Current frameworks on violence against women and existing evidence

Gaps in research and evidence stem in part from conflicting theoretical approaches, definitions and conceptual frameworks concerning violence against older women. The dominant theoretical frameworks are the older adult mistreatment framework and older adult protection framework [7, 13, 14]. The older adult mistreatment framework conceptualizes violence against older women as a form of elder abuse, focusing on age as the primary factor influencing vulnerability to exposure to violence. The older adult protection framework specifically understands violence within the context of care-giving and institutional arrangements, where older adults’ often be gender neutral, and the adult protection framework can result in a framing of older adults as inherently impaired and vulnerable. In addition, the IPV framework primarily understands vulnerability to violence in terms of gender inequality and partnership dynamics, which may neglect analysis of how ageing and partner violence intersect. These differing frameworks inform multiple aspects of research, including study design, data collection and analysis, and reporting, resulting in fragmented data and evidence. For example, some research utilizing the older adult mistreatment framework lacks a focus on the gendered dimensions of violence [14, 15], and other studies have solely focused on women in institutional settings, neglecting measurement of violence perpetrated by intimate partners and other family members [13].

Existing syntheses of evidence on violence against older women often reflect these differing conceptual frameworks. Employing an older adult mistreatment framework, a systematic review of quantitative studies of elder abuse (against men and women aged 60+) found that the global prevalence of elder abuse in community settings is 15.7% in the past year, with psychological abuse and financial abuse as the most prevalent forms of abuse reported [16]. This review reported prevalence by type of violence, but did not report on perpetrators. Analysis of studies conducted in institutional settings found women, aged 60 and above, to be significantly more vulnerable to abuse, with psychological abuse as the most prevalent form of violence, followed by physical violence, neglect, financial and sexual abuse [17]; this analysis included data reporting staff-to-resident abuse. Analysis of quantitative data of women aged 60 and above in the systematic review of quantitative studies of elder abuse found a global prevalence of elder abuse against women of 14.1% in the past year, with psychological abuse reported as the most prevalent form of violence, followed by neglect [16]. The focus of this review was prevalence of different sub-types of violence, and type of perpetrator was not considered. Another systematic review of quantitative data on interpersonal violence (physical and/or sexual violence) against older women (aged 65 and above) in community dwellings primarily employed an IPV framework, finding prevalence of reported interpersonal violence ranged from 6 to 59% over a lifetime, from 6 to 18% since turning 50, and 0.8 to 11% in the past year, however, results indicated that definitions of violence vary widely and affect prevalence estimates [18]. Syntheses of quantitative literature have identified prevalent forms of violence against older women, highlighting limitations in the evidence-base due to variations in definitions and methodology, and a primary emphasis on populations in high-income, Western countries. These reviews have captured a wide range of types of violence, however, have not considered type of perpetrators or patterns of co-occurring types of violence.

Alongside these systematic reviews of quantitative data, some reviews have included qualitative and mixed methods studies. An empirical review of IPV in later life examined 27 quantitative, 22 qualitative and 7 mixed-methods studies, finding that forms of IPV amongst older women in later life shifted from a higher prevalence of physical and sexual abuse during reproductive years, to a higher prevalence of forms of psychological abuse [19]. A review of qualitative research on IPV amongst older women identified a number of relevant themes, including patterns of abuse that were continuous and consistent with previous experiences of abuse in families of origin and previous relationships [20]. A systematic review and meta-synthesis of qualitative studies of IPV and older women focused on how previous exposure to IPV influenced health-seeking behaviours, specifically mental health care [21]. An empirical review of quantitative and qualitative studies of sexual violence against older people identified widespread variation in prevalence rates across studies, and a range of perpetrators, primarily intimate partners or adult children [22]. A recent narrative review of quantitative, qualitative and mixed methods studies of IPV against women aged 45 and above concluded that women’s “age and life transitions mean that they may experience abuse differently to younger women. They also face unique barriers to accessing help, such as disability and dependence on their partners” [23].

However, amongst these existing systematic reviews of qualitative literature, none have focused specifically on older women, while also being inclusive of any form of violence. In order to improve understanding of violence against older women, it is important to explore patterns, dynamics and experiences through examination of the qualitative literature. Qualitative data on violence against older women complements quantitative evidence not only by offering insight into lived experiences of older women subjected to violence, but also by expanding and clarifying types of violence, perpetrators, linkages to particular risk factors, and physical, mental and social impacts of violence against older women.

In the present review, we aimed to build on previous systematic reviews and strengthen the evidence-base by i) including studies and evidence focused specifically on women; ii) including any form of violence against women, rather than adopting a specific theoretical framework on what types of violence or perpetrators should be included from the outset; iii) focusing on women aged 50 and above (as many surveys often specifically focus on women of reproductive age, which is considered to be up to 49 years of age); and iv) focusing specifically on qualitative studies, to explore the nature and dynamics of violence against older women from the perspective of women. We aimed to identify, evaluate and synthesize qualitative studies from all countries, exploring violence against women aged 50 and above, identifying types and patterns of violence, perpetrators of violence, and impacts of violence on various outcomes for older women, including physical and mental health and social support, and women’s responses to experiences of violence. We include the following forms of violence: elder abuse, family violence and intimate partner violence. Elder abuse is defined as “single or repeated act, or lack of appropriate action, occurring within any relationship where there is an expectation of trust which causes harm or distress to an older person” [24]. Intimate partner violence is defined as “behaviour by an intimate partner or ex-partner that causes physical, sexual or psychological harm, including physical aggression, sexual coercion, psychological abuse and controlling behaviours” [25]. Family violence is often used interchangeable with intimate partner violence, however, also encompasses abuse and violence perpetrated by other family members, for example, adult children or in-laws. While there is no universal agreed-upon definition of older women, for the purposes of this review, we define older women as women aged 50 and above, while recognizing that aging and age are social phenomenon, and definitions vary across organizations, cultures and communities. The protocol was pre-registered with PROSPERO, Registration Number: CRD42019119467, https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42019119467 (see also [26]).

Search strategy

In this systematic review, we searched 11 electronic databases–PubMed, PsycINFO, Embase, CINAHL, PILOTS, ERIC, Social Work Abstracts, International Bibliography of the Social Sciences, Social Services Abstracts, ProQuest Criminal Justice and Dissertations & Theses Global, from 1990. We conducted searches that combined the following domains as part of the research question: 1) age (50 and above); AND 2) women; AND 3) violence; AND 4) qualitative methodology. For each of these domains, we identified the relevant keywords and search terms, which varied by database; the search strategy was appropriately modified for each database, including syntax and specific terms, topics and/ or headings. The search strategy for PubMed is included in S1 File. Searches were conducted in April 2018 and updated in July 2019. We did not limit the search by year of publication or language.

We also hand searched reference lists of relevant existing systematic reviews, which we identified both through background research and through the formal database searches, and reviewed relevant references (44 identified). We consulted with 49 experts on violence against older women or older adults, including researchers, practitioners and policy makers, from all regions globally. All experts were contacted and followed-up with a minimum of 2 contacts. 26 experts responded with 424 articles, 64 of which were duplicates. We reviewed the full text of 43 articles and ultimately included 2 in the full review. Grey literature was not systematically searched; grey literature submitted by experts was initially considered for inclusion, however, conducting comparable data extraction and quality assessment for grey literature alongside the peer-reviewed literature was not possible.

We identified 18 non-English language articles for full-text review. For 17 of these articles, we identified a native speaker external reviewer who was provided with inclusion and exclusion criteria and consulted with authors regarding final inclusion (4 Portuguese, 7 Spanish, 1 Hebrew, 1 Dutch, 1 German, 1 Danish, 2 French). One non-English article (in Farsi) was not reviewed as the research team could not engage a Farsi speaker to review the article. The external reviewers consulted with SRM to decide on inclusion of full texts, and conducted data extraction and quality assessment on 3 articles identified for inclusion (2 Spanish, 1 Portuguese) [27–29].

Study selection and data extraction

After removing duplicates, study selection proceeded in two stages: in the first stage, two authors (SRM and MEL) reviewed titles and abstracts of all identified manuscripts. We included studies that met the following criteria: i) focused on women aged 50 and older, ii) employed qualitative methodology, and iii) focused on women’s experiences of any type of violence perpetrated by any type of perpetrator. Studies including men or also including women aged younger than 50 were included if specific and separate sex and age-specific analyses were included. We included studies employing any type of qualitative methodology, and mixed methods studies were included if qualitative data was presented separately. Studies were excluded if the whole sample was children, adolescents or adults under the age of 50; if the sample only included men; if the methodology was quantitative, or in the case of mixed methods studies, if the qualitative results were not separately presented, and if the data only included the perspectives on violence against women as reported by care providers, health professionals, legal professionals and nursing home managers.

After the first stage of title and abstract review, we reviewed the full text of any manuscript considered relevant by either of the authors. In the second stage, two authors (SRM and MEL) independently reviewed all articles selected for full text review for eligibility, to reach consensus on inclusion. Any discrepancies were resolved with the input of an external reviewer. Fig 1 indicates the full search and study selection process.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Identification of included studies.

From: Moher D, Liberati A, Tetzlaff J, Altman DG, The PRISMA Group (2009). Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement. PLoS Med 6(6): e1000097. doi:10.1371/journal.pmed1000097. For more information, visit www.prisma-statement.org.

https://doi.org/10.1371/journal.pone.0239560.g001

We designed a data extraction Excel spreadsheet specifically for the purposes of the review, including characteristics of included studies (location of the research, research question), methodology (conceptual framework or theoretical approach, data collection methods, data analysis methods, sampling), characteristics of the sample (inclusion and exclusion criteria, brief description of the sample), types and nature of violence (context of violence, perpetrator and brief description of impacts of violence). We extracted main findings, participant quotations where possible, and study limitations, if reported. Data extraction was conducted by one author (MEL), and checked for accuracy by a second author (SRM), with discrepancies resolved by discussion to reach consensus.

Quality assessment

All included studies were assessed for quality using an adapted version of the Critical Appraisal Skills Programme [CASP] scale. The adapted scale included the following questions [30]:

1. Was there a clear statement of the aims of the research?
2. Is a qualitative methodology appropriate?
3. Are the setting(s) and context described adequately?
4. Was the research design appropriate to address the aims of the research?
5. Is the sampling strategy described, and is this appropriate?
6. Is the data collection strategy described and justified?
7. Is the data analysis described, and is this appropriate?
8. Are the claims made/findings supported by sufficient evidence?
9. Is there evidence of reflexivity?
10. Does the study demonstrate sensitivity to ethical concerns?

Two authors (SRM, MEL) assessed the quality of the studies, assigning a 1 for each affirmative response and 0 for each negative response, for a final score out of 10. Disagreement was resolved by discussion between the two authors. Quality assessment was not used to determine if any studies should be excluded, but rather to assess the strength of each study.

Synthesis

An Excel spreadsheet to compile all relevant findings and quotations from the studies for thematic analysis was developed. Two of the authors (SRM and MEL) coded the main findings extracted from each study. We used line-by-line coding on a sub-set of articles, developing a set of over-arching themes and sub-themes for a draft codebook. The coding proceeded as an iterative process, with the two authors each separately coding the main findings using the draft codebook, discussing coding results, and refining the codebook based on overlap and redundancies identified. After all data were coded, we tallied all occurrences of each code and further explored areas of overlap and merged sub-themes with low numbers of codes, finalizing the broad themes and focused sub-themes. For non-English articles included, the external reviewer translated primary quotations into English and thematic analysis on these articles was conducted alongside the English language articles.

Reporting

The synthesis and all aspects of the systematic review process are reported following the 21-item checklist provided in the Enhancing Transparency in Reporting the Synthesis of Qualitative Research statement [31] and the PRISMA checklist [S2 and S3 Files].

Studies identified and characteristics

Our searches of 11 databases yielded 9318 articles, with an additional 468 articles identified through cross- referencing and expert recommendation. After removing duplicates, 7834 articles remained. We identified 417 articles that were potentially eligible and included in full text screening. Two of these articles had not yet been published. Additionally, 1 Farsi language study was unable to be translated and assessed against the selection criteria. Fifty-two articles met criteria for inclusion in this systematic review (Fig 1). The 52 included articles represent data from 31 studies.

Overview of study characteristics

Study setting (Table 1).

Most studies were conducted in high-income countries (HIC), including the United States of America (n = 16), Israel (n = 12), Canada (n = 7), the United Kingdom (n = 4), Hong Kong (n = 2), and Australia (n = 1). Six articles were from upper-middle income countries–Brazil (n = 3), Mexico (n = 2) and Iran (n = 1); and three articles were from low-income countries–Uganda (n = 1) and Ethiopia (n = 2). One article came from India, a lower-middle income country.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Characteristics of included studies.

https://doi.org/10.1371/journal.pone.0239560.t001

Quality assessment

Application of the adapted version of CASP scale yielded variable results across the 52 articles assessed [see Table 2]. Ratings of research methodology, statement of research aims and selection of appropriate research design were overall high. The majority (46 articles) [29, 32–76] gave support for research findings with references to primary data (participant quotations, case study vignettes, case file excerpts). Ten articles [41, 46, 49, 50, 59, 65, 77–80] lacked data analysis descriptions.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2.

https://doi.org/10.1371/journal.pone.0239560.t002

Only 12 articles [29, 35, 38–40, 45, 51, 58, 72–74, 79] reflected on the relationship between the researchers and the participants (reflexivity). Procedures for ethical research were described in 36 articles [27, 29, 33–37, 39–41, 45–52, 54, 55, 58, 60, 61, 63, 64, 66–72, 74, 76, 77, 79]. Five articles [43, 44, 56, 73, 75] described obtaining consent, but lacked descriptions of ethical approval, and 10 articles [32, 38, 42, 53, 57, 59, 62, 65, 78, 80] lacked descriptions of both ethical approval and obtaining consent. A significant number of articles [32, 34–42, 45–47, 49–51, 53, 54, 56, 57, 59, 61–66, 69, 70, 72–75, 77–79] lacked adequate descriptions of the study setting and context.

Descriptions and patterns of types of violence

Older women described IPV, family violence and elder abuse of various types, perpetrated by a range of perpetrators [Table 1]. Among the specific types of violence reported in the articles in this review, across IPV, elder abuse and family violence, physical violence was most frequently reported [27, 32–54, 57, 60–63, 66, 69–80], followed by emotional/ psychological [28, 32, 36–39, 41–51, 53, 54, 56, 60–62, 66–80], economic/ financial [34–36, 39, 41, 43–45, 48–50, 61, 62, 64, 68, 71, 72, 74–78], sexual [27, 33, 34, 40, 42, 49, 50, 54, 57, 59, 60, 67, 72, 74–76, 80], verbal [32, 40, 45, 52, 62, 68–70, 73, 77], controlling behaviors [45, 48, 49, 51, 53, 64, 67, 70, 76, 77], and lastly, neglect [28, 35, 39, 61, 62, 68, 71, 76].

Older women’s experience of IPV was the most frequent form of violence reported (42 articles) [27, 29, 33, 34, 36–38, 40–42, 45–56, 59–65, 67–80]. Older women described on-going instances of neglect, verbal abuse and financial exploitation in a study conducted in India [41], in other cases, physical violence characterized earlier and on-going experiences of violence within intimate partner relationships [37, 40, 47, 54]. IPV in particular was described by older women as occurring throughout different stages in the relationship, spanning their youth and into older age. Older women often experienced an escalation of IPV and controlling behaviors despite the age and/ or illness of their partner [36, 40, 46, 61, 69, 77]. Changing relationship dynamics due to ageing–including a husband’s retirement, children leaving the home, women wanting to engage in activities outside of the home, or diagnosis of a chronic or terminal illness–triggered escalating IPV [36, 40, 46, 47, 56, 69]. Shifts in types of violence, from predominantly physical violence to predominant psychological abuse and neglect, were commonly described in studies that encompassed previous and on-going IPV [34, 51]. Studies focused on IPV commonly described both previous and on-going violence, and a smaller number described only or primarily violence experienced while aged 50 or above [64, 69].

Violence occurring within the family was discussed in 15 articles [35, 39, 43, 44, 46, 48, 49, 55, 63, 66, 68, 69, 71, 76, 79], with perpetrators including family members not including children [27, 43, 44, 48, 49, 59, 63, 68, 69, 71] and adult children [35, 39, 48, 66, 76, 79]. Studies captured instances of physical and verbal aggression by mentally ill adult children against older women [66], violence enacted by elderly with dementia against older women who were caregivers [46], and forms of neglect, financial exploitation and emotional abuse enacted by family members, including children [44]. The majority of these studies focused on violence experienced in older age, while one study explored dynamics of abuse between children and mothers across the lifespan [35, 39].

10 articles reported on experiences of elder abuse [28, 32, 35, 43, 44, 55, 57–59, 78] with perpetrators including community members [43, 44], caregivers [57, 59], nursing home residents [58] and health care providers [28]. Types of elder abuse included verbal abuse, physical assault and inappropriate sexual advances [58] and sexual assault [59].

Financial control spanned instances of elder abuse, family violence and IPV [43, 53, 62, 64, 76, 78], and was described as co-occurring with and resulting in other forms of violence. Financial exploitation could result in emotional and/ or physical violence if older women resisted control [62, 64]. An older woman explained that in the context of her relationship with her husband, “If I did not follow his control [over money], he would be verbally abusive” [64].

Themes and sub-themes identified through coding are displayed in Table 3.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Themes, sub-themes and quotations.

https://doi.org/10.1371/journal.pone.0239560.t003

Intersection of ageing and violence

A number of sub-themes emerged emphasizing the interconnections between the experience of ageing amongst older women, and dynamics, impacts, experiences and perceptions of violence.

Perpetrator-related factors

Some included studies reported on perpetrator-related factors that initiated or exacerbated forms of violence against older women.

Social and gender norms regarding response to violence

Older women described the ways in which social and gender norms shaped their experiences of and responses to violence.

Lifelong IPV

Many older women described experiences of IPV throughout their life-course. Several sub-themes were identified related to lifelong patterns of violence, cumulative consequences of IPV, and linkages of violence in older age to earlier experiences of violence.

Needs of older women affected by violence

Older women who reported exposure to violence described various needs in terms of social support, access to services, and issues accessing these services due to their age.

Coping mechanisms

Older women reported various approaches to coping with the experience and impacts of different forms of violence, often employing several different coping mechanisms such as leaving relationship with an abuser and emotion-based coping strategies such as alcohol or drug usage, in order to navigate difficult decisions, maintain their health and well-being, and protect other family members in the context of exposure to violence.

Limitations and strengths

Several strengths and limitations should be considered while interpreting the findings of this systematic review. In contrast to previous systematic reviews, we included all qualitative evidence concerning violence against older women, regardless of type of violence and perpetrator, allowing insights into the overall focus of the evidence-base, which revealed limited engagement with elder abuse against women and family violence perpetrated by non-partners, for example, children. Additionally, we followed a rigorous protocol, adhering to a preregistration protocol in line with ENTREQ guidelines [31]. We carried out an extensive systematic review across 11 databases, supplemented by hand searched references lists and article recommendations from 49 experts on violence against women or older adults, and therefore it is unlikely that published articles would have been overlooked in this review. We reviewed all articles in any language, apart from Farsi.

In order to minimize selection bias or for relevant articles to be missed, two authors screened all titles and abstracts and all articles at the full text review stage. At the data extraction phase, only MEL extracted relevant data, introducing the possibility of transcription errors. Despite this limitation, all extracted data was double checked by SRM to minimize potential of missing descriptive data, and both completed independent quality appraisals to minimize potential for biased assessments. Additionally, during the analysis phase, both authors coded article main findings and key quotes, and developed descriptive and analytical themes to strengthen the interpretation and synthesis of findings.

Another limitation of the findings of this review is the concentration of studies in higher-income contexts, which greatly limits the transferability of findings to low- and middle-income populations. In addition, the small number of studies conducted in low and middle-income countries entailed that comparison of patterns between high-income and low and middle-income contexts was not possible. This review was also limited by the quality of included articles. Many articles did not clearly report on study setting and context, sampling procedures, data analysis, reflexivity, and research ethics. Moreover, many articles included samples of older women across wide age ranges (e.g. 65–85 years old). The available evidence does not disaggregate findings to enable understanding of whether or how women in different age groups experience violence differently, despite significant variation in living conditions, employment and health status of women aged 50–64 vs. 65 and up, for example. This lack of specificity limited our ability to understand the differential causes, experiences and impacts of violence among specific age groups of older women.

Implications for future research

In light of the findings from this review, as noted above, there is an urgent need to address the scarcity of research on violence against older women in low and middle-income contexts, and to expand research in high-income contexts to diverse populations and age groups who may have different risk profiles for violence in older age.

Our results indicate that the focus of the existing qualitative evidence-base is primarily on IPV in older age. It is unclear whether this research focus reflects the actual burden of IPV compared to other forms of violence against older women, and if the evidence-base currently adequately includes accounts of types of violence and perpetrators that are most significant for older women. As noted, assessment of polyvicimization in the evidence-base is lacking. As such, further studies of violence against older women that are inclusive of any type of violence, by any perpetrator, or take an open-ended approach to older women’s accounts of violence, are needed. In the quantitative evidence-base, systematic reviews have focused on elder abuse and on IPV. A review of quantitative evidence on IPV identified 19 studies [15] and the review of elder abuse against women included 50 studies; as such, the quantitative evidence-base appears to capture more in terms of forms of violence against older women.

Our findings indicate that for older women who had experienced violence throughout the life-course, aspects of ageing, such as frailty, injuries, chronic disease, and cognitive decline, make coping with different forms of violence more difficult than earlier in life. Qualitative and quantitative research does not currently shed light on associations between types of violence, chronicity of violence, and physical and mental health outcomes for older women, and additional research in this area is warranted. Other themes that emerged in our review call for further research. Regret and hopelessness were commonly described as significant issues for older women; these factors appear to significantly influence well-being, psychosocial health and physical and mental health. However, these outcomes are rarely measured, and these may further impact other specific mental health and psychosocial issues for older women subjected to violence. Further research could explore if and how regret and hopelessness amongst older women differs from shame and stigma as currently measured and reported amongst women of reproductive age, and further elucidate its impacts on psychosocial well-being. In addition, economic and financial abuse appeared to be correlated and interlinked with older women’s experiences of violence, and barriers to leaving abusive relationships; terminology and definitions of these forms of abuse are varied and often unclear, and measures often cover several constructs [109]. While there is some growing consistency in how economic and financial aspects of abuse are conceptualized and measured, there is more work needed on how to assess economic or financial abuse, and understand its linkages with physical and mental health outcomes.

Conclusion

The current qualitative data available on violence against older women has important limitations, including that it is predominantly derived from high-income countries, often does not address context, is focused on IPV to the exclusion of other types of violence and perpetrators, and does not disaggregate by age group. However, our findings highlight some important issues. IPV persists into older age, and shares characteristics and impacts as in younger age groups. In some cases, there may be factors, such as a partner’s retirement or illness such as dementia, which can precipitate or increase violence. As shown in quantitative reviews, physical violence tends to decrease with age while psychological abuse and controlling behaviours increase, and financial and economic abuse are important elements of older women’s experiences of violence and control. Older women described being strongly influenced by social norms that dictate a sense of duty to stay in a relationship with an abusive partner, a desire to protect children, and shame and silence surrounding experiences of violence. More research is needed, particularly from LMICs to fill in the many gaps in the evidence-base. However, it is clear that action to support older women in abusive relationships is needed. Services for older people need to be aware of the prevalence and forms of violence against older women and know when to identify and respond in a sensitive and non-judgmental way, to improve prevention of and response to violence against older women.