Rural-Urban Differences in the Prevalence and Correlates of Elder Mistreatment

Abstract

Rural and urban areas differ in terms of age structure, health outcomes, and access to resources, but there is limited research on rural-urban differences in elder mistreatment. This study addresses that gap. We used data from Round 3 of the National Social Life, Health & Aging Project (NSHAP) survey (n=2,333) to examine rural-urban differences in rates of 11 indicators of elder mistreatment, overall and by individual indicators. We conducted logistic regression models and generated predicted probabilities adjusting for sociodemographic and social well-being characteristics. Elder mistreatment was less common among rural than urban older adults (adjusted predicted probability 41.5% vs. 49.5%, p<0.01). Rates were also higher for urban older adults for nearly every individual indicator of mistreatment. Rates of elder mistreatment were high (>40%) for both rural and urban older adults, but urban older adults faced a greater risk, requiring attention to urban risk and rural resilience.

Introduction

Elder mistreatment, defined by the U.S. Department of Justice as “an intentional or negligent act by any person that causes harm or a serious risk of harm to an older adult” (U.S. Department of Justice, n.d.), is a widespread, yet underreported, issue with far-reaching implications for the health and well-being of older adults. Official definitions of elder mistreatment differ, reflecting the complicated nature of the issue (Dong, 2015; Yon et al., 2017). For example, the U.S. Department of Justice defines five categories of elder mistreatment (U.S. Department of Justice, n.d.): physical, sexual, psychological, financial, and neglect/abandonment; while the U.S. Department of Health and Human Services views neglect and abandonment as separate issues, and also includes self-neglect as a seventh category (U.S. Department of Health and Human Services, 2022). Federal agencies and U.S. states also vary in the age criteria used to define an ‘older adult.’ For example, the Centers for Disease Control and Prevention (CDC) defines an older adult as someone age 60 or older (CDC, 2024), while the National Institute on Aging (NIA) defines an older adult as age 65 or older (NIA, 2024), though their official Elder Abuse page refers to “adults over the age of 60” (NIA, 2023). State elder abuse and mistreatment statutes also vary widely, with some covering dependent or vulnerable adults, older adults of varying ages, or both (Elder Abuse Guides for Law Enforcement, n.d.; Jirik & Sanders, 2014).

Conducted more than a decade ago, the National Elder Mistreatment Study (Acierno et al., 2010) found that 10% of older adults age 60 and older self-reported experiencing some form of elder mistreatment, however the actual prevalence may be much higher, as only an estimated one out of every 24 cases are reported (U.S. Department of Justice & U.S. Department of Health and Human Services, 2014). There has also been growing concern about elder mistreatment in recent years, especially during the COVID-19 pandemic, with elder mistreatment appearing to increase both in prevalence (Ansberry, 2021; Chang & Levy, 2021) and severity (Weissberger et al., 2022). Such increases are likely due to a combination of factors, including an aging U.S. population, greater isolation among older adults and their caregivers, more financial strain, and other COVID-19 related stressors (Chang & Levy, 2021; Chokkanathan & Mohanty, 2025; Makaroun et al., 2022).

The Contextual Theory of Elder Abuse (Roberto & Teaster, 2017) conceptualizes that risk for elder mistreatment is impacted by multiple interrelated factors, including at the individual level (e.g., an older adult’s personal characteristics, such as their health, cognitive status, and socioeconomic status), the relational level (e.g., social network density, relationship quality, social cohesion), and the broader community level (e.g., geographic location and community characteristics such as informal supports and formal support services). These individual, relational, and community contexts are embedded within a broader societal context that encompasses ideological values, laws, and systems governing the response to elder mistreatment.

Most research on elder mistreatment focuses on identifying individual and relational-level risk factors despite the importance of considering the broader community context. There is currently limited research on how the prevalence and risk of elder mistreatment vary by rurality (Zhang et al., 2022), in spite of well-documented differences in age structure and the aging experience between rural and urban areas (Henning-Smith, 2020; Tuttle et al., 2020). The limited research in this area suggests competing perspectives on how elder mistreatment may differ by rurality. On the one hand, rural older adults experience worse health care access and outcomes and possess fewer financial resources than their urban counterparts (Tuttle et al., 2020). They also live in more geographically isolated contexts, which may make mistreatment more easily undetected (Warren & Blundell, 2019). Indeed, qualitative research has shown that isolation is one of several key challenges to preventing and addressing elder abuse (Marrs et al., 2025), especially in rural communities (Lahr et al., 2024). Availability and resourcing of formal social services have also been found to vary across counties and states (Adult Protective Services Technical Assistance Resource Center, 2023; Steinman & Anetzberger, 2022). For instance, variation in adult protective services funding affects case worker-to-client ratios as well as resources they can offer across rural and urban populations (Administration for Community Living, 2024). However, rural communities tend to be more socially cohesive than urban communities (Moss et al., 2023; Sood et al., 2023), and rural older adults may be more closely connected with those around them. Social cohesion may increase the likelihood of detecting the signs of elder mistreatment and even prevent mistreatment from occurring. Indeed, prior research indicates that social support is a protective factor against adult maltreatment (Fettig et al., 2023; Marzbani et al., 2023).

Despite the high and underreported prevalence of elder mistreatment, combined with rural-urban differences in aging, little research has examined rural-urban differences in rates of elder mistreatment in the U.S. This research addresses that gap by investigating rural-urban differences in reports of elder mistreatment. Given the multiple elevated risk factors for elder mistreatment faced by rural older adults, combined with rural-urban differences in the community characteristics emphasized in the Contextual Theory of Elder Abuse, we hypothesize that rates of elder mistreatment will be higher among rural older adults than among urban older adults, both overall and for specific measures of elder mistreatment. Results from this study will provide baseline foundational information for health care systems, law enforcement, policymakers, and researchers to improve the prevention, detection, and treatment of elder mistreatment within different geographic contexts.

Methods

Data and sample

Data for this study come from the National Social Life, Health & Aging Project (NSHAP) Round 3 (2015–2016) survey (Waite et al., 2017), merged with a dichotomous rural/urban identifier from Round 2 (2010–2011) (Waite et al., 2014). The NSHAP is a longitudinal, population-based survey of U.S. older adults’ health and social relationships. Round 3 was the most recent NSHAP data available at the time of this study. Our total sample size includes 2,333 community-dwelling older adults ages 60 and older who were present in both Rounds 2 and 3 of the NSHAP, which was required in order to merge the data. Of those, 551 resided in rural areas. NSHAP bases rurality on 2010 Rural-Urban Commuting Area (RUCA) codes.

Measures

We examined 11 self-reported indicators of elder mistreatment, for which survey respondents gave yes/no responses: “Since you turned 60… have you been afraid of anyone in your family?” “…has anyone borrowed your money without paying you back?” “…have you felt uncomfortable with anyone in your family?” “…has there been a family conflict at home?” “… has anyone forced you to do things you didn’t want to do?” “…has anyone close to you tried to hurt or harm you?” “…has anyone close to you called you names, put you down, or made you feel badly?” “…has someone in your family made you stay in bed or told you that you are sick when you know you are not?” “…has anyone taken things that belong to you without your OK?” “…has anyone told you that you gave them too much trouble?” and “…have you felt that nobody wanted you around?” We also created a dichotomous elder mistreatment indicator variable indicating whether the respondent responded yes to at least one elder mistreatment indicator variable. For a detailed description of the development of the elder mistreatment questions in the NSHAP, see Wong et al. (2021).

We examined several additional sociodemographic and social well-being variables. Sociodemographic variables included age (in years); sex (male/female); highest educational degree earned (less than high school, high school degree or equivalent, vocational school/some college, and bachelor’s degree or higher); race and ethnicity (non-Hispanic white, non-Hispanic Black, other); self-rated health (excellent/very good/good vs fair or poor); current employment status (employed/homemaker, retired/disabled/unemployed, or other/unknown). To account for the fact that elder mistreatment involves negative interactions between social network members, we also included several variables to measure the social context. These included (1) social support (high/low), measured as whether respondents reported often being able to open up to and/or rely on at least one of the following: spouse, friends, or family; (2) social engagement (high/low), measured as whether respondents reported getting together with one or more of the following groups at least once a week: friends/relatives, volunteer commitments, or organized group meetings; and (3) social cohesion (high/low), measured as whether respondents reported “yes” to, agreed with, or strongly agreed with at least one of the following about their area: people often visit each other, people often do favors for each other, the area is close-knit, people are willing to help their neighbors, people in the area generally get along, people in the area can be trusted. Social support, engagement, and cohesion were defined similarly to recent research using NSHAP data to evaluate the effects of social circles on older adult quality of life (Larnyo et al, 2024), although the exact construction of these variables differed.

Analysis

We performed descriptive statistics and chi-square tests (significance threshold: p<.05) examining rural/urban differences in sociodemographics and social well-being covariates, the dichotomous elder mistreatment indicator, and overall counts of elder mistreatment indicators. We also conducted sensitivity analyses removing three of the mistreatment variables that could arguably not be focused specifically on the older adult. Those were “feeling afraid” of someone in the family, “feeling uncomfortable” around someone in the family, and “conflict in the family.” These may be measuring general sentiment around family dynamics, rather than specific harm toward the older adult themselves. However, because we aim to take a broad and inclusive view of elder mistreatment, we report results on the more expansive measuring using all 11 indicators.

We then built multivariate logistic regressions adjusted for sociodemographics and social well-being variables to determine the odds ratios (rural vs. urban) and predicted probabilities (stratified by rurality) of the dichotomous indicator of elder mistreatment and each of the 11 specific indicators of mistreatment individually. Firth penalized logistic regressions were performed as posthoc sensitivity analyses to account for quasi-complete separation of data points. These models were more robust for certain elder mistreatment indicators, and so we report the Firth regression numbers for these particular indicators as our final results instead of the multivariate regression numbers. All analyses were conducted using SAS 9.4 (SAS Institute, Cary, NC) and Stata 18.0 (StataCorp LLC, College Station, TX). Sample sizes of N < 10 are suppressed due to reidentification concerns, considering the rural context and sensitive nature of the data.

Results

Table 1 describes the sociodemographic and social well-being characteristics of the sample by rurality. There were almost no significant differences by rurality; the average age of both subsamples was around 76 years old, there were more females than males, marital statuses and social well-being rates were similar across the two samples, the majority of both reported being in good health, and most were not currently employed. The only significant differences between rural and urban older adult samples were in race and ethnicity, where rural older adults were more likely to identify as non-Hispanic white and less likely to identify as non-Hispanic Black or Hispanic (p<0.001), and in education status, where urban residence was associated with higher educational attainment (p=0.004).

Table 1:

Sociodemographic and Social Well-Being Characteristics by Rurality

		Rural (N=551)	Urban (N=1,782)	P-value
Sociodemographic Variables
Age (mean, SD)		76.2 (6.8)	75.9 (7.1)	0.513
Sex				0.678
	Female	55.9	56.9
	Male	44.1	43.1
Highest Degree Earned				0.004
	Less than high school	15.6	17.7
	High school/ equivalent	29.4	22.2
	Vocational/associate’s/ some college	31.6	32.0
	Bachelor’s or higher	23.4	28.1
Race/ethnicity				<0.001
	White non-Hispanic	86.8	68.2
	Black non-Hispanic	8.5	15.7
	Hispanic	3.5	13.3
	Other/unknown non-Hispanic	N<10	2.8
Marital Status				0.845
	Married/partnered	66.8	66.2
	Separated/divorced	8.4	9.6
	Widowed	23.2	22.6
	Never married	N<10	1.7
Employment Status				0.409
	Employed/homemaker	28.7	27.7
	Retired/disabled/unemployed	70.8	71.1
	Other/unknown	N<10	1.2
Self-rated Physical Health				0.240
	Excellent/Very Good/Good	73.9	76.3
	Fair/Poor	26.1	23.7
Social Well-Being Variables
High Social Support		89.5	86.7	0.087
High Social Engagement		60.8	58.9	0.433
High Social Cohesion		89.3	87.8	0.333

Data are from NSHAP Round 3 (2015–16) merged with RUCA codes from NSHAP Round 2 (2010–11). P-values are results of Chi-square tests. Cell sizes were suppressed at N<10.

Overall, rural residents were less likely to report experiencing any elder mistreatment compared to urban older adults (42.5% [95% confidence interval [CI]: 38.3–46.6] vs 49.2% [95% CI: 46.8–51.5], p<0.01). Figure 1 shows rural-urban differences in the overall number of elder mistreatment indicators reported. Rural older adults reported zero indicators at a higher rate than urban older adults (57.5% vs. 50.8%), and urban older adults reported 1, 2, 3, and 4+ indicators at slightly higher rates than rural older adults, although none of these differences were statistically significant (p=0.05). In sensitivity analyses excluding the three family dynamics indicators, we found that overall rates of mistreatment were lower overall, and that rural residents still reported lower rates compared with urban, although the difference was no longer statistically significant (33.4% vs. 36.8%, p=0.14).

Figure 1: Counts of Elder Mistreatment Indicators by Rurality.

Overall Chi-square P-value: 0.0500. Data are from NSHAP Round 3 (2015–16) merged with RUCA codes from NSHAP Round 2 (2010–11).

Table 2 shows the results of regression analyses estimating the odds of elder mistreatment after adjustment for sociodemographic and social well-being characteristics. Compared to urban older adults, rural older adults had significantly lower odds of reporting at least one indicator of elder mistreatment (adjusted odds ratio [aOR]: 0.72, 95% CI: 0.59 – 0.87). They also had lower odds of reporting several specific indicators, including feeling uncomfortable with anyone in their family (aOR: 0.77, 95% CI: 0.60 – 0.99); a family conflict at home (aOR: 0.62, 95% CI: 0.48 – 0.81); and being told that they were too much trouble (aOR: 0.57, 95% CI: 0.34 – 0.94).

Table 2:

Adjusted Odd Ratios of Elder Mistreatment Indicators, Rural vs. Urban Older Adults

Elder Mistreatment Indicator	aOR (95% CI), Rural vs urban older adults	Pr > ChiSq
At least one indicator reported**	0.72 (0.59–0.87)	0.001
Afraid of anyone in family?	0.57 (0.28–1.17)	0.400
Anyone borrowed money without paying it back?	0.81 (0.63–1.03)	0.089
Felt uncomfortable with anyone in family?*	0.77 (0.60–0.99)	0.040
Been a family conflict at home?***	0.62 (0.48–0.81)	<0.001
Anyone forced you to do things you didn’t want to do?	0.60 (0.33–1.08)	0.086
Anyone close to you tried to hurt or harm you?	1.38 (0.72–2.63)	0.334
Anyone close to you called you names/put you down?	0.80 (0.59–1.09)	0.155
Anyone in your family made you stay in bed?	0.71 (0.30–1.69)	0.442
Anyone taken things that belong to you without permission?	0.93 (0.67–1.30)	0.685
Anyone told you that you give them too much trouble?*	0.57 (0.34–0.94)	0.027
Felt that nobody wanted you around?	0.99 (0.69–1.42)	0.974

Data are from NSHAP Round 3 (2015–16) merged with RUCA codes from NSHAP Round 2 (2010–11). Results are adjusted for all sociodemographic and social involvement covariates described in Table 1. aOR 95% CIs containing 1.00 are not statistically significant. Results for “At least one indicator reported”, “Afraid of anyone in family”, “Anyone forced you to do things you didn’t want to do?”, “Anyone in your family made you stay in bed?”, “Anyone told you that you give them too much trouble?” and “Felt that nobody wanted you around?” are from Firth penalized logistic regressions due to quasi-complete separation of data points; all other results are from multivariate logistic regressions. P-values are results of Wald chi-square analyses from standard maximum likelihood estimates for standard models and penalized likelihood ratio tests for Firth regression models.

Figure 2 shows the adjusted predicted probabilities of elder mistreatment indicators stratified by rurality, adjusting for all sociodemographic and social well-being covariates. The predicted probability of at least one indicator of elder mistreatment was 41.5% among rural and 49.5% among urban older adults. Predicted probabilities of the 11 indicators ranged from 1.3% to 18.5% among rural and urban older adults, and from 1.8% to 21.9% among urban older adults. The least common indicator for both groups was reporting that anyone in their family had made them stay in bed. The most common for both was that someone had borrowed money without paying it back. Among almost all indicators, the predicted probability was higher in the urban population. The only exception was for the “Anyone close to you tried to hurt or harm you?” indicator, where the rural probability was 2.9% and the urban probability was 2.1%, although this difference was not statistically significant.

Figure 2: Adjusted Predicted Probabilities of Elder Mistreatment Indicators by Rurality.

Data are from NSHAP Round 3 (2015–16) merged with RUCA codes from NSHAP Round 2 (2010–11). Results are adjusted for all sociodemographic and social involvement covariates described in Table 1. Results for “At least one indicator reported”, “Afraid of anyone in family”, “Anyone forced you to do things you didn’t want to do?”, “Anyone in your family made you stay in bed?”, “Anyone told you that you give them too much trouble?” and “Felt that nobody wanted you around?” are from Firth penalized logistic regressions due to quasi-complete separation of data points; all other results are from multivariate logistic regressions. Asterisks indicate statistically significant P-values resulting from Wald chi-square tests from standard maximum likelihood estimates for standard models and penalized likelihood ratio tests for Firth regression models where * = P<0.05, ** = P<0.01, and *** = P<0.001 (see Table 2).

Discussion

Using Round 3 of the NSHAP, we found that elder mistreatment is relatively common among older adults, regardless of geographic location. More than 40% of older adults reported experiencing at least one type of elder mistreatment, with the most common types including having someone borrow money without paying it back, feeling uncomfortable with anyone in the family, and being in a family conflict at home. These findings should raise concerns about the well-being of older adults.

Alongside the high prevalence overall, we also found differences by rurality, although the results were contrary to our hypothesis that rural older adults were at greater risk for elder mistreatment. Our bivariate analyses found that rural older adults had a lower unadjusted prevalence of elder mistreatment than urban adults (42.5% vs. 49.2%, p<0.01). After controlling for sociodemographic and social well-being characteristics, our adjusted regression analyses similarly revealed that rural older adults had significantly lower odds of elder mistreatment reports than urban older adults (aOR: 0.72, 95% CI: 0.59–0.87), and the adjusted predicted probabilities of elder mistreatment for rural and urban older adults were similar to the unadjusted prevalences (41.5% and 49.5%). These are surprising results; the prevailing narrative is that rural older adults face greater risks for poor health outcomes, including risks specific to elder abuse. It is possible that rurality and/or social cohesion could impact the reporting of abuse, which we describe in more detail below. Despite the fact that these results did not support our hypothesis, the overall finding that rates of elder mistreatment varied by rurality were in alignment with the Contextual Theory of Elder Abuse (Roberto & Teaster, 2017) and its focus on the importance of broader community level (e.g., geographic) factors.

Certain geographic characteristics and social dynamics may affect the occurrence, recognition, and disclosure of elder mistreatment in rural versus urban areas. While we found that rates of reported mistreatment were lower among rural older adults overall, rural residents may be more reluctant to disclose abuse to law enforcement or social service providers due to shame or ongoing contact with abusers in smaller communities (Lahr et al., 2024; Warren & Blundell, 2019). More isolated older adults who lack alternative options for care and support may choose to tolerate mistreatment by a close friend or relative instead of reporting abuse to authorities and risking leaving their homes or communities to receive care elsewhere (Warren & Blundell, 2019). A previous study conducted by the National Institute of Justice (2020) using data from the National Elder Mistreatment Study found that 89.9% of older adults who experienced emotional abuse and 87.5% who experienced financial exploitation by a family member, friend, or acquaintance did not report this to authorities, citing “not wanting to get the person who committed the crime in trouble” and “not wanting publicity” as the main reasons for failing to report abuse; nonreporting rates for abuse by strangers, although still high, were lower, particularly for cases of financial exploitation (83.5% for emotional abuse and 33% for financial exploitation). Alternatively, increased social cohesion in rural areas could act as a protective factor and decrease the likelihood of mistreatment. We adjusted for several measures of social well-being in our regression analyses to account for this. Even after this adjustment, rural older adults still reported lower odds and predicted probabilities for several elder mistreatment indicators. This may indicate that social well-being is a protective factor, but not the main factor, as to why rural and urban older adults experience differing rates of elder mistreatment.

We found that rural older adults were significantly less likely to report conflict and discomfort with family members, suggesting that they may face a lower risk of abuse perpetrated by relatives or that they may be particularly disinclined to “get [their family members] in trouble” and want to avoid publicity (National Institute of Justice, 2020). However, indicators of financial exploitation were not significantly different between rural and urban older adults after adjusting for covariates, indicating that the risk of financial abuse is not strongly affected by geography. Rates of financial exploitation for both groups are concerning, as this type of mistreatment was the most commonly reported, with over 18% of rural and 21% of urban older adults reporting, even after controlling for sociodemographic and social well-being characteristics. Future research should focus on financial exploitation specifically in order to identify whether certain populations of rural or urban residents (e.g., low-income individuals) face higher risk, and whether that risk differs by rurality.

Regardless of the significant difference in prevalence of reports, the finding that 41.5% of rural and 49.5% of urban older adults reported at least one indicator of elder mistreatment is very concerning. These estimates are much higher than the U.S. Department of Justice’s current elder abuse prevalence estimate of 10%, although they have acknowledged that the true prevalence was likely higher due to underreporting (U.S. Department of Justice & U.S. Department of Health and Human Services, 2014). Our findings using the NSHAP elder mistreatment measures also found slightly higher rates compared with a recent meta-analysis on the prevalence of elder mistreatment in rural communities. Zhang and colleagues (2023) found that the pooled prevalence of elder abuse and neglect in rural areas was 33%, although their meta-analysis also included data from studies outside of the U.S.

NSHAP’s Round 3 elder mistreatment items capture broader and arguably less severe forms of interpersonal conflict compared to measures of elder mistreatment that have been used in earlier prevalence studies. For example, other elder mistreatment instruments, such as the Conflict Tactics Scale (Straus et al., 1996), ask the older adult if they were hit, kicked, yelled at, burned, confined in a room, and subjected to other specific and violent behaviors, rather than the ‘hurt or harm’ language used by NSHAP to assess physical abuse. These differences may help explain why rates of reported elder mistreatment are higher in this study than in previous studies. We also found higher rates of mistreatment compared to previous studies using earlier rounds of the NSHAP (Laumann, Leitsch, & Waite, 2008), which can be explained by having a broader and more expansive set of measures. This is arguably a strength of this study, however, as we cast a wider net, and may notice risks that more severe measures would not. Still, any indication of mistreatment should raise concern, and the severity and importance of various indicators is somewhat subjective. Future research should investigate whether rural-urban differences in rates of elder mistreatment hold up using different measures.

Community Education and Policy Implications

Elder mistreatment is a persistent social justice and public health issue that affects older adults across geographies. Although we do not find a greater reported prevalence of abuse and neglect in rural areas, the high overall prevalence (more than 40%) suggests that many rural older adults are exposed to harm, potentially exacerbating rural health concerns. Prior research highlights the lack of services and resources to respond to abuse in rural and remote areas (Warren & Blundell, 2019) and the wide variation across geographies since adult protective services (APS) agencies are administered at the state and county level (Steinman & Anetzberger, 2022). Some regions have very few adult protection workers who must respond to abuse reports across vast territories. Steinman and Anetzberger (2022) found that APS agencies located in smaller and less populated counties had fewer opportunities to receive assistance on challenging cases from elder abuse multidisciplinary teams comprised of law enforcement, criminal justice, social service, and health agencies. Relative to larger and more populated counties, smaller counties were also less likely to report having “excellent” working relationships with their local Area Agency on Aging, county prosecutor, and probate court. Lack of collaboration to address elder mistreatment could lead to worse outcomes for older adults residing in rural areas. Together, these studies suggest a need for increased APS staffing levels, training, and more cross-agency collaboration in rural and remote areas.

Additionally, there is a clear need for community education for elder mistreatment prevention in both rural and urban contexts, as well as a need for social services to reduce risks after abuse is detected. Our analysis does not explicitly examine policy or practice solutions. However, prior research suggests that educating health care practitioners increases their knowledge and awareness of elder mistreatment (Mydin et al., 2021). Other studies indicate that caregiver support interventions may reduce the likelihood of abuse and neglect, but more evidence is needed (Fearing et al., 2017). We found relatively high rates of social support and cohesion among both rural and urban adults in this analysis. While adjusting for social well-being did not impact the overall risk of elder mistreatment in our results, tapping into the social networks of rural and urban older adults may be an important way to mitigate future mistreatment. Drawing again on the Contextual Theory of Elder Abuse (Roberto & Teaster, 2017), elder mistreatment must be viewed within the broader relational, community, and geographic contexts in which older adults live.

Limitations

Our study has several limitations that should be explored in future research. This study is cross-sectional, and so we cannot establish causality. Another limitation is that responses to the elder mistreatment items in the NSHAP are self-reported, so older adults with serious cognitive impairment, who are perhaps the most vulnerable to abuse and neglect, are not included in the survey. (Older adults with mild cognitive impairment were included.) Future surveys may incorporate responses from proxy respondents or administrative records from adult protective service agencies. Furthermore, we are relying on respondents’ recall to report various incidents, which may bias the results toward recent or more severe events. We also did not include income in our analysis, although we did control for education, marital status, and employment status, which are all correlated with income. Lastly, we are unable to determine if reported instances of elder mistreatment represent persistent interpersonal conflict within households or if the reported behaviors arose after the respondent became an older adult. Future research should explore how household and aging dynamics differ by rurality with respect to elder mistreatment.

We used the most recent NSHAP data at the time of conducting our analysis, but the data are from 2015–2016. The COVID-19 pandemic likely impacted many of the risk factors and outcomes we were examining. Future research should build on these results with more recent data to see how the results may have changed over time. Still, this is the first study of its kind to use the NSHAP module on elder mistreatment to illuminate rural-urban differences, and these results provide an important foundation for future study. Additionally, we relied on the most recently available geographic information from 2010–2011 to categorize individuals as rural or urban, so we assume that individuals did not move from rural to urban areas between 2010–2016. However, older adults, especially those in rural areas, have relatively low rates of residential mobility (Henning-Smith, Tuttle, Swendener, Lahr, & Yam, 2023), which provides some assurances in the validity of our results.

Conclusion

In this study, we found high rates of elder mistreatment for older adults, regardless of rurality, with more than four in 10 older adults reporting that they had experienced at least one type of mistreatment. After adjusting for sociodemographic and social well-being characteristics, we found that rural older adults had a lower predicted probability of experiencing elder mistreatment than their urban counterparts; this was true overall and also for nearly every individual mistreatment measure. We also found variation in the specific types of mistreatment reported. Some of the most common elder mistreatment indicators for both rural and urban older adults were asset exploitation (financial/borrowed money without paying it back, taking things without permission) and interpersonal and family conflict/abuse (felt uncomfortable, family conflict, name-calling). These results provide useful information on where to target efforts to address elder mistreatment. Overall, results from this study should raise concerns about mistreatment risk among urban older adults, as well as questions about protective factors for rural older adults that require additional research. Still, the high rates overall call for further prevention and mitigation efforts across levels of rurality; such efforts will need to take the unique geographic contexts into account.

What this paper adds.

• More than 40% of all older adults report experiencing some form of mistreatment, regardless of rurality.

• The most common forms of mistreatment included family conflict and having money borrowed without being paid back.

• Rural older adults are less likely to report mistreatment overall and within specific types of mistreatment.

Applications of study findings to practice, policy, and/or research.

• Urban older adults may require particular attention to prevent and address elder mistreatment.

• More research is needed to understand resilience among rural older adults, as well as to identify potential risks of underreporting mistreatment.

Data Availability Statement:

The National Social Life, Health and Aging Project are publicly available from the NORC at the University of Chicago: https://www.norc.org/research/projects/national-social-life-health-and-aging-project.html.