Association between social determinants of health and erectile dysfunction: insights from a nationally representative US sample

Abstract

Background

Beyond traditional risk factors, public health has increasingly focused on the impact of social determinants of health (SDoHs) on disease risk. However, no studies have systematically evaluated the effects of both individual and cumulative SDoHs on the risk of erectile dysfunction (ED), representing a critical gap that our study aims to address.

Methods

Data were collected from a representative sample of adult men participating in the U.S. National Health and Nutrition Examination Survey (NHANES) from 2001 to 2004. Self-reported SDoHs were categorized based on Healthy People 2030 criteria, with a cumulative score of favorable SDoHs created for analysis. ED was assessed using a validated single-item measure. Multivariable regression models were used to examine the effects of individual and cumulative SDoHs on ED risk. Subgroup and sensitivity analyses were conducted to enhance the robustness of the results.

Results

A total of 3,489 eligible participants were included, with 1,001 diagnosed with ED. Unfavorable SDoHs showed significant associations with increased ED risk: being unemployed [odds ratio (OR): 1.97, 95% confidence interval (CI): 1.40, 2.77], family income-to-poverty ratio <300% (OR: 1.65, 95% CI: 1.32, 2.06, P<0.001), marginal-to-very-low food security (OR: 1.50, 95% CI: 1.02, 2.20, P=0.04), high school education or lower (OR: 1.41, 95% CI: 1.09, 1.81, P=0.01), and non-homeownership (OR: 1.23, 95% CI: 1.01, 1.53, P=0.04). A strong inverse relationship was observed between the cumulative count of favorable SDoHs and ED risk (OR: 0.83, 95% CI: 0.75, 0.93, P=0.002). When treated as categorical variables, lower-middle [3–4] and low levels [0–2] of SDoHs were associated with higher ED risk (ORs: 1.76 and 2.62, P<0.05). Additionally, sensitivity analysis showed that the low-level group (0–2 SDoHs) had a significantly higher risk of severe ED compared to the high-level group (OR =3.78, 95% CI: 1.26, 11.34, P=0.02).

Conclusions

The findings suggest that unfavorable SDoHs, particularly when accumulated, are associated with increased risk of ED among U.S. men. Addressing unfavorable SDoHs should be a public health priority, not only to improve male sexual health, but also to promote health equity and overall population well-being. Large-scale population cohort studies are underway to provide higher-level evidence supporting our conclusions.

Highlight box.

Key findings

• Experiencing unfavorable social determinants of health (SDoHs) significantly elevates the risk of erectile dysfunction (ED).

What is known and what is new?

• Previous studies have only examined a few unfavorable SDoHs in relation to ED, lacking a comprehensive and integrated assessment.

• Our study demonstrates that both individual and cumulative unfavorable SDoHs significantly increase the risk of ED.

What is the implication, and what should change now?

• It is worth recognizing that addressing unfavorable SDoHs is a key public health strategy for ED prevention.

Introduction

Erectile dysfunction (ED), historically referred to as impotence, is defined as the persistent inability to achieve or maintain an erection sufficient for satisfactory sexual intercourse (1). According to the Massachusetts Male Aging Study, approximately 52% of men aged 40 to 70 years are affected by varying degrees of ED, with over 30 million cases occurring in the United States alone (2). Recent projections estimate that the global population affected by ED will reach 322 million by 2025 (3). Beyond its high prevalence, ED imposes a substantial public health burden, reflected in escalating healthcare expenditures—from $185 million in 1994 to $330 million by 2000 in the U.S. alone (4). ED not only significantly diminishes the quality of life of those affected but also serves as a key indicator of subclinical cardiovascular disease (CVD), which can surpass ED in posing a potential threat to patient survival (5). Even more unfortunately, the highly complex pathophysiology of ED involves psychogenic, organic, or combined mechanisms that adversely affect the vascular, neuronal, and endocrine systems related to erection (6). Despite such complexity and severity, treatment options for ED remain highly limited, with generally modest efficacy (7). Thus, identifying modifiable risk factors for ED is essential to reducing its burden, underscoring the need for greater attention from researchers and public health policymakers globally.

The U.S. Healthy People 2030 initiative identifies social determinants of health (SDoHs) across five dimensions: economic stability, education access and quality, healthcare access and quality, neighborhood and built environment, and social and community context (8). These dimensions encompass the broader environment in which individuals are born, live, learn, work, play, worship, and age (9). While many well-established risk factors contribute significantly to ED onset (10,11), SDoHs have also emerged as crucial contributors. Previous studies have explored the effects of food security and socioeconomic status on ED risk (5,12). However, these studies often overlook the interactions and complex interconnections between various SDoH factors, leading to findings that are inconclusive and sometimes contradictory. Moreover, many individuals face multiple unfavorable SDoH simultaneously compared to their peers, which increases their risk of developing ED (13). This underscores the need for a systematic investigation of the cumulative burden of unfavorable SDoH on ED risk, a gap that current research has yet to adequately address.

Specifically, we seek to determine which SDoHs are most strongly associated with ED risk, whether there is a cumulative effect of SDoHs on ED, and how these relationships persist after adjusting for potential confounding variables, such as age, lifestyle, and comorbid health conditions. By providing a comprehensive understanding of the interplay between SDoHs and ED, this study aims to inform public health strategies targeting the broader social environment to improve men’s sexual health outcomes. We present this article in accordance with the STROBE reporting checklist (available at https://tau.amegroups.com/article/view/10.21037/tau-2025-454/rc).

Methods

Data resource and study population

The National Health and Nutrition Examination Survey (NHANES) database (https://www.cdc.gov/nchs/nhanes/), managed by the National Center for Health Statistics (NCHS) under the Centers for Disease Control and Prevention (CDC), is a population-based program designed to assess the nutritional and health status of the U.S. population to inform public health policy development and regulation. Since the 1999–2000 cycle, NHANES has been conducted biennially, collecting data through structured interviews, physical examinations, laboratory assessments, and questionnaires across five domains: demographic characteristics, dietary data, physical examinations, laboratory tests, and health-related questionnaires. A multistage, complex probability sampling design is employed to ensure that the sample accurately represents the U.S. population. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The NHANES database was approved by the NCHS Ethics Review Board (Protocol No. #98-12).

The assessment of male erectile function in NHANES was limited to the 2001–2002 and 2003–2004 cycles, restricting our sample selection to 21,161 participants from these two consecutives 2-year cycles. Initially, we excluded 10,860 female participants and 6,185 males lacking ED assessment information. Subsequently, 392 participants without sufficient SDoH data were excluded, followed by the exclusion of 235 participants missing covariate data to ensure the accuracy of statistical results. The detailed exclusion criteria and selection process are illustrated in Figure 1. Ultimately, a total of 3,489 adult male participants with complete exposure, outcome, and covariate data were included, consisting of 1,001 ED cases and 2,488 normal controls.

Figure 1.

Flowchart of sample selection process from the NHANES 2001–2004 cycles. BMI, body mass index; CKD, chronic kidney disease; CVD, cardiovascular disease; ED, erectile dysfunction; NHANES, National Health and Nutrition Examination Survey; SDoHs, social determinants of health.

Measurement of social determinants of health

Healthy People 2030 categorizes SDoHs into five primary domains: economic stability; access to and quality of education; access to and quality of healthcare; neighborhood and built environment; and social and community context. In the 2001–2004 NHANES cycles, eight variables were selected to represent these SDoH subdomains, each aligned with the Healthy People 2030 framework. Participants provided information on the following aspects: (I) employment status, (II) family income-to-poverty ratio, (III) food security, (IV) education level, (V) health insurance status, (VI) type of insurance, (VII) home ownership, and (VIII) marital status. Detailed descriptions, coding, and classifications of these variables are presented in Table S1.

The eight variables were dichotomized using predefined cut-off points to assess the cumulative effects of SDoHs, with 1 indicating favorable conditions and 0 representing unfavorable conditions (14). For the analysis of individual SDoHs, each of the eight SDoH factors was separately included in the regression models. In terms of cumulative SDoHs, the total number of unfavorable SDoHs was calculated by summing the scores of the dichotomized variables and then categorized into four levels: high level [7–8], upper-middle level [5–6], lower-middle level [3–4], and low level [0–2]. These categories were incorporated into the regression analysis, using the high level [7–8] as the reference group.

Assessment of ED

ED was assessed using a single survey item, which asked: “How would you describe your ability to get and keep an erection adequate for satisfactory sexual intercourse?” Although simple, this question was adapted from the Massachusetts Male Aging Study and has been shown to accurately predict clinician-diagnosed ED, making it a valid tool for ED screening (15,16). The question offered four response options: “always or almost always able”, “usually able”, “sometimes able”, and “never able”. In our analysis, male participants who answered “sometimes able” or “never able” were classified as having ED, while those who answered “always or almost always able” or “usually able” were considered normal (17,18). Notably, in our sensitivity analysis, we also adopted a stricter definition of ED, as suggested by previous literatures (19,20). Under this definition, only participants who responded “never able” to maintain an erection were classified as having ED, or more specifically, severe ED.

Selection of potential covariates

Covariate selection was guided by prior studies that identified factors potentially affecting the relationship between SDoH and ED. The covariates included various demographic variables such as age, race, and body mass index (BMI), as well as health-related factors like smoking, alcohol use, and comorbidities, including diabetes mellitus (DM), hypertension, CVD, hypercholesterolemia, asthma, and chronic kidney disease (CKD). Additional SDoH-related factors are detailed in Table S1. Race was classified into five groups: Mexican American, Non-Hispanic White, non-Hispanic Black, other Hispanic, and other races. BMI was analyzed both as a continuous variable and in three categories: <25, ≥25 and <30, and ≥30 kg/m². Age was considered as both a continuous and binary variable, with 60 years as the cutoff. Smoking was divided into never, former, and current categories. Participants were considered yes-drinkers if they had consumed at least 12 alcoholic drinks in their lifetime, or in any single year, and at least one drink in the past 12 months; otherwise, they were classified as no-drinkers.

The diagnosis of comorbidities was established when participants met at least one of three criteria: a self-reported history of diagnosis, current use of relevant medication, or meeting objective examination standards (21). CKD was confirmed by self-reported history or estimated glomerular filtration rate (eGFR) <60 mL/min/1.73 m², calculated using the Chronic Kidney Disease Epidemiology Collaboration equation (CKD-EPI equation). Due to data limitations, CVD and asthma were determined solely by self-reported history, with CVD including congestive heart failure, coronary heart disease, angina, or heart attack, while asthma was defined by prior diagnosis alone (22).

Statistical analysis

All statistical analyses in this study were conducted in accordance with the CDC guidelines for NHANES data. We used appropriate Mobile Examination Center (MEC) weights, adjusted by *1/2 to account for the two combined survey cycles. Baseline characteristics of the study population were described first. Continuous variables were presented as mean ± standard error (SE), while categorical variables were expressed as weighted percentages. Weighted regression analysis was employed for continuous variables, and weighted Chi-squared tests were used for categorical variables. The effects of individual and cumulative SDoHs on ED risk were evaluated using weighted multivariable regression models. Individual SDoHs were analyzed separately, while cumulative SDoHs were treated as a categorical variable, with the high-level group serving as the reference. Three models were constructed to examine the influence of covariates on the relationship between individual and cumulative SDoHs and risk of ED. Model 1 was unadjusted and included only the exposure variables. Model 2 incorporated minimal adjustments by adding age and BMI to the exposure variables. Model 3 was fully adjusted, accounting for potential confounders such as age, BMI, smoking, alcohol use, and the history of hypertension, DM, CVD, CKD, hypercholesterolemia, and asthma. In Model 3, when individual SDoHs were the primary exposure, non-exposure SDoHs were also included for adjustment. Results from the regression analyses were reported as odds ratios (ORs) with 95% confidence intervals (CIs).

To further analyze the impact of cumulative SDoHs on ED risk, subgroup analyses were performed using Model 3, adjusting for all potential covariates except for the stratification variable. Subgroups were defined based on factors that could potentially influence the association between cumulative SDoHs and ED, including age, race, poverty income ratio (PIR), BMI, smoking status, DM, CVD, and CKD. Interaction tests were conducted in all subgroup analyses to evaluate heterogeneity across subgroups. Additionally, for regression analyses using categorical cumulative SDoHs as the exposure, trend tests were performed to assess dose-response relationships. Throughout all statistical analyses, a two-tailed P value <0.05 was considered indicative of statistical significance. Data extraction, cleaning, and analysis were primarily conducted using R version 4.2.1 (R Foundation for Statistical Computing) and EmpowerStats (http://www.empowerstats.com, X&Y Solutions, Inc.).

Results

Characteristics of study participants

A total of 3,489 participants, with a mean age of 45.10±0.38 years, were included in the final analysis, among whom 1,001 were diagnosed with ED. When comparing participants with and without ED, the mean age in the ED group was higher at 61.06±0.55 years, compared to 41.29±0.32 years in the control group. Except for asthma, the prevalence of comorbidities differed significantly between the ED and non-ED groups, with P values <0.001 for each comparison (Table 1). Further analysis of the eight SDoH variables, presented in Table 2, revealed significant differences between the two groups across most factors, except for food security. For instance, the ED group had higher rates of unemployment, lower income levels, and poorer educational attainment compared to the control group. Additionally, the cumulative SDoHs score was notably lower in the ED group (4.58±0.07) than in the control group (4.99±0.05). Similarly, the proportion of participants in the high-level [7–8] and upper-middle-level [5–6] cumulative SDoH categories was lower in the ED group than in the control group.

Table 1. Characteristics of study participants based on ED status, weighted.

Characteristics	Total participants (n=3,489)	History of ED		P value
		No (n=2,488)	Yes (n=1,001)
Age, years	45.10±0.38	41.29±0.32	61.06±0.55	<0.0001
BMI, kg/m2	28.10±0.11	27.86±0.13	29.11±0.28	<0.001
<25, %	29.57	31.05	23.38	<0.001
≥25 and <30, %	41.08	41.06	41.14
≥30, %	29.35	27.89	35.48
Age, %				<0.0001
<60 years	81.18	90.31	42.97
≥60 years	18.82	9.69	57.03
Race, %				0.05
Mexican American	7.82	8.02	6.96
Non-Hispanic White	74.50	73.91	76.95
Non-Hispanic Black	9.37	9.62	8.33
Other Hispanic	4.39	4.10	5.58
Other races	3.93	4.35	2.18
Alcohol intake, %				<0.0001
No	23.29	20.07	36.77
Yes	76.71	79.93	63.23
Smoking, %				<0.0001
Never	42.54	45.20	31.39
Former	29.05	24.96	46.16
Now	28.41	29.83	22.45
History of DM, %				<0.0001
No	89.64	93.91	71.77
Yes	10.36	6.09	28.23
History of CVD, %				<0.0001
No	90.95	94.84	74.64
Yes	9.05	5.16	25.36
History of hypertension, %				<0.0001
No	65.30	70.96	41.58
Yes	34.70	29.04	58.42
History of hypercholesterolemia, %				<0.0001
No	28.03	29.90	20.20
Yes	71.97	70.10	79.80
History of asthma, %				0.08
No	88.43	88.00	90.25
Yes	11.57	12.00	9.75
History of CKD, %				<0.0001
No	87.93	92.91	67.05
Yes	12.07	7.09	32.95

Continuous variables are presented as mean ± standard error, while categorical variables are shown as weighted percentages. Group comparisons were conducted using weighted regression analysis and Chi-squared tests for continuous and categorical variables, respectively. BMI, body mass index; CKD, chronic kidney disease; CVD, cardiovascular disease; DM, diabetes mellitus; ED, erectile dysfunction.

Table 2. Individual and cumulative SDoHs of study participants stratified by ED status, weighted.

Characteristics	Total participants (n=3,489)	History of ED		P value
		No (n=2,488)	Yes (n=1,001)
Employment status, %				<0.001
Employed, student, or retired	88.35	89.44	83.77
Unemployed	11.65	10.56	16.23
Family income-to-poverty ratio, %				<0.0001
≥300%	55.48	57.63	46.47
<300%	44.52	42.37	53.53
Food security, %				0.61
Full security	85.11	84.98	85.62
Marginal, low, or very low security	14.89	15.02	14.38
Educational level, %				<0.0001
College or above	56.66	58.81	47.67
High school or lower	43.34	41.19	52.33
Routine access to health care, %				<0.0001
Routine place	79.34	76.51	91.19
No routine place	20.66	23.49	8.81
Type of health insurance, %				<0.0001
Private insurance	67.63	69.41	60.15
Government or no insurance	32.37	30.59	39.85
Home ownership, %				<0.0001
Own home	70.63	68.46	79.74
Rent or other arrangements	29.37	31.54	20.26
Marital status, %				<0.0001
Married or living with a partner	70.55	68.87	77.59
Not married nor living with a partner	29.45	31.13	22.41
SDoH scores	4.91±0.05	4.99±0.05	4.58±0.07	<0.0001
SDoH categories, %				<0.0001
High level [7–8]	9.93	11.28	4.24
Upper-middle level [5–6]	55.45	56.34	51.76
Lower-middle level [3–4]	27.90	25.85	36.50
Low level [0–2]	6.72	6.53	7.50

Categorical variables are presented as weighted percentages, and comparisons between groups were performed using weighted Chi-squared tests. Statistical significance was set at P<0.05. ED, erectile dysfunction; SDoHs, social determinants of health.

Association between individual and cumulative SDoHs and ED

Table 3 and Figure 2 present the associations between individual SDoH and ED. In the fully adjusted model (Model 3), several unfavorable SDoHs showed positive associations with ED. Specifically, being unemployed had an OR of 1.97 (95% CI: 1.40, 2.77, P<0.001), having a family income-to-poverty ratio below 300% had an OR of 1.65 (95% CI: 1.32, 2.06, P<0.001), experiencing marginal, low, or very low food security had an OR of 1.50 (95% CI: 1.02, 2.20, P=0.04), having a high school education or lower had an OR of 1.41 (95% CI: 1.09, 1.81, P=0.01), and renting or other non-homeownership arrangements had an OR of 1.23 (95% CI: 1.01, 1.53, P=0.04). Table 4 and Figure 3 show a strong inverse relationship between the cumulative count of favorable SDoH and ED risk. When cumulative SDoHs was treated as a continuous variable, an increase in favorable SDoHs significantly reduced ED risk, with an OR of 0.83 (95% CI: 0.75, 0.93, P=0.002). Similarly, when treating cumulative SDoHs, both the lower-middle level (3–4 SDoHs) and the low level (0–2 SDoHs) were significantly associated with increased ED risk, with ORs of 1.76 (95% CI: 1.05, 2.98, P=0.04) and 2.62 (95% CI: 1.36, 5.04, P=0.01), respectively. The trend test also confirmed the statistical significance of this dose-response relationship (P for trend =0.004).

Table 3. Associations between individual SDoHs and ED risk across different models, weighted.

Single SDoH	Model 1			Model 2			Model 3
	OR (95% CI)	P value		OR (95% CI)	P value		OR (95% CI)	P value
Employment status
Employed, student, or retired	Ref			Ref			Ref
Unemployed	1.64 (1.27, 2.12)	<0.001		2.45 (1.74, 3.46)	<0.0001		1.97 (1.40, 2.77)	<0.001
Family income-to-poverty ratio
≥300%	Ref			Ref			Ref
<300%	1.57 (1.33, 1.84)	<0.0001		1.84 (1.45, 2.33)	<0.0001		1.65 (1.32, 2.06)	<0.001
Food security
Full security	Ref			Ref			Ref
Marginal, low, or very low security	0.95 (0.78, 1.16)	0.61		1.75 (1.24, 2.47)	0.003		1.50 (1.02, 2.20)	0.04
Educational level
College or above	Ref			Ref			Ref
High school or lower	1.57 (1.33, 1.85)	<0.0001		1.56 (1.25, 1.94)	<0.001		1.41 (1.09, 1.81)	0.01
Routine access to health care
Routine place	Ref			Ref			Ref
No routine place	0.31 (0.23, 0.43)	<0.0001		0.83 (0.58, 1.19)	0.29		0.90 (0.62, 1.33)	0.58
Type of health insurance
Private insurance	Ref			Ref			Ref
Government or no insurance	1.50 (1.27, 1.78)	<0.0001		1.37 (1.06, 1.77)	0.02		1.18 (0.90, 1.55)	0.22
Home ownership
Own home	Ref			Ref			Ref
Rent or other arrangements	0.55 (0.45, 0.68)	<0.0001		1.32 (1.09, 1.62)	0.01		1.23 (1.01, 1.53)	0.04
Marital status
Married, or living with a partner	Ref			Ref			Ref
Not married nor living with a partner	0.64 (0.52, 0.78)	<0.001		1.10 (0.83, 1.46)	0.50		1.00 (0.71, 1.41)	0.79

Model 1 includes only the exposure variables without adjustments; Model 2 adjusts for age and BMI; Model 3 is fully adjusted, accounting for age, BMI, smoking, alcohol use, histories of DM, CVD, CKD, hypertension, hypercholesterolemia, asthma, and other non-exposure SDoHs. Statistical significance was defined as P<0.05. BMI, body mass index; CI, confidence interval; CKD, chronic kidney disease; CVD, cardiovascular disease; DM, diabetes mellitus; ED, erectile dysfunction; OR, odds ratio; SDoHs, social determinants of health.

Figure 2.

Logistic regression models of the weighted association between individual SDoHs and ED in NHANES 2001–2004. All SDoHs are referenced against favorable SDoHs, with the x-axis representing the OR and 95% CI. Model 1 includes only the exposure variables without adjustments; Model 2 adjusts for age and BMI; Model 3 is fully adjusted, accounting for age, BMI, smoking, alcohol use, histories of DM, CVD, CKD, hypertension, hypercholesterolemia, asthma, and other non-exposure SDoHs. The results indicate that, after adjusting for all covariates, unfavorable employment status, PIR, food security, and education level are positively associated with ED risk. BMI, body mass index; CI, confidence interval; CKD, chronic kidney disease; CVD, cardiovascular disease; DM, diabetes mellitus; ED, erectile dysfunction; NHANES, National Health and Nutrition Examination Survey; OR, odds ratio; PIR, poverty income ratio; SDoH, social determinants of health.

Table 4. Associations between cumulative SDoHs and ED risk across different models, weighted.

SDoH	Model 1			Model 2			Model 3
	OR (95% CI)	P value		OR (95% CI)	P value		OR (95% CI)	P value
SDoH, continuous	0.82 (0.78, 0.87)	<0.0001		0.80 (0.72, 0.88)	<0.001		0.83 (0.75, 0.93)	0.002
SDoH, categories
High level [7–8]	Ref
Upper-middle level [5–6]	2.45 (1.53, 3.91)	<0.001		1.36 (0.82, 2.27)	0.22		1.28 (0.74, 2.23)	0.35
Lower-middle level [3–4]	3.76 (2.35, 6.02)	<0.0001		2.06 (1.24, 3.41)	0.01		1.76 (1.05, 2.98)	0.04
Low level [0–2]	3.06 (1.85, 5.05)	<0.0001		3.30 (1.78, 6.13)	<0.001		2.62 (1.36, 5.04)	0.01
P for trend	<0.0001			<0.001			0.004

Model 1 includes only the exposure variables without adjustments; Model 2 adjusts for age and BMI; Model 3 is fully adjusted, accounting for age, BMI, smoking, alcohol use, and histories of DM, CVD, CKD, hypertension, hypercholesterolemia, and asthma. Statistical significance was defined as P<0.05. for trend indicates the dose-response relationship. BMI, body mass index; CI, confidence interval; CKD, chronic kidney disease; CVD, cardiovascular disease; DM, diabetes mellitus; ED, erectile dysfunction; OR, odds ratio; SDoHs, social determinants of health.

Figure 3.

Logistic regression models of the weighted association between cumulative SDoHs and ED in NHANES 2001–2004. (A) Associations between cumulative SDoHs and ED; the top figure shows the relationship between the continuous cumulative SDoH score and ED risk, while the bottom figure displays the relationship between the categorical cumulative SDoH score and ED risk. The x-axis represents OR (95% CI). The results demonstrate that both continuous and categorical cumulative SDoH scores are associated with an increased risk of ED. (B) Associations between cumulative SDoHs and severe ED; the top figure shows the relationship between the continuous cumulative SDoH score and severe ED risk, while the bottom figure displays the relationship between the categorical cumulative SDoH score and severe ED risk. The x-axis represents OR (95% CI). The results demonstrate that both continuous and categorical cumulative SDoH scores are associated with an increased risk of severe ED. Model 1 includes only the exposure variables; Model 2 adjusts for age and BMI; Model 3 is fully adjusted, accounting for age, BMI, smoking, alcohol use, histories of DM, CVD, CKD, hypertension, hypercholesterolemia, asthma, and other non-exposure SDoHs. BMI, body mass index; CI, confidence interval; CKD, chronic kidney disease; CVD, cardiovascular disease; DM, diabetes mellitus; ED, erectile dysfunction; NHANES, National Health and Nutrition Examination Survey; OR, odds ratio; SDoHs, social determinants of health.

Subgroup analysis and sensitivity analysis

When cumulative SDoHs were treated as a continuous variable, the association between cumulative SDoHs and ED was consistent across most subgroups, as shown in Figure 4. Regardless of subgroup classifications by age, PIR, race, BMI, smoking status, DM, CVD, or CKD history, an increase in cumulative favorable SDoHs was consistently associated with a reduced risk of ED. Specifically, in the subgroup with PIR <300%, cumulative unfavorable SDoHs significantly increased the risk of ED (OR: 0.83, 95% CI: 0.71–0.97, P=0.02), whereas in the subgroup with PIR ≥300%, this relationship was not significant (OR: 1.06, 95% CI: 0.90–1.25, P=0.46). In the subgroup analysis by race, the above regression results remained statistically significant only in the Mexican American, non-Hispanic Black, and other races groups. However, some variations in the strength of association were observed, particularly among participants with different BMI and smoking statuses. Table 5 presents the results of subgroup analysis with cumulative SDoHs as categorical variable. Compared to the high-level group (7–8 SDoHs), the low-level group showed a significantly higher ED risk in most subgroups. For instance, in participants aged <60 years, the OR for ED in the low-level group was 3.40 (95% CI: 1.38, 8.37, P=0.01), whereas in those aged ≥60 years, the association was weaker and not statistically significant (OR =0.95, 95% CI: 0.38, 2.33, P=0.18). The interaction tests for subgroup heterogeneity indicated no significant differences across all subgroups (P for interaction >0.05). Trend tests for increasing SDoHs consistently showed significant associations with reduced ED risk across most subgroups (P for trend <0.05), supporting a dose-response relationship.

Figure 4.

Subgroup analysis of cumulative SDoHs and ED risk, weighted. Regression analyses were adjusted using Model 3, which accounts for age, BMI, smoking, alcohol use, and histories of DM, CVD, CKD, hypertension, hypercholesterolemia, and asthma, excluding the grouping variable. ORs with 95% CIs are presented for each subgroup, and interaction tests were conducted to assess heterogeneity, with statistical significance set at P<0.05. BMI, body mass index; CI, confidence interval; CKD, chronic kidney disease; CVD, cardiovascular disease; DM, diabetes mellitus; ED, erectile dysfunction; OR, odds ratio; SDoHs, social determinants of health.

Table 5. Subgroup analysis of cumulative SDoHs and ED risk, weighted.

Subgroup	SDoH categories				P for trend	P for interaction
	High level [7–8]	Upper-middle level [5–6]	Lower-middle level [3–4]	Low level [0–2]
Age						0.06
<60 years	Ref	1.88 (0.78, 4.55)	2.06 (0.89, 4.78)	3.40 (1.38, 8.37)	0.01
≥60 years	Ref	0.93 (0.50, 1.73)	1.29 (0.61, 2.76)	0.95 (0.38, 2.33)	0.18
PIR						0.21
≥300%		1.92 (0.71, 5.18)	2.35 (0.78, 7.09)	3.20 (0.92, 5.10)	0.06
<300%		1.05 (0.55, 2.00)	1.00 (0.49, 2.08)	1.23 (0.50, 2.67)	0.95
Race						0.05
Mexican American	Ref	0.80 (0.14, 4.65)	1.81 (0.35, 9.44)	2.16 (0.41, 11.41)	0.01
Non-Hispanic White	Ref	1.00 (0.49, 2.01)	1.11 (0.58, 2.10)	1.69 (0.65, 4.39)	0.23
Non-Hispanic Black	Ref	2.57 (0.81, 8.14)	5.43 (1.28, 23.11)	2.68 (0.57, 3.56)	0.02
Other Hispanic	Ref	1.15 (0.89, 2.44)	1.67 (1.28, 3.05)	1.69 (1.09, 3.34)	0.04
Other races	Ref	1.29 (0.36, 3.74)	1.71 (1.20, 2.56)	2.76 (1.94, 5.08)	0.01
BMI						0.10
<25 kg/m2	Ref	3.28 (0.88, 12.25)	7.10 (1.96, 25.65)	8.85 (1.97, 39.77)	0.002
≥25 and <30 kg/m2	Ref	1.02 (0.44, 2.33)	1.51 (0.72, 3.18)	3.47 (1.12, 10.71)	0.01
≥30 kg/m2	Ref	1.09 (0.36, 3.29)	1.03 (0.35, 3.06)	1.28 (0.37, 4.40)	0.85
Smoking						0.09
Never	Ref	0.91 (0.31, 2.65)	1.52 (0.48, 4.80)	2.20 (0.46, 10.54)	0.13
Former	Ref	0.87 (0.37, 2.01)	0.88 (0.36, 2.12)	0.90 (0.29, 2.82)	0.94
Now	Ref	5.59 (1.84, 6.97)	2.72 (2.41, 3.59)	4.17 (3.66, 5.86)	0.001
History of DM						0.07
No	Ref	1.02 (0.54, 1.92)	1.45 (0.82, 2.58)	2.16 (1.07, 4.38)	0.004
Yes	Ref	10.84 (2.99, 39.28)	13.08 (3.04, 56.37)	14.92 (3.67, 60.75)	0.14
History of CVD						0.29
No	Ref	1.24 (0.65, 2.35)	1.61 (0.89, 2.88)	2.67 (1.29, 5.53)	0.01
Yes	Ref	1.74 (0.51, 5.97)	2.94 (0.76, 11.32)	2.36 (0.58, 9.58)	0.07
History of CKD						0.43
No	Ref	1.35 (0.63, 2.91)	2.03 (1.01, 4.09)	3.10 (1.36, 7.09)	<0.001
Yes	Ref	0.92 (0.18, 4.61)	0.96 (0.20, 4.62)	1.25 (0.18, 8.75)	0.74

All results in this table are adjusted using Model 3, which accounts for age, BMI, smoking, alcohol use, and histories of DM, CVD, CKD, hypertension, hypercholesterolemia, and asthma. ORs with 95% CIs are presented for each subgroup, with statistical significance defined as P<0.05. P for trend indicates the dose-response relationship within each subgroup, while P for interaction assesses heterogeneity across subgroups. BMI, body mass index; CI, confidence interval; CKD, chronic kidney disease; CVD, cardiovascular disease; DM, diabetes mellitus; ED, erectile dysfunction; OR, odds ratio; PIR, poverty income ratio; SDoHs, social determinants of health.

In the sensitivity analysis, a stricter ED definition was applied to evaluate the robustness of the association between cumulative SDoHs and severe ED. As shown in Table 6, the results were consistent with the main analysis. When cumulative SDoHs were treated as a continuous variable, the association remained significant, with an OR of 0.84 (95% CI: 0.71, 1.00, P=0.05) in the fully adjusted Model 3. Similarly, when cumulative SDoHs were treated as categorical variables, the low-level group (0–2 SDoHs) had a significantly higher risk of severe ED compared to the high-level group (OR =3.78, 95% CI: 1.26, 11.34, P=0.02). The trend tests also remained significant across models (P for trend <0.001), indicating a persistent dose-response relationship even under the severe ED.

Table 6. Sensitivity analysis of cumulative SDoHs and severe ED risk, weighted.

SDoH	Model 1			Model 2			Model 3
	OR (95% CI)	P value		OR (95% CI)	P value		OR (95% CI)	P value
SDoH, continuous	0.82 (0.75, 0.89)	<0.0001		0.81 (0.69, 0.96)	0.02		0.84 (0.71, 1.00)	0.05
SDoH, categories
High level [7–8]	Ref
Upper-middle level [5–6]	3.05 (1.57, 5.92)	0.002		1.58 (0.73, 3.39)	0.23		1.44 (0.62, 3.32)	0.36
Lower-middle level [3–4]	4.55 (2.17, 9.55)	<0.001		1.90 (0.86, 4.20)	0.11		1.65 (0.70, 3.88)	0.23
Low level [0–2]	4.18 (1.98, 8.80)	<0.001		4.68 (1.64, 13.33)	0.01		3.78 (1.26, 11.34)	0.02
P for trend	<0.001			<0.001			<0.001

Model 1 includes only the exposure variables without adjustments; Model 2 adjusts for age, and BMI; Model 3 is fully adjusted, accounting for age, BMI, smoking, alcohol use, and histories of DM, CVD, CKD, hypertension, hypercholesterolemia, and asthma. ORs with 95% CIs are presented for each model, with statistical significance defined as P<0.05. P for trend indicates the dose-response relationship. BMI, body mass index; CI, confidence interval; CKD, chronic kidney disease; CVD, cardiovascular disease; DM, diabetes mellitus; ED, erectile dysfunction; OR, odds ratio; SDoHs, social determinants of health.

Discussion

Our study is the first to systematically analyze the effects of both individual and cumulative SDoHs on ED risk, supporting our initial hypothesis. Unfavorable individual SDoHs, such as employment status, family income-to-poverty ratio, food security, education level, and home ownership, were significantly linked to an increased risk of ED. However, factors like routine healthcare access, type of health insurance, and marital status did not show a significant impact. More importantly, our findings revealed a cumulative effect of SDoHs, showing that individuals with multiple social disadvantages had a higher risk of developing ED.

Firstly, our findings show that individual SDoHs such as unemployment, low income, food insecurity, lower educational attainment, and non-homeownership are significantly linked to an increased risk of ED. These results align with existing literature. For instance, Macdonald et al. found a strong association between low socioeconomic status and higher ED risk in their analysis of NHANES data, suggesting that factors like income and employment contribute to ED through psychological stress and life instability (12). Additionally, Wang et al. explored the relationship between food security and ED risk, finding that food insecurity could serve as an independent risk factor for ED among U.S. men, a result that aligns closely with our findings (5). While some previous studies have examined individual SDoH in relation to ED, few have systematically accounted for the interactions and cumulative burden of multiple unfavorable SDoHs. Our study observed a strong dose-response relationship between cumulative unfavorable SDoHs and ED risk, suggesting that men facing multiple adverse SDoH factors have a significantly higher likelihood of developing ED than those with higher levels of favorable SDoH. This supports the theory of “cumulative inequality”, whereby the additive effect of multiple unfavorable SDoH may further exacerbate health outcomes through psychological stress, altered health behaviors, and biological pathway disruptions (8). Our study’s contribution lies in emphasizing the cumulative effect of SDoHs on ED, shifting the focus from single risk factors to a broader consideration of social environmental interventions. Finally, we must emphasize that the cross-sectional nature of our study limits our ability to establish causal relationships between these factors. Particularly for variables like marital status and PIR, it is possible that ED and its associated disease burden could lead to divorce and decreased household income, rather than the reverse. The presence of reverse causality may further influence the formulation of public health policies. If ED is not only a result of SDoHs but also exacerbates these adverse factors, then interventions solely targeting SDoH may not effectively reduce the risk of ED. Comprehensive treatment for ED is crucial to break this cycle and, in turn, reverse the negative impact of SDoH. Moreover, while our study demonstrates a robust association between unfavorable SDoHs and ED, the underlying mechanisms are likely multifactorial. Limited healthcare access among disadvantaged individuals may reduce the likelihood of diagnosis and treatment with PDE5 inhibitors (PDE5is) or testosterone replacement therapy. Economic strain, psychosocial stress, and lack of stable relationships may also impact sexual activity and erectile function. Therefore, future research should explore these pathways in more detail, ideally using longitudinal or mixed-methods approaches to clarify both directionality and context.

Our findings, consistent with previous research, aim to inform public health policy development standards. We emphasize that healthcare providers should consider social health inequities alongside common risk factors when addressing ED. Healthcare systems should enhance insurance coverage and accessibility of ED screening and treatment services, particularly for socioeconomically disadvantaged populations. First, community programs should focus on vocational retraining for unemployed men to improve their reemployment prospects. Secondly, widespread education on the prevention and treatment of ED should be conducted, particularly targeting populations with lower educational attainment. Additionally, for socioeconomically disadvantaged populations, providing healthcare subsidies and ensuring regular screening for ED and its associated risk factors are essential. Moreover, initiatives to address food insecurity should be implemented to reduce the risk posed by unhealthy dietary habits. The cumulative effect of multiple SDoH factors on ED risk indicates that effective interventions require multi-level strategies combining healthcare access, education, and economic support. Policy makers should consider these findings when designing interventions to address ED-related health disparities and ensure equitable access to care.

In our subgroup analysis and interaction tests, we found no interaction effect between SDoH and ED in relation to PIR and race. This result should be interpreted with caution. In fact, individuals in different PIR groups experience distinct SDoH, and PIR also influences the selection of social support, which is a significant factor in ED (18). A study examining the relationship between SDoH and depression found an interaction effect between PIR and SDoH on depression, with a sample size of 30,372 (13). Therefore, one potential factor influencing the detection of meaningful interactions in our study could be the sample size, which may have limited statistical power. Future research should include larger sample sizes to validate these findings. In our subgroup analyses, the association between unfavorable SDoHs and ED appeared stronger in participants without DM, CVD, or CKD. This may be due to higher marginal effects of SDoHs in individuals without significant baseline risk, or greater statistical power in these subgroups. However, none of the interaction terms were statistically significant, and these findings should be interpreted with caution.

The link between unfavorable SDoHs and an increased risk of ED can be explained through multiple intersecting pathways, including psychological stress, health-risk behaviors, and biological disruptions. Psychological stress, stemming from socioeconomic disadvantages such as low income, food insecurity, and unemployment, plays a critical role in this relationship. Chronic stress activates the hypothalamic-pituitary-adrenal (HPA) axis, resulting in elevated cortisol levels, which adversely affect vascular health (23). Elevated cortisol has been linked to endothelial dysfunction and arterial stiffness (24), both of which are critical in the pathophysiology of ED (25). Additionally, studies have shown that chronic stress contributes to poor sleep quality, which can lead to lower testosterone levels, further exacerbating the risk of ED (26). Behavioral mechanisms are another pathway through which SDoH influences ED risk. Individuals in lower socioeconomic positions may be more likely to engage in health-risk behaviors, such as smoking, excessive alcohol consumption, and poor dietary choices, as coping mechanisms in response to chronic stress. Smoking, for instance, is associated with vasoconstriction and reduced nitric oxide availability, both of which impair erectile function by limiting blood flow to the penile tissue (27). Meta-analyses confirm alcohol consumption leads to ED through mechanisms including endothelial inflammation and testosterone deficiency from chronic use (28-30). Studies have demonstrated the relationship between diet and ED, showing that it can accelerate ED development through mechanisms such as endothelial cell inflammation and oxidative stress (31-33). Biologically, unfavorable factors have been associated with systemic inflammation, a recognized pathway linking socioeconomic disadvantage to adverse health outcomes (34). Low socioeconomic status has been shown to correlate with elevated levels of inflammatory markers, including C-reactive protein (CRP), interleukin-6 (IL-6), and tumor necrosis factor-alpha (TNF-α), which are implicated in the development of endothelial dysfunction (35). These inflammatory responses, triggered by chronic stress and compounded by poor lifestyle choices, are known to impair the vascular and smooth muscle function necessary for achieving and maintaining an erection (36). Limited healthcare access reduces early screening opportunities and treatment options for patients. Due to ED’s sensitive nature, unprofessional reception or questioning can further decrease patients’ willingness to seek care; additionally, choosing inappropriate healthcare providers may intensify patients’ disease-related shame (37). These issues highlight the future need for establishing specialized men’s health hospitals and training medical professionals to ensure proper patient reception and professional diagnosis and treatment.

However, several limitations warrant caution in interpreting the results. First, the NHANES database used in this study is over 20 years old, which may limit its relevance to the current U.S. population. This highlights the need for future large-scale, multi-center clinical studies to validate our findings and strengthen the evidence base. Second, the diagnosis of ED was based on a single survey question, which, although validated, is less comprehensive than the more widely used International Index of Erectile Function-5 (IIEF-5). Additionally, due to the sensitive nature of ED, self-reported diagnoses may be subject to recall bias or inaccuracies. Third, many SDoHs were assessed through face-to-face interviews, which inherently introduces recall bias. Similarly, several comorbidities in this study were based on self-reported medical history, which may introduce recall bias or misclassification. Although such methods are widely used in population-based surveys, they can underestimate or inaccurately capture the presence of chronic conditions. Furthermore, while chronic conditions such as hypertension and CVD were considered in our models, detailed medication data were not included. Some medications, particularly antihypertensives, are known to have sexual side effects, including ED. The lack of information on medication use may therefore introduce residual confounding. Fourth, the absence of data on mental health status and PDE5i usage, which could not be included as covariates due to NHANES data limitations, represents another potential limitation. In addition, although we have adjusted for a wide range of potential confounders, residual confounding cannot be entirely ruled out. For example, human immunodeficiency virus (HIV) infection—an established risk factor for ED—was not included due to lack of available data in our analytic sample (38,39). Moreover, sexual orientation was not assessed in our study, and the ED measure used—while validated—was primarily developed and validated among heterosexual populations. As such, its applicability to non-heterosexual individuals may be limited, and this should be considered when interpreting our findings. It is also worth noting that the mean BMI in our study population exceeded the normal range, with a high proportion of participants being overweight or obese. While this reflects the broader epidemiologic trends in the U.S. adult male population during the 2001–2004 period, it may limit the generalizability of our findings to populations with lower average BMI. Moreover, given the established association between obesity and ED, residual confounding may persist despite statistical adjustment. Finally, the lack of longitudinal follow-up data prevents us from establishing temporal relationships between SDoHs and ED. Future multi-center, large-scale clinical and cohort studies are needed to confirm these findings, update NHANES data, and further investigate ED risk factors and prevention strategies.

Conclusions

Our findings suggest that experiencing multiple unfavorable SDoHs may be associated with a higher risk of ED; however, these associations should be interpreted with caution due to several study limitations. Our results underscore the importance of addressing unfavorable social determinants of health should be a public health priority, not only to improve male sexual health, but also to promote health equity and overall population well-being. Further clinic-based studies are needed to validate our conclusions and strengthen the evidence base.

Supplementary

The article’s supplementary files as

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The NHANES database was approved by the NCHS Ethics Review Board (Protocol No. #98-12).