Cumulative social determinants of health and frailty risk in older adults

Abstract

Background

Frailty is a growing global health challenge, leading to increased rates of adverse outcomes and higher healthcare costs. Its prevalence is particularly high among older adults. This study examined the relationship between social determinants of health (SDoH) and frailty in older adults.

Method

This study included 9,933 participants aged ≥ 65 years from the repeated cross-sectional National Health and Nutrition Examination Survey (NHANES), using data from 2003 to 2018, with different individuals sampled in each cycle. Weighted logistic regression models were used to examine the association between frailty and eight SDoH variables: education level, employment status, poverty-income ratio, food security, health insurance, access to healthcare, homeownership, and marital status. Stratified analyses were also conducted. Nonlinear associations were assessed using restricted cubic spline models. As a sensitivity analysis, we additionally examined the association between cumulative SDoH and pre-frailty (frailty index 0.10–0.20), using non-frail individuals as the reference group.

Result

Of the participants, 3,868 were classified as frail, with a weighted total of 12,619,578 cases. Logistic regression revealed a 37% increase in the odds of prevalent frailty risk for each additional SDoH risk factor (AOR = 1.37; P < 0.001). The cumulative number of unfavorable SDoH was positively associated with frailty risk (P for trend < 0.001). Stratified analyses indicated that the impact of cumulative unfavorable SDoH on frailty was more pronounced in those aged 65–75 years (P for interaction < 0.001). In a sensitivity analysis using pre-frailty as the outcome, the associations with cumulative SDoH were weaker and not consistently significant.

Conclusion

This study suggests that unfavorable SDoH are significantly associated with higher odds of frailty in older adults (≥ 65 years), with a more pronounced association in those aged 65–75 years.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12877-025-06582-5.

Introduction

It is estimated that the global ageing population will rise from 771 million (10%) in 2022 to 1.6 billion (16%) in 2050. This increase will make the elderly population twice as large as the number of children under 5 years of age and nearly equal to the number of children under 12 years. However, there is little evidence that increased longevity is accompanied by improved health. As the population ages, there may be a gradual decline in physical, mental, and cognitive function, leading to impaired abilities and heightened vulnerability to morbidity and mortality [1, 2]. With the growing elderly population, frailty has emerged as a significant global health burden [3]. Previous studies have shown that frailty prevalence increases with age [4–6]. Frailty is a multidimensional and dynamic condition that reflects the age-related decline of multiple physiological systems, making individuals more vulnerable to health changes triggered by minor stressors [7]. It is linked to impaired quality of life, depression [8, 9], dementia [10, 11], falls, and an increased risk of hospital admissions, mortality, and long-term care [12–14].

Social determinants of health (SDoH) refer to the non-medical factors that influence health outcomes, including the conditions in which people are born, grow, live, work, and age. SDoH are increasingly recognized as key drivers of health inequalities and significantly influence our ability to maintain health as we age [15]. Some studies have shown that SDoH are strongly associated with the prevalence of neurological diseases [16–18], coronary heart disease [19], and cancer [20]. In the United States, factors such as lower educational attainment, poverty, and lack of health insurance contribute to the widening mortality gap between different population groups [21–23].

Given the dynamic nature of frailty, including the ability of individuals to move in and out of frail states, and its higher prevalence in older adults, timely identification and intervention could slow its progression and reduce the incidence of adverse outcomes and healthcare costs [24]. SDoH screening is increasingly common in healthcare settings; however, its impact on frailty among older adults remains insufficiently explored. While previous studies have investigated the association between individual social factors and frailty, few have systematically examined the cumulative effect of multiple unfavorable SDoH. Moreover, much of the existing evidence is based on small or localized cohorts, which limits the generalizability of the findings. To address these gaps, the present study uses nationally representative data from the U.S. NHANES survey to assess both the cumulative burden of adverse SDoH and the individual effects of specific SDoH domains on frailty in older adults.

Method

Study population

The National Health and Nutrition Examination Survey (NHANES) is a cross-sectional national survey that collects data on demographics, health, and health behaviours. NHANES employs a repeated cross-sectional design, with different participants sampled in each 2-year survey cycle. Detailed information on the study design, data collection protocols, and publicly available datasets can be found at www.cdc.gov/nchs/nhanes/. This study used data from 2003 to 2018. All study protocols were approved by the Institutional Review Board of the National Center for Health Statistics, and participants provided written informed consent. The study included 9,933 participants aged ≥ 65 years, excluding those without data on frailty index (FI) variables or SDoH.

Baseline characteristics

Sociodemographic and lifestyle data were collected through standardized self-administered questionnaires. Sociodemographic variables included sex (male or female), age (recorded as a continuous variable), and race (categorized as non-Hispanic white, non-Hispanic Black, Mexican American, or other). Lifestyle factors encompassed alcohol consumption and smoking status: participants were classified as never drinkers (defined as consuming fewer than 12 alcoholic drinks in their lifetime) or current drinkers, while smoking status was categorized into nonsmokers (individuals who had never smoked or smoked fewer than 100 cigarettes in their lifetime), former smokers (those with a lifetime consumption of ≥ 100 cigarettes but abstinent at the time of interview), and current smokers. Dietary intake was assessed using 24-hour recalls conducted by trained staff at mobile centers via the USDA automated multiple-pass method. These data informed the calculation of the Healthy Eating Index-2010 (HEI-2010), selected for its alignment with the study’s data collection timeline to ensure contemporaneous dietary guideline adherence. The HEI-2010 comprises 12 components (maximum score: 100), with nine quantifying adequate intake of fruits, vegetables, whole grains, dairy, proteins, and fatty acids, and three evaluating moderation in refined grains, sodium, and empty calories (SoFAAS: solid fats, alcohol, added sugars). Lower moderation scores reflect higher consumption. Total HEI scores (0–100) correlate positively with diet quality.

Physical activity (PA) was evaluated using a 30-day retrospective questionnaire documenting activity type, duration, intensity, and frequency across transportation, occupational, and leisure domains. Metabolic equivalent task (MET) scores were calculated as: Total MET (minutes per week) = number of days *number of minutes* MET scores.

SDoH assessment

To explore the relationship between SDoH and frailty, we selected eight SDoH variables based on prior research [25]. These variables, defined by Health 2030 [26] and the World Health Organization [27] included employment status, household poverty-to-income ratio, food security, education level, access to healthcare, type of health insurance, homeownership, and marital status. Cumulative adverse SDoH counts (range: 0–8) were categorized as 0, 1–2, 3–5, or ≥ 6, with the latter group representing extreme disadvantage. Variable definitions are detailed in eTable 1.

Definition of frailty

Frailty levels were determined using the FI, calculated from 49 variables in the NHANES dataset related to signs, symptoms, disabilities, diseases, and laboratory measurements [28]. Each variable was recoded as 0 (absence of a deficit) or 1 (presence of a deficit), and the FI was computed as the ratio of present deficits to the total number of considered items (49) [29]. Based on established thresholds [30, 31], participants were categorized as non-frail (FI < 0.21) or frail (FI ≥ 0.21). For sensitivity analysis, we further applied a three-level frailty classification: non-frail (FI < 0.10), pre-frail (FI 0.10–0.20), and frail (FI ≥ 0.21). As NHANES is a cross-sectional survey, our outcome reflects the prevalence of frailty at the time of assessment rather than the incidence over time.

Statistical analysis

To account for oversampling in NHANES, we applied recommended survey weights. Continuous variables were expressed as mean ± SE, while categorical variables were presented as counts (percentages). Baseline characteristics were compared using t-tests for continuous variables and χ² tests for categorical variables. To examine the relationship between SDoH and FI, we conducted both unadjusted and adjusted multivariable logistic regression analyses. Model 1 was unadjusted (crude model), Model 2 adjusted for demographic factors (age, gender, and race/ethnicity), and Model 3 additionally incorporated body mass index (BMI), smoking status, drinking status, HEI2010 total score, and PA total MET. To explore the early influence of social determinants, we conducted weighted logistic regression to assess the association between cumulative SDoH and pre-frailty status, using non-frail individuals as the reference group. We also assessed the relationship between individual SDoH components (e.g., employment, poverty-to-income ratio) and FI. Nonlinear relationships between SDoH and FI were explored using restricted cubic splines. Stratified analyses were performed by age, sex, race, smoking, alcohol use, and obesity, with interaction terms to formally test for age-related differences in the impact of SDoH on FI in older adults.

All statistical analyses were conducted using R software (version 4.3.1; R Foundation for Statistical Computing, Vienna, Austria), utilizing the “survey” package to account for NHANES’ complex design. Statistical significance was set at P < 0.05.

Result

This study included 9,933 participants aged ≥ 65 years, of whom 6,065 were non-frail and 3,868 were frail (Table 1). On average, the non-frail group was younger (mean age: 73.1 ± 0.1 years) compared to the frail group (mean age: 76.1 ± 0.2 years). Frailty was more prevalent among women (60.7% [n = 2,116] vs. 39.3% [n = 1,752]). Non-frail was more common among non-Hispanic whites (82.3% [n = 3,513] vs. 77.0% [n = 2,202]). Additionally, frailty was associated with a higher mean BMI (30.0 ± 0.2 vs. 28.1 ± 0.1, P < 0.001), lower HEI-2010 score (52.4 ± 0.4 vs. 55.5 ± 0.3, P < 0.001), and lower PA (MET: 2,064.4 ± 121.9 vs. 2,614.8 ± 82.3, P < 0.001). Lifestyle factors also differed, with non-frail having a higher proportion of alcohol use (85.4% vs. 79.7%, P < 0.001) and never smokers (50.0% vs. 46.1%, P = 0.020). Moreover, frailty was associated with lower levels of education, economic instability (e.g., unemployment, low poverty-to-income ratio, food insecurity), and social disadvantages (e.g., lack of homeownership, being unmarried), as well as higher exposure to unfavorable SDoH risk factors (all P < 0.001; Table 2).

Table 1.

Characteristics of participants according to frailty

Variable	Total N = 9,933	Non-frail N = 6,065	Frail N = 3,868	P value
Age, years	74.2(0.1)	73.1(0.1)	76.1(0.2)	< 0.001
Gender, n (%)				< 0.001
Female	4,996(55.8)	2,880(53.2)	2,116(60.7)
Male	4,937(44.2)	3,185(46.8)	1,752(39.3)
Race, n (%)				< 0.001
Mexican American	1,083(3.4)	645(3.1)	438(4.1)
Non-Hispanic Black	1,776(8.1)	1,060(7.2)	716(9.8)
Non-Hispanic White	5,715(80.4)	3,513(82.3)	2,202(77.0)
Other Hispanic	687(2.9)	419(2.8)	268(3.3)
Other Race	672(5.1)	428(4.7)	244(5.8)
BMI, kg/m2	28.7(0.1)	28.1(0.1)	30.0(0.2)	< 0.001
Drinking, n (%)				< 0.001
Never	1,608(14.1)	920(14.6)	688(20.3)
Now	6,726(71.4)	4,314(85.4)	2,412(79.7)
Smoke status, n (%)				0.020
Former	4,147(42.6)	2,441(41.5)	1,706(44.8)
Never	4,795(48.6)	3,026(50.0)	1,769(46.1)
Now	983(8.7)	592(8.5)	391(9.1)
SDoH	2.1(0.0)	1.7(0.0)	2.7(0.0)	< 0.001
HEI-2010	54.5(0.3)	55.5(0.3)	52.4(0.4)	< 0.001
PA total MET	2,486.4(74.3)	2,614.8(82.3)	2,064.4(121.9)	< 0.001
SDoH group, n (%)				< 0.001
0	1,075(17.2)	883(22.2)	192(7.8)
1	1,940(25.8)	1,356(28.4)	584(20.7)
2	2,116(22.6)	1,365(22.6)	751(22.5)
3	1,860(15.5)	1,105(14.2)	755(18.0)
4	1,418(10.0)	726(7.3)	692(15.2)
5	907(5.6)	399(3.6)	508(9.4)
≥ 6	617(3.4)	231(1.7)	386(6.5)

Data are presented as mean ± SE or n (%)

BMI Body mass index, SDoH Social Determinants of Health, HEI Healthy Eating Index, PA Physical activity, MET Metabolic equivalent task

Table 2.

SDoH variables of participants according to frailty

Variable	Total N = 9,933	Non-frail N = 6,065	Frail N = 3,868	P value
Employment, n (%)				< 0.001
Employed	8,437(88.0)	5,483(92.6)	2,954(79.2)
Not employed	1,496(12.0)	582(7.4)	914(20.8)
PIR, n (%)				< 0.001
< 3	6,780(56.2)	3,783(49.0)	2,997(69.8)
≥ 3	3,153(43.8)	2,282(51.0)	871(30.2)
Food security, n (%)				< 0.001
Full food security	8,082(87.8)	5,254(92.3)	2,828(79.2)
Marginal	1,851(12.2)	811(7.7)	1,040(20.8)
Health insurance, n (%)				< 0.001
Government or no insurance	5,037(42.7)	2,854(39.3)	2,183(49.1)
Private insurance	4,896(57.3)	3,211(60.7)	1,685(50.9)
Housing instability, n (%)				< 0.001
Own home	7,653(83.2)	4,926(87.2)	2,727(75.7)
Rent or other arrangement	2,280(16.8)	1,139(12.8)	1,141(24.3)
Marital status, n (%)				< 0.001
Married or living with a partner	5,413(59.7)	3,594(64.1)	1,819(51.4)
Not married nor living with a partner	4,520(40.3)	2,471(35.9)	2,049(48.6)
Access to healthcare, n (%)				0.126
No routine place	550(4.1)	378(4.4)	172(3.6)
Routine place to go for healthcare	9,383(95.9)	5,687(95.6)	3,696(96.4)
Education, n (%)				< 0.001
High school or more	6,603(77.9)	4,340(82.7)	2,263(68.8)
Less than high school	3,330(22.1)	1,725(17.3)	1,605(31.2)

Data are presented as mean ± SE or n (%)

PIR Poverty/income ratio

Overall, frailty was present in 3,868 cases, with a weighted count of 12,619,578 cases. Logistic regression analysis indicated that each additional SDoH risk factor increased the likelihood of prevalent frailty by 37% (adjusted odds ratio [AOR] = 1.37; 95% CI: 1.30–1.45; P < 0.001; Table 3). Model 1 showed that individuals with ≥ 6 unfavorable SDoH variables had an 11-fold increased risk of frailty (OR = 11.01, 95% CI: 8.36–14.50; P < 0.001). In the total population, all models demonstrated an increasing trend in the odds of prevalent frailty as the number of cumulative unfavorable SDoH variables increased (P for all trends < 0.001, Table 3). Restricted cubic spline regression indicated no significant nonlinear relationship between FI and cumulative unfavorable SDoH variables (overall P < 0.001, nonlinear P = 0.734; Fig. 1). In a sensitivity analysis focusing on pre-frailty, we observed a positive association between cumulative SDoH and pre-frailty status. While the continuous SDoH variable was not significantly associated in the fully adjusted model (OR = 1.06, 95% CI: 0.99–1.13, P = 0.10; eTable 2), categorical analysis showed significantly increased odds among individuals with four or more unfavorable SDoH. For example, participants with four SDoH had 42% higher odds of being pre-frail (OR = 1.42, 95% CI: 1.02–1.98). Restricted cubic spline regression for the sensitivity analysis using pre-frailty as the outcome showed a significant overall association with cumulative unfavorable SDoH, but no evidence of a nonlinear relationship (overall P < 0.001, nonlinear P = 0.114; eFigure 1).

Table 3.

Odds ratios for the association between SDOH and frailty

Variable	Event/All population	Weight event/Weight number	Model 1		Model 2		Model 3
			OR (95% CI)	P value	OR (95% CI)	P value	OR (95%CI)	P value
SDoH (per 1 risk number)	3,868/9,933	12,619,578/36,559,721	1.44(1.40,1.49)	< 0.001	1.43(1.38,1.48)	< 0.001	1.37(1.30,1.45)	< 0.001
Cumulative SDoH variable
0	192/1,075	984,058/6,293,340	Reference		Reference		Reference
1	584/1,940	2,608,821/9,427,917	2.06(1.68, 2.54)	< 0.001	1.83(1.48, 2.27)	< 0.001	1.90(1.38, 2.62)	< 0.001
2	751/2,116	2,838,763/8,249,436	2.83(2.28, 3.52)	< 0.001	2.38(1.90, 2.98)	< 0.001	2.20(1.61, 3.00)	< 0.001
3	755/1,860	2,269,143/5,657,468	3.61(2.94, 4.44)	< 0.001	2.92(2.36, 3.62)	< 0.001	2.63(1.90, 3.63)	< 0.001
4	692/1,418	1,913,360/3,670,425	5.88(4.77, 7.24)	< 0.001	4.96(3.98, 6.19)	< 0.001	4.47(3.20, 6.24)	< 0.001
5	508/907	1,182,479/2,034,780	7.49(5.84, 9.60)	< 0.001	6.75(5.15, 8.84)	< 0.001	5.91(4.08, 8.57)	< 0.001
≥ 6	386/617	822,953/1,226,357	11.01(8.36,14.50)	< 0.001	11.21(8.29,15.14)	< 0.001	8.64(5.24,14.23)	< 0.001
P for trend			< 0.001		< 0.001		< 0.001

Model 1: Not adjusted

Model 2: Adjusted for age, gender and race

Model 3: Adjusted for age, race, gender, drink, smoke, BMI, HEI-2010 total score and PA total MET

SDoH Social Determinants of Health, OR Odd ratio, CI Confidence interval

Fig. 1.

Nonlinear relationship between unfavorable SDoH and frailty

Table 4 presents the relationship between the eight SDoH sub-items and frailty. Most factors were significantly associated with increased odds of frailty, including economic instability (unemployment [AOR = 2.78, 95% CI: 1.94–3.98, P < 0.001]; family income to poverty ratio < 3 [AOR = 1.83, 95% CI: 1.48–2.27, P < 0.001]; food insecurity [AOR = 3.04, 95% CI: 2.43–3.81, P < 0.001]), education below high school level (AOR = 1.65, 95% CI: 1.33–2.05, P < 0.001), lack of health insurance (AOR = 1.43, 95% CI: 1.22–1.67, P < 0.001), lack of homeownership (AOR = 1.87, 95% CI: 1.48–2.38, P < 0.001), and being unmarried or not living with a partner (AOR = 1.22, 95% CI: 1.02–1.45, P = 0.03). Lack of regular healthcare access was not significantly associated with frailty (AOR = 1.12, 95% CI: 0.73–1.73, P = 0.060).

Table 4.

Odds ratios for the association between different components of SDoH and frailty

Variable	Event/All population	Weight event/Weight number	Model 1		Model 2		Model 3
			OR (95% CI)	P value	OR (95% CI)	P value	OR (95% CI)	P value
Employment
Employed, student, retired	2,954/8,437	9,996,810/32,166,318	Reference	-	Reference	-	Reference	-
Not employed	914/1,496	2,622,768/4,393,403	3.28(2.74,3.94)	< 0.001	3.63(2.98,4.42)	< 0.001	2.78(1.94,3.98)	< 0.001
Family income to poverty ratio
≥ 3	871/3,153	3,817,304/16,030,855	Reference	-	Reference	-	Reference	-
< 3	2,997/6,780	8,802,274/20,528,866	2.40(2.13,2.70)	< 0.001	2.07(1.84,2.32)	< 0.001	1.83(1.48,2.27)	< 0.001
Food security
Full security	2,828/8,082	9,991,581/32,093,413	Reference	-	Reference	-	Reference	-
Marginal, low, or very low security	1,040/1,851	2,627,996/4,466,308	3.16(2.79,3.58)	< 0.001	3.40(2.96,3.91)	< 0.001	3.04(2.43,3.81)	< 0.001
Education level
High school graduate or higher	2,263/6,603	8,688,428/28,486,723	Reference	-	Reference	-	Reference	-
Less than high school	1,605/3,330	3,931,150/8,072,999	2.16(1.88,2.48)	< 0.001	1.91(1.66,2.19)	< 0.001	1.65(1.33,2.05)	< 0.001
Regular health care access
≥One regular health care facility	3,696/9,383	452,685/1,497,158	Reference	-	Reference	-	Reference	-
None or emergency room	172/550	12,166,893/35,062,563	0.82(0.63,1.06)	0.130	0.81(0.62,1.06)	0.130	1.12(0.73,1.73)	0.600
Type of health insurance
Private	1,685/4,896	6,194,379/20,961,449	Reference	-	Reference	-	Reference	-
Government or none	2,183/5,037	6,425,198/15,598,272	1.49(1.35,1.64)	< 0.001	1.46(1.32,1.61)	< 0.001	1.43(1.22,1.67)	< 0.001
Home ownership
Own home	2,727/7,653	9,547,060/30,411,845	Reference	-	Reference	-	Reference	-
Rent home or other arrangement	1,141/2,280	3,072,518/6,147,876	2.18(1.90,2.51)	< 0.001	1.98(1.70,2.30)	< 0.001	1.87(1.48,2.38)	< 0.001
Marital status
Married or living with a partner	1,819/5,413	6,482,887/21,823,286	Reference	-	Reference	-	Reference	-
Not married nor living with a partner	2,049/4,520	6,136,691/14,736,435	1.69(1.52,1.88)	< 0.001	1.33(1.19,1.50)	< 0.001	1.22(1.02,1.45)	0.030

Model 1: Not adjusted

Model 2: Adjusted for age, gender and race

Model 3: Adjusted for age, race, gender, drink, smoke, BMI, HEI-2010 total score and PA total MET

SDoH Social Determinants of Health, OR Odd ratio, CI Confidence interval

We first examined the prevalence of frailty across levels of cumulative SDoH, stratified by age group. As shown in Fig. 2, frailty prevalence increased steadily with the number of unfavorable SDoH in both age groups, and was consistently higher among participants aged ≥ 75 years. Notably, the rate of increase was more pronounced in the younger-old group (age 65–75), suggesting a differential sensitivity to social disadvantage. We further assessed whether the association between cumulative SDoH and frailty differed across subgroups using stratified analyses. As shown in Fig. 3, a significant association was observed in all subgroups, including gender, race, smoking status, alcohol use, and obesity. A statistically significant interaction was found between age and cumulative SDoH (P for interaction < 0.001), confirming that the impact of SDoH on frailty was more pronounced among adults aged 65–75 years.

Fig. 2.

Prevalence of frailty by age and cumulative unfavorable SDoH

Fig. 3.

Stratified associations between SDoH and frailty OR, odd ratio; CI, Confidence interval eFigure 1. Nonlinear relationship between unfavorable SDoH and pre-frailty

Discussion

Our findings indicate that the risk of frailty increases with the number of cumulative unfavorable SDoH variables. Each SDoH was significantly associated with frailty, except for lack of regular healthcare access. Additionally, we observed an age-related difference in the impact of cumulative unfavorable SDoH on frailty, with a more pronounced effect in those aged 65–75 years. By contrast, the association with pre-frailty was weaker and not consistently significant.

SDoH is increasingly recognized as a key contributor to disease incidence, mortality, and health inequalities, strongly influencing a wide range of diseases. SDoH is an important factor in cardiovascular disease (CVD) and CVD mortality [32], and one study found diabetes to be highly influenced by SDoH factors [33]. SDoH intervention has also been shown to increase overall cancer screening rates by 8.4% points (IQI = 1.8–18.8) [20]. Additionally, SDoH has been linked to neurological diseases such as multiple sclerosis (MS), neuromyelitis optica spectrum disorder (NMOSD), and stroke [16, 18, 34]. Beyond physical health, SDoH affects psychosomatic conditions like anxiety and depression and can influence their progression and recovery [35, 36]. Furthermore, studies suggest that SDoH may be correlated with sleep health [37]. Our study further affirms the significant association between SDoH and frailty in older adults, particularly in the 65–75 age group, and indicates that the impact of SDoH early in life can have lasting effects on future health outcomes.

Previous research has illustrated the significant associations between frailty and various factors such as sociodemographic, physiological, biological, lifestyle, and psychological factors. A 13-year longitudinal study found that older adults with lower educational attainment were more likely to experience frailty compared to those with higher education [38]. Our study corroborates this, confirming that education below high school is associated with frailty in older adults. Economic factors have also been linked to frailty in patients with myocardial infarction [39]. A prospective cohort studyrevealed that protein intake, particularly animal protein and medium-chain fatty acids, was inversely associated with frailty incidence [40]. Additionally, low serum micronutrient concentrations have been identified as an independent risk factor for frailty in disabled older women, with the risk increasing as the number of micronutrient deficiencies rises [41]. Moreover, physical activity [42] and poor mental health [43, 44] were also found affected the frailty status. Our study supports these findings, showing that frailty in older adults is strongly linked to economic instability (unemployment, low poverty-to-income ratio, and food insecurity), as well as low physical activity and poor mental health. Importantly, this study advances the existing evidence base by using nationally representative data from NHANES (2003–2018) to assess both the cumulative burden and individual effects of eight distinct SDoH domains on frailty prevalence. Unlike previous studies that focused on single factors or were limited to local or clinical cohorts, our approach allows for broader generalizability and population-level interpretation. The construction of a cumulative SDoH index based on the Healthy People 2030 framework provides a more comprehensive view of social disadvantage in older adults. Therefore, these findings reinforce the need to include SDoH in frailty risk assessment strategies for older adults. This approach can help with risk stratification, ensuring that older individuals receive the best possible management and support, reducing frailty incidence and health inequalities, and ultimately lowering healthcare costs and the global health burden.

This study has several strengths. First, it uses data from NHANES, a nationally representative and rigorously designed survey, which enhances the generalizability of our findings to older adults in the United States. Second, we constructed a cumulative index of eight SDoH domains aligned with the Healthy People 2030 framework, allowing a multidimensional assessment of social disadvantage. This cumulative approach offers a more realistic view of the complex interplay of social risks compared to analyses that focus on single factors. However, there are several limitations to this study. First, due to NHANES’s repeated cross-sectional design—where different individuals are sampled in each cycle—causal or temporal inferences cannot be made. Second, SDoH data were collected through self-report questionnaires, which may be subject to recall or reporting bias. Additionally, frailty assessment tools vary widely, including the Physical Frailty Phenotype (PFP), Deficit Accumulation Index (DAI), Gill Frailty Measure, and Frailty/Vigor Assessment. This study used the FI exclusively, and future studies should incorporate other frailty assessment tools to validate these findings. Finally, due to the observational nature of the study, it could not account for all potential confounders. Emerging evidence suggests that environmental factors may interact with social disadvantage to affect frailty [45, 46]. However, environmental exposures could not be assessed in our study due to the lack of geographic identifiers in NHANES.

Conclusion

Our study shows that unfavorable SDoH significantly increase the risk of frailty in older adults, with a cumulative effect. This association was particularly pronounced in those aged 65–75 years. Among the various SDoH examined, factors such as food insecurity, unemployment, lack of homeownership, and low income-to-poverty ratio emerged as particularly impactful. These findings support the inclusion of SDoH in frailty risk assessment strategies. Health policymakers should consider incorporating SDoH screening into routine geriatric care and addressing key social vulnerabilities through targeted interventions.

Authors’ contributions

Ying Sun and Chun Yang had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Concept and design: All authors. Acquisition, analysis, or interpretation of data: Mingyue Xu, Weihua Chen. Drafting of the manuscript: Mingyue Xu, Zijin Li. Critical revision of the manuscript for important intellectual content: All authors. Statistical analysis: Weihua Chen. Administrative, technical, or material support: Mingyue Xu. Supervision: All authors.

Funding

This study was supported by the National Natural Science Foundation of China (82100351 to CY).

Data availability

All data are available as publicly accessible datasets through NHANES. It is open and publicly accessible through the following link; https://wwwn.cdc.gov/nchs/nhanes/.

Declarations

Ethics approval and consent to participate

Ethics approval and consent to participate

The ethical approval of NHANES was granted by the US National Center for Health Statistics Research Ethics Review Board (Protocol No. 98 − 12, Protocol No. 2011-17, Continuation of Protocol No. 2011-17, Protocol No. 2018-01).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.