Skip to main content
NCBI home page
The Journals of Gerontology Series B: Psychological Sciences and Social Sciences logoLink to The Journals of Gerontology Series B: Psychological Sciences and Social Sciences
. 2021 Apr 17;77(1):249–259. doi: 10.1093/geronb/gbab066

Life Course Stressors and Functional Limitations in Later Life Among White, Black, and Hispanic Adults: Deleterious, Hardening, or Benign?

Madison R Sauerteig 1,2,✉, Kenneth F Ferraro 1,2, Shawn Bauldry 1,2
Editor: Deborah S Carr
PMCID: PMC8755903  PMID: 33864079

Abstract

Objectives

Although striking racial and ethnic disparities in health are manifest during later life, they may be rooted in early-life exposures. Drawing from cumulative inequality theory, we investigate whether life course stressors are associated with the risk of later-life functional limitations and whether this relationship differs by race and ethnicity.

Methods

We utilize longitudinal data from the Health and Retirement Study to test whether child and adult stressors predict trajectories of the occurrence and severity of functional limitations among a diverse sample of older adults.

Results

Child and adult stressors are associated with greater occurrence and severity of functional limitations during later life. Mediation analyses reveal the indirect influence of child stressors via adult stressors on occurrence and severity of functional limitations; however, the indirect effects are slightly stronger for Black and Hispanic adults than their White counterparts.

Discussion

Child stressors, in and of themselves, do not increase functional limitations among Black and Hispanic people but are associated with greater adult stress exposure, predisposing them to more functional limitations. Results suggest that childhood stressors are associated with distinct social pathways to functional limitations among White, Black, and Hispanic older adults.

Keywords: Disability, Health disparities, Stress accumulation, Trauma, Weathering

Decades of research provide evidence of glaring racial and ethnic disparities in various health indicators, ranging from life expectancy to functional limitations. Scholars have shown that the gravity and historical persistence of health disparities are likely due to structural disadvantages associated with racism in the social, economic, and political spheres of life (Phelan & Link, 2015; Williams et al., 2019). As such, underrepresented minorities are more likely to live their lives in the face of multiple stressors including, but not limited to, poverty, victimization, and substance abuse (Forrester et al., 2019; Jackson et al., 2011). Although scores of studies of racial and ethnic disparities in health have examined proximal stressors, we study health disparities in functional limitations by examining the accumulation of stressors over the life course. Stressors refer to the contextual threats or demands that tax the “individual’s ordinary capacity to adapt” as well as the absence of desired resources to reach personal goals (Aneshensel & Mitchell, 2014, p. 54).

Functional limitations are restrictions in an individual’s ability to perform physical tasks and hinder many people from aging optimally. Although functional limitations are a threat to well-being among all adults, we examine the relationship between life course stressors and functional limitations among three U.S. racial–ethnic groups: non-Hispanic White, non-Hispanic Black, and Hispanic Americans (hereafter, White, Black, and Hispanic adults, respectively). Our reasons for doing so are twofold. First, evidence reveals that Black and Hispanic adults are at notably higher risk of functional limitations than White adults (Boen & Hummer, 2019; Brown, 2018; Brown et al., 2012; Kail et al., 2020; Lin & Kelley-Moore, 2017).

Second, Black and Hispanic adults are more likely than White adults to report more stressful experiences, both during childhood and later life (Brown et al., 2020; Lee & Chen, 2017; Maguire-Jack et al., 2020; Sternthal et al., 2011; Umberson et al., 2014). Given their higher lifetime exposure to stressful living conditions and higher rates of functional limitations, we investigate whether lifetime stress accumulation is implicated in racial–ethnic disparities in health. In doing so, we also examine the direct influence of childhood stressors on functional limitations as well as their potential indirect influence via greater adult stress (i.e., mediation). This research is designed to enable the development of effective upstream interventions to reduce U.S. health disparities.

Health Disparities and the Life Course

Stressors and Disparities in Functional Limitations

Many health disparities manifest during later life but have their origins in early-life exposures. The life course perspective emphasizes that humans move through a series of transitions from birth to death, interacting with various institutions along the way (Elder & Caspi, 1988). Transitions are discrete changes in status that punctuate life course trajectories (Foner & Kertzer, 1978; George, 1993). When compared to White people, transitions by Black and Hispanic people may occur on somewhat different chronological clocks and be complicated by racism and its structural disadvantages (Forrester et al., 2019; Jackson et al., 2011). For example, Black and Hispanic youth are less likely than White youth to complete high school by age 18. Also, the age of a cancer diagnosis for Black adults in the United States is about 3 years earlier than for White adults. Serious stressors that involve a status change may have long-term consequences and manifest as functional limitations or other health problems (Krause et al., 2004; Lantz et al., 2005). There is also some evidence that the repercussions of traumatic events remain longer for Black than White people (Jackson et al., 2011). Could it be that racial–ethnic disparities in functional limitation are due in part to greater accumulated stress exposure among Black and Hispanic people?

The literature on functional limitations is substantial, and a growing number of studies use longitudinal data, shedding light on whether the racial–ethnic gap in functional limitations widens, shrinks, or remains fairly constant over time (Kail et al., 2020; Kelley-Moore & Ferraro, 2004; Lin & Kelley-Moore, 2017; Lin, 2020). Among studies that consider stress and functional limitations, some examine stressors during childhood but not adulthood, with the childhood stressors limited to socioeconomic status (SES) (Brown et al., 2012; Kelley-Moore & Huang, 2017) or SES and health (Luo & Waite, 2005). There also are studies that examine adult stressors but not child stressors (Lantz et al., 2005) or combine child and adult stressors into a measure of lifetime trauma or adversity (Boen & Hummer, 2019). Empirical generalizations from these studies include (a) Black and Hispanic older adults have more functional limitations than their White counterparts, (b) stressors raise the risk of functional limitations, and (c) the disparity remains similar over time (persistent inequality).

Among the few studies that investigate the influence of child and adult stressors separately on functional limitations and test for racial–ethnic disparities, some examine SES and health stressors (Haas & Rohlfsen, 2010), extensive inventories of trauma (Krause et al., 2004), or stress (Sternthal et al. 2011). Krause et al. (2004) found no differences in functional limitations between White and non-White respondents after accounting for childhood and midlife trauma, but Sternthal et al. (2011) reported substantial disparities between White, Black, and Hispanic adults. Despite these exemplary studies, a key life course question merits greater attention. Do childhood stressors predispose one to a lifetime of greater stress exposure, thereby raising the risk of greater functional limitation? To the best of our knowledge, only one study has addressed this question directly. Haas and Rohlfsen (2010) examined a limited range of stressors (SES and health) over the life course and reported that the “impact of childhood health was strongly mediated through subsequent health and socioeconomic attainment in adulthood” (p. 248). We build on the findings of this study, but do so with a wider range of stressors and focus on the endogeneity of adult stressors. Instead of viewing stress accumulation as a simple count of exposures, we view it as a social pathway. We anticipate that childhood stressors increase adult stressors and that the indirect influence on functional limitations via adult stressors will be stronger for Black and Hispanic adults (i.e., moderated mediation).

Theoretical Framework

Several concepts and models are useful for this investigation, and we integrate them with cumulative inequality (CI) theory, a middle-range theory focusing on how stratification unfolds throughout the life course. This theory emphasizes how early disadvantage and advantage lead to risks, resources, and opportunities over the life course via demographic and developmental processes (Ferraro et al., 2009). Two elements of CI theory are central to this study.

First, CI theory emphasizes that the magnitude and duration of disadvantages are related to greater exposure to additional risks. The severity of disadvantage sets the stage for life course continuity and discontinuity, and prior studies reveal that the quantity of exposure elevates the risk of functional limitation (Montez & Hayward, 2014). Furthermore, CI theory emphasizes the importance of when and what type of stress occurred. Child stressors may predispose one to adult stressors, increasing the risk of physiological dysregulation and functional limitations (Boen & Hummer, 2019). By examining both child stressors and adult stressors, this study identifies how early insults set the stage for later-life disparities (Brown et al., 2012; Haas & Rohlfsen, 2010).

Second, CI theory emphasizes that social status leads to distinct lived experiences that increase exposure to stressors and accentuate how social context shapes human agency and resource activation to modify life trajectories (Ferraro et al., 2017). Although CI theory specifies that human agency can alter later-life trajectories, it also highlights that “inequality is not primarily the result of individual choices and actions but is structurally generated” (Ferraro & Shippee, 2009, p. 334).

An abundance of literature reveals that racism is embedded into the structure of society, whereby minority groups face more disadvantages across life domains, constrained choices, and limited resources (Omi & Winant, 2014; Williams et al., 2019). As such, this social structure has historically limited racial–ethnic minorities’ opportunities to attain socioeconomic resources and enact health-protective behaviors (August & Sorkin, 2011; Hales et al., 2020). Living in a social structure that constrains opportunities also channels Black and Hispanic adults into stigmatized life paths and greater stress exposure. Research by Geronimus et al. (2006) refers to this as weathering, a stress hypothesis, because racial–ethnic minorities are more likely to “experience early health deterioration as a consequence of the cumulative impact of repeated experience with social or economic adversity and political marginalization” (p. 826). Similarly, as Wheaton (1990) observed decades ago, “chronic stress is analogous to the stress model of physical mechanics” involving a sustained load of wear-and-tear (p. 210). Based on CI theory, we expect accumulated stressors and limited resources to be deleterious to physical function, and that Black and Hispanic adults will have greater stress exposure than White adults.

A competing hypothesis, referred to as the hardening effect, suggests that when faced with difficult stressors, the long-term effects may be beneficial rather than harmful (Bleuler, 1974; Elder & Caspi, 1988). In the nineteenth century, Friedrich Nietzsche argued “that which does not kill us makes us stronger,” and the idea is popularized in lay culture. Some of the earliest research on the phenomenon appeared in Bleuler’s (1974) examination of the children of schizophrenics; he found that despite the young people’s “miserable childhoods … pain and suffering can have a steeling—a hardening—effect on some children, rendering them capable of mastering life with all its obstacles” (p. 106). Elder and Caspi (1988) found that children under stress during the Great Depression were able to overcome the hardships without much strain or damage, and Seery et al. (2010) reported that experiencing some lifetime adversity predicted relatively lower distress and functional impairment compared to high levels of adversity. These writers did not suggest that adverse childhood experiences are somehow salubrious, only that they provide a “problem” from which individuals are able to adapt in constructive ways (Chen & Miller, 2012). Although weathering specifies greater risk to minorities due to greater stress exposure, hardening specifies the opposite: stressors, except perhaps at very high levels, may reduce health risks because people learn to adapt by developing skills to overcome their plight. Based on CI theory, hardening is more likely when stressor accumulation is not appraised as severe and/or the person has sufficient resources to “reframe” the stressors. By corollary, we expect stress accumulation to be benign when it is low and/or the person has ample resources to mitigate the consequences.

Research Questions

To extend the previous literature, we pose three research questions. First, are life course stressors associated with more functional limitations in later life? Second, does race–ethnicity influence the relationship between life course stressors and trajectories of functional limitation? Finally, do adult stressors mediate the relationship between child stressors and functional limitation trajectories for White, Black, and Hispanic older adults?

Method

Sample

This study utilizes longitudinal data from the Health and Retirement Study (HRS)—a multistage, probability study of adults aged 50 years and older, with oversampling of Black and Hispanic Americans and Florida residents. A lag structure was specified to assure temporal order in predictive models while optimizing the use of variables collected from half-sample surveys. Most demographic variables, life course stressors, and covariates are collected at baseline (2006/2008), with the first wave of functional limitations gathered 2 years later (2008/2010).

We limited our analysis sample to respondents who met the following criteria: (a) participated and had non-zero weights in the 2006 or 2008 wave including the psychosocial module (N = 13,805), (b) scored within 2 SD of the average cognition score or higher (N = 13,326), (c) did not have proxy reports for child stressors (N = 12,593), (d) identified as either non-Hispanic White, non-Hispanic Black, or Hispanic (N = 12,346), and (e) provided information on functional limitations for at least one wave between 2008 and 2016 (N = 11,132). These selection criteria yielded an analysis sample of 11,132 older adults who are distinctive in that they were not cognitively impaired and were able to answer questions on childhood stressors.

Demographic Variables

Models adjust for baseline demographic characteristics including age, gender, and race–ethnicity. We also adjust for region born and marital status, as these variables may confound racial–ethnic differences in functional limitation trajectories (Haas & Rohlfsen, 2010).

Functional Limitations

Up to four waves of functional limitations are used for each respondent. For each wave, we calculated the number of reported limitations across 10 standard measures (e.g., walking several blocks); see Supplementary Material for a full list of indicators. Figure 1 illustrates how we analyzed multiple waves of data with the half-sample measurement design.

Figure 1.

Figure 1.

Open in a new tab

Illustration of construction of outcome variables. Notes: In order to maintain temporal ordering, for respondents who reported adult stressors in 2006 (Sample 06), we constructed functional limitation measures from 2008, 2010, 2012, and 2014 waves and for respondents who reported adult stressors in 2008, we constructed functional limitation measures from 2010, 2012, 2014, and 2016 waves. The control variables (e.g., age, marital status, physical activity) that may vary from wave to wave are adjusted for to match when adult stressors are reported. The other variables that stay stable over time (e.g., gender, race–ethnicity, education level, child stressors) are accounted for when respondents first responded to those questions.

Life Course Stressors

We measure life course stressors covering childhood and adulthood. Child stressors are derived from self-reported exposures to six experiences prior to age 18 (e.g., physically abused by a parent). Each indicator is coded dichotomously and summed to create a count variable, ranging from 0 to 6.

Adult stressors are self-reported exposures to 11 experiences (e.g., received food stamps). Each indicator is coded dichotomously and summed to create a cumulative adult stressor measure, ranging from 0 to 11.

Adult SES and Lifestyle Factors

We adjust for two indicators of adult SES (education and wealth) and four indicators of adult lifestyle factors (smoking status, alcohol consumption, physical activity, and obesity). Additional details about the sample and measures are given in Table 1 and Supplementary Material.

Table 1.

Weighted Descriptive Statistics by Race–Ethnicity, Health and Retirement Study

Total sample White adults Black adults Hispanic adults
N = 11,132 N = 8,925 N = 1,354 N = 853
Demographics
Age (years) 65.49 (0.09) 65.82 (0.10) 63.89 (0.24)a 63.53 (0.29)b
Woman, % 55.97 55.38 61.03a 56.73
Race–ethnicity
 Non-Hispanic White, % 80.17 — — —
 Non-Hispanic Black, % 12.16 — — —
 Hispanic, % 7.66 — — —
Born in South, % 28.39 24.38 73.51a 21.39c
Marital status
 Married, % 67.65 70.07 46.42a 64.79bc
 Divorced/separated, % 13.79 12.30 25.5a 17.14bc
 Widowed, % 14.81 14.40 20.18a 13.19c
 Never married, % 3.75 3.23 7.9a 4.88bc
Outcome
T1 functional limitations 2.62 (0.03) 2.51 (0.03) 3.32 (0.10)a 3.12 (0.12)b
T2 functional limitations 2.89 (0.03) 2.75 (0.03) 3.64 (0.10)a 3.59 (0.13)b
T3 functional limitations 2.84 (0.03) 2.72 (0.04) 3.67 (0.11)a 3.29 (0.12)bc
T4 functional limitations 2.97 (0.03) 2.80 (0.04) 4.07(0.12)a 3.60 (0.13)bc
Life course stressors
Childhood stressors (0–6) 0.77 (0.01) 0.75 (0.01) 0.83 (0.03)a 0.86 (0.04)b
Adulthood stressors (0–11) 1.59 (0.01) 1.52 (0.01) 2.05 (0.05)a 1.97 (0.06)b
Adult SES and lifestyle factors
Education (years) 13.06 (0.03) 13.44 (0.03) 12.28 (0.08)a 9.67 (1.53)bc
Wealth (∛) 2.92 (0.02) 3.14 (0.02) 1.67 (0.04)a 1.89 (0.06)bc
Physical activity (0–17.6) 8.21 (0.04) 8.38 (0.04) 7.16 (0.11)a 7.44 (0.14)b
Obese, % 41.07 39.3 54.2a 45.92bc
Smoking status
 Never smoker, % 43.11 43.02 41.19 46.38c
 Former smoker, % 42.84 43.45 38.54a 40.92
 Current smoker, % 14.05 13.53 20.27a 12.70c
Alcohol consumption
 Does not drink, % 62.41 60.40 73.65a 72.24b
 Moderate drinker, % 31.05 32.41 22.78a 25.13b
 Heavy drinker, % 6.54 7.19 3.57a 2.63b
Open in a new tab

Notes: SES = socioeconomic status. Weighted descriptive statistics (mean or percentage) with standard errors in parentheses, based on 10 multiple-imputed data sets with the exception of functional limitations. Numbers are rounded to the nearest hundredth. Significance indicated at p < .05.

aSignificant difference between Black and White adults.

bSignificant difference between Hispanic and White adults.

cSignificant difference between Hispanic and Black adults.

Missing Data

We address missing data in the predictors using Stata’s implementation of multiple imputation with chained equations to construct 10 complete data sets (StataCorp, 2019). We stratify the imputation models by race–ethnicity to account for potential interactions and use the individual indicators of child and adult stressors as auxiliary variables to incorporate all available information for imputing values for those measures. Rather than imputing values for functional limitations, we address missing data in the estimation process via full information maximum likelihood (FIML) as is standard in the latent growth modeling framework (Bollen & Curran, 2006).

Analysis

Prior studies highlight the importance of distinguishing whether individuals report any functional limitations (occurrence) from the number of functional limitations (severity) (Haas & Rohlfsen, 2010; Taylor, 2008). Following Haas and Rohlfsen (2010), for each time point, we construct one series of binary indicators for respondents who reported any functional limitations and another series of continuous measures for the log of the count of functional limitations, excluding respondents with zero functional limitations. We specify bivariate latent growth models that simultaneously fit trajectories in the probability of reporting any functional limitations and trajectories in the severity of functional limitations while also allowing us to include predictors of the respective latent intercepts and slopes (Bollen & Curran, 2006). Missing data for the respondents with zero functional limitations for the repeated measures of log counts of functional limitations are addressed in the estimation process via FIML.

Our analysis proceeds in three steps. First, we fit a series of bivariate latent growth models that specify sets of predictors for the latent intercepts and slopes of occurrence and severity of functional limitations. Second, we develop our models by adding (a) child and adult stressors, (b) adult SES and lifestyle factors, and (c) interaction terms between race–ethnicity and child and adult stressors. To help address mortality selection, we estimated the probability of mortal attrition during the waves we used for functional limitation and include the inverse as a covariate in all models (Heckman, 1979). Third, we examine whether adult stressors mediate the relationship between child stressors and trajectories of functional limitations. For this analysis, we add adult stressors as an additional outcome predicted by child stressors and demographic variables and then use the resulting estimates to calculate the direct, indirect, and total effects of childhood stressors on the latent intercepts and latent slopes. We assess mediation by testing the significance of the indirect effects with standard errors calculated using the delta method (Muthén & Muthén, 1998–2017). In addition to fitting the model on the whole sample, we stratify the analysis by race–ethnicity to determine whether estimates of direct, indirect, and total effects of childhood stressors vary by race–ethnicity, using standard tests for differences in coefficients for independent samples (Clogg et al., 1995). Significant differences in the indirect effects across race–ethnicity provide evidence of moderated mediation.

We fit all the bivariate growth models in Mplus 8 (Muthén & Muthén, 1998–2017). Mplus enables us to incorporate sample weights and adjust standard errors for the multistage cluster sampling design.

Results

Descriptive Statistics

Descriptive statistics and differences by race–ethnicity for all study variables are presented in Table 1. As expected, Black and Hispanic adults report more functional limitations on average than White adults across all four time points. Figure 2 presents racial–ethnic trends in the occurrence and severity of functional limitations across the four time points. This figure reveals clear racial–ethnic differences in the proportion of having any limitation at each time point and the average number of limitations over time. Black and Hispanic adults are more likely than White adults to have functional limitations at each time point (approximately 0.70, 0.68, and 0.66, respectively, at Time 1). Among respondents who have functional limitations, Black and Hispanic adults report a higher number at each time point.

Figure 2.

Figure 2.

Open in a new tab

Proportion of respondents with any functional limitations and mean number of functional limitations (if report anything other than 0) by race-ethnicity at each time point. Solid line = White, dotted line = Black, and dashed line = Hispanic.

Examining the mean number of stressors, we find that compared to White adults, Black and Hispanic adults reported more child stressors (0.75, 0.83, and 0.86, respectively) and adult stressors (1.52, 2.05, and 1.97, respectively).

Predictors of Trajectories

We first fit an unconditional bivariate latent growth model to characterize average trajectories of occurrence and severity and to determine the extent of variation in the latent intercepts and slopes. On average, trajectories of occurrence are essentially stable across the four time points, while trajectories of severity show a modest increase over time (Supplementary Table A1). We also note that there is substantially more variance in the latent intercepts when compared with the latent slopes for occurrence and severity.

Table 2 reports selected estimates from a series of bivariate latent growth models. We report estimates only for predictors of the latent intercepts as we found little evidence of associations with the latent slopes (see Supplementary Table A2 for the complete set of estimates). We note that the reported estimates for the occurrence of functional limitations are log odds that can be exponentiated to obtain odds ratios and the reported estimates for the severity of functional limitations are standard regression coefficients, which with a logged outcome can be interpreted as percentage changes. Adjusting for demographic characteristics (Model 1), we see large disparities in both occurrence and severity of functional limitations. Hispanic adults have higher odds (log odds b = 0.19, p < .001) relative to White adults of reporting any functional limitations at Time 1; among those who report any functional limitations, Black and Hispanic adults report more than White adults at baseline (b = 0.17, p < .001 and b = 0.21, p < .001, respectively).

Table 2.

Selected Parameter Estimates from Bivariate Latent Growth Curve Models for Functional Limitations (N = 11,132)

Occurrence latent intercept Severity latent intercept
1 2 3 4 5 1 2 3 4 5
Demographics
Age 0.03*** 0.03*** 0.04*** 0.04*** 0.04*** 0.01*** 0.01*** 0.01*** 0.01*** 0.01***
Womana 0.34*** 0.38*** 0.39*** 0.40*** 0.39*** 0.15*** 0.17*** 0.17*** 0.16*** 0.16***
Blackb 0.01 −0.03 −0.20*** −0.25*** −0.30** 0.17*** 0.14*** 0.05 0.02 0.01
Hispanicb 0.19** 0.12* −0.22*** −0.21** −0.15 0.21*** 0.18*** 0.00 0.03 −0.04
Born in Southc 0.23*** 0.20*** 0.18*** 0.15** 0.15** 0.14*** 0.13*** 0.11*** 0.08*** 0.07***
Marital status
 Divorced/separatedd 0.21*** 0.14* 0.00 0.00 0.00 0.21*** 0.15*** 0.07* 0.07* 0.07*
 Widowedd 0.20*** 0.01 −0.11 −0.11 −0.11* 0.16*** 0.04 −0.03 −0.03 −0.03
 Never marriedd 0.01 0.03 −0.04 −0.05 −0.05 0.07 0.09 0.02 0.00 −0.01
Life course stressors
Childhood stressors 0.12*** 0.09*** 0.09*** 0.11*** 0.06*** 0.04*** 0.04*** 0.05***
Adulthood stressors 0.13*** 0.12*** 0.12*** 0.10*** 0.08*** 0.07*** 0.06*** 0.05***
Adult SES and lifestyle factors
Education (years) −0.06*** −0.05*** −0.05*** −0.03*** −0.02*** −0.01***
Wealth (∛) −0.12*** −0.08*** −0.08*** −0.08*** −0.04*** −0.04***
Physical activity −0.07*** −0.07*** −0.05*** −0.05***
Obesee 0.48*** 0.48*** 0.13*** 0.13***
Former smokerf 0.09* 0.09* 0.06** 0.06**
Current smokerf 0.21*** 0.21*** 0.14*** 0.14***
Moderate drinkerg −0.16*** −0.16*** −0.13*** −0.13***
Heavy drinkerg −0.08 −0.08 −0.11** −0.12**
Interactions
Child stressors × Black −0.10* −0.05
Child stressors × Hispanic −0.09 0.00
Adult stressors × Black 0.07 0.03
Adult stressors × Hispanic 0.01 0.03
R 2 0.10 0.13 0.21 0.36 0.36 0.09 0.13 0.19 0.31 0.31
Open in a new tab

Notes: SES = socioeconomic status. Unstandardized estimates based on 10 multiple imputation data sets. Estimates incorporate sample weights and adjust standard errors for survey design. The estimates for occurrence are log odds; the estimates for severity are regression coefficients.

aReference group is men.

bReference group is White.

cReference group is non-South.

dReference group is married.

eReference group is not obese (body mass index <30 kg/m2).

fReference group is never smoker.

gReference group is does not drink.

***p < .001, **p < .01, *p < .05.

Model 2 introduces child and adult stressors as predictors. We find that both are associated with higher odds of occurrence (childhood: log odds b = 0.12, p < .001; adult: log odds b = 0.13, p < .001) and greater severity (b = 0.06, p < .001 and b = 0.08, p < .001, respectively, for childhood and adulthood). In addition, the inclusion of both slightly attenuates the relationships between race–ethnicity and occurrence and severity of functional limitations.

Models 3 and 4 add adult SES and lifestyle factors. Adjusting for SES attenuates slightly the associations of child and adult stressors for both occurrence and severity. In addition, adjusting for SES reveals a net negative association for Black and Hispanic respondents for the odds of occurrence and completely attenuates the positive associations observed with the severity of functional limitations. This pattern persists after including adult lifestyle factors.

Variability by Race-Ethnicity

Table 2, Model 5 reports estimates from a model that permits the associations of child and adult stressors for both occurrence and severity to vary by race–ethnicity. We find the odds of occurrence of functional limitation based on child stressors vary by race–ethnicity. In particular, we find higher odds of occurrence based on child stressors for Whites (log odds b = 0.11, p < .001) but essentially no relationship between child stressors and the odds of occurrence of functional limitations for Blacks (interaction log odds b = −0.10, p < .05). We do not find evidence for any other interactions.

Mediation Analysis

Table 3 reports the estimates from our mediation analysis. We include only estimates for mediation with respect to the occurrence and severity latent intercepts because we found no evidence of mediation for the latent slopes. Beginning with the overall sample, we estimate the net total effect of child stressors on the occurrence latent intercept as log odds b = 0.12 (p < .001) with a net indirect effect of log odds b = 0.03 (p < .001), accounting for about 25% of the total effect. Similarly, for the severity latent intercept, we estimate a net total effect of b = 0.06 (p < .001) with a net indirect effect of b = 0.01 (p < .001), accounting for about 17% of the total effect.

Table 3.

Moderated Mediation Results for Child Stressors and Functional Limitations

Parameter Occurrence latent intercept Severity latent intercept
All respondents (N = 11,132)
Child stressors–Functional limitations 0.09*** 0.05***
Child stressors–Adult stressors 0.24*** 0.24***
Adult stressors–Functional limitations 0.11*** 0.05***
Total indirect effect 0.03*** 0.01***
Total effect 0.12*** 0.06***
White respondents (N = 8,925)
Child stressors–Functional limitations 0.11*** 0.05***
Child stressors–Adult stressors 0.24*** 0.24***
Adult stressors–Functional limitations 0.10*** 0.05***
Total indirect effect 0.02*** 0.01***
Total effect 0.13*** 0.06***
Black respondents (N = 1,354)
Child stressors–Functional limitations −0.02 −0.02
Child stressors–Adult stressors 0.30*** 0.30***
Adult stressors–Functional limitations 0.16*** 0.09***
Total indirect effect 0.05*** 0.03***
Total effect 0.03 0.01
Hispanic respondents (N = 853)
Child stressors–Functional limitations 0.03 0.03
Child stressors–Adult stressors 0.27*** 0.27***
Adult stressors–Functional limitations 0.12* 0.08***
Total indirect effect 0.03* 0.02**
Total effect 0.06 0.05
Open in a new tab

Notes: Unstandardized estimates based on 10 multiple imputation data sets. Estimates incorporate sample weights and adjust for standard errors for survey design. The estimates for occurrence are log odds; the estimates for severity are regression coefficients. Models adjust for demographics and adult socioeconomic status and lifestyle factors. Each dash (–) in the first column refers to a relationship between the predictor variable and the outcome variable.

***p < .001, **p < .01, *p < .05.

Table 3 also reports estimates stratified across the three racial–ethnic groups. Consistent with the results reported above, when we stratify the analysis, we find no evidence of net direct effects of child stressors on the occurrence and severity of functional limitations at Time 1 for either Black or Hispanic respondents. In other words, for Black and Hispanic respondents, we find that child stressors do not influence later-life functional limitations directly; however, they are indirectly related via adult stressors. In fact, the indirect effects of childhood stressors on the occurrence of functional limitations are slightly larger for Black and Hispanic respondents primarily due to stronger relationships between child and adult stressors for minorities with b = 0.24 (p < .001) for Whites, b = 0.30 (p < .001) for Blacks, and b = 0.27 (p < .001) for Hispanics.

Discussion

We used two vantage points—life course analysis and racial–ethnic health disparities—to advance understanding of the relationship between life course stressors and functional limitations in later life. Related to our first research question, longitudinal analyses reveal that stressful experiences, during both childhood and adulthood, are associated with greater occurrence and severity of functional limitations in later life, even after adjusting for adult SES and lifestyle factors. These findings corroborate previous literature demonstrating the influence of stressful experiences on later-life functional health (Brown et al., 2012; Haas & Rohlfsen, 2010; Luo & Waite, 2005). Some studies report that childhood factors are significant by themselves but not after adjusting for adult resources (Haas & Rolfsen, 2010) while others examine lifetime exposure only (Boen & Hummer, 2019; Shrira & Litwin, 2014). By distinguishing child and adult stressors, the present study reveals that stressors at each life stage elevate both the occurrence and severity of functional limitations in later life. Although adult stressors are more consequential, even after adjusting for a long vector of adult resources and lifestyle factors, the influence of child stressors is long-lasting.

Related to our second research question, results reveal that most relationships between life course stressors and functional limitations are similar across racial–ethnic groups, meaning that stress exposure is associated with the occurrence and more severe limitations for White, Black, and Hispanic adults. However, we found that Black adults with more child stressors are slightly less likely than White adults to experience baseline occurrence of functional limitations. This interaction effect for the occurrence latent intercept provides partial support for the hardening thesis pertaining to childhood exposures, but among Black people only (Elder & Caspi, 1988), perhaps because of how Black people appraise stressors (Brown et al., 2020).

By contrast, there is no evidence of hardening due to adult stressors or SES for Black, Hispanic, or White adults. Not only is there strong evidence that adult stressors are deleterious to functional ability, but the influence of adult SES on functional limitations is substantial and consistent with prior studies (Boen & Hummer, 2019; Haas & Rohlfsen, 2010). Indeed, after adjusting for adult SES, we found that the occurrence of functional limitations for Black and Hispanic adults is negative, revealing that Black and Hispanic adults are less likely than their White counterparts to have any limitation at baseline (Haas & Rohlfsen, 2010). Both the weathering hypothesis and CI theory specify SES resources as essential to reducing inequality, not only because they are directly related to health outcomes but also because they correlate with modifiable risk factors such as smoking and obesity, each of which accelerates functional decline (Ferraro & Shippee, 2009; Geronimus et al., 2006).

Related to our third research question, we initially found that Black and Hispanic adults experience more child and adult stressors than White adults, providing support for differential exposure. Our mediation analysis, however, reveals that child stressors are associated with later-life functional ability differently for White, Black, and Hispanic adults. For White adults, child stressors have a long-lasting and direct influence on the occurrence and severity of functional limitations. For Black and Hispanic adults, however, child stressors indirectly influence functional ability through experiences of adult stress.

Without considering mediation, one might conclude that child stressors are benign to the health of Black and Hispanic adults. The mediation analyses, however, reveal that childhood stressors predispose one to more functional limitations via greater stress exposure for Black and Hispanic adults. The findings help advance a view that stress accumulation is more than a simple sum of life course stressors; instead, it is a social pathway whereby childhood stressors channel many Black and Hispanic people into higher levels of adult stressors—a lifetime of greater stress exposure. Although Haas and Rohlfsen (2010) found that childhood health was strongly mediated by adult SES and health, we are unaware of any prior study that examined the mediating effect of adult stress on the relationship between childhood stress and later-life health. Consistent with prior studies, our findings reveal large racial–ethnic disparities in functional health, but we show that these disparities result from different social pathways of stress exposure. We provide evidence for moderated mediation (conditional indirect effects) and highlight how the process of stress accumulation predisposes Black and Hispanic persons to more functional limitations.

Findings from this study should be interpreted with several limitations in mind. First, child stressors are assessed retrospectively. Many scholars argue for the validity of these measures because evidence reveals that autobiographical experiences and salient events from childhood can be recalled accurately (Conway, 2009). Nevertheless, we sought to reduce potential bias by adjusting for demographic variables and SES as well as excluding respondents who fell 2 SD below the mean cognition score (Vuolo et al., 2014).

Second, there is some unevenness in the measurement of adult stressors. Seven items from the HRS psychosocial questionnaire ask if the respondent “ever experienced” the event. Four items from the HRS core survey probe experiences “since the last wave,” which may lead to underestimating the prevalence of adult stressors. Relatedly, our analysis focused on the use of acute stressful events as opposed to chronic stressors occurring during childhood and adulthood. Whereas research suggests that Black and White people may experience distinct stressors (Jackson et al., 2011), future research on functional limitations should consider differentiating chronic and acute stressors and types of stressors.

Third, although there is a lag structure in place to ensure that stressors and other adult SES resources and lifestyle factors are gathered prior to functional limitations, it is unknown precisely when the adult stressors occurred. Greater specificity in the timing and duration of stressors will greatly advance future research.

Finally, many of the variables in this study are self-reported. Of particular interest, the outcome is based on self-reports that may underestimate the race difference in functional limitations (Simonsick et al., 2008). Most of these limitations, however, probably lead to conservative bias, potentially underestimating the influence of life course stress on later-life functional limitations.

Conclusions

Consistent with CI theory, the accumulation of life stressors, both during childhood and adulthood, influences functional limitations in later life for White, Black, and Hispanic adults. Child stressors, in and of themselves, do not increase functional limitations among Black and Hispanic people. The deleterious influence of child stressors among Black and Hispanic people is indirect: Child stressors channel Black and Hispanic adults into a lifetime of stressful experiences that, in turn, result in greater functional limitations. By contrast, child stressors have both a direct and an indirect effect on functional limitation for White adults. The weight of the evidence is that life course stressors are deleterious to functional health, but adult resources, especially education, wealth, and health-protective behaviors, can reduce racial health disparities in functional limitations.

Funding

This work was supported by grants from the National Institute on Aging (R01-AG043544, RF1-AG043544) to K. F. Ferraro.

Conflict of Interest

None declared.

Author Contributions

M. R. Sauerteig drafted the manuscript, prepared the data, and supervised the statistical analyses. K. F. Ferraro planned the study and wrote sections of the manuscript. S. Bauldry performed the statistical analyses needed for the revision of this manuscript and wrote sections of the revised manuscript pertaining to analytic methods and the interpretation of results. All authors contributed to revisions of the manuscript.

Supplementary Material

gbab066_suppl_Supplementary_Materials
Click here for additional data file. (179KB, pdf)

References

  1. Aneshensel, C. S., & Mitchell, U. A. (2014). The stress process: Its origins, evolution, and future. In Johnson R. J., Turner R. J., & Link B. G. (Eds.), Sociology of mental health: Selected topics from forty years 1970s–2010s (pp. 53–74). Springer. doi: 10.1007/978-3-319-07797-03 [DOI] [Google Scholar]
  2. August, K. J., & Sorkin, D. H. (2011). Racial/ethnic disparities in exercise and dietary behaviors of middle-aged and older adults. Journal of General Internal Medicine, 26(3), 245–250. doi: 10.1007/s11606-010-1514-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bleuler, M. (1974). The offspring of schizophrenics. Schizophrenia Bulletin, 1(8), 93–107. doi: 10.1093/schbul/1.8.93 [DOI] [PubMed] [Google Scholar]
  4. Boen, C. E., & Hummer, R. A. (2019). Longer—but harder—lives?: The Hispanic health paradox and the social determinants of racial, ethnic, and immigrant–native health disparities from midlife through late life. Journal of Health and Social Behavior, 60(4), 434–452. doi: 10.1177/0022146519884538 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Bollen, K. A., & Curran, P. J. (2006). Latent curve models: A structural equation perspective. John Wiley & Sons. doi: 10.1002/0471746096 [DOI] [Google Scholar]
  6. Brown, L. L., Mitchell, U. A., & Ailshire, J. A. (2020). Disentangling the stress process: Race/ethnic differences in the exposure and appraisal of chronic stressors among older adults. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 75(3), 650–660. doi: 10.1093/geronb/gby072 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Brown, T. H. (2018). Racial stratification, immigration, and health inequality: A life course-intersectional approach. Social Forces, 96(4), 1507–1540. doi: 10.1093/sf/soy013 [DOI] [Google Scholar]
  8. Brown, T. H., O’Rand, A. M., & Adkins, D. E. (2012). Race–ethnicity and health trajectories: Tests of three hypotheses across multiple groups and health outcomes. Journal of Health and Social Behavior, 53(3), 359–377. doi: 10.1177/0022146512455333 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Chen, E., & Miller, G. E. (2012). “Shift-and-persist” strategies: Why low socioeconomic status isn’t always bad for health. Perspectives on Psychological Science, 7(2), 135–158. doi: 10.1177/1745691612436694 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Clogg, C. C., Petkova, E., & Haritou, A. (1995). Statistical methods for comparing regression coefficients between models. American Journal of Sociology, 100(5), 1261–1293. doi: 10.1086/230638 [DOI] [Google Scholar]
  11. Conway, M. A. (2009). Episodic memories. Neuropsychologia, 47(11), 2305–2313. doi: 10.1016/j.neuropsychologia.2009.02.003 [DOI] [PubMed] [Google Scholar]
  12. Elder Jr, G. H., & Caspi, A. (1988). Human development and social change: An emerging perspective on the life course. In Bolger N., Caspi A., Downey G., & Moorehouse M. (Eds.), Persons in context: Developmental processes (pp. 77–113). Cambridge University Press. doi: 10.1017/CBO9780511663949.005 [DOI] [Google Scholar]
  13. Ferraro, K. F., Kemp, B. R., & Williams, M. M. (2017). Diverse aging and health inequality by race and ethnicity. Innovation in Aging, 1(1), 1–11. doi: 10.1093/geroni/igx002 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Ferraro, K. F., & Shippee, T. P. (2009). Aging and cumulative inequality: How does inequality get under the skin? The Gerontologist, 49(3), 333–343. doi: 10.1093/geront/gnp034 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Ferraro, K. F., Shippee, T. P., & Schafer, M. H. (2009). Cumulative inequality theory for research on aging and the life course. In Bengston V. L., Gans D., N. M. Pulney, & Silverstein M. (Eds.), Handbook of theories of aging (pp. 413–433). Springer Publishing Co. [Google Scholar]
  16. Foner, A., & Kertzer, D. (1978). Transitions over the life course: Lessons from age-set societies. American Journal of Sociology, 83(5), 1081–1104. doi: 10.1086/226675 [DOI] [Google Scholar]
  17. Forrester, S. N., Gallo, J. J., Whitfield, K. E., & ThorpeJr, R. J. (2019). A framework of minority stress: From physiological manifestations to cognitive outcomes. The Gerontologist, 59(6), 1017–1023. doi: 10.1093/geront/gny104 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. George, L. K. (1993). Sociological perspectives on life transitions. Annual Review of Sociology, 19(1), 353–373. doi: 10.1146/annurev.so.19.080193.002033 [DOI] [Google Scholar]
  19. Geronimus, A. T., Hicken, M., Keene, D., & Bound, J. (2006). “Weathering” and age patterns of allostatic load scores among blacks and whites in the United States. American Journal of Public Health, 96(5), 826–833. doi: 10.2105/AJPH.2004.060749 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Haas, S., & Rohlfsen, L. (2010). Life course determinants of racial and ethnic disparities in functional health trajectories. Social Science & Medicine, 70(2), 240–250. doi: 10.1016/j.socscimed.2009.10.003 [DOI] [PubMed] [Google Scholar]
  21. Hales, C. M., Carroll, M. D., Fryar, C. D., Ogden, C. L. (2020) Prevalence of obesity and severe obesity among adults: United States, 2017–2018. NCHS Data Brief, 360, 1–8. PMID: 32487284. [PubMed] [Google Scholar]
  22. Heckman, J. J. (1979). Sample selection bias as a specification error. Econometrica, 47(2), 153–161. doi: 10.2307/1912352 [DOI] [Google Scholar]
  23. Jackson, J. S., Hudson, D., Kershaw, K., Mezuk, B., Rafferty, J., & Tuttle, K. K. (2011). Discrimination, chronic stress, and mortality among black Americans: A life course framework. In Rogers R. G. & Crimmins E. M. (Eds.), International handbook of adult mortality (pp. 311–328). Springer. doi: 10.1007/978-90-481-9996-9_15 [DOI] [Google Scholar]
  24. Kail, B. L., Taylor, M. G., & Rogers, N. (2020). Double disadvantage in the process of disablement: Race as a moderator in the association between chronic conditions and functional limitations. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 75(2), 448–458. doi: 10.1093/geronb/gby027 [DOI] [PubMed] [Google Scholar]
  25. Kelley-Moore, J. A., & Ferraro, K. F. (2004). The black/white disability gap: Persistent inequality in later life? The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 59(1), S34–S43. doi: 10.1093/geronb/59.1.S34 [DOI] [PubMed] [Google Scholar]
  26. Kelley-Moore, J., & Huang, W.(2017). The “good times” cohort in later-life: Black–white differences in pathways to functional limitations. Research on Aging, 39(4), 526–548. doi: 10.1177/0164027516655582 [DOI] [PubMed] [Google Scholar]
  27. Krause, N., Shaw, B. A., & Cairney, J. (2004). A descriptive epidemiology of lifetime trauma and the physical health status of older adults. Psychology and Aging, 19(4), 637–648. doi: 10.1037/0882-7974.19.4.637 [DOI] [PubMed] [Google Scholar]
  28. Lantz, P. M., House, J. S., Mero, R. P., & Williams, D. R. (2005). Stress, life events, and socioeconomic disparities in health: Results from the Americans’ Changing Lives Study. Journal of Health and Social Behavior, 46(3), 274–288. doi: 10.1177/002214650504600305 [DOI] [PubMed] [Google Scholar]
  29. Lee, R. D., & Chen, J. (2017). Adverse childhood experiences, mental health, and excessive alcohol use: Examination of race/ethnicity and sex differences. Child Abuse & Neglect, 69, 40–48. doi: 10.1016/j.chiabu.2017.04.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Lin, J. (2020). Inter-individual variability in trajectories of functional limitations by race/gender. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 75(5), 1082–1092. doi: 10.1093/geronb/gby156 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Lin, J., & Kelley-Moore, J. (2017). Intraindividual variability in late-life functional limitations among White, Black, and Hispanic older adults: Implications for the weathering hypothesis. Research on Aging, 39(4), 549–572. doi: 10.1177/0164027516655583 [DOI] [PubMed] [Google Scholar]
  32. Luo, Y., & Waite, L. J. (2005). The impact of childhood and adult SES on physical, mental, and cognitive well-being in later life. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 60(2), S93–S101.doi: 10.1093/geronb/60.2.S93 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Maguire-Jack, K., Lanier, P., & Lombardi, B. (2020). Investigating racial differences in clusters of adverse childhood experiences. The American Journal of Orthopsychiatry, 90(1), 106–114. doi: 10.1037/ort0000405 [DOI] [PubMed] [Google Scholar]
  34. Montez, J. K., & Hayward, M. D. (2014). Cumulative childhood adversity, educational attainment, and active life expectancy among U.S. adults. Demography, 51(2), 413–435. doi: 10.1007/s13524-013-0261-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Muthén, L. K. & Muthén, B. O. (1998–2017). Mplus user’s guide (8th ed.). Muthén & Muthén. [Google Scholar]
  36. Omi, M., & Winant, H. (2014). Racial formation in the United States (3rd ed.). Routledge. doi: 10.4324/9780203076804 [DOI] [Google Scholar]
  37. Phelan, J. C., & Link, B. G. (2015). Is racism a fundamental cause of inequalities in health? Annual Review of Sociology, 41, 311–330. doi: 10.1146/annurev-soc-073014-112305 [DOI] [Google Scholar]
  38. Seery, M. D., Holman, E. A., & Silver, R. C. (2010). Whatever does not kill us: Cumulative lifetime adversity, vulnerability, and resilience. Journal of Personality and Social Psychology, 99(6), 1025–1041. doi: 10.1037/a0021344 [DOI] [PubMed] [Google Scholar]
  39. Shrira, A., & Litwin, H. (2014). The effect of lifetime cumulative adversity and depressive symptoms on trajectories of functional status among older adults. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 69(6), 953–965. doi: 10.1093/geronb/gbu056 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Simonsick, E. M., Newman, A. B., Visser, M., Goodpaster, B., Kritchevsky, S. B., Rubin, S., Nevitt, M.C., & Harris, T.B. (2008). Mobility limitation in self-described well-functioning older adults: Importance of endurance walk testing. The Journals of Gerontology, Series A: Biological Sciences and Medical Sciences, 63(8), 841–847. doi: 10.1093/gerona/63.8.841 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. StataCorp . (2019). Stata statistical software: Release 16. StataCorp LP. [Google Scholar]
  42. Sternthal, M. J., Slopen, N., & Williams, D. R. (2011). Racial disparities in health: How much does stress really matter? Du Bois Review, 8(1), 95–113. doi: 10.1017/S1742058X11000087 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Taylor, M. G. (2008). Timing, accumulation, and the Black/White disability gap in later life: A test of weathering. Research on Aging, 30(2), 226–250. doi: 10.1177/0164027507311838 [DOI] [Google Scholar]
  44. Umberson, D., Williams, K., Thomas, P. A., Liu, H., & Thomeer, M. B. (2014). Race, gender, and chains of disadvantage: Childhood adversity, social relationships, and health. Journal of Health and Social Behavior, 55(1), 20–38. doi: 10.1177/0022146514521426 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Vuolo, M., Ferraro, K. F., Morton, P. M., & Yang, T. Y. (2014). Why do older people change their ratings of childhood health? Demography, 51(6), 1999–2023. doi: 10.1007/s13524-014-0344-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Wheaton, B. (1990). Life transitions, role histories, and mental health. American Sociological Review, 55(2), 209–223. doi: 10.2307/2095627 [DOI] [Google Scholar]
  47. Williams, D. R., Lawrence, J. A., & Davis, B. A. (2019). Racism and health: Evidence and needed research. Annual Review of Public Health, 40, 105–125. doi: 10.1146/annurev-publhealth-040218-043750 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

gbab066_suppl_Supplementary_Materials
Click here for additional data file. (179KB, pdf)

Articles from The Journals of Gerontology Series B: Psychological Sciences and Social Sciences are provided here courtesy of Oxford University Press

RESOURCES