Loss of Autonomy: Likely Dementia and Living Arrangement Transitions Among Mexicans and Mexican Americans

Abstract

Background and Objectives

To examine the role of probable dementia on changes in living arrangements and mortality among very old Mexicans and Mexican Americans in 2 different nations.

Research Design and Methods

We employ the Hispanic Established Population for the Epidemiologic Study of the Elderly and the Mexican Health and Aging Study, 2 comparable longitudinal data sets, to identify predictors of changes in living arrangements using multinomial logistic regression, controlling for cognitive status, demographic characteristics, and resources.

Results

In Mexico, women with dementia who lived alone at baseline were more likely to become part of an extended family household than men with similar levels of cognitive impairment. A similar pattern emerges for the oldest Mexican-American women. Spousal loss increases the likelihood of living alone for women in the United States regardless of dementia. Although dementia elevates the risk of mortality for men living alone in the United States, in both countries, women in their 90s who lived alone with dementia had a lower risk of mortality relative to men.

Discussion and Implications

Longer life spans increase the risk of living alone with dementia in both countries, especially for women. Older individuals in both countries face financial hardships. Mexicans have limited formal options in dementia care. Mexican Americans with dementia continue to live alone despite low income although, unlike the Mexicans, they have access to Medicaid long-term care. For Mexico and the United States, the growing number of older individuals with dementia represents a growing public health concern.

In this paper, we compare changes in living arrangements associated with dementia among older Mexicans in Mexico and older Mexican Americans in the United States. These two countries have very different long-term care systems and are at different levels of development. In both countries, older Mexican-origin adults (Mexican Americans and Mexicans) are highly reliant on family members for late-life care and support (Cantu et al., 2022). Such dependence is inevitable given few individual and family resources among these groups and the scarcity of government-funded alternatives in Mexico (Angel et al., 2016; López-Ortega & Aranco, 2019). The situation in both countries reflects not only the high costs of formal care but also the greater inaccessibility of formal long-term care in Mexico. It may also reflect a culturally based rejection of any type of institutionalization (Angel et al., 1996; Rote et al., 2017). Additionally, as dementia progresses one’s capacity to care for oneself, and thus the capacity to live alone decreases making understanding the causes and consequences of changes in living arrangements highly relevant.

Although Mexico’s old-age welfare state remains less developed than that of the United States, similar social and demographic factors are affecting family composition and the living arrangements of older individuals in that country (Angel et al., 2017). Lower average incomes and fewer resources clearly limit the possibility of living alone in Mexico (Huffman et al., 2019) and people continue to live with their families to a greater extent than in the United States (Montes de Oca et al., 2014). Poverty is a major determinant of living arrangements in Mexico where approximately 47% of individuals over the age of 60 live in moderate or extreme poverty. This reality constrains independence and the possibility of living alone (Mejia-Arango et al., 2020; Montes de Oca et al., 2014).

In the context of high reliance of older adults on unpaid family care and support, living arrangements have become a fundamental factor in maintaining older adult health and well-being. Universally, this is especially relevant for older adults with cognitive impairment and dementia who present highly demanding care and need instrumental support. Rapid aging in the United States and Mexico as elsewhere has increased the number of older adults living with serious cognitive impairments (Agudelo-Botero et al., 2022; Angel et al., 2021).

Dementia is one of the most common causes of disability and dependence in the world and is a frequent reason that older people require supportive living situations (World Health Organization, 2022). In the United States, dementia rates for older adults were 3%, 22%, and 33%, respectively, and observed for people in three age groups: 70–74, 75–84, and 85–89 (Freedman et al., 2021). Rates are higher among Latinos—one and a half times greater among Latinos than non-Latino Whites among those 65 and older (Chen & Zissimopoulos, 2018; Manly et al., 2022). The burden of dementia is high in the Mexican-American population over 80 years. In Mexico, the most recent study shows an estimated global prevalence of dementia among people 60 and older of 14%. In this study, 7.8% participants met the criteria for Alzheimer’s disease, 4.3% for vascular dementia, and 2.1% for mixed dementia, and as seen in other countries, the prevalence of dementia was higher in women than in men with figures of 15.3% and 12.5%, respectively (Juarez-Cedillo et al., 2022). Dementia is now among the five main causes of disability-adjusted life-years in Mexicans 60 years and older (Parra et al., 2021), and among women 85 and over (Agudelo-Botero et al., 2022). In both countries, high dementia rates suggest greater family dependence. However, a recent study showed that more than 20% of older Mexican Americans and Mexicans living alone have dementia and are at risk of dependency and social isolation (Cantu et al., 2022).

In the following analyses, we employ the Hispanic Established Epidemiologic Study of the Elderly (HEPESE) in the United States and the Mexican Health and Aging Study (MHAS) to present a detailed longitudinal analysis of the way in which dementia impacts late-life transitions in household living arrangements. Both surveys provide useful comparative data and include detailed information on demographics, socioeconomic characteristics, activities of daily living, and cognitive functioning. We build on previous research of the panel studies to focus on Mexican-origin populations with similar sources of vulnerability, including low education, inadequate income, and few assets in their relative contexts (Mejia-Arango et al., 2020; Prickett & Angel, 2017). These contexts include different housing practices, different family availability, and differential social program access (Angel et al., 2021).

A Transitional Model of Living Arrangements

Mexico and the United States differ dramatically in the level of development, housing stock, home ownership patterns, patterns of co-residence among family and nonfamily members, and health systems, all of which have profound implications for the living arrangements of frail older adults (Angel et al., 2016). In the study, we operationalize living arrangements in terms of marital status and household size. Figure 1 is based on previous research on factors associated with late-life living arrangements that have implications for the extent of social and economic support and assistance needed when one’s capacity to live independently declines (Portacolone et al., 2022). The model depicts three possible options in living arrangements (alone, spouse only, and with others) at two different times, and the determinants that might cause a transition in living arrangements for older adults, such as demographic characteristics, resources, and cognitive function. As it shows, those factors may lead to changes in living arrangements over time.

Figure 1.

Model of dementia and changes in living arrangements.

Previous studies based on cross-sectional research designs have documented that economic and health factors are major correlates of household extension among those living with dementia (Cantu et al., 2021). Age, gender, low socioeconomic status, and childlessness are also associated with those living alone in Mexico and the United States (Mejia-Arango et al., 2020; Read & Grundy, 2016; Torres et al., 2021). Another study shows that the majority of Mexican Americans who move from independent living into extended family households experience declines in physical capacity (Prickett & Angel, 2017).

Our transition model proposes three categories of predictors of changes in living arrangements: cognitive status, demographic characteristics, and resources. Cognitive status indicators consist of instrumental activities of daily life (IADL). IADL deficits may be present in cognitive impairment given its association with neuropsychological functioning, and as we discuss later in detail, are incorporated into the algorithm used to define likely dementia (Mejía-Arango et al., 2015). Among older people of Mexican descent in the United States, for those with an ADL difficulty, higher income increases the likelihood of living with family (Prickett & Angel, 2017). Other studies investigating contexts of disability in late life in the United States have demonstrated that needs for assistance with instrumental activities of daily living (IADL disability) predict living arrangements among older adults. For example, individuals who live alone or in extended family households without a spouse had greater increases in IADL disability than individuals who lived with their spouse only (Henning-Smith et al., 2018; Portacolone et al., 2018). Mexican Americans with IADL disability and low cognitive functioning scores measured by the Mini-Mental Status Exam (MMSE) are less likely to live alone than in extended family households (Monserud, 2019). However, Mexican-American women with dementia had a greater tendency to live alone than their male counterparts (Cantu et al., 2022).

Previous work on living arrangements transitions from independent living into an extended household using HEPESE from 1993 through 2010 found that disability is a better predictor than finances for these transitions (Prickett & Angel, 2017). Event history analyses revealed that, of the 10.2% who transitioned during the observation period, late-life migration, physical disability, and cognitive impairment were key predictors of moving in with others. Financial strain did not signal a move to co-residential living as shown in earlier years of the life course; rather, the cohort’s aging has made health declines much more salient.

Demographic factors include age, gender, and education. Age and lower educational attainment are associated with a greater risk of dementia (National Research Council, 2004), and gender elevates the incidence rate because women live longer than men do. Resources are defined as the number of children, which is associated with the propensity to either live alone or co-reside with family, and also considers household poverty, which is negatively related to serious cognitive impairment (Cantu et al., 2022). Research also shows that financial resources are linked to living arrangements (Hughes & Waite, 2002).

Based on this risk profile, the current study examines the consequences of the potential loss of autonomy that accompanies dementia in forcing older individuals who live alone to move in with others, controlling for cognitive status, individual characteristics, and resources. We expect that because women frequently live longer than men, women in the last stage of the life course (over 80 years old) will be more likely to live alone or to move in with adult children due to the death of a spouse or disabilities related to increased age. We use the HEPESE (Markides et al., 1999) and the MHAS (Wong et al., 2015) to determine the extent to which the onset of dementia forces an older individual to move in with others rather than continue to live alone.

Data and Methods

In the present study of living arrangement transitions, we use the HEPESE and MHAS, two large longitudinal surveys of older adults, which include information demographic, socioeconomic, ADL, and cognitive status. The HEPESE is a prospective cohort household-based sample that, at baseline, was representative of Mexican Americans aged 65 years and above living in the Southwestern states of Arizona, California, Colorado, New Mexico, and Texas. The original baseline sample of 3,050 was interviewed between September 1993 and June 1994, and a secondary refreshing sample of 902 was interviewed during the fifth follow-up between September 2004 and June 2005. Interviews took place both in person and via proxy. For this study, we used a cohort from the seventh (2010–2011) through ninth wave (2015–2016) of the HEPESE with a follow-up of three waves for a total of 6 years.

The MHAS is a prospective panel study of adults 50 years and older in Mexico. The baseline survey was conducted in 2001 with a nationally representative sample of adults born in 1951 or earlier. Follow-up interviews were conducted in 2003, 2012, 2015, and 2018. In 2012, a new sample of adults was added to ensure representation of adults born between 1952 and 1962. In this study, we used data from the 2012 MHAS followed for two waves, 2015 and 2018 (Wong et al., 2015).

To make the MHAS and HEPESE samples comparable, the current study limits both samples to noninstitutionalized, Mexicans (n = 1,623) and Mexican Americans aged between 80 and 90 years old in 2010 (n = 744) with no missing covariates.

Dependent Variable

Living arrangements are defined by the total number of household members and marital status (married vs nonmarried) in both the HEPESE and MHAS. Using these indicators, we classified living arrangements into three categories: (1) live alone, (2) live with spouse only, and (3) live with others. We also examine the role of mortality to examine the probability of death in each living arrangement.

Independent Variables

The main predictor variable in this study is cognitive status. The HEPESE and MHAS contain different measures of cognitive functioning and disability that we used to construct a measurement of probable dementia (hereafter dementia). In the HEPESE, we use the MMSE to assess cognitive function and dementia among Mexican Americans in seven domains of mental health: orientation to time (5 points), orientation to place (5 points), three-word registration (3 points), attention and calculation (5 points), three-word recall (3 points), language (8 points), and visual construction (1 point). MMSE is not a clinical diagnosis, but a screening instrument used for research purposes to discriminate reliably between the highly functioning nondemented and poorly functioning demented, regardless of etiology. We use the different MMSE domains and IADL information to classify individuals as having dementia, two impaired domains and an IADL, and nondementia. Additionally, proxy-interviewed individuals were classified as dementia if the informant reported them as mentally incapacitated, defined as being unable to think properly in carrying out basic activities of daily life, including giving informed consent (Folstein et al., 1975). The MMSE ranges from 0 to 30 with a higher score representing respondents with normal cognition. To insure comparable measures of dementia between the two surveys, we use a protocol for determining dementia based on education-standardized cognitive functioning scores, IADL disability, and proxy report of cognitive impairment (Mejia-Arango et al., 2020).

In the MHAS, we use the modified Cross-Cultural Cognitive Examination to assess impairment based on two cognitive domains and an interference with the ability to function at work or to perform the usual activities of daily living (Mejía-Arango et al., 2015). The Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE) assesses cognitive status through a proxy respondent. In the MHAS a cutoff score of 3.4 with two cognitive domains and an IADL disability or a score of >3.4 in the IQCODE classifies subjects with dementia (Cherbuin & Jorm, 2013). To equalize the classification between the HEPESE and MHAS, we measure individuals who are impaired in two cognitive domains and one IADL disability or who score >3.4 in the IQCODE to distinguish respondents living with dementia from the nondementia group (Mejia-Arango et al., 2020). We operationalize dementia as high cognitive impairment combined with at least one type of IADL disability. Although these individuals might not receive a formal diagnosis of dementia in a rigorous examination, they clearly suffer from seriously impaired autonomy.

For demographic and functional status predictors, we included the respondent’s age (coded continuously in years), gender (coded as 1 = female, 0 = male), education (total number of years in formal education), total number of living children, and poverty as other predictors of living arrangement transitions. Poverty data from the MHAS are based on income quintiles for the entire 65 and older population at baseline, with older adults in the lower quintile of wealth being coded as living in poverty. It should be noted that there is no “poverty” measure in the MHAS but it is estimated that 20% of people over 65 in Mexico live in poverty (Cantu et al., 2022). Similarly, for the HEPESE, we use income data to determine poverty (1 = yes, 0 = otherwise). Older adults with income below $10,000 are considered as living in poverty. ADL items were used to measure self-care tasks needed to live independently (i.e., bathing, dressing, eating, walking, getting from a bed to a chair, and using the toilet). The six-item scale was dichotomized as 1 = difficulty with at least one or more ADLs versus 0 = no difficulty as the reference category.

Analysis

We first present descriptive statistics of both the total sample and the sample stratified by gender. We then use multinomial logistic regressions to examine the extent to which differences exist in living arrangement transitions for individuals living alone with dementia in each country. Based on this risk profile, the current study examines the probability of changes in living arrangements as the result of dementia controlling for demographic correlates of dementia. We employ a multinomial logistic regression method to estimate age-dementia-specific transition probability between three living arrangements (live alone, live with spouse only, and live with others) at baseline and at follow-up the transition to death as an additional possible outcome. All analyses are stratified by gender and control for poverty, years of education, and number of children. The regression model is represented by the following equation:

ln (p_ij/p_ii) = β_0+ β_1·age+ β_2·dementia+ β_3·poverty+ β_4·years of education+ β_5·number of children, stratified by gender.

Where p_ij is defined as the transition probability from living arrangement state i to living arrangement state j, and p_ii is defined as the probability that staying in the same living arrangement state i. Living arrangements state i includes living alone, living with spouse only, and living with extended family while living arrangements state j includes living alone, living with spouse only, living with extended family, and death. For all the possible transitions, we include the results where there are significant gender and age differences in probabilities in living arrangements transitions, and consider four types of transitions in living arrangements, including death: live alone to alone (no change), live with spouse only to alone, live alone to live with extended family, and alone to death. A bootstrapping technique is applied to obtain 95% empirical intervals of our transition probability estimates to perform statistical testing.

Results

Table 1 shows the descriptive statistics at the baseline (seventh wave for HEPESE, 2010–2011, and 2012 cohort for MHAS) stratified by gender in both the HEPESE (Mexican Americans) and MHAS (Mexicans). Mexican Americans are older (2.7 years); they have higher educational attainment (2 years) and fewer children than Mexicans. The Mexican-American sample is 65% women, and the Mexican sample is 57% women. Close to a third of each sample lives in poverty (32% for the HEPESE vs 42% for the MHAS). The Mexican sample has a lower prevalence of dementia than their Mexican American counterparts (17% vs 24%). Table 1 also shows the living arrangements for the total sample. Extended households are common in both countries with almost three-quarters of Mexicans (73%) and one-half (49%) of Mexican Americans living with others.

Table 1.

Descriptive Statistics for HEPESE (United States) and MHAS (Mexico) at Baseline

Characteristics	HEPESE (aged 80–90)						MHAS (80–90)
	Men		Women		Total		Men		Women		Total
	Mean	SD	Mean	SD	Mean	SD	Mean	SD	Mean	SD	Mean	SD
Age	84.0	2.1	84.1	2.2	84.0	2.2	81.2	2.2	81.3	2.3	81.3	2.3
Years of education	4.9	4.1	5.5	4.0	5.3	4.1	3.5	3.9	3.1	3.6	3.3	3.7
Number of children	5.1	2.9	4.6	3.0	4.7	3.0	6.4	3.3	5.8	3.3	6.1	3.3

	Freq	%	Freq	%	Freq	%	Freq	%	Freq	%	Freq	%
Dementia	49	18.9	126	26.0	175	23.5	86	12.2	195	21.3	281	17.3
Poverty	47	18.2	190	39.2	237	31.9	286	40.5	402	43.8	688	42.4
ADL	62	23.9	225	46.4	287	38.6	185	26.2	336	36.6	521	32.1
Living arrangement at baseline
Alone	48	18.5	160	33.0	208	28.0	68	9.6	146	15.9	214	13.2
With spouse only	97	37.5	73	15.1	170	22.8	138	19.6	86	9.4	224	13.8
With others	114	44.0	252	52.0	366	49.2	500	70.8	685	74.7	1185	73.0
Mortality rates
Alive	179	69.1	320	66.0	499	67.1	568	80.5	764	83.3	1332	82.1
Deatha	80	30.9	165	34.0	245	32.9	138	19.5	153	16.7	291	17.9
Total N	259		485		744		706		917		1,623
(%)	34.8%		65.2%		100.0%		43.5%		56.5%		100%

Notes: The table represents data at baseline (2010 HEPESE and 2012 MHAS). ADL = activities of daily living; Freq = frequency; HEPESE = Hispanic Established Epidemiologic Study of the Elderly; MHAS = Mexican Health and Aging Study; N = number of observations; SD = standard deviation.

^aIndividuals who died during the observation period of this study.

In the HEPESE, more women live in poverty (39% vs 18%) and have dementia than men (26% vs 19%). A higher percentage of women live alone compared with men (33% vs 19%). More than twice as many men live with a spouse only compared with alone (38% vs 15%). For the MHAS, men and women report equivalent poverty prevalence. Dementia prevalence among the oldest Mexicans is considerably lower than in their Mexican-American peers; however, gender patterns are similar in both cohorts. One out of five Mexican women (21%) report dementia, whereas only one out of eight (12%) men do. In Mexico, a majority of both men and women live in extended households, but a higher fraction of women live alone than men do (16% vs 10%). Among older Mexican Americans, a higher percentage of both men and women live alone than their Mexican peers, but almost twice the proportion of Mexican-American women live alone than Mexican-American men (33% vs 19%).

Figure 2 presents patterns in transitions in living arrangements by gender. Time 1 represents the baseline, 2010 HEPESE, and 2012 MHAS. For Time 2, we consider the subsequent two cohorts (i.e., 2012 and 2016 HEPESE and 2015 and 2018 MHAS). Data on those who remained living alone at Time 2 are shown in Figure 2. At baseline, the probability of living alone is higher for Mexican-American men than for Mexican men. In general, this probability decreases with age for Mexican-American men but not for Mexican men. Conversely, the data also reveal that Mexican-American men with normal cognition had a higher propensity to live alone than their Mexican counterparts. For women with dementia between 80 and 84 years old, the probability of continuing to live alone is higher for Mexican Americans than Mexicans. The likelihood of Mexican women with dementia remaining alone doubled with age from 0.33 for the young old to 0.74 for the oldest old. For both Mexican and Mexican-American women with normal cognition, the probability of remaining alone is higher than for those with dementia, with values ranging from about 0.49 to 0.82.

Figure 2.

Probability of continuing to live alone by dementia and gender. (A) HEPESE = Hispanic Established Epidemiologic Study of the Elderly; (B) MHAS = Mexican Health and Aging Study.

Figure 3 estimates the probability of a change from living alone to living with others by dementia status. Mexican women have a higher likelihood of a transition to moving in with others than men do (0.21 for women vs 0.11 for men on average). Mexican Americans present a different pattern with a higher likelihood observed only for women older than 87 relative to men in the same age group.

Figure 3.

Probability of transitioning from living alone to living with others by dementia and gender. (A) HEPESE = Hispanic Established Epidemiologic Study of the Elderly; (B) MHAS = Mexican Health and Aging Study.

Among those living with dementia, Mexican men under 83 have a higher likelihood of moving in with others than Mexican-American men. However, men with dementia in both cohorts follow a similar pattern, the likelihood of moving with others decreases steadily with age from 0.14 to 0.1 for Mexican-American men and from 0.26 to 0.03 for Mexican men. Mexican-American women with and without dementia gradually increase the likelihood of moving in with others, with values ranging from 0.02 to 0.22. Becoming part of an extended family household did not change significantly for Mexican women across their later life course. The oldest Mexican-American women experienced a slightly greater likelihood of moving in with others than their Mexican counterparts (Figure 3).

Figure 4 depicts a noticeable difference in transition from living with a spouse to living alone for women. Mexican-American women with and without dementia had an increased probability of changes from living with a spouse to living alone. Mexican women, regardless of cognitive status, experience a slight decrease in the likelihood of remaining with a spouse only. Spousal loss has a modest effect on living alone with dementia for Mexican-American men, with values under 0.01. The pattern varies slightly for Mexican men with dementia, with values as high as 0.08.

Figure 4.

Probability of transitioning from living with spouse only to living alone by dementia and gender. (A) HEPESE = Hispanic Established Epidemiologic Study of the Elderly; (B) MHAS = Mexican Health and Aging Study.

Figure 5 shows the transition from living alone to death. For men with dementia, we observe a dramatic difference in mortality associated with living alone. Among men in Mexico, the probability of death increases dramatically with age. Men 90 years old are about four times more likely than men at the age of 80 to experience a transition from living alone to death. Mexican-American women with dementia, however, display the greatest probability of this transition (0.46 for women at age 80). For women with dementia living alone at baseline in both cohorts, the odds of death slightly decreases with age for Mexican Americans and dramatically decreases for Mexicans. For those without dementia, the transition to death increases with age for men in both cohorts, but not for women. Older Mexican women who lived alone had the lowest probability of death (0.03 for women at age 90).

Figure 5.

Probability of transitioning from living alone to death by dementia and gender. (A) HEPESE = Hispanic Established Epidemiologic Study of the Elderly; (B) MHAS = Mexican Health and Aging Study.

Discussion

Our findings demonstrate that in Mexico and the United States dementia is clearly associated with living with others, but also that a large fraction of older individuals with cognitive impairment continue to live alone. In Mexico, as in much of the developing world, the new reality of rapidly aging populations presents governments and families with serious challenges that require novel solutions (Angel & Montes de Oca, 2022). The more highly developed old-age welfare state in the United States provides more potential support than is the case in Mexico which lacks programs such as Medicaid or community long-term services and supports (Angel et al., 2021).

To be sure, profound demographic forces, including low fertility rates coupled with high out-migration of the young, have shrunk the family caregiver network, creating unmet needs in long-term services and supports. Our study shows how serious the problem is in low-income countries, it also underscores that it is serious enough in affluent countries, despite a generous old-age welfare state. A caring state means that long-term care for those living with dementia is relied on only as a last resort. Living alone is the only viable alternative if extended households are unavailable as depicted in Table 1. The current study provides a template to examine this issue.

In this binational study, we examined the impact of dementia among the oldest Mexican Americans and Mexicans on changes in living arrangements separately for men and women with differing social and material resources. Older Mexican-American men living alone, regardless of cognitive status, had a greater increase in risk of death with age compared with men with other living arrangements. This segment of the Mexican-American population is especially vulnerable to unmet care needs, falls, and loneliness. Intervention and prevention efforts such as increasing access to home-based dementia care services and outreach and support should, therefore, target older Mexican-American men who are living alone (Angel et al., 2021). The COVID-19 pandemic, which limited social engagement and potentially social support, may have had a particularly detrimental impact on older Mexican-American men living alone (Portacolone et al., 2018). Still, future research on this topic should address the unique predictors and explanations for this increased risk in mortality by age.

For Mexican-American women, the transition from living alone to living with others was more common than the transition from living with a spouse to living alone and this was slightly higher for women living with dementia. Mejia-Arango and colleagues reached a similar conclusion based on cross-sectional research of the oldest Mexican Americans (Mejia-Arango et al., 2020). These findings point to stark gender differences of changes in living arrangements with more movement to larger living arrangements for women living with and without dementia than for men. This could reflect gendered expectations in the need for caregiving and support structures for Mexican-American women as well as the early mortality of men. A particularly important finding for the Mexican-American cohort was that while the probability of transitioning from living with a spouse to living alone increases throughout late life for women, the opposite pattern is shown for older men. Widowed women are especially at risk of living alone with cognitive impairment which stems from the need to manage living alone and challenges associated with difficulties in carrying out basic activities of daily living as well as the lack of appropriate services to foster independence (Portacolone et al., 2022). These results bolster previous research, suggesting a protective effect of family living arrangements as long-term care options diminish during late life. As a result, socioeconomic resources may become more important for this vulnerable population (Prickett & Angel, 2017).

In Mexico, a different pattern emerged, with more older women than men without dementia living alone throughout later life. For men, the most common transition in living arrangements was from living alone to living with others. For men living alone, there was an increased probability of death in both groups living with and without dementia. Much as in the United States, living alone is a major risk factor of death for men, and living alone with dementia is particularly precarious. Dementia and cognitive impairment affect the ability to carry out daily and instrumental activities as well as manage complex medical and financial tasks, especially in the event of a transition to living alone (Portacolone et al., 2022). Older men with dementia were also at greater risk of death than women with dementia, which could be attributable to similar risk factors including cardiovascular disease as well as social isolation or lack of support (Mejia-Arango et al., 2020).

Finally, as discussed earlier, we observed that men living with and without dementia had an increased probability of continuing to live with a spouse throughout late life. This reflects the important role of spousal care for Mexican-American men and the extended longevity of Mexican women. For older women in Mexico, the most common transition was from living alone or with the spouse only to living with others. Much like their U.S. counterparts, older women with and without dementia living alone in Mexico were also more likely to transition to living with extended family than men. In Mexico, men as a group are not moving in with extended family partly because few outlive their spouses or lose autonomy (Montes de Oca et al., 2014).

Limitations

This study has several limitations that future research should address. Although the results presented here provide important insights into the complex interplay among socioeconomic resources, living arrangements, and the aging process in two nations growing old, future analyses need to assess changes in the increased role of socioeconomic resources in determining whether older Mexicans and Mexican Americans living with dementia can depend on their family. A lingering and significant issue is whether older adults with few resources can afford to live in neighborhoods with access to comprehensive health care and better amenities. Older Mexicans, and to some extent, Mexican Americans must rely on their own human capital to age in place in the community. Socioeconomic resources that allow a certain level of autonomy would provide fresh insights into how institutions mobilize support for older adults in both countries.

One particularly important objective of future research is furthering our understanding of the association between changes in marital status over time, especially widowhood on the course of dementia given the longer life expectancy of women generally. Widowed older women in our study were more likely than married women to live with adult children or other relatives, especially in Mexico given that country’s higher fertility regime. Additionally, filial piety, defined as cultural norms of family caregiving, supports co-residence as a preferred arrangement for a widowed parent, regardless of the number of children. Understanding different marital transitions over time and the relationship to living arrangements will lead to a better understanding of how changes in family structure influence pathways to dementia care and needs for specific social care support strategies for those living alone.

In sum, both countries living alone with dementia pose a serious public health risk and without adequate support undermines the arc of the quality of life for older adults. Future investigations should examine the availability of children and their expectations in taking care of aging parents. Motivations for forming an extended household to address older adult cognitive impairment merits particular attention. Lastly, most of the older adults in both countries rely on unpaid family carers for support and care whose needs must also be addressed.

Funding

This work was supported by The National Institute on Aging (Grant R03AG063183), Center on Aging and Population Sciences (P30 center grant, P30AG066614), and Population Research Center (P2C center grant, P2CHD042849, NICHD). Dr. Cantu is supported by a research career development award (K12HD052023: Building Interdisciplinary Research Careers in Women’s Health Program-BIRCWH; Berenson, PI) from the National Institutes of Health/Office of the Director (OD), National Institute of Allergy and Infectious Diseases (NIAID), and Eunice Kennedy Shriver National Institute of Child Health & Human Development (NICHD). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Conflict of Interest

None.

Data Availability

Publicly available datasets were analyzed in this study. These data can be found at the National Archive of Computerized Data on Aging, located within ICPSR, The University of Michigan [https://www.icpsr.umich.edu/web/NACDA/series/546]. This study was not pre-registered.