Determining Prevalence and Correlates of Elder Abuse Using Promotores: Low Income Immigrant Latinos Report High Rates of Abuse and Neglect

Abstract

Low-income Latino immigrants are understudied in elder abuse research. Limited English proficiency, economic insecurity, neighborhood seclusion, a tradition of resolving conflicts within the family, and mistrust of authorities may impede survey research and suppress abuse reporting. To overcome these barriers, we recruited and trained promotores, local Spanish-speaking Latinos, to interview a sample of Latino adults age 66 and older residing in low-income communities. The promotores conducted door-to-door interviews in randomly selected census tracts in Los Angeles to assess the frequency of psychological, physical, and sexual abuse, financial exploitation, and caregiver neglect. Overall, 40.4% of Latino elders experienced some form of abuse and/or neglect within the previous year. Nearly 25% reported psychological abuse, 10.7% indicated physical assault, 9% reported sexual abuse, 16.7% indicated financial exploitation, and 11.7% were neglected by their caregivers. Younger age, higher education, and experiencing sexual or physical abuse before age 65 were significant risk factors for psychological, physical, and/or sexual abuse. Years lived in the United States, younger age, and prior abuse were associated with increased risk of financial exploitation. Years spent living in the U.S. was a significant risk factor for caregiver neglect. Abuse prevalence was much higher in all mistreatment domains than findings from previous research on community-dwelling elders, suggesting that low-income Latino immigrants are highly vulnerable to elder mistreatment, or that respondents are more willing to disclose abuse to promotores who represent their culture and community.

INTRODUCTION

Current research suggests that more than one in ten adults aged 60 and over are victims of elder abuse every year,¹ yet only a small proportion of the estimated 5 million cases of non-institutional elder abuse annually are reported to Adult Protective Services (APS), the state agencies charged with investigating abuse. Risk factors for abuse include low income, low education, living with others, isolation, and less use of formal services.^1,2 These characteristics are prominent among Latino immigrants, suggesting that they may be a group that is particularly vulnerable. In 2010, Latinos made up 6.7% of the U.S. population aged 65 and older,³ up from 5.0% in 2000.⁴ Among Latino older adults, 55.0% were immigrants, 41.1% had limited English proficiency and 20.4% lived below the federal poverty level.⁵

Cultural norms indicate that immigrant Latinos may be less likely to identify as victims of abuse for several reasons. Older Latinos often reside with their families and rely on them for long-term care.⁶ Although close kin networks and values of filial responsibility have been found to be protective,^7,8 familism, which emphasizes the needs of the family over the needs of the individual, may conceal mistreatment and inhibit formal help-seeking.^8,9 Preserving la familia to avoid vergüenza, or shame, promotes tolerance of family violence and suppresses reporting.^7,8 Citizenship status may also be a barrier if undocumented Latinos are less likely to report abuse for fear of deportation of themselves or their family members.^7,9 General mistrust of government could exacerbate these fears,¹⁰ and limited English proficiency may inhibit reporting even when fears are overcome. Furthermore, cultural beliefs about the acceptability of sharing money and resources among family members⁹ and economic interdependence within extended families may blur the boundaries of financial exploitation, especially in households where the only reliable source of income is the elder’s benefits (e.g., Social Security, Supplemental Security Income [SSI]).

Recently, several population-based studies have systematically explored the prevalence and correlates of elder mistreatment in domestic settings in the U.S. A nationally representative study¹¹ of 3,005 adults aged 57–85 (6.8% Latino) found past year prevalence rates of 9.0% for psychological aggression, less than 1% for physical assault, and 3.5% for financial exploitation. Subjects were informed that incidents of abuse would be disclosed to their state’s reporting hotline, including personal identifying information. Latinos were the racial/ethnic group least likely to acknowledge being abused. In a nationwide telephone survey of community-residing adults aged 60 and above (N=5,777) (4.3% Latino), Acierno and colleagues found the following one-year prevalence rates¹²: 4.6% psychological; 1.6% physical; 0.6% sexual; 5.2% financial; and 0.5% caregiver neglect. Respondents who were of Latino or Hispanic ethnicity^a were not independently associated with higher rates of abuse in any of these domains.

While national prevalence studies have included Latinos, it is possible that those hardest to reach—low-income, non-English speaking immigrants—may have been under-represented. Possible barriers include challenges with accessing Latino immigrants living in ethnically segregated communities, issues of trust impeding disclosure of sensitive information, and cultural norms of keeping problems within the family. Furthermore, Latinos are less likely to have access to landline telephones according to research by the National Center for Health Statistics.¹³

Given the potential of under representing low-income Latino immigrants, the goal of this study was twofold. First, we sought to identify the overall prevalence of five types of elder abuse in a sample of community-residing immigrant Latinos. Second, we examined correlates of abuse and compared these to national prevalence studies.

METHODS

Sample and Interview Protocol

Research suggests that face-to-face recruitment in high-density Latino neighborhoods is the most effective method for recruiting the target population.¹⁴ Therefore, we employed monolingual Spanish-speaking promotores to conduct door-to-door interviews. Traditionally, the role of promotores, Latino community health workers, is to provide culturally sensitive linkages between communities and health and/or social services. More recently, promotores have been recognized as an effective tool in research to collect data among hard-to-reach populations.¹⁵ In our study, promotores were recruited through a partnership with a local agency that serves the target area. The agency had well over a decade of experience hiring, training, and employing promotores for a number of health-related programs in the Latino community. Our community partner recruited three promotores who had recently completed 240 hours of training; our fourth interviewer was their supervisor. We provided two days of additional training, which covered protocols for screening potential respondents for participation, ensuring informed consent, and administering the questionnaire. The training included practice administering the instrument to a bilingual member of the team. Promotores were provided with community resource packets to give to each participant and were trained to help respondents identify appropriate services if the interview suggested such a need. After a review of California’s statutes defining mandated reporters for elder abuse¹⁶ and consultation with the study’s 13-member advisory committee, it was determined that promotores did not qualify as mandatory reporters. Although they did not collect any personally identifiable information from respondents (e.g., name, address, phone number), promotores were instructed to contact APS or law enforcement directly if they encountered respondents who faced immediate danger. To maximize reliability and ensure consistency in survey administration, we conducted regular debriefing sessions with promotores, their bilingual supervisor, and a bilingual member of the research team.

We aimed to measure the prevalence of abuse/neglect among Spanish-speaking Latinos aged 65 and older, residing in predominantly low-income, racial/ethnic minority neighborhoods in Los Angeles. Because the survey asked about abuse prevalence in the previous 12 months, eligible subjects were aged 66 and older. Subjects were determined to have capacity to provide informed consent based on a modified version of the Evaluation to Sign Consent form.¹⁷ In an effort to maximize reliability, this study did not interview proxies for cognitively impaired subjects unable to provide consent.

The population of Los Angeles County (9.8 million) is separated into eight geographic Service Planning Areas (SPAs). From among the 705 block groups in SPA 6, we randomly selected six block groups that included at least 100 residents using Summary File 1 data from the 2000 U.S. Census. From February to June of 2010, promotores went door-to-door within the selected block groups to identify residents who met inclusion criteria and request their participation. Based on considerable demographic shifts within SPA 6 since the 2000 census, concentrations of Latino elders were far from the levels expected; therefore, we worked with the promotores to identify three additional survey areas in late May to reach the target sample of 200 respondents. Individuals encountered outside their homes who met inclusion criteria were also interviewed. Promotores offered potential subjects ten dollars to participate. Those who agreed (n=200) were asked to give verbal consent. Two subjects’ records were later excluded due to missing data.

Promotores used tracking forms to identify those who declined to participate. Based on a sample of these tracking forms, we conservatively estimate the study’s response rate at 65%. Interviews lasted approximately 60 to 90 minutes; all were conducted in Spanish. The study was approved by the University of Southern California Institutional Review Board.

Measures and Variables

Elder abuse prevalence was measured in the aggregate and within five sub-types: psychological aggression, physical assault, sexual coercion, financial exploitation, and caregiver neglect. Surveys were conducted using a 63-item abuse instrument developed for the study, the University of Southern California Older Adult Conflict Scale (USC-OACS). The instrument included questions derived from the Revised Conflict Tactics Scales (CTS2 and CTSPC)¹⁸ and the Conflict Tactics Scales for Older Adults (UC Irvine, NIH Grant #R21AG028060) with permission from both sources. Input was sought from older adult focus groups conducted in English and Spanish and a multidisciplinary group of experts. The instrument was first constructed in English and translated into Spanish by two members of the research team from different Latin American countries of origin, to avoid the inclusion of words or phrases unique to any region. It was then reviewed by the three promotores and two supervisors to identify areas that could be restated to provide clarity and avoid misinterpretation. The instrument was pilot tested in English with five volunteers aged 65 and older, followed by cognitive interviews with 12 participants aged 65 and older from a senior center. These two steps helped to ensure that the language was easily understandable and that the participants interpreted the questions as intended. Advisory committee meetings were held quarterly to provide community input on the instrument and the approach to data collection.

The USC-OACS included 9 items of psychological aggression (α=0.80), 15 of physical assault (α=0.85), and 4 of sexual coercion (α=0.65). To identify caregiver neglect (12 items), we first asked about need for assistance with activities of daily living (ADL) and instrumental activities of daily living (IADL) and then asked those who needed assistance if adequate support was provided (α=0.76). We measured financial exploitation using 14 items (α=0.49). To minimize the pressure of social desirability, we prefaced each mistreatment section with statements that conveyed acceptance and normalized the questions that followed.¹⁹ Survey items asked about concrete abusive behaviors, designed to prompt closed-ended responses. For example, “Did someone close to you threaten to put you in a nursing home [in the last 12 months]?” For all five abuse areas, promotores asked if these behaviors had happened in the last 12 months (‘yes’ or ‘no’) and, if so, how often. Subjects were considered victims of mistreatment if they endorsed any item within a particular domain. Based on convention, we classified psychological, physical, and sexual abuse as either minor or severe depending on the type of behavior¹⁸ and drew on the work of Conrad to extend this classification paradigm to financial abuse²⁰; neglect severity was determined based on the degree to which the elder’s needs were unmet on each domain. Promotores also collected socio-demographic data, including age, sex, race/ethnicity, country of origin, years in the U.S., living arrangement, marital status, level of education, and employment status. A question of whether individual monthly income exceeded $902.00 (California’s SSI threshold at the time) was excluded from analyses because 49% did not answer.

Measures of Vulnerability

With increased dependency on caregivers, physical impairment is commonly cited as a risk factor for elder abuse and neglect.^21,22 Using a scale validated by Spector and Fleishman,²³ six ADL and six IADL impairments were combined into a needs-based impairment scale (range 0–12; α=0.74). Lack of social support also increases risk of abuse,^1,2,22 so five questions were derived from the UCLA Loneliness Scale.²⁴ These questions were summed to yield a total social isolation score (range=0–5; α=0.67), with (0) indicating the lowest level of social isolation and (5) representing the highest level of isolation. Several studies have indicated that prior abuse is another risk factor for abuse.^1,25 Therefore, after the physical and sexual abuse sections, a global question was asked about the experience of abuse in each of these domains prior to age 65.

Statistical Analysis

We measured prevalence using frequencies for each of the five types of abuse, separated into mild and severe, and aggregated them to assess the overall prevalence of abuse/neglect. Next we identified risk factors associated with abuse using logistic regression. Due to their conceptual similarity as willful actions intended to cause physical or emotional harm, psychological aggression, physical assault, and sexual coercion were combined into a single “conflict domain” (α=0.91). For conflict abuse and financial exploitation, subjects were considered victims of abuse and assigned a value of 1 if they responded affirmatively to one or more of the items within the domain. A similar approach was used for neglect, with subjects considered victims of neglect if they indicated any one or more ADL/IADL impairments for which a potential caregiver was present but not meeting their needs. We incorporated neglect in the total prevalence rate for elder abuse/neglect; however, the denominator for neglect used in our regression model included only those who needed caregiver support (as indicated by self-reported ADL/IADL impairment) and therefore had the potential to be neglected. Independent variables in the logistic regression analysis were: age, sex, education (dichotomized as completing ninth grade or above, the end of compulsory education in Mexico), years lived in the U.S., marital status, living arrangement (3 categories), functional impairment (excluded from the neglect model), physical/sexual abuse prior to age 65, and social isolation. Cases with missing data on the independent variables were included in the analysis sample, and all available data were analyzed in Mplus 6.11 using the full-information maximum likelihood (FIML) approach.²⁶

RESULTS

Forty percent (N=198) indicated that they had experienced at least one type of abuse; 21% had experienced multiple types. As shown in Table 1, the mean age was 72, 56% were female, 50% were married or partnered, 13% were single, and 20% were widowed. One third lived with a spouse, 46% lived with children and/or grandchildren, and 17% lived alone. Ninety-five percent were immigrants, primarily from Mexico (64%), with an average age at immigration of 41 years (SD=18). Only 13% graduated high school/earned a GED equivalent. Of those remaining, average education was 4.4 years; only 20% of the sample had completed at least a ninth grade education. Among those who reported income (n=101), 89% indicated less than $902/month. Twenty-four percent indicated at least one IADL limitation, 16% reported at least one ADL limitation; 70% were not impaired. On a scale of 0–5, the mean level of self-perceived loneliness/isolation was 1.8 (SD=0.7). Twenty-three percent reported physical or sexual abuse before age 65.

Table 1.

Sample characteristics

	n	Mean (SD) or %
Age	194	72.3 (± 6.9)
Gender (female)	192	56.3
Ethnicity
Hispanic, Latino, or Spanish origin	192	99.5
Country of Birth	198
U.S.		5.1
Mexico		64.1
El Salvador		15.7
Guatemala		10.6
Other Spanish/Latin Country		4.5
Years Lived in U.S.	194	30.9 (± 17.1)
Highest Level of Education	176
Less than high school		86.9
High school or GED Equivalent		8.5
Some College		2.3
College Graduate		2.3
Years of education (if < high school)	151	4.4 (± 3.1)
Marital Status	196
Single		13.3
Married/Partnered		50.5
Widowed		20.4
Divorced/Separated		15.8
Living Arrangement*	195
Live alone		17.4
Live with spouse		33.3
Live with child/children		46.7
Live with grandchild/grandchildren		5.1
Live with sibling/other relative		14.9
Live with friend		4.1
Work Status*	192
Retired		50.0
Paid Employment		10.4
Volunteer		8.9
Homemaker		31.8
Monthly Income	101
Above $902.00		10.9
Average Monthly Income	60	$344.32 (± $365.39)
Physical/Sexual abuse before age 65	193	23.8
Functional Impairment
ADL Impairment	196	15.8
IADL Impairment	195	24.6

^*Participant could choose multiple responses

Specific Prevalence Findings

As shown in Table 2, nearly 25% reported psychological abuse, (14.2% minor, 10.6% severe). Of the 10.7% that indicated physical abuse, half reported at least one incident of severe physical assault. Nine percent reported sexual abuse (1.6% severe) and 16.7% reported financial exploitation (10.7% severe). Caregiver neglect was reported by 11.7%. Of these, 83% suffered from minor neglect (caregiver sometimes provided assistance); 17% indicated severe neglect. Six percent reported having impairments with no caregiver available to help. In the combined conflict domains, 30.7% reported one or more incident of psychological, physical, and/or sexual abuse (minor or severe). Three study respondents (1.5%) indicated that they had reported abuse of any kind to APS in the last year.

Table 2.

Presence of Abuse by Domain and Severity

Abuse Domain	n	Frequency (%)
		Any Abuse	Severe Abuse
Total	198	40.4	22.7
Conflict Abuse	198	30.7	13.6
Psychological	198	24.8	10.6
Physical	197	10.7	5.6
Sexual	189	9.0	1.6
Financial Abuse	180	16.7	10.6
Caregiver Neglect	197	11.7	2.0

Correlates of Abuse

Table 3 presents the logistic regression for conflict and financial abuse (N=198), and for neglect among those who needed caregiver assistance (n=59). In the combined conflict domain, age was inversely related (OR=0.90, p<.01), and education was positively related (OR=5.58, p<.01). Functional impairment trended toward significance (OR=1.28, p<.1). Physical and/or sexual abuse prior to age 65 showed a strong relationship with conflict abuse after age 65 (OR=13.1). Prior abuse was also related to financial exploitation (OR=2.93), as was age (OR=0.92). Years lived in the U.S. trended toward significance (OR=1.02, p<.10), such that each additional year lived in the U.S. increases the risk of financial abuse by 2%. For caregiver neglect, only years lived in the U.S. was significant (OR=1.05, p<.05).

Table 3.

Estimates of logistic regression analysis of selected covariates with conflict abuse (combined psychological, physical, and sexual abuse), financial abuse, and caregiver neglect

Independent Variables	Conflict Abuse		Financial Abuse		Caregiver Neglect
	OR (95% CI)	p	OR (95% CI)	p	OR (95% CI)	p
Sociodemographic Characteristics
Age	0.90 (0.84–0.97)	0.007 **	0.92 (0.85–1.01)	0.068 †	1.00 (0.91–1.10)	0.944
Sex (Female=1)	1.45 (0.62–3.41)	0.389	0.71 (0.28–1.80)	0.466	1.52 (0.29–8.08)	0.624
More than 9th grade education	5.58 (2.08–14.98)	0.001 **	2.13 (0.82–7.18)	0.124	1.30 (0.23–7.52)	0.767
Years lived in U.S.	1.00 (0.97–1.02)	0.797	1.02 (1.00–1.05)	0.078 †	1.05 (1.01–1.09)	0.016 *
Married/Partnered	1.34 (0.56–3.22)	0.513	1.26 (0.47–3.36)	0.648	2.00 (0.41–9.71)	0.388
Other living arrangement (reference)	----	----	----	----	----	----
Living Alone	1.02 (0.29–3.57)	0.979	0.91 (0.25–3.28)	0.886	0.45 (0.05–4.13)	0.477
Lives with Children/Grandchildren	2.01 (0.81–5.01)	0.132	1.41 (0.53–3.79)	0.493	1.78 (0.45–7.01)	0.412
Functional Status
Functional Impairment	1.28 (1.00–1.65)	0.050 †	1.10 (0.82–1.47)	0.531	----	----
Individual Characteristics
History of Abuse	13.1 (5.24–32.83)	<.001 ***	2.93 (1.18–7.30)	0.021 *	1.31 (0.29–5.88)	0.728
Isolation	0.64 (0.35–1.19)	0.157	0.63 (0.30–1.29)	0.205	0.45 (0.15–1.31)	0.142

Note: OR = odds ratio; CI = confidence interval;

^†p <.10,

^*p <.05,

^**p <.01,

^***p <.001

N=198 for conflict abuse and financial abuse; N=59 for caregiver neglect (impaired subjects only)

DISCUSSION

The high rates of abuse, neglect, and financial exploitation in this sample (40.4% total; 22.7% severe) exceed those reported in previous community-based studies. There are several explanations for these striking findings. First, our sample may represent a particularly high-risk subset of the Latino population that has been under-represented in elder abuse research. Challenges faced by this community—limited English proficiency, economic insecurity, neighborhood seclusion, fear of crime and mistrust of authorities—may present barriers to survey research and suppress abuse reporting. The approach of going door-to-door within the target neighborhoods may have recruited Latinos who do not typically participate in studies of elder mistreatment.

Second, it is possible that we did not capture a hidden population, but rather our interview method uncovered abuse that is often concealed. Respondents may have felt more comfortable discussing mistreatment with promotores who represent their culture and community. Furthermore, promotores did not collect identifying information and were not mandated to report abuse, potentially increasing willingness to disclose. To our knowledge, no new reports of abuse were filed as a result of our survey, despite offers from promotores to assist respondents.

Third, our instrument may have been more sensitive and/or our threshold may have been lower than other studies. Comparing the USC-OACS to the telephone survey used by Acierno revealed that both instruments assess similar abusive behaviors, although the telephone survey combines them into a single item, e.g., “Has anyone ever hit you with their hand or object, slapped you, or threatened you with a weapon?”¹² The USC-OACS asks about each behavior separately, potentially capturing a wider range of harmful experiences. Further validation and psychometric testing with other populations is needed to determine whether the USC-OACS is more sensitive to mistreatment.

Few studies have examined the prevalence of elder abuse in Latin America; among them, the only studies we are aware of that used large community-based samples took place in Mexico. In 2007, a study of 1,078 rural Mexican elders reported an overall abuse prevalence of 8.1%.²⁷ A second study reported a past-year abuse prevalence of 16% among urban-dwelling Mexican elders.²⁸ As with our study’s findings, psychological mistreatment was the most commonly reported abuse-type, yet our overall frequency was much higher. Higher rates of abuse from smaller studies have been reported in other Latin American countries²⁹; however, prevalence findings in Mexico are more consistent with nationally representative surveys conducted in the U.S.¹

Based on population statistics from the 2010 U.S. Census³⁰ and our analysis of data obtained from Los Angeles County, approximately 2.0% of the residents within the target community of SPA 6 made a report to APS in the year prior to our final survey being administered (July 2009-June 2010). This number is very close to the 1.5% of our respondents who indicated reporting abuse of any kind to APS in the year prior to taking our survey, but is significantly lower than abuse rates detected by the USC-OACS instrument. This comparison suggests that Latino immigrants considerably underreport mistreatment; perhaps stemming from a cultural tendency to resolve conflict within the family or from fear that contacting authorities constitutes a greater risk to the victim and the family.

Correlates of Abuse

Consistent with existing research,¹ we found that the risk of conflict abuse decreases with age. Our finding that conflict abuse is higher among those with more education contradicts earlier research.² We also found that respondents who have lived longer in the U.S. face a higher risk of financial exploitation. Although the underlying reasons are not apparent from this study, these associations suggest that Latinos who are more acculturated to the U.S. and more educated are more likely to acknowledge abuse.

Experiencing physical and/or sexual abuse before age 65 was a considerable risk factor for mistreatment in our sample. This finding informs the debate that some elder abuse is domestic abuse that persists into old age.²¹ Our findings support previous research that suggests that other types of elder abuse, particularly financial exploitation and neglect, are associated with different vulnerabilities, including functional and cognitive impairment.^31,32

Limitations and Future Research

Several caveats should be noted when considering these findings. First, the study focused on a specific population within a specific geographic region of Los Angeles County. As such, the goal was to provide greater detail about a group that may be underrepresented in national studies, however, results may not generalize to all Latinos. Second, although there are trade-offs with any survey approach, we believe that because of their cultural knowledge, promotores likely obtained a higher response rate and perhaps more valid results. This should be tested by comparing promotores with other survey methods. Because promotores worked off-site in a community-based agency, it was important to ensure that they were well trained and acquainted with the goals of the study.

Our instrument had good internal consistency with the exception of financial exploitation. In analyzing responses, it appears that some items were less relevant in this population (e.g., forced to change will/power of attorney), and some were less likely to be perceived as harmful. Cultural differences in the perception of what constitutes a misuse of funds⁹ suggest that normative assumptions of proper/improper resource exchange may not be sufficiently culturally-nuanced to identify financial exploitation in our population.

Finally, we did not include respondents with moderate or severe cognitive impairment and chose not to interview proxy respondents. Previous research has found that proxies may report different levels of mistreatment than victims.^12,21 Measuring neglect using self-report also presents a challenge, and we may have missed elders who were highly dependent and unable to be surveyed. We believe, however, that our approach of identifying need for caregiver support as a criterion for neglect is important to the conceptual framework of this abuse type.

Health practitioners that serve older Latinos should recognize that elder abuse is considerably underreported in these populations. With this in mind, practitioners should be aware of the reporting requirements in their community and actively seek to identify abuse in their older patients. Strategies such as a brief screening for abuse coupled with careful attention to physical and emotional warning signs of mistreatment should be considered (e.g., bruising, malnutrition, depression).³³ In addition, from a public health perspective, using promotores to educate and advocate on behalf of older adults in their communities may bring greater awareness to the problem and normalize the practice of reporting. The last several years have seen groundbreaking national studies measuring the incidence of elder abuse among community-residing older adults in the U.S.^1,11 To continue to improve our understanding of how elder abuse affects older adults and communities, more efforts are needed to include minorities into our current models of abuse prevalence and vulnerability.