses of heterosexual men and women, as well as GB men. For LB women reporting EA, male grandchildren who lived with them were most common perpetrators. However, more than half of LB women experiencing physical abuse did not report the relationship of their perpetrators nor the perpetrator’s sex, perhaps because they were embarrassed or afraid of characterizing their perpetrators, even within a highly confidential national research platform like the CLSA.

Although financial abuse was most often perpetrated by strangers, family member perpetrators were most often adult children of heterosexual men and women, male partners/spouses of GB men, and somewhat to our surprise, male siblings of LB women who did not live with them.

The other unexpected finding of our research was the high proportion of individuals who reported the perpetrator to be ‘other’ rather than a spouse/partner, family member, or friend. This raises questions about whether traditional categories in commonly used EA surveys are sufficient to distinguish EA from newer forms of crimes against older adults and their differing perpetrator characteristics: for example, for the purposes of differentiating financial EA from mail, phone, and internet-based fraud and financial scams targeting older adults, or for the purpose of distinguishing psychological or physical EA from stranger-based harassment. Psychological or physical assaults are now rising in numbers online and in communities worldwide.

With regard to poly-victimization, Weissberger et al’s. findings based on calls to the National Center on Elder Abuse (NCEA) resource line showed that 18.2% of calls reported multiple abusers with family members more likely to be involved in multiple types of abuse compared to other relationship types [30]. Our results for psychological abuse are similar, with 15.0% of psychological abuse victims reporting abuse by more than one perpetrator. Importantly, 31.1% of LB women who were psychologically abused in the year preceding data collection were abused by more than one perpetrator.

Living arrangement with the perpetrator can also influence elder abuse in multiple ways, including having impacts on the chronicity of the abuse [32]. While previous studies suggest that perpetrators tend to be dependent on and living with their victim(s) [29,33], we observed different patterns across elder abuse types and SOGI groups. Most GB men and LB women reporting psychological or physical abuse indicated their perpetrators did not live with them at the time abuse occurred. It was also noteworthy that more than half of LB women experiencing physical abuse did not respond to the question whether their abuser lived with them. Traditional gender role stereotyping characterizing women as being nonviolent, as well as the strongly held belief that ‘girls don’t hit other girls’ can interfere with the ability of LB women to recognize abuse in their relationships, even when physical abuse is evident. Hassouneh and Glass [34] have demonstrated that in same-sex relationships, the female perpetrator is frequently aware of the potential of using traditional gender role stereotypes to create confusion and uses these to their advantage, playing to be the victim [34]. Other research has indicated that detection and management of EA is more challenging when the perpetrator lives with the victim, or when a close relationship exists between the victim and perpetrator [32]. Older men who are abused by a co-habiting spouse, partner or child are reluctant to report abuse, because of the feeling of shame in the context of gender-role socialization, in particular when a man’s abuser is a woman [28,32]. The results of the present study lead us to conclude that gender-role socialization may play a role in the reluctance of LB women to respond to questions about the sex and living arrangement of their abusers.

Our study found that most financial abuse, across all SOGI groups, was perpetrated by abusers who did not live with their victims, including perpetrators who were identified as family, spouses, children, or siblings. Financial abusers of both heterosexual and GB men were most often non-family ‘others,’ heterosexual women were financially abused most often by a child or spouse, and LB women by other family or siblings. Previous studies have found that perpetrators of financial abuse tend to be more distant relatives who may be financially dependent on their victims [29].

While some previous studies have elaborated on individual-level perpetrator risk factors (e.g. age, mental health problems and alcohol dependence), as well as relationship-level factors (dependence of the perpetrator on the victim) [9,29,35], unfortunately the CLSA surveys to date do not collect this information.

One of the important findings in this study was the distinct pattern of victimization experienced by LB women compared to other SOGI groups. More than 30% of LB women experiencing psychological abuse reported they were victimized by multiple (more than one) perpetrators. It is important to repeat here that our criteria for coding an individual as having experienced psychological abuse in 3 of its 4 forms (criticized, insulted, or excluded/ignored)- required a response of “many times” or “everyday”, versus “once” or “a few times”. Recoding these responses reduced the possibility of including those who may be responding to or recalling an isolated encounter which was quick to resolve, or other encounters perpetrated by individuals whose relationship to the participant does not include an expectation of trust. In other words, our recoded data capture psychological abuse that is more chronic in nature, and likely to be more harmful over time. The exception was psychological abuse form “threatened or intimidated” which was considered to have occurred if experienced at least once in the past year. Our revised threshold for coding an exposure to psychological abuse, combined with the high percentage of LB women reporting psychological abuse by multiple perpetrators, sheds light on a significant and serious vulnerability of LB women to mental and physical health impacts of chronic, high “dose” psychological abuse. We also found more than half of LB women reporting physical abuse chose not to disclose details (relationship and/or living arrangement, or sex) about their perpetrators, implying they live in fear of repercussions of disclosing their abuse, even where it was anonymous disclosure in the context of CLSA’s confidential and highly secure research platform.

4.2 Implications for Policy, Practice and Future Research

The finding that LB women and GB men less commonly lived with their perpetrators, has practical implications for prevention, specifically to be cautious of strangers or their offers of assistance or companionship. However, different policies and practice recommendations for prevention or response to elder abuse should be considered not just based on partnered vs. non-partnered status of abused older adults or their living arrangement (with vs. without the perpetrators) as has traditionally been the case but also taking into consideration the victim’s SOGI and life experience of homophobia and transphobia.

Further, while trauma-informed training should be required for all levels of care providers to older adults -especially it should be a required for those working with LB older adults - to promote timely detection of abuse and appropriate interventions, including providing safe shelters. Countering stereotypical thinking and gendered-biased socialization as well as homophobia and transphobia also need to be a key component of worker training. Additionally, the LB population warrants more study, including community-based participatory research, to identify the full extent and impacts of EA and taking into consideration other forms of abuse experienced over their life course.

4.3 Study Limitations

The primary limitation of this study is the small proportion of SOGI minority participants included in the CLSA (1.8%). This restricted our ability to stratify and analyze subgroups beyond heterosexual, homosexual, and bisexual. Similarly, gender identity could not be stratified beyond men and women (e.g. transman, transwoman, genderqueer, other) given the small number of gender minorities overall in the CLSA who were old enough to be administered the elder abuse module (age 65 or more). We acknowledge that combining bisexual individuals with those who had similar gender orientation may have obscured their unique abuse experiences. We also recognize that perpetrator characteristics and patterns will not be identical among cisgender and transgender older adults. There may be limited potential for such research using CLSA data. However, future CLSA cycles will include more individuals who meet the age 65 threshold for eligibility to answer its elder abuse questions, and this may include more sexual and gender minority participants as well. While findings from the current small sample of sexual minority men and women should be interpreted with caution, the value of repeat administration of an elder abuse module in CLSA follow-up 3, currently in progress, should not be underestimated. It represents an opportunity to fill knowledge gaps concerning incidence of elder abuse among minority SOGI groups who have historically been marginalized, but who now form an important and strong community within the Canadian population.

A further limitation is that the CLSA elder abuse module only looks at psychological, physical and financial abuse, and not sexual abuse, neglect, or frauds and scams, as noted above.

5. Conclusion

Our study provides insight concerning the distinct characteristics of perpetrators across the three main types of elder abuse and among SOGI minority and majority victims. These findings should be taken into consideration when considering abuse dynamics experienced by community-dwelling older adults and in the development of preventive strategies [6,9,31,32]. While most of our findings concern individuals who experienced victimization by only one relationship type, the finding that 15% of LB women experienced psychological abuse from more than one relationship type identifies a topic that should be the subject of future investigation.

Acknowledgments

This research was made possible using data collected by the Canadian Longitudinal Study on Aging (CLSA). Opinions expressed in this manuscript are the authors’ own and do not reflect the views of the Canadian Longitudinal Study on Aging (CLSA). Funding for the Canadian Longitudinal Study on Aging (CLSA) is provided by the Government of Canada through the Canadian Institutes of Health Research (CIHR) under grant reference: LSA 94473 and the Canada Foundation for Innovation, as well as the following provinces, Newfoundland, Nova Scotia, Quebec, Ontario, Manitoba, Alberta, and British Columbia. The current research was funded through the British Columbia Council to Reduce Elder Abuse and uses the CLSA dataset [Baseline Tracking Dataset - Version 3.7, Baseline Comprehensive Dataset - Version 6.0, Follow-up 1 Tracking Dataset - Version 2.2, Follow-up 1 Comprehensive Dataset - Version 3.0 and Vital Status] under Application ID #2104013. The CLSA is led by Drs. Parminder Raina, Christina Wolfson and Susan Kirkland.

Author Contributions

Conceptualization: GG, MK, HS; Data Curation, analytical methodology and analysis: MK; Original draft preparation: MK, GG; Manuscript writing, review and editing: GG, MK, HS.

Competing Interests

The authors have no conflict of interest to disclose.