Amputation Rates and Associated Social Determinants of Health in the Most Populous US Counties

This cross-sectional study uses the Healthcare Cost and Utilization Project State Inpatient Database to assess which social determinants of health are associated with high amputation rates in the most populous US counties.

Key Points

Question

Which social determinants of health (SDOH) are associated with high amputation rates in the most populous US counties?

Findings

In this cross-sectional study of the Healthcare Cost and Utilization Project State Inpatient Database, amputation rates in the most populous US counties were associated with individual components of SDOH, such as African American race, diabetes, smoking, and food insecurity. Counties with high amputation rates were associated with higher rates of residents with physical distress, greater degree of Black and White racial segregation, and higher proportion of African American inhabitants.

Meaning

County amputation rates were associated with SDOH, some of which are modifiable and may be targets for intervention, which may include creation of community-level preventive measures, particularly in communities with higher levels of Black and White racial segregation.

Abstract

Importance

Social Determinants of Health (SDOH) have been found to be associated with health outcome disparities in patients with peripheral artery disease (PAD). However, the association of specific components of SDOH and amputation has not been well described.

Objective

To evaluate whether individual components of SDOH and race are associated with amputation rates in the most populous counties of the US.

Design, Setting, and Participants

In this population-based cross-sectional study of the 100 most populous US counties, hospital discharge rates for lower extremity amputation in 2017 were assessed using the Healthcare Cost and Utilization Project State Inpatient Database. Those data were matched with publicly available demographic, hospital, and SDOH data. Data were analyzed July 3, 2022, to March 5, 2023.

Main outcome and Measures

Amputation rates were assessed across all counties. Counties were divided into quartiles based on amputation rates, and baseline characteristics were described. Unadjusted linear regression and multivariable regression analyses were performed to assess associations between county-level amputation and SDOH and demographic factors.

Results

Amputation discharge data were available for 76 of the 100 most populous counties in the United States. Within these counties, 15.3% were African American, 8.6% were Asian, 24.0% were Hispanic, and 49.6% were non-Hispanic White; 13.4% of patients were 65 years or older. Amputation rates varied widely, from 5.5 per 100 000 in quartile 1 to 14.5 per 100 000 in quartile 4. Residents of quartile 4 (vs 1) counties were more likely to be African American (27.0% vs 7.9%, P < .001), have diabetes (10.6% vs 7.9%, P < .001), smoke (16.5% vs 12.5%, P < .001), be unemployed (5.8% vs 4.6%, P = .01), be in poverty (15.8% vs 10.0%, P < .001), be in a single-parent household (41.9% vs 28.6%, P < .001), experience food insecurity (16.6% vs 12.9%, P = .04), or be physically inactive (23.1% vs 17.1%, P < .001). In unadjusted linear regression, higher amputation rates were associated with the prevalence of several health problems, including mental distress (β, 5.25 [95% CI, 3.66-6.85]; P < .001), diabetes (β, 1.73 [95% CI, 1.33-2.15], P < .001), and physical distress (β, 1.23 [95% CI, 0.86-1.61]; P < .001) and SDOHs, including unemployment (β, 1.16 [95% CI, 0.59-1.73]; P = .03), physical inactivity (β, 0.74 [95% CI, 0.57-0.90]; P < .001), smoking, (β, 0.69 [95% CI, 0.46-0.92]; P = .002), higher homicide rate (β, 0.61 [95% CI, 0.45-0.77]; P < .001), food insecurity (β, 0.51 [95% CI, 0.30-0.72]; P = .04), and poverty (β, 0.46 [95% CI, 0.32-0.60]; P < .001). Multivariable regression analysis found that county-level rates of physical distress (β, 0.84 [95% CI, 0.16-1.53]; P = .03), Black and White racial segregation (β, 0.12 [95% CI, 0.06-0.17]; P < .001), and population percentage of African American race (β, 0.06 [95% CI, 0.00-0.12]; P = .03) were associated with amputation rate.

Conclusions and Relevance

Social determinants of health provide a framework by which the associations of environmental factors with amputation rates can be quantified and potentially used to guide interventions at the local level.

Introduction

The need for major lower extremity amputation is often the result of advanced comorbid medical conditions causing an end-stage atherosclerotic process wherein arterial perfusion does not meet the basal metabolic requirements of the extremity, resulting in ischemic pain or tissue loss.¹ The implications of undergoing amputation are significant, resulting in decreased quality of life, economic potential, and life expectancy, with the annual costs of amputation in the US estimated to be over $10 billion annually.^2,3,4

The rates of amputation in the US are unequally distributed, disproportionately affecting patients from racial and ethnic minority groups, patients with lower socioeconomic status and increased comorbidities, and residents of specific geographic locations.^5,6 Of particular interest are the metropolitan areas of the US, where groups who are disproportionately affected by amputation predominantly reside and where, despite having the highest concentration of hospitals and specialty and primary care practitioners, patients continue to experience the highest amputation rates in the country on an absolute numbers basis.^7,8,9

The discordance between the availability of specialized medical care and elevated rates of amputation within these regions suggests the influence of factors beyond quality of care driving amputation rates.¹⁰ Social determinants of health (SDOH) provide a framework to evaluate the economic and social variables that contribute to the health of individuals and communities. In this cross-sectional study, we assess the association of SDOH with amputation rates in the most populous counties of the US.

Methods

Study Design

We performed a cross-sectional study using amputation rates obtained from the Healthcare Cost and Utilization Project (HCUP) State Inpatient Database (SID) linked with demographic and SDOH data. The HCUP consists of a series of US health care databases containing hospital care data with encounter-level information for all payers.¹¹ We followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.¹² The institutional review board at Cambridge Health Alliance does not consider analysis of publicly available, deidentified data to be human subjects research.

Patient Population

We queried the HCUP SID for the rates of hospital discharge for amputation in the 100 most populous counties in the US for 2017, the most recent year that data were available. Data were retrieved as HCUP SID county hospital discharge for amputation divided by the Claritas county population estimate, producing a rate of hospital discharge for amputation per 100 000 population. Amputation data were linked to their respective county Federal Information Processing Standard codes, a unique identifier for US state counties used by the US Census Bureau and other databases. The HCUP SID search was performed using Diagnosis Related Groups for amputation for circulatory system disorders except upper limb and toe, with or without major complications or comorbid conditions (Diagnosis Related Groups 239, 240, and 241).

Variables

Information on race and ethnicity was obtained from US Census Bureau and included percentages of the county population that are African American, Asian, Hispanic, and non-Hispanic White. Racial disparities in health care outcomes have been well described, and their inclusion with residential segregation data allowed for potential correlation with amputation data. Using the Federal information Processing Standard county codes, the resulting amputation rates were matched with county US Census Bureau demographic and SDOH data for 2017 using the data sources given in this section. In order to allow for a higher degree of granularity in this study, a total of 28 individual SDOH metrics were assessed, based on the 5 groups defined by the Centers for Disease Control and Prevention: health care access and quality, education access and quality, social and community context, economic stability, and neighborhood and built environment. The SDOH metrics were obtained from the University of Wisconsin Population Health Institute County Health Rankings & Roadmaps (CHR&R),¹³ Dartmouth Atlas of HealthCare,¹⁴ and US Census Bureau¹⁵ websites, with transportation scores assessed from the Center for Neighborhood Technology AllTransit¹⁶ websites. The SDOH metrics assessed in this study and their original derivations are included in Table 1.^16,17,18,19

Table 1. Unadjusted Analysis of Associations Between Social Determinants of Health Component and Amputation Rate.

SDOH component	Unadjusted resulta
	β Coefficient (95% CI)	P value
Health care access and quality
Primary care practitioner ratiob	0.002 (0 to 0.004)	.03
Health care costc	0.0009 (0.0003 to 0.0014)	.002
Diabetes prevalenced	1.73 (1.33 to 2.15)	<.001
Diabetes monitoringc	−0.04 (−0.29 to 0.21)	.76
Percentage of uninsured adultse	0.18 (0.04 to 0.32)	.02
Drug overdose mortality ratef	0.14 (0.02 to 0.27)	.03
Mental distressg	5.25 (3.66 to 6.85)	<.001
Physical distressg	1.23 (0.86 to 1.61)	<.001
Percentage of obesityd	0.51 (0.35 to 0.66)	<.001
Percentage of adult smokersf	0.69 (0.46 to 0.92)	.002
Total beds per countyc	0.00019 (−0.000074 to 0.0004485)	.16
Major teaching institutions per countyc	0.41 (−0.09 to 0.90)	.11
Education access and quality
English proficiencyh	−0.027 (−0.23 to 0.18)	.80
Percentage of high school graduatesi	−0.67 (−1.39 to 0.05)	.07
Percentage of some collegeh	−0.18 (−0.27 to −0.10)	<.001
Economic stability
Percentage of povertyj	0.46 (0.32 to 0.60)	<.001
Median household incomee	−0.0013 (−0.0010 to −0.0009)	<.001
Percentage of unemploymentk	1.16 (0.59 to 1.73)	.03
Neighborhood and built environment
Homicide rate (per 100 000)f	0.61 (0.45 to 0.77)	<.001
Percentage of severe housing probleml	0.13 (−0.058 to 0.31)	.17
Air pollution particulate matterm	0.10 (−0.32 to 0.52)	.63
Physical inactivityd	0.74 (0.57 to 0.90)	<.001
Social and community context
Racial segregation Black and Whiteh	0.19 (0.13 to 0.26)	<.001
Racial segregation Non-White and Whiteh	0.19 (0.13 to 0.25)	.004
Percentage with limited access to healthy foodsn	0.29 (0.074 to 0.50)	.009
Food insecurityo	0.51 (0.30 to 0.72)	.04
Percentage of single-parent householdh	0.29 (0.22 to 0.36)	<.001
Transit scorep	0.21 (−0.22 to 0.64)	.34

Abbreviations: CDC, Centers for Disease Control and Prevention; SDOH, social determinants of health.

^a β Coefficient, percentage difference in amputation rate.

^b Data source: Area Health Resource File from the American Medical Association.¹⁷

^c Data source: Dartmouth Atlas of Health Care.¹⁸

^d Data source: CDC Diabetes Interactive Atlas.¹⁷

^e Data source: Small Area Health Insurance Estimates.¹⁷

^f Data source: CDC WONDER mortality data.¹⁷

^g Data source: Behavioral Risk Factor Surveillance System.¹⁷

^h Data source: American Community Survey.¹⁷

ⁱ Data source: US Department of Education EDFacts.¹⁷

^j Data source: US Census Bureau, Small Area Income and Poverty Estimates Program.¹⁹

^k Data source: Bureau of Labor Statistics.¹⁷

^l Data source: Comprehensive Housing Affordability Strategy data.¹⁷

^m Data source: Environmental Public Health Tracking Network.¹⁷

ⁿ Data source: US Department of Agriculture Food Environment Atlas.¹⁷

^o Data source: Map the Meal Gap.¹⁷

^p Data source: Alltransit–Center for Neighborhood Technology.¹⁶

Statistical Analysis

The distribution of county amputation rates was assessed with histograms and then quartiles. Differences in demographic characteristics and SDOH according to amputation rate quartile were assessed using analysis of variance. Given the linearity of the reported county amputation rate, regression analysis was then performed between county amputation rates and each of the 28 SDOH metrics. Results are reported as standard coefficients representing change in amputation rate per 100 000 per each unit change of the respective independent variable. Multivariate linear regression analysis was performed to adjust for potential confounding variables based on associations found on unadjusted analysis and literature to preserve adequate power.

County-level arterial bypass rates were obtained from the Dartmouth Atlas of Health Care for 2017 and matched where available to the counties in the present study. The bypass to amputation ratio was determined by dividing the available county bypass rates by the amputation rates derived from the SID. Data assembly and analysis were performed from July 3, 2022, to March 5, 2023, with Microsoft Excel (Microsoft Corp) and Stata, version 16.1 (StataCorp LP). A 2-side P ≤ .05 was considered statistically significant.

Results

Amputation discharge data were available for 76 of the 100 most populous counties in the United States. Within these counties, 15.3% were African American, 8.6% were Asian, 24.0% were Hispanic, and 49.6% were non-Hispanic White.

County Amputation Rates

The mean (SD) rate of amputation in 76 of 100 most populous US counties with available data was 9.6 (3.7) per 100 000 population vs 11.2 per 100 000 population across all US counties. The median (range) rate of amputation across the largest US counties was 8.9 (3.1-21.2) per 100 000 population. Rates of amputation in the most populous counties in the US ranged from 5.5 per 100 000 in quartile 1 to 14.5 per 100 000 in quartile 4. There was considerable variability in rates of amputation in the most populous counties in the US, which was attributable to outlier counties with higher rates of amputation (Figure) as well as geographical variability (eFigure in Supplement 1).

Figure. Overall Distribution of Amputation Rate per 100 000 Inhabitants in the Most Populous US Counties.

Distribution of counties by amputation rate.

Demographic Data

Compared with counties in lowest amputation quartile, counties in the highest quartile had higher percentages of African American individuals (27.0% vs 7.9%, P < .001); lower percentages of Asian individuals (4.4% vs 12.5%, P = .002); higher rates of diabetes (10.6% vs 7.9%, P < .001), smoking (16.5% vs 12.5%, P < .001), physical inactivity (23.1% vs 17.1%, P < .001), unemployment (5.8% vs 4.6%, P = .01), poverty (15.8% vs 10.0%, P < .001), single-parent households (41.9% vs 28.6%, P < .001), and food insecurity (16.6% vs 12.9%, P = .04); and lower rates of having some college education (63.3% vs 71.1%, P = .006) (Table 2).

Table 2. Patient Characteristics by Amputation Quartile.

Characteristic	Quartile,a patients, %b				P value
	1st	2nd	3rd	4th
Amputation rate per 100 000, mean (SD)	5.5 (2.5)	7.9 (3.3)	10.4 (5.0)	14.5 (5.4)	<.001
Absolute No. of amputationsc
Hospital amputation discharge, No.	1607	2536	2979	3476
Demographic data
Age ≥65 y	12.6	14.1	12.5	14.4	.17
Race and ethnicity
African American	7.9	12.0	14.2	27.0	<.001
Asian	12.5	11.4	5.9	4.4	.002
Hispanic	24.8	20.2	30.9	19.9	.20
Non-Hispanic White	51.3	53.4	46.7	46.9	.54
Uninsured	12.7	12.4	15.3	15.9	.13
Diabetes	7.9	8.9	9.4	10.6	<.001
Current smoker	12.5	13.3	15.3	16.5	<.001
Physical inactivity	17.1	19.5	21.2	23.1	<.001
Rural	2.9	3.0	3.2	3.6	.91
Unemployed	4.6	5.2	5.7	5.8	.01
Poverty rate	10.0	10.6	14.7	15.8	<.001
Single-parent household	28.6	29.4	36.7	41.9	<.001
Housing insecurity	20.6	20.5	22.0	21.1	.73
Food insecurity	12.9	13.4	14.8	16.6	.04
High school graduate	81.9	85.2	81.2	79.5	.12
Some college	71.1	69.6	63.6	63.3	.006
Not proficient in English	6.7	5.7	8.1	5.3	.15

^a Quartiles corresponding to amputation rate per 100 000 in 76 US counties, with the first quartile representing the lowest amputation rate and the fourth quartile, the highest amputation rate.

^b Percentages reported correspondent to county averages.

^c Calculated by multiplying amputation rate per 100 000 by county population estimate obtained from US Census Bureau data for 2017.

Unadjusted Associations Between Amputation Rate and SDOH

County amputation rates were significantly associated with prevalence of mental distress (β, 5.25 [95% CI, 3.66-6.85]; P < .001), diabetes prevalence (β, 1.73 [95% CI, 1.33-2.15], P < .001), physical distress (β, 1.23 [95% CI, 0.86-1.61]; P < .001), unemployment (β, 1.16 [95% CI, 0.59-1.73]; P = .03), physical inactivity (β, 0.74 [95% CI, 0.57-0.90]; P < .001), adult smokers (β, 0.69 [95% CI, 0.46-0.92]; P = .002), higher homicide rate (β, 0.61 [95% CI, 0.45-0.77]; P < .001), food insecurity (β, 0.51 [95% CI, 0.30-0.72]; P = .04), and poverty (β, 0.46 [95% CI, 0.32-0.60]; P < .001) (Table 1). Additional variables, such as the ratio of primary care practitioners; health care cost; rate of drug overdose mortality, obesity, and college attendance; median household income; and racial segregation, had statistically significant associations with amputation (Table 1).

Adjusted Association Between Amputation Rate and SDOH

Multivariable regression analysis found that physical distress (β, 0.84 [95% CI, 0.16-1.53]; P = .03), Black and White racial segregation (β, 0.12 [95% CI, 0.06-0.17]; P < .001), and percentage of African American race (β, 0.06 [95% CI, 0.00-0.12; P = .03) were significantly associated with amputation rate (Table 3).

Table 3. Adjusted Analysis of Associations Between Amputation Rate and Social Determinants of Health.

SDOH	β Coefficient (95% CI)a	P value
Age	0.11 (−0.09 to 0.31)	.27
African American race	0.06 (0.00 to 0.12)	.03
Black and White racial segregation	0.12 (0.06 to 0.17)	<.001
Diabetes prevalence	0.26 (−0.41 to 0.92)	.44
Smoker	−0.19 (−0.47 to 0.10)	.20
Unemployed	−0.11 (−0.59 to 0.37)	.66
Physical inactivity	0.20 (−0.08 to 0.49)	.16
Physical distress	0.84 (0.16 to 1.53)	.03
Food insecurity	0.10 (−0.14 to 0.33)	.43
Obese	0.10 (−0.12 to 0.32)	.38
Mental distress	0.24 (−0.53 to 1.00)	.54
Uninsured	0.40 (−0.24 to 1.00)	.21
Homicide rate	−0.09 (−0.34 to 0.16)	.49

Abbreviation: SDOH, social determinants of health.

^a β Coefficient, percentage difference in amputation rate.

Surgical Bypass Patterns by County Amputation Quartile

Surgical bypass data were available for 56 counties and were included in a secondary analysis (Table 4). In absolute terms, there was a statistically significant higher proportion of bypasses in the quartiles with the higher amputation rates. When rates of bypass surgical procedures were adjusted for the number of amputations performed per quartile, resulting in a surgical bypass to amputation ratio, the pattern was reversed, with the highest rates of surgical bypass to amputation ratios being present in the quartiles with the lowest amputation rates, although this result was not statistically significant.

Table 4. Surgical Bypass Patterns Across Amputation Quartiles^a.

Variable	No. per 100 000 inhabitants (95% CI), by quartile				P value
	1st	2nd	3rd	4th
Discharged for amputation	5.5 (4.8-6.1)	7.9 (7.2-8.5)	10.4 (9.7-11.0)	14.5 (13.8-15.2)	<.001
Bypass procedure	32.9 (22.2-43.8)	41.4 (32.4-50.3)	53.5 (43.9-63.1)	64.1 (54.8-73.4)	<.001
Bypass procedure per amputation ratio	5.9 (4.7-7.1)	5.3 (4.3-6.3)	5.2 (4.1-6.2)	4.5 (3.5-5.6)	.36

^a Counties grouped according to amputation rate per 100 000 inhabitants.

Discussion

In this cross-sectional study, we evaluated amputation rates in the most populous counties of the US, with counties serving as a proxy for major metropolitan areas, which is where, in absolute numbers, the majority of amputations in the US take place. Persons who are disproportionately affected by amputation, African American and Hispanic individuals, predominantly reside in these metropolitan areas, with 59% and 71% residing there, respectively.²⁰ Although the incidence of amputation had been decreasing for decades, this pattern has recently been reversing, and there is a risk of compounded increase as the proportions of the population that are elderly or have diabetes or obesity increase.²¹ Without additional preventive efforts dedicated to individuals at highest risk, the rate of amputation in the US has the potential to further increase in general and in particular within the most vulnerable groups.

We observed wide variation in the rate of major amputation across the most populous counties in the US, ranging from 3.1 to 21.2 amputations per 100 000. Assessed by county quartile, rates varied from 5.5 to 14.5 amputations per 100 000. Evaluating county-level amputation rates by quartile allowed for population-level analysis of demographic characteristics, revealing significant differences in the racial composition, socioeconomic status and health-related characteristics of high vs lower amputation rate counties. Quartiles with higher rates of amputation were more likely to have greater proportions of persons who were African American, smoked, had diabetes, and/or were living in poverty, consistent with studies demonstrating race and socioeconomic status as independent risk factors for major amputation.^{6,22,23,24,25} With major urban centers having a disproportionate number of health care practitioners and specialists, it would suggest that factors outside of health care quality are contributing to these elevated rates of amputation.

In the health care access and quality domain of SDOH, a large number of factors were associated with amputation in unadjusted and adjusted analyses. We found that community mental and physical distress, a measure of adults reporting 14 or more days of poor physical or mental health a month, were important factors associated with amputation rates that may have been previously underappreciated. The association between amputation and mental and physical distress reinforces described associations between environmental stressors and cardiovascular health associated with pathophysiological pathways related to upregulation of stress hormones, oxidative damage and inflammation.²⁶ Environmental stressors may be associated with depression, which in turn has been shown to be associated with increased risk of amputation in patients with peripheral artery disease (PAD).²⁷ Reported neighborhood stress levels have been associated with failure to comply with preventive health initiatives, such as smoking cessation, whereas low general health, frequent physical distress, frequent mental distress, and frequent activity limitation have been associated with a higher prevalence of chronic conditions.^28,29 Interventions addressing community physical and mental stress levels may be of particular importance in reducing disparities in amputation rates experienced by these communities.

There were associations between Black and White segregation and amputation rate in both unadjusted and adjusted analyses. Interestingly, these associations were independent of a number of other factors, including those associated with patient health and socioeconomic factors, such as employment and insurance status. In a study by Kershaw et al,³⁰ communities with higher African American segregation were found to have higher incidences of cardiovascular disease, an association that persisted when adjusting for neighborhood characteristics, socioeconomic position, and risk factors. Our findings echo their study and suggest inherent characteristics of segregated communities that may be associated with amputation. Further studies are warranted to assess factors that may contribute to progression of PAD resulting in amputation, which may include socioeconomic variables, health care access issues, exposure to community endemic health harming factors, or patient health behaviors.^5,31 Powell et al³² assessed the role of medical mistrust, perceived racism, and everyday racism and its association with delays in preventive care and found that individuals with higher levels of these factors were more likely to delay screenings and routine checkups. Community outreach has been suggested as a means to overcome distrust of health care and medical institutions, with 1 study implementing community-based PAD screenings in parts of the community with high social cohesion, which resulted in increased awareness of disease processes and the willingness to undergo screening.³³ More research is required to determine how segregation contributes to risk factors that contribute to PAD and its progression and the role of targeted interventions to preempt them.

In unadjusted analysis, uninsured status was associated with amputation. The Affordable Care Act (ACA) went into effect in 2014, resulting in the expansion of health care coverage for millions of individuals in the US, particularly those of African American or Hispanic race and ethnicity or with low incomes.³⁴ The ACA has been associated with increased access to primary care, fewer skipped medications due to costs, reduced likelihood of emergency department visits, increased outpatient visits, and increased routine care for chronic conditions.³⁵ Despite these improvements, the number of uninsured individuals in the US remains substantial, and barriers to care for disadvantaged populations persist,³⁶ including long waiting times for appointments, inconvenient office hours for workers lacking paid sick leave, transportation difficulties, and copayments for some Medicaid enrollees and most enrollees in non-Medicaid insurance plans. In a study assessing reported barriers to care during the preceding 2 decades, such barriers to care increased among all racial groups, although it increased disproportionately among African American or Hispanic individuals.³⁷

In the unadjusted analysis, diabetes prevalence had a positive association with amputations rates, but the percentage of diabetes monitoring was not significantly associated. The testing of hemoglobin A_1c (HbA_1c) as part of an initial quality measure for patients with diabetes has been endorsed by the Ambulatory Quality Alliance, a multistakeholder quality organization, for the purpose of improving patient outcomes.³⁸ While some studies^39,40 suggest the importance of routine surveillance because of an association between progressively increasing levels of HbA_1c and lower extremity amputation rates, other studies^41,42 have found only a moderate association between the speed of diffusion of HbA_1c testing and rates of lower extremity amputation among patients with diabetes, indicating that increased testing may not immediately translate to improvement in patient outcomes.^39,40,41,42 This result is also replicated in our findings, indicating the possibility that diagnostic results may not have changed management due to issues with integration of results into care pathways or medical noncompliance. Focusing simply on HbA_1c as a measure of diabetes control does not appear to be sufficient to reduce diabetes-associated complications, such as amputation, and extension of quality of care metrics in communities with high amputation rates should be implemented.^43,44 The importance of best medical management of risk factors for PAD has been well established and has been incorporated into societal guidelines, including the Global Vascular Guidelines on the Management of Chronic Limb Threatening Ischemia.¹ Without appropriate medical management, PAD will invariably progress to chronic limb-threatening ischemia, which carries a 5-year risk of mortality and amputation of 29% and 57%, respectively.^45,46 Studies have found decreased adherence to best medical management principles among patients with vascular disease, with our findings demonstrating smoking, obesity, and physical inactivity as areas that warrant attention.^47,48

In all 3 of the measures in the economic stability domain, there were statistically significant associations between amputation rates, including factors such as unemployment and poverty rates, in unadjusted analysis. These findings reinforce calls for addressing employment opportunities to improve community health.^24,49 Unemployment has been associated with increased diabetes prevalence and atherosclerosis, both risk factors for progression of PAD to chronic limb-threatening ischemia.^50,51

Considering the SDOH domain neighborhood and built environment, physical inactivity was associated with amputation rates in the unadjusted analysis. Multiple systematic reviews have demonstrated the benefits of supervised exercise programs in improving maximal walking times compared with patients undergoing no exercise or nonsupervised exercise regimens.^52,53 While those studies were inconclusive on benefits to mortality and amputation reduction, meta-analyses assessing the benefits of exercise-based cardiac rehabilitation programs among patients with coronary artery disease have demonstrated reduction in cardiovascular mortality and risk of hospital admission.⁵⁴ Studies also suggest a potential benefit of mobile-based exercise therapy programs in adherence, with resulting physical and psychological benefits.^55,56 Additionally, in the present study, homicide rates also were associated with amputation rates in unadjusted analysis, suggesting a potential association of safety perception and engagement with physical exercise.²⁶ The association between physical activity and amputation may also be attributable to surgeon decision-making, as ambulatory status is an important component of patient-centered limb preservation risk-benefit analysis.⁵⁷ The absence of an association between physical inactivity and amputation in the adjusted analysis also suggested that the findings in the unadjusted analysis may be associated with the presence of confounding variables, such as neighborhood safety and physical stress levels, which were found to be associated with amputation rate. Overall, these findings highlight the importance of implementation of policies in these communities to target physical activity engagement (eg, parks, sidewalks, bike lanes, and overall public safety), ensuring a safe environment for such activities while reducing stressors that may be inhibitory.

Although rates of surgical bypass were found to be greater in counties with higher amputation rates, analysis of the rates of surgical bypasses per amputation showed a pattern in the opposite direction. These discrepancies may reflect patients with PAD presenting with later stage disease or comorbidities that preclude surgical bypass, counties having a disproportionate number of physicians performing endovascular therapies as opposed to bypass, and a regional preference or bias against bypass.^58,59 Further studies are warranted to determine if greater access to bypass surgery in the counties with the highest rates could reduce amputation given the results of the BEST-CLI (Best Endovascular vs Best Surgical Therapy in Patients with Critical Limb Ischemia) study, which demonstrated superiority of surgical bypass over endovascular interventions in decreasing major adverse limb events and all-cause mortality when a suitable vein bypass conduit is available.⁶⁰

Limitations

The limitations of the present study are that only a single year was evaluated and that data for amputation rates in the 100 largest US counties was limited to approximately three-quarters of the counties, restricting our ability to draw conclusions about geographic variability. The University of Wisconsin Population Health Institute CHR&R, from which we obtained most of our SDOH data, drew its information from multiple sources with disparate methods. The data sets in this study did not have information regarding disease severity, which could help determine if increased amputation was associated with progression of disease or if patients had adequate autologous vein bypass conduits, which may influence patient eligibility for surgical bypass.

Conclusion

This cross-sectional study highlights the associations between amputation rates and several SDOH factors that are beyond traditionally evaluated comorbidities and clinical care metrics. The findings suggest that improvements in community health outreach and services, employment opportunities, and education may have important associations with PAD outcomes. Additionally, the technique of identifying amputation risk factors at a county level might be incorporated into a grading system to determine county-level risk for amputation, facilitating evidenced-based targeted interventions to reduce the risk of this devastating complication. Further research is necessary to assess the rates of bypass in the US counties with the highest rates of amputation.

Supplement 1.

eFigure. Map with amputation rates of the most populous US counties

Supplement 2.

Data Sharing Statement