Resiliency Among Women’s Health Initiative Women Aged 80 and Older by Race, Ethnicity, and Neighborhood Socioeconomic Status

Abstract

Objectives

A comprehensive examination of resilience by race, ethnicity, and neighborhood socioeconomic status (NSES) among women aged ≥80 is needed, given the aging of the U.S. population, increasing longevity, and growing racial and ethnic diversity.

Methods

Participants were women aged ≥80 enrolled in the Women’s Health Initiative. Resilience was assessed with a modified version of the Brief Resilience Scale. Descriptive statistics and multiple linear regression examined the association of demographic, health, and psychosocial variables with resilience by race, ethnicity, and NSES.

Results

Participants (n = 29,367, median age = 84.3) were White (91.4%), Black (3.7%), Hispanic (1.9%), and Asian (1.7%) women. There were no significant differences by race and ethnicity on mean resiliency scores (p = .06). Significant differences by NSES were observed regarding mean resiliency scores between those with low NSES (3.94 ± 0.83, out of 5) and high NSES (4.00 ± 0.81). Older age, higher education, higher self-rated health, lower stress, and living alone were significant positive correlates of resilience in the sample. Social support was correlated with resilience among White, Black, and Asian women, but not for Hispanic women. Depression was a significant correlate of lower resilience, except among Asian women. Living alone, smoking, and spirituality were significantly associated with higher resilience among women with moderate NSES.

Discussion

Multiple factors were associated with resilience among women aged ≥80 in the Women’s Health Initiative. Despite some differing correlates of resilience by race, ethnicity, and NSES, there were many similarities. These results may aid in the design of resilience interventions for the growing, increasingly diverse population of older women.

Resilience has been defined as an individual’s ability to “bounce back,” as well as a process of adapting well in the face of adversity, trauma, tragedy, threats, or significant sources of stress (American Psychological Association, 2015). The Staudinger model of resilience (Staudinger & Greve, 2015) and Kumper’s resilience framework (Kumpfer, 1999) both theorize resilience as a multifaceted, biopsychosocial construct affected by personal issues, environmental factors, education, and quality of life. In other words, resilience is a biopsychosocial construct that considers interactive causes and contributing factors that influence resilience. For example, mental/emotional, social, and physical characteristics associated with resilience include happiness, optimism, adaptive coping styles, lower depression, and greater life satisfaction (Felix et al., 2019; MacLeod et al., 2016; Springfield et al., 2022; Wells, 2010; Woods et al., 2016). High resilience has also been associated with several correlates of physical health including activities of daily living, independence, high mobility, fewer cognitive complaints, longevity, successful aging, and high self-reported physical health (MacLeod et al., 2016; Taylor & Carr, 2020). Resilience has also been associated with social factors including community involvement, sense of purpose, social support and engagement, and contact with family and friends (Gooding et al., 2012; MacLeod et al., 2016). Despite these known correlates among the general population, limited research has examined these associations within older adults, specifically older women aged ≥80 years.

Examining resilience within older populations is appropriate considering the increasing population of older adults in the United States. Older adults are currently the fastest-growing age group in the United States, dominated by “Baby Boomers,” with adults aged ≥80 years growing most rapidly. Considering women typically live longer than their male counterparts, it may be especially important to investigate the association between psychological resilience and associated factors among older women to understand longevity in this group. Despite the negative stereotypes of aging, older adults often report high levels of well-being, quality of life, and resilience and consider themselves aging successfully despite chronic disease (MacLeod et al., 2016; Woods et al., 2016). Prior research (Gooding et al., 2012; Nygren et al., 2005) found adults aged ≥85 to have the same or greater capacity for resilience compared to their younger counterparts, suggesting that resiliency and longevity are correlated.

Exploring potential differences in resilience among adults aged ≥80 years by the social constructs of race and ethnicity is warranted (Garcia et al., 2022). Not only is the oldest-old population growing the fastest but also it is becoming increasingly diverse. Thus, responding to their health needs in a culturally sensitive manner, while considering inequalities over the life course, is more important than ever before. Members of minoritized racial and ethnic groups experience discrimination and systematic racism (i.e., the ways in which societies foster discrimination through mutually reinforcing inequitable systems) in many aspects of American society (Bailey et al., 2017). This historical and present struggle affects their mental and physical health and contributes to lower economic, social, and political status. In the same vein, individuals living in lower neighborhood socioeconomic status (NSES) disproportionately experience psychological stressors, such as financial difficulties, familial instability, discriminatory acts, and limited access to quality health care. NSES takes into account not only income but also neighborhood disadvantage (Chrisinger et al., 2022). Differences in NSES are, in part, a result of racial segregation, which is associated with poor health among members of minoritized racial and ethnic groups (Ungar, 2011; Williams et al., 2019). Assessing NSES also provides a strong measure of environmental risk, which has been shown to influence resilience (Shuey & Leventhal, 2017). Although older adults from minoritized racial and ethnic groups and/or low NSES are at risk for poor physical and mental outcomes, there are many who are able to overcome the negative consequences of their environments, develop resilience, and experience a healthy quality of life (Kok et al., 2018; Lee et al., 2022). Therefore, research findings on differences in resilience by minoritized racial and ethnic groups and NSES are unclear (Springfield et al., 2022; Wells, 2010; Woods et al., 2016).

Racial and ethnic differences in resilience may demonstrate a cross-over effect in older age, related to racial disparities in life expectancy. Research has determined that older Black adults have higher mortality in younger old age because of structural racism, but Black older adults have lower mortality risk than White older adults after about age 80, resulting in the “cross-over effect” (Sautter et al., 2012). Thus, the population of older Black adults aged ≥80 may be more robust than their White peers, creating a survivorship bias in research. As older age is associated with resilience (Gooding et al., 2012; Woods et al., 2016), the potential cross-over effect observed in resilience by race and ethnicity within the ≥80-year-old population is worth investigating. Thus, the purpose of this study was to examine factors associated with higher resilience in women 80 years and older in the Women’s Health Initiative (WHI) by race and ethnicity (e.g., Asian, Black, Hispanic, and White) and NSES. This study also sought to assess differences and similarities in factors associated with resilience among minoritized racial and ethnic groups, as well as women with lower NSES.

Method

Study Population

The WHI is a longitudinal study of 161,808 postmenopausal women originally recruited at 40 clinical centers across the United States from 1993 to 1998. Women were enrolled in either the WHI observational study or randomized clinical trials that could include, in various combinations, estrogen plus progestin or estrogen alone therapy, dietary modification, and calcium and vitamin D supplementation (Women’s Health Initiative Study Group, 1998). Women were between the ages of 50–79 years at recruitment and had to be postmenopausal at baseline. In 2005, after the completion of all WHI clinical trials, women in both the observational study and the clinical trial components were reconsented for the first WHI extension (2005–2010), second extension (2010–2015), and then a third extension (2015 onward).

For these analyses, we used a data set from the WHI, which was created for a special issue in 2016 on women’s health at age 80 and older (Journals of Gerontology: Series A, 2016). WHI participants were included if they were (a) originally enrolled in the WHI observational study or one of the WHI clinical trials; (b) consented to the first (2005–2010) and second WHI extensions (2010–2015); (c) 80 years of age or older on September 17, 2012; and (d) completed WHI follow-up Form 155, when they were age 80 years or greater, on participants’ reports of their physical and mental functioning, psychosocial concerns, and health-related quality of life.

Measures

Primary outcome—resilience

Resilience to life’s stressors was measured by a modified version of the Brief Resilience Scale (Smith et al., 2008). It included three items (i.e., “I tend to bounce back quickly after hard times.”; “It does not take me long to recover from a stressful event.”; “I have a hard time making it through stressful events.”). Response categories were as follows: strongly disagree, disagree somewhat, disagree slightly, agree slightly, agree somewhat, and agree strongly. This scale is scored using the mean of the three items, with total scores ranging from 1 to 5, where higher scores indicate better resilience.

Demographic characteristics

The following demographic variables were used in these analyses: Age in years; self-reported race (White, Black, and Asian); self-reported ethnicity (Hispanic or non-Hispanic); the highest level of education achieved (<high school, high school graduate; some/college or associate’s degree, 4-year college degree; and >4-year college degree); marital status (married/living as married, divorced/separated/widowed, and single/never married); and current living situation (live alone and living with someone).

Neighborhood socioeconomic status

The NSES index ranges from 0 to 100 across U.S. census tracts, with higher scores indicating more affluent tracts (Dubowitz et al., 2012). NSES was calculated at the level of census tracts using an index of six variables collected in the 2010 Census: (1) percent of adults older than 25 with less than a high school education; (2) percent of male participants who were unemployed; (3) percent of households with incomes below the poverty line; (4) percent of households receiving public assistance; (5) percent of households with children headed by a woman; and (6) median household income. NSES indices were assigned to the WHI participants based on their tract of residence, and were grouped as low (bottom 25%), moderate (middle 50%), or high (top 25%).

Self-rated health

The participants’ appraisal of their current health status was obtained from a single item (Hays & Morales, 2001). Response categories were excellent, very good, good, fair, or poor. Higher scores indicate lower self-reported health.

Physical Health subscale of the Research and Development Corporation

The Physical Health subscale of the Research and Development Corporation (RAND) 36-Item Health Survey (Zimmerman et al., 2006) is comprised of 10 items that assess activities of daily living: engaging in vigorous and moderate activities, lifting or carrying groceries, climbing several flights of stairs, climbing one flight of stairs, bending and stooping, walking more than a mile, walking several blocks, walks one block, and bathing and dressing themselves. Scores are transformed to a scale from 0 to 100 with higher scores indicating higher levels of functioning. The RAND-36 has been used extensively in studies of older adults and distinguishes between older adults with and without chronic disease (Hays & Morales, 2001).

Center for Epidemiologic Studies—Depression Scale Short Form (Burnham scoring) depression

Depressive symptoms were measured by the eight-item Burnham short form of the Center for Epidemiologic Studies—Depression Scale (CES-D; Burnam et al., 1988). Responses were scored according to the Burnham algorithm with a final range from 0 to 1. Scores >0.06 are considered indicative of significant depressive symptoms, and those ≤0.06 are indicative of no/minimal depressive symptoms.

Perceived Stress Scale

A four-item version of the 14-item Perceived Stress Scale (Cohen et al., 1983) was used in the WHI to assess the participants’ level of stress. This scale ranges from 0 to 16, with a higher score indicating more perceived stress.

Spirituality

The spiritual beliefs of the participants were assessed by a single item: “Religion gives me strength and comfort.” Response categories were none, a little, or a great deal with higher scores indicating higher spiritual beliefs.

Social support

To evaluate the social support available to the participants, nine questions from the Medical Outcomes Study Social Support Survey (Sherbourne & Stewart, 1991) were included. Responses to these questions ranged from 1 = none of the time to 5 = all of the time. The summary scores range from 9 to 45, and a higher score indicates more social support.

Activities of daily living

Four items were used to assess the participant’s ability to perform four primary activities of daily functioning (i.e., whether they could eat, dress, get in and out of bed, and take a bath). Responses to each function were 1 = without help, 2 = some help, and 3 = completely unable to do. Responses to the four items were summed, with a lower score indicating a greater ability to perform basic activities of daily living.

Life events

To evaluate the impact of life events on comorbidity, the WHI included 11 items from the Alameda County Epidemiologic Study (Berkman & Syme, 1979), which were later modified for the Beta Blocker Heart Attack Trial Study (Ruberman et al., 1984). These items include the death of a spouse/partner; a close friend/family member who died or had a serious illness; major problems with money; a divorce or breakup with a spouse/partner; a family member who had a divorce/breakup; major conflict with children or grandchildren; major accidents, disasters or assaults; loss of a job or retiring; being physically abused; being verbally abused; and the death of a pet. One point is given for any occurrence of the 11 events. Scores range from 0 to 11, with high scores indicating a greater number of life events.

History of comorbidities

Outcomes and major illnesses, experienced by the participants, are collected annually in WHI. The participants’ history of heart disease, stroke, cancer, diabetes mellitus, and hip fracture after age 55 reported at any point in WHI was included in the analysis.

Symptom burden

A 17-item symptom checklist was used to assess the impact of physical and psychological symptoms on participant well-being. This is an abbreviated version of the scale used in the past WHI assessments, as well as other studies (Greendale et al., 1998). Participants were asked to respond to the presence or absence of the symptom, and if present, whether the symptom was mild, moderate, or severe. Symptoms included night sweats, general aches and pains, breast tenderness, hot flashes, mood swings, irritability, feeling tired, forgetfulness, skin dryness or scaling, headaches or migraines, difficulty concentrating, joint pain or stiffness, uncontrolled leaking of urine, uncontrolled leaking of feces, vaginal itching, and vaginal dryness. Scores from each item were summed and then divided by the total number of items. Higher scores indicate higher symptom severity.

Lifestyle variables assessed

Smoking status (no/yes, and how many cigarettes were smoked per day if they were a smoker); body mass index (BMI; weight [kg]/divided by height [m²], categorized as ≤24.9 [normal weight], 25.0–29.99 [overweight], or ≥30.0 [obese]); and total minutes of recreational physical activity per week, which included walking, and mild, moderate, and strenuous physical activity.

Healthy Eating Index 2015

Healthy Eating Index 2015 (HEI-2015) scores were calculated from the Dietary Health Questionnaire completed by the participants. Total HEI scores range from 0 to 100, with 100 being in complete compliance with the Dietary Guidelines for Americans (U.S. Department of Agriculture, 2022). Generally, HEI scores >80 indicate a “good” diet, scores ranging from 51 to 80 reflect a diet that “needs improvement,” and HEI scores <51 imply a poor diet (U.S. Department of Agriculture, 2022).

Statistical Analysis

Descriptive statistics were summarized by race, ethnicity, and NSES using means and standard deviations for continuous variables and frequencies and percentages for categorical variables. These characteristics were compared among race, ethnicity, and NSES groups using Chi-square or Fisher’s exact tests for categorical variables and analysis of variance (ANOVA) for continuous variables. To handle missing data, five multiple imputed data sets were created with an arbitrary missing pattern (SAS PROC MI). Discrimination function methods were used to impute missing values in categorical variables and regression methods in continuous variables. For each imputation, multivariate linear regression models were used to examine the association of demographic, health, and psychosocial risk factors with resilience for the entire sample and stratified by racial (White, Black, and Asian) and ethnic group (Hispanic and non-Hispanic) and NSES. Results of the imputations were combined and then generated valid statistical inferences (SAS PROC MIANALYZE). Beta values and their 95% confidence intervals (CIs) were reported. All analyses were conducted using SAS v9.4 (SAS Institute, Cary, NC).

Results

Sample Characteristics

Tables 1 and 2 show the demographic, health, and psychosocial characteristics of WHI participants aged 80 years and older by race and ethnicity and NSES, respectively. Participants (n = 29,367, median age = 84.30 ± 3.44) were White (91.4%), Black (3.7%), Hispanic (1.9%), and Asian (1.7%) women. Nearly two-thirds of the women had comorbidities (66%), and 63.2% were overweight or obese. The majority of participants reported their health to be “very good” (36.3%) or “good” (40.4%). Average perceived stress scores were 4.43 ± 3.02 (out of 16) and the average resilience score was 3.96 ± 0.82 (out of 5). A majority of the participants lived in the area with moderate NSES (48.6%), followed by high (24.4%) and low NSES (24.3%).

Table 1.

Demographic, Health, and Psychosocial Characteristics of WHI Participants Aged 80 and Older by Race and Ethnicity

Variable	White, n (%), n = 26,843	Black, n (%), n = 1,091	Asian, n (%), n = 511	Hispanic, n (%), n = 586	Overalla, n (%), n = 29,367
Age, mean (SD)***	84.34 (3.44)	83.76 (3.37)	83.82 (3.33)	83.73 (3.20)	84.30 (3.44)
Education***
<High school	733 (2.73)	84 (7.7)	15 (2.94)	61 (10.41)	912 (3.11)
High school graduate	4,586 (17.08)	151 (13.84)	82 (16.05)	104 (17.75)	4,983 (16.97)
Some college/associate’s degree	9,909 (36.91)	366 (33.55)	165 (32.29)	228 (38.91)	10,801 (36.78)
4-year college	3,255 (12.13)	96 (8.8)	71 (13.89)	50 (8.53)	3,501 (11.92)
>4-year college degree	8,237 (30.69)	384 (35.2)	176 (34.44)	136 (23.21)	9,028 (30.74)
Marital status***
Married/living as married	17,164 (63.94)	445 (40.79)	348 (68.10)	328 (55.97)	18,471 (62.90)
Divorced/separated/widowed	8,637 (32.18)	604 (55.36)	145 (28.38)	230 (39.25)	9,752 (33.21)
Single/never married	960 (3.58)	36 (3.30)	16 (3.13)	23 (3.92)	1,049 (3.57)
Neighborhood socioeconomic status***
Low	5,988 (22.31)	707 (64.80)	98 (19.18)	14 (2.39)	7,121 (24.25)
Moderate	13,389 (49.88)	255 (23.37)	227 (44.42)	215 (36.69)	14,269 (48.59)
High	6,723 (25.05)	84 (7.70)	184 (36.01)	256 (43.69)	7,163 (24.39)
Living situation***
With someone (spouse, children)	286 (48.81)	451 (41.34)	299 (58.51)	286 (48.81)	12,743 (43.39)
Alone	246 (41.98)	562 (51.51)	184 (36.01)	246 (41.98)	14382 (48.97)
Social support, mean (SD)*	36.71 (7.89)	36.9 (7.44)	35.63 (8.51)	35.54 (8.87)	36.67 (7.91)
Self-rated health***
Excellent	2,108 (7.85)	35 (3.21)	26 (5.09)	30 (5.12)	2,220 (7.56)
Very good	9,891 (36.85)	282 (25.85)	181 (35.42)	191 (32.59)	10,647 (36.25)
Good	10,717 (39.92)	534 (48.95)	219 (42.86)	243 (41.47)	11,852 (40.36)
Fair	3,072 (11.44)	190 (17.42)	67 (13.11)	94 (16.04)	3,484 (11.86)
Poor	301 (1.12)	20 (1.83)	5 (0.98)	11 (1.88)	340 (1.16)
Comorbidity (% yes)***	17,858 (66.53)	638 (58.48)	268 (52.45)	398 (67.92)	19,391 (66.03)
Symptoms, mean (SD)***	0.57 (0.32)	0.58 (0.34)	0.46 (0.33)	0.60 (0.36)	0.57 (0.33)
Physical functioning, mean (SD)***	57.80 (27.13)	55.45 (27.65)	64.95 (26.63)	59.45 (27.16)	57.80 (27.19)
Assistance with activities, mean (SD)***	6.43 (1.35)	6.39 (1.09)	6.29 (1.02)	6.44 (1.38)	6.43 (1.34)
BMI; mean [SD]***	27.26 (5.16)	30.1 (5.69)	24.21 (3.66)	27.85 (4.94)	27.34 (5.21)
Likelihood of depression, mean (SD)***	0.02 (0.09)	0.02 (0.09)	0.02 (0.09)	0.05 (0.15)	0.02 (0.10)
Perceived stress, mean (SD)***	4.41 (3.06)	4.49 (3.02)	4.53 (3.01)	4.94 (3.07)	4.43 (3.02)
Major life stressors, mean (SD)	1.27 (1.20)	1.50 (1.31)	1.16 (1.16)	1.49 (1.32)	1.29 (1.21)
Religion gives strength and comfort***
None	7,991 (29.77)	15 (1.37)	27 (5.28)	15 (2.56)	1,485 (5.06)
A little	2,546 (9.48)	42 (3.85)	76 (14.87)	37 (6.31)	2,727 (9.29)
A great deal	1,419 (5.29)	554 (50.78)	101 (19.77)	214 (36.52)	8,990 (30.61)
Currently a smoker (% yes)**	359 (1.34)	24 (2.20)	2 (0.39)	10 (1.71)	399 (1.36)
Alcohol consumption***
Never	9,002 (33.54)	608 (55.73)	355 (69.47)	260 (44.37)	10,377 (35.34)
Less than 1 per week	8,252 (30.74)	331 (30.34)	91 (17.81)	190 (32.42)	8,965 (30.53)
1 or 2 times per week	2,921 (10.88)	68 (6.23)	21 (4.11)	46 (7.85)	3,082 (10.49)
3 or 4 times per week	1,831 (6.82)	35 (3.21)	12 (2.35)	33 (5.63)	1,927 (6.56)
5 or 6 times per week	1,876 (6.99)	14 (1.28)	12 (2.35)	23 (3.92)	1,944 (6.62)
Every day	2,420 (9.02)	17 (1.56)	9 (1.76)	21 (3.58)	2,482 (8.45)
Total HEI-2015 score, mean (SD)***	68.99 (9.84)	67.94 (10.26)	69.64 (9.29)	67.33 (10.04)	68.92 (9.86)
Brief resilience scale score, mean (SD)	3.96 (0.81)	3.98 (0.89)	3.92 (0.80)	3.88 (0.87)	3.96 (0.82)

Notes: Some percentages will not add up to 100 due to missingness. BMI = body mass index; HEI-2015 = Healthy eating index 2015; SD = standard deviation; WHI = Women’s Health Index.

^aOverall includes women with reported races other than White, Black, Asian, or Hispanic (American Indian/Alaska Native, n = 31; Native Hawaiian/Other Pacific Islander, n = 9; Other race category as reported by participants, n = 296).

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

Table 2.

Demographic, Health, and Psychosocial Characteristics of WHI Participants Aged 80 and Older by Neighborhood Socioeconomic Status

Variable	Low neighborhood SES, n = 7,121	Moderate neighborhood SES, n = 14,269	High neighborhood SES, n = 7,163	Overall, n = 28,553
Age, mean (SD)	84.27 (3.42)	84.3 (3.44)	84.35 (3.46)	84.3 (3.44)
Race and ethnicity***
White	5,988 (84.09)	13,389 (93.83)	6,723 (93.86)	26,100 (91.41)
Black	707 (9.93)	255 (1.79)	84 (1.17)	1,046 (3.66)
Asian	98 (1.38)	227 (1.59)	184 (2.57)	509 (1.78)
Hispanic	215 (3.02)	256 (1.79)	101 (1.41)	572 (2.00)
Education***
< High school	397 (5.58)	413 (2.89)	67 (0.94)	877 (3.07)
High school graduate	1,499 (21.05)	2,632 (18.45)	675 (9.42)	4,806 (16.83)
Some college/associate’s degree	2,871 (40.32)	5,344 (37.45)	2,279 (31.82)	10,494 (36.75)
4-year college	626 (8.79)	1,668 (11.69)	1,137 (15.87)	3,431 (12.02)
>4-year college degree	1,698 (23.84)	4,138 (29.00)	2,971 (41.48)	8,807 (30.84)
Marital status***
Married/living as married	3,986 (55.98)	9,056 (63.47)	4,879 (68.11)	17,921 (62.76)
Divorced/separated/widowed	2,800 (39.32)	4,648 (32.57)	2,066 (28.84)	9,514 (33.32)
Single/never married	312 (4.38)	518 (3.63)	195 (2.72)	1,025 (3.59)
Living situation***
Alone	3,675 (51.61)	6,940 (48.64)	3,350 (46.77)	13,965 (48.91)
With someone (spouse, children)	2,935 (41.22)	6,192 (43.39)	3,269 (45.64)	12,396 (43.41)
Social support, mean (SD)*	36.42 (8.00)	36.72 (7.89)	36.78 (7.85)	36.66 (7.91)
Self-rated health***
Excellent	403 (5.66)	1,065 (7.46)	699 (9.76)	2,167 (7.59)
Very good	2,335 (32.79)	5,251 (36.8)	2,778 (38.78)	10,364 (36.3)
Good	3,063 (43.01)	5,717 (40.07)	2,727 (38.07)	11,507 (40.3)
Fair	1,007 (14.14)	1,659 (11.63)	718 (10.02)	3,384 (11.85)
Poor	98 (1.38)	169 (1.18)	64 (0.89)	331 (1.16)
Comorbidity (% yes)***	4,847 (68.07)	9,432 (66.10)	4,590 (64.08)	18,869 (66.08)
Symptoms, mean (SD)***	0.59 (0.34)	0.57 (0.33)	0.54 (0.31)	0.57 (0.33)
Physical functioning, mean (SD)***	54.69 (27.47)	57.67 (27.09)	61.22 (26.7)	57.83 (27.18)
Assistance with activities, mean (SD)***	6.46 (1.34)	6.43 (1.34)	6.39 (1.32)	6.43 (1.33)
BMI; mean [SD]***	28.17 (5.45)	27.32 (5.16)	26.5 (4.96)	27.32 (5.22)
Likelihood of depression, mean (SD)*	0.03 (0.11)	0.02 (0.10)	0.02 (0.10)	0.02 (0.10)
Perceived stress, mean (SD)***	4.61 (3.02)	4.43 (3.01)	4.24 (3.00)	4.42 (3.02)
Major life stressors, mean (SD)***	1.34 (1.25)	1.28 (1.21)	1.23 (1.16)	1.28 (1.21)
Religion gives strength and comfort***
None	245 (3.44)	665 (4.66)	550 (7.68)	1,460 (5.11)
A little	541 (7.6)	1,289 (9.03)	839 (11.71)	2,669 (9.35)
A great deal	2,612 (36.68)	4,371 (30.63)	1,709 (23.86)	8,692 (30.44)
Currently a smoker (% yes)*	114 (1.60)	195 (1.37)	77 (1.07)	386 (1.35)
Alcohol consumption***
Never	3,226 (45.3)	4,931 (34.56)	1,870 (26.11)	10,027 (35.12)
Less than 1 per week	2,079 (29.2)	4,497 (31.52)	2,157 (30.11)	8,733 (30.59)
1 or 2 times per week	622 (8.73)	1,512 (10.6)	878 (12.26)	3,012 (10.55)
3 or 4 times per week	337 (4.73)	956 (6.7)	581 (8.11)	1,874 (6.56)
5 or 6 times per week	312 (4.38)	919 (6.44)	669 (9.34)	1,900 (6.65)
Every day	395 (5.55)	1160 (8.13)	876 (12.23)	2,431 (8.51)
Total HEI-2015 score, mean (SD)***	67.82 (10.18)	68.95 (9.75)	69.97 (9.63)	68.94 (9.85)
Brief resilience scale score, mean (SD)***	3.94 (0.83)	3.95 (0.82)	4.00 (0.81)	3.96 (0.82)

Notes: Some percentages will not add up to 100 due to missingness. BMI = body mass index; HEI-2015 = Healthy eating index 2015; SD = standard deviation; WHI = Women’s Health Initiative.

^aRace groups include American Indian/Alaska Native, n = 31; Native Hawaiian/Other Pacific Islander, n = 9; Other race category as reported by participants, n = 296.

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

There were no significant differences by race on mean resiliency scores (p = .06). Mean resilience was higher among women with higher NSES (low NSES = 3.94 ± 0.83, moderate NSES = 3.95 ± 0.82, high NSES = 4.00 ± 0.81; p < .001). There were significant differences for all other demographic, health, and psychosocial characteristics by race, ethnicity, and NSES with the exception of needing assistance with activities and age, respectively.

Regression Results for Whole Sample

Table 3 shows multivariate-adjusted correlates of resilience in the overall sample. Analyses found resilience was inversely associated with older baseline age (β = −0.004, 95% CI: −0.006, 0.001) and positively associated with higher levels of education. Resilience was positively associated with self-rated health (β comparing excellent to poor = −0.338, 95% CI: −0.433, −0.244), social support (β = 0.012, 95% CI: =0.011, 0.013), lower perceived stress (β = −0.092, 95% CI: −0.095, −0.089), and living alone (β = 0.035, 95% CI: 0.016, 0.053). Counterintuitively, greater spirituality was inversely associated with resilience. Smoking (β = 0.109, 95% CI: 0.037, 0.180) and higher BMI (β = 0.008, 95% CI: 0.006, 0.009) were positively associated with resilience, while higher alcohol intake (β comparing 5–6 drinks per week vs never = −0.046, 95% CI: −0.082, −0.010), higher likelihood of depressive symptoms (β = −0.672, 95% CI: −0.765, −0.579), and higher symptom burden (β = −0.328, 95% CI: −0.360, −0.297) were inversely related to resilience.

Table 3.

Multivariate Correlates of Resilience Among of WHI Participants Aged 80 and Older by Race and Ethnicity

Variable	White (n = 26,843), β (95% CI)	Black (n = 1,091), β (95% CI)	Asian (n = 511), β (95% CI)	Hispanic (n = 586), β (95% CI)	Whole samplea, β (95% CI)
Age, mean (SD)	−0.004 (−0.006, −0.001)**	0.000 (−0.015, 0.015)	−0.008 (−0.028, 0.011)	−0.005 (−0.026, 0.016)	−0.004 (−0.006, −0.001)**
Education
< High school	Ref.	Ref.	Ref.	Ref.	Ref.
High school graduate	0.035 (−0.019, 0.090)	0.173 (−0.040, 0.385)	0.201 (−0.192, 0.594)	0.009 (−0.238, 0.256)	0.047 (−0.004, 0.098)
Some college/associate’s degree	0.079 (0.026, 0.131)**	0.130 (−0.062, 0.323)	0.153 (−0.225, 0.532)	0.112 (−0.238, 0.341)	0.090 (0.041, 0.138)***
4-year college	0.061 (0.004, 0.118)*	−0.042 (−0.277, 0.193)	0.204 (−0.196, 0.604)	0.146 (−0.149, 0.441)	0.069 (0.016, 0.122)*
>4-year college degree	0.099 (0.045, 0.153)***	0.182 (−0.012, 0.377)	0.258 (−0.127, 0.642)	0.212 (−0.036, 0.461)	0.112 (0.063, 0.162)***
Marital status
Married/living as married	−0.036 (−0.082, 0.011)	0.129 (−0.143, 0.402)	−0.027 (−0.419, 0.364)	0.042 (−0.295, 0.378)	−0.028 (−0.073, 0.017)
Divorced/separated/widowed	−0.003 (−0.050, 0.044)	0.164 (−0.100, 0.429)	0.006 (−0.385, 0.397)	0.059 (−0.276, 0.394)	0.005 (−0.040, 0.050)
Single/never married	Ref.	Ref.	Ref.	Ref.	Ref.
Neighborhood socioeconomic status
Low	Ref.	Ref.	Ref.	Ref.	Ref.
Moderate	−0.026 (−0.048, −0.005)*	−0.038 (−0.154, 0.078)	−0.041 (−0.212, 0.130)	−0.039 (−0.181, 0.103)	−0.023 (−0.044, −0.003)*
High	−0.011 (−0.037, 0.015)	0.088 (−0.099, 0.402)	0.051 (−0.128, 0.231)	−0.143 (−0.334, 0.047)	−0.009 (−0.034, 0.016)
Living alone	0.035 (0.016, 0.055)***	0.054 (−0.052, 0.161)	0.063 (−0.081, 0.207)	−0.034 (−0.188, 0.119)	0.05 (0.016, 0.053)***
Social support	0.012 (0.011, 0.013)***	0.012 (0.005, 0.019)***	0.012 (0.003, 0.020)**	0.006 (−0.002, 0.014)	0.012 (0.011, 0.013)***
Self-rated health
Excellent	Ref.	Ref.	Ref.	Ref.	Ref.
Very good	−0.110 (−0.436, −0.238)***	0.128 (−0.165, 0.421)	−0.110 (−0.410, 0.190)	−0.166 (−0.469, 0.136)	−0.112 (−0.145, −0.079)***
Good	0.145 (−0.310, −0.215)***	−0.041 (−0.333, 0.251)	−0.161 (−0.467, 0.145)	−0.304 (−0.612, 0.003)	−0.218 (−0.253, −0.183)***
Fair	0.261 (−0.310, −0.215)***	−0.075 (−0.402, 0.252)	−0.101 (−0.462, 0.261)	−0.420 (−0.791, −0.049)*	−0.267 (−0.313, −0.220)***
Poor	0.782 (−0.436, −0.238)***	−0.188 (−0.721, 0.345)	0.001 (−0.780, 0.782)	−0.576 (−1.200, 0.049)	−0.338 (−0.433, −0.244)***
Comorbidity	−0.005 (−0.025, 0.014)	−0.035 (−0.140, 0.070)	−0.024 (−0.081, 0.207)	−0.086 (−0.240, 0.068)	−0.009 (−0.027, 0.010)
Symptoms	−0.340 (−0.373, −0.306)***	−0.161 (−0.332, 0.009)	−0.381 (−0.630, −0.131)**	−0.206 (−0.445, 0.033)	−0.328 (−0.360, −0.297)***
Physical functioning	0.000 (0.000, 0.001)	0.000 (−0.002, 0.003)	−0.001 (−0.004, 0.002)	0.000 (−0.003, 0.003)	0.000 (0.000, 0.001)
Assistance with activities	−0.003 (−0.010, 0.004)	0.008 (−0.041, 0.058)	−0.011 (−0.077, 0.055)	0.017 (−0.040, 0.074)	−0.003 (−0.010, 0.004)
BMI	0.008 (0.006, 0.010)***	0.004 (−0.005, 0.013)	0.010 (−0.007, 0.028)	0.011 (−0.003, 0.025)	0.008 (0.006, 0.009)***
Depression	−0.670 (−0.768, −0.573)***	−0.623 (−1.202, −0.043)*	−0.503 (−1.253, 0.246)	−0.572 (−1.058, −0.086)*	−0.672 (−0.765, −0.579)***
Perceived stress	−0.090 (−0.093, −0.087)***	−0.111 (−0.128, −0.094)***	−0.100 (−0.124, −0.075)***	−0.095 (−0.118, −0.071)***	−0.092 (−0.095, −0.089)***
Major life stressors	−0.005 (−0.012, 0.003)	−0.036 (−0.077, 0.004)	0.016 (−0.049, 0.081)	−0.023 (−0.075, 0.029)	−0.005 (−0.012, 0.002)
Religion gives strength and comfort
None	Ref.	Ref.	Ref.	Ref.	Ref.
A little	−0.130 (−0.185, −0.075)***	−0.503 (−1.046, 0.040)	−0.114 (−0.469, 0.241)	0.104 (−0.446, 0.654)	−0.132 (−0.186, −0.079)***
A great deal	−0.056 (−0.113, 0.001)	−0.233 (−0.731, 0.264)	−0.335 (−0.734, 0.065)	0.058 (−0.488, 0.604)	−0.060 (−0.115, −0.005)*
Current smoker	1.551 (0.042, 0.191)**	−0.115 (−0.449, 0.219)	0.547 (−0.458, 1.551)	0.259 (−0.223, 0.742)	0.109 (0.037, 0.180)***
Alcohol consumption
Never	Ref.	Ref.	Ref.	Ref.	Ref.
Less than 1 per week	0.002 (−0.019, 0.023)	0.004 (−0.104, 0.112)	−0.055 (−0.218, 0.108)	−0.077 (−0.230, 0.076)	0.000 (−0.021, 0.020)
1 or 2 times per week	−0.020 (−0.050, 0.010)	0.020 (−0.179, 0.219)	−0.129 (−0.435, 0.177)	0.045 (−0.207, 0.297)	−0.019 (−0.048, 0.009)
3 or 4 times per week	−0.026 (−0.063, 0.011)	−0.125 (−0.395, 0.145)	0.030 (−0.373, 0.432)	0.028 (−0.260, 0.315)	−0.027 (−0.063, 0.009)
5 or 6 times per week	−0.041 (−0.077, −0.005)*	−0.393 (−0.815, 0.029)	−0.188 (−0.614, 0.239)	0.025 (−0.314, 0.364)	−0.046 (−0.082, −0.010)*
Every day	−0.023 (−0.055, 0.008)	−0.056 (−0.447, 0.335)	0.441 (−0.033, 0.915)	0.058 (−0.290, 0.406)	−0.023 (−0.054, 0.009)
HEI-2015 total score	0.001 (0.000, 0.002)	0.000 (−0.007, 0.006)	0.002 (−0.005, 0.009)	−0.005 (−0.013, 0.002)	0.000 (−0.001, 0.002)

Notes: BMI = body mass index; CI = confidence interval; HEI-2015 = Healthy eating index 2015; Ref. = reference; SD = standard deviation; WHI = Women’s Health Initiative.

^aOverall includes women with reported races other than White, Black, Asian, or Hispanic (American Indian/Alaska Native, n = 31; Native Hawaiian/Other Pacific Islander, n = 9; Other race category as reported by participants, n = 296).

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

Regression Results by Race and Ethnic Group

Table 3 shows the multivariate-adjusted correlates of resilience by race and ethnicity. Analyses found resilience was positively associated with social support among older White (β = 0.012, 95% CI: 0.011, 0.013), Black (β = 0.012, 95% CI: 0.005, 0.019), and Asian women (β = 0.012, 95% CI: 0.003, 0.020). Higher education and self-rated health were positively associated with resilience among White women.

Resilience was inversely associated with perceived stress across all race and ethnic groups. The likelihood of depression was also inversely associated with resilience among White (β = −0.670, 95% CI: −0.768, −0.573), Black (β = −0.623, 95% CI: −1.202, −0.043), and older Hispanic women (β = −0.572, 95% CI: −1.058, −0.086). Resilience was inversely associated with symptom burden among White (β = −0.340, 95% CI: −0.373, −0.306) and older Asian women (β = −0.381, 95% CI: −0.630, −0.131). Last, NSES was inversely related to resilience among White women (β = −0.026, 95% CI: −0.048, −0.005).

Regression Results by NSES

Table 4 shows the multivariate-adjusted correlates of resilience by NSES. Resilience was positively associated with higher education (>4-year college degree) among low (β = 0.107, 95% CI: 0.027, 0.186), moderate (β = 0.102, 95% CI: 0.031, 0.173), and high NSES (β = 0.191, 95% CI: 0.010, 0.373). Resilience was positively associated with self-rated health among all NSES categories. Social support and higher BMI were also positively associated with resilience across all NSES categories. Among women with moderate and high NSES, resilience was positively associated with some spirituality (β comparing none to a little = −0.104, 95% CI: −0.182, −0.025; −0.183, 95% CI: −0.247, −0.093, respectively).

Table 4.

Multivariate Correlates of Resilience Among of WHI Participants Aged 80 and Older by Neighborhood Socioeconomic Status

Variable	Low neighborhood SES, n = 7,327, β (95% CI)	Moderate neighbourhood SES, n = 14,681, β (95% CI)	High neighbourhood SES, n = 7,359, β (95% CI)
Age	−0.002 (−0.007, 0.003)	−0.004 (−0.008, −0.001)*	−0.005 (−0.010, 0.000)
Race and ethnicity
White	Ref.	Ref.	Ref.
Black	0.008 (−0.048, 0.065)	0.017 (−0.072, 0.105)	0.082 (−0.067, 0.232)
Asian	−0.032 (−0.175, 0.110)	−0.073 (−0.166, 0.020)	−0.003 (−0.109, 0.103)
Hispanic	0.053 (−0.044, 0.150)	0.037 (−0.049, 0.123)	−0.096 (−0.234, 0.042)
Education
<High school	Ref.	Ref.	Ref.
High school graduate	0.051 (−0.028, 0.130)	0.048 (−0.024, 0.120)	0.089 (−0.100, 0.278)
Some college/associate’s degree	0.064 (−0.011, 0.139)	0.100 (0.030, 0.169)**	0.157 (−0.026, 0.339)
4-year college	0.065 (−0.024, 0.155)	0.084 (0.009, 0.159)*	0.114 (−0.070, 0.298)
>4-year college degree	0.107 (0.027, 0.186)**	0.102 (0.031, 0.173)**	0.191 (0.010, 0.373)*
Marital status
Married/living as married	−0.003 (−0.086, 0.081)	−0.060 (−0.124, 0.004)	0.011 (−0.088, 0.111)
Divorced/separated/widowed	0.022 (−0.062, 0.106)	−0.027 (−0.091, 0.038)	0.054 (−0.047, 0.155)
Single/never married	Ref.	Ref.	Ref.
Living alone	0.018 (−0.021, 0.056)	−0.003 (0.016, 0.074)	−0.009 (−0.002, 0.070)
Social support	0.011 (0.009, 0.013)***	0.011 (0.010, 0.013)***	0.013 (0.010, 0.015)***
Self-rated health
Excellent	Ref.	Ref.	Ref.
Very good	−0.120 (−0.195, −0.045)**	−0.097 (−0.144, −0.051)***	−0.128 (−0.188, −0.069)***
Good	−0.226 (−0.306, −0.146)***	−0.215 (−0.264, −0.165)***	−0.214 (−0.278, −0.150)***
Fair	−0.290 (−0.384, −0.196)***	−0.263 (−0.329, −0.197)***	−0.238 (−0.327, −0.149)***
Poor	−0.376 (−0.565, −0.186)***	−0.388 (−0.526, −0.250)***	−0.164 (−0.39, 0.041)
Comorbidity	−0.018 (−0.056, 0.020)	−0.003 (−0.029, 0.023)	−0.009 (−0.046, 0.027)
Symptom composite score	−0.270 (−0.332, −0.208)***	−0.357 (−0.401, −0.314)***	−0.330 (−0.410, −0.249)***
Physical functioning	0.000 (0.000, 0.001)	0.000 (−0.001, 0.000)	0.001 (0.000, 0.001)
Assistance with activities	0.000 (−0.016, 0.016)	−0.008 (−0.018, 0.002)	0.004 (−0.010, 0.019)
BMI	0.008 (0.004, 0.011)***	0.007 (0.005, 0.010)***	0.008 (0.005, 0.012)***
Depression	−0.610 (−0.784, −0.436)***	−0.696 (−0.830, −0.562)***	−0.700 (−0.896, −0.504)***
Perceived stress	−0.096 (−0.102, −0.090)***	−0.089 (−0.093, −0.085)***	−0.093 (−0.099, −0.087)***
Major life stressors	−0.012 (−0.026, 0.003)	−0.006 (−0.017, 0.005)	0.006 (−0.010, 0.021)
Religion gives strength and comfort
None	Ref.	Ref.	Ref.
A little	−0.106 (−0.233, 0.020)	−0.104 (−0.182, −0.025)**	−0.183 (−0.274, −0.093)***
A great deal	−0.012 (−0.140, 0.115)	−0.032 (−0.112, 0.049)	−0.183 (−0.195, 0.008)
Current smoker	0.016 (−0.122, 0.153)	0.172 (0.073, 0.272)***	0.080 (−0.078, 0.239)
Alcohol consumption
Never	Ref.	Ref.	Ref.
Less than 1 per week	0.004 (−0.035, 0.043)	−0.001 (−0.030, 0.027)	−0.006 (−0.050, 0.038)
1 or 2 times per week	0.039 (−0.022, 0.100)	−0.024 (−0.065, 0.016)	−0.056 (−0.113, 0.002)
3 or 4 times per week	−0.046 (−0.128, 0.036)	−0.041 (−0.090, 0.008)	0.008 (−0.058, 0.073)
5 or 6 times per week	−0.077 (−0.162, 0.009)	−0.059 (−0.111, −0.007)*	−0.020 (−0.083, 0.044)
Every day	0.010 (−0.035, 0.043)	−0.012 (−0.058, 0.034)	−0.056 (−0.050, 0.038)
HEI-2015 total score	−0.001 (−0.003, 0.001)	0.000 (−0.001, 0.002)	0.001 (0.000, 0.003)

Notes: BMI = body mass index; CI = confidence interval; HEI-2015 = Healthy eating index 2015; Ref. Reference; SES = socioeconomic status; WHI = Women’s Health Initiative.

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

Resilience was negatively associated with perceived stress, greater likelihood of depression, and higher symptom burden across all NSES categories. Among women with moderate NSES, age (β = −0.004, 95% CI: −0.008, −0.001), and higher alcohol intake (β comparing 5–6 drinks per week vs never = −0.059, 95% CI: −0.111, −0.007) were inversely related with resilience. Among this group, living alone (β = 0.045, 95% CI: 0.016, 0.074) and smoking (β = 0.051, 95% CI: 0.073, 0.272) were positively associated with resilience.

Discussion

This study examined factors associated with higher resilience among participants in the WHI aged 80 and older by race and ethnicity (e.g., Asian, Black, Hispanic, and White) and NSES. Results indicated that resilience was similar across all racial groups within this cohort. One potential reason for this similarity may be the use of a modified resilience scale, which does not examine resilience using the suggested cutoffs for high, medium, and low resilience. Another reason may be an observed ceiling effect, supported by recent resilience theoretical and empirical evidence (Staudinger & Greve, 2015; Woods et al., 2016), in which medium–high resilience is observed among older women across all racial groups.

Significant differences in mean resiliency scores by NSES were observed between those with low NSES (3.94 ± 0.83, out of 5) and high NSES (4.00 ± 0.81), corresponding with previous literature showing that current lower NSES is associated with lower resilience among older adults (Lau et al., 2018; Phillips et al., 2016). However, the statistical difference observed between the two groups may be driven by the study sample size, rather than a practical difference in resilience. Impoverished neighborhoods may have a culture of ongoing mutual support, which may be reflected in the high mean resilience scores reported by the women with low NSES (Kok et al., 2018; Lee et al., 2022). This study also found resilience was associated with physical, psychosocial, and behavioral factors, which is consistent with studies conducted in study populations that include individuals across the life span (MacLeod et al., 2016; Taylor & Carr, 2020). However, no study had examined correlates of resilience by race, ethnicity, and NSES among older women. Given the increasing racial and ethnic diversity among older adults in the United States and widening wealth inequality (Federal Interagency Forum on Aging-Related Statistics, 2020), this research is warranted.

Higher education and lower perceived stress were consistently associated with higher resiliency among older women, across race, ethnicity, and NSES groups. Higher education, a personal resource, has been positively associated to resilience, coping skills, healthy lifestyles, and health later in life (MacLeod et al., 2016; Taylor & Carr, 2020). Lower perceived stress was associated with higher resilience among the entire sample and by race, ethnicity, and NSES. Similarly, da Silva-Sauer et al. (2021) found a strong inverse association between perceived stress and resilience among community-dwelling older adults. Given the nonsignificant association between resilience and major stressful life events, these results suggest that resilience does not protect older adults from feeling stressed, but helps them cope with stress more satisfactorily.

Assessing resilience in different racial and ethnic populations is important, as group members often share similar experiences with structural racism, which in turn, affect internal (psychological) and external (social and physical) resources (Ungar, 2011). Groups experiencing structural racism experience greater stressors and restricted access to material and health resources (Williams et al., 2019). There were some differing correlates of resiliency by the racial group in this study. Increased likelihood of depression was a significant correlate of lower resilience for older women, except for Asian women. This finding adds to the limited research on this subject in the older adult population. A systematic review and meta-analysis found a significant association between greater resilience and lower depressive symptomatology (Wermelinger Ávila et al., 2017). However, the review included studies on both men and women. Vahia et al. (2010) investigated 1,979 women and found those not depressed had higher levels of resilience than those depressed. This association was not observed among older Asian women in this study, which may be related to our small sample size of Asian women. However, Sorkin et al. (2009) found that Asian older adults were less likely to report needing help dealing with their mental health problems and accessing mental health services than non-Hispanic White older adults. More research is needed in this area including consideration of English language proficiency and acculturation, which can affect the measurement of reported depression.

Social support was another consistent correlate of resilience among the White, Black, and Asian women in this study, supporting prior literature (Gooding et al., 2012; MacLeod et al., 2016; Southwick et al., 2016). This finding was not observed among Hispanic women. Potential reasons for this finding can include the nature of older Hispanic women’s networks may differ from other minoritized racial and ethnic groups, as well as our sample size of Hispanic women being smaller, which can hinder statistical power.

Close social connections may promote psychological resilience. One key mechanism involved in “being resilient” likely relates to interpersonal (social support and social engagement) resources. Individuals do not become or remain resilient completely on their own, and strong social bonds facilitate adaptation in the face of adversity (Southwick et al., 2016). Social support and social connectivity are essential for older adults’ quality of life and physical and mental health (Phillips et al., 2016). Future health interventions to improve resilience among older adults should highlight the role of social support, yet tailor them to the individual’s needs, preferences, and contexts.

Physical symptoms and self-rated health were significantly associated with resilience among non-Hispanic White and Asian women. Historically, and in this study, older adults identifying as Black experience structural racism and report lower self-rated health, compared with White older adults (Bell et al., 2018). Similarly, symptom burden was higher among Black and Hispanic women compared to their White counterparts. However, symptom burden and self-rated health were not significantly associated with resilience among these women. This suggests the need for further research on identifying the sources of resilience, despite worse health status, among older women of minoritized racial and ethnic groups.

Although the majority of correlates of resilience were similar across NSES, some correlates of resilience were unique to the women with moderate NSES. The correlates of living alone, smoking, alcohol use, and spirituality were significantly and positively associated with resilience among those women with moderate NSES. More information is needed regarding the participants’ preferences for their living situations and the length of time living alone. Although smoking is associated with negative health outcomes and mortality in the general population and older adults (U.S. Department of Health and Human Services, 2014), our study suggests that smoking can also support the resilience of older women, which is supported by a limited number of studies (van Schoor et al., 2022).

Although the influence of neighborhood-level socioeconomic status (SES) on health is smaller than individual-level SES, it has been demonstrated that systematic health inequalities exist between neighborhoods differing in SES (Williams et al., 2019; Yao & Robert, 2008). Neighborhoods of low SES are disproportionally communities of color created by racist housing policies and practices (e.g., redlining, predatory lending policies, racially restrictive covenants, exclusionary zoning, and blockbusting) that have fundamentally shaped housing tenure, the built environment, and residents’ health (Lynch et al., 2021). Previous studies have identified associations between NSES and health behaviors (e.g., exercise, healthy eating, and weight gain) in older adults (Rachele et al., 2017; Stringhini et al., 2010). Future research should consider how older women move across neighborhood contexts and how neighborhood contexts change over time to better understand this association (Yao & Robert, 2008).

The wide range of correlates identified in this study suggests that further examination of resilience among women over 80 is needed. Future studies would benefit from employing qualitative methods to explore the context of the psychosocial, health, and demographic factors associated with psychological resilience. Information about one’s culture and neighborhood would be essential to continue this research to discern the similarities and differences of resilience within diverse older populations. In addition, further research is needed to identify how these correlates of psychological resilience translate to a multisystem of resilience, providing increased buffering from life stressors and shaping the rate of biological aging. Together, this multisystem of resilience can support successful aging in terms of both quality and quantity of years among older women.

This study has a number of notable strengths including a large sample size, a focus on understudied women, an exploration of differences in resilience by race, ethnicity, and NSES, and a comprehensive assessment of demographic, health, and psychosocial factors. A study limitation is survivor bias, such that we included WHI participants who continued their long-term participation in the WHI and were alive during survey periods. The main outcome of resilience is based on a modified version of the Brief Resilience Scale, using three items from the original six-item scale, which may not recapitulate the validated scale. However, the concept of “resilience” is incommensurable, and lacks standard measurement (Olsson et al., 2015). Moreover, researchers have used different resilience scales across studies, which limits our ability to make direct comparisons. Establishing a standard measure of resilience would enhance study comparisons, and facilitate evaluations of interventions to improve resilience among older adults (MacLeod et al., 2016). Despite this limitation, it is important to publish this data, using the WHI, which is the largest investigation of women’s health, in an understudied, yet growing population of older women aged 80 years and older. A limitation of the WHI and this study is that the majority of the sample are well-educated, non-Hispanic White women. However, the WHI cohort of women by race and ethnicity is for the most part reflective of the distribution of older women in the United States. The NSES indices were grouped low/moderate/high based on the WHI sample, which may not be reflective of the NSES indices of the general U.S. population.

Conclusion

This study found that higher education and lower perceived stress were associated with higher resilience among women aged 80 and older regardless of race, ethnicity, and NSES. Furthermore, higher social support, self-rated health, as well as lower risk of depression and lower symptom burden, were associated with higher resilience across NSES indices, as well as among most minoritized racial and ethnic groups. Future studies on the psychological resilience of this population would benefit from the utilization of repeated measures, as well as mixed methods to better assess the context of resilience among older women.

Funding

The WHI program is funded by the National Heart, Lung, and Blood Institute; National Institutes of Health; and U.S. Department of Health and Human Services through contracts 75N92021D00001, 75N92021D00002, 75N92021D00003, 75N92021D00004, and 75N92021D00005.

Conflict of Interest

None declared.

Author Contributions

J. L. Krok-Schoen: Conceptualization, data curation, methodology, writing—original draft, writing—review and editing. M. J. Naughton: Conceptualization, methodology, writing—original draft, writing—review and editing. A. S. Felix: Writing—original draft, writing—review and editing. C. W. Cené: Conceptualization, writing—original draft, writing—review and editing. S. Springfield: Conceptualization, writing—original draft, writing—review and editing. M. Yu: Data curation, formal analysis, methodology, writing—review and editing. E. M. McLaughlin: Data curation, formal analysis, methodology, writing—review and editing. A. H. Shadyab: Writing—original draft, writing—review and editing. T. S. Nolan: writing—original draft, writing—review and editing. C. H. Kroenke: Writing—review and editing. L. Garcia: Writing—review and editing. S. Follis: Writing—review and editing. R. D. Jackson: conceptualization, supervision, writing—original draft, writing—review and editing.