Social Vulnerability Index as a Tool to Evaluate the Distribution of Head and Neck Oncology Surgeons

ABSTRACT

Objective

To assess the geographic distribution of head and neck oncology surgeons (HNS) in the U.S. and to evaluate the association of this distribution with neighborhood‐level social determinants of health as measured by the Social Vulnerability Index (SVI).

Methods

U.S.‐based HNSs and their practice addresses were obtained from publicly available sources. The geographic distribution of HNSs was analyzed at the state, county, and metropolitan statistical area levels. U.S. census tracts were scored on a continuous scale of increasing social vulnerability (0–1) across Overall SVI and four subthemes: Socioeconomic Status, Household Composition‐Disability Status, Racial‐Ethnic Minority Status, and Housing‐Transportation Status. The distance from each census tract to the nearest HNS was calculated, and univariable linear regression analyses assessed associations between SVI scores and distances.

Results

This study included 609 HNSs that were disproportionately distributed at the state, county, and MSA levels. Higher vulnerability scores in Overall SVI (β = 12.9 [95% CI: 11.05, 14.69]), Socioeconomic Status (β = 11.5 [95% CI: 9.67, 13.32]), Household Composition‐Disability Status (β = 10.4 [95% CI: 8.61, 12.27]), and Housing‐Transportation Status (β = 18.2 [95% CI: 16.42, 20.06]) were associated with significantly increased distance to the nearest HNS, while higher vulnerability scores among Racial‐Ethnic Minority Status were associated with significantly decreased distance to the nearest HNS (β = −25.8 [95% CI: −27.64, −24.01]).

Conclusion

Inequities in the distribution of HNSs highlight the need for targeted strategies to improve access to head and neck cancer care. SVI may help identify especially vulnerable areas lacking access to this care.

This study examines the geographic distribution of head and neck oncology surgeons in the U.S. and its association with neighborhood‐level social determinants of health using the Social Vulnerability Index (SVI). Findings indicate that higher social vulnerability within specific themes of social determinants of health correlates with increased distances to the nearest surgeon, highlighting disparities in access to specialized care.

1. Introduction

Head and neck cancer (HNC) accounts for 3% of all malignancies in the United States (U.S.), affecting over 60,000 new patients every year [1]. HNC mortality varies depending on presentation, with local and metastatic HNC having 5‐year survival rates of 84% and 39%, respectively [1]. These differences in HNC outcomes make timely diagnosis and intervention valuable, especially in patients with limited healthcare access who are more likely to present with advanced HNC [2]. Prior studies have found that patients with a heavy neighborhood‐level burden of social determinants of health (SDoH) were more likely to have advanced HNC presentation and decreased surveillance and survival across a number of HNCs [3, 4, 5, 6, 7]. The complex interplay between neighborhood‐level SDoH and prognosis in HNC is multifactorial and requires a deeper understanding to mitigate such HNC disparities. One of the potential contributors may include limited access to care secondary to geographic proximity to specialized healthcare providers. Previous studies have shown worse survival outcomes among patients with increased distance to treatment facilities in several malignancies such as lung, colon, and esophageal cancer [8, 9, 10]. In regard to HNC, a study by Morse et al. found that longer travel distance to care was associated with increased stage at presentation and a higher rate of laryngectomy in patients with laryngeal squamous cell carcinoma [11]. These studies emphasize that distance and accessibility to specialized physicians may delay diagnosis and lead to late presentation with more advanced disease. This emphasizes the utility of studies highlighting the geographic distribution of specialists in the U.S. whose geographic distributions are especially distinct [12, 13, 14].

Existing literature suggests that most otolaryngologists practice in metropolitan areas and areas with more college graduates, higher socioeconomic status, and lower poverty levels [15, 16]. A study by Sannes et al. looked specifically at the geographic distribution of facial plastic and reconstructive surgeons (FPRS) in the U.S. and found that FPRS had a predominant urban preference, consistent with existing literature [17]. In regard to head and neck oncology, previous studies have characterized the distribution of head and neck surgeons (HNSs) in the U.S. with national and state‐level population densities, identifying relative shortages at the state level. [18, 19] While these studies provide a broader understanding of the geographic distribution of the HNS workforce, it remains unclear how the geographic distribution of HNSs is associated with urban metropolitan status as well as with neighborhood‐level SDoH factors.

Neighborhood‐level SDoH factors have increasingly been studied using SDoH indices, most notably the Social Vulnerability Index (SVI) [3, 4, 5, 6, 20, 21, 22, 23, 24, 25, 26, 27, 28]. The Centers for Disease Control and Prevention (CDC) developed the SVI as a tool to identify disadvantaged communities for the allocation of disaster prevention and recovery services [29]. U.S. census data measures 15 social factors, which the SVI uses to identify at‐risk neighborhoods through four subthemes: Socioeconomic Status, Household Composition‐Disability Status, Racial‐Ethnic Minority Status, and Housing‐Transportation Status [30]. Initially designed to improve resource distribution during natural disasters, the SVI has also been used to evaluate neighborhood‐level disparities in healthcare outcomes, including COVID‐19, cardiovascular disease, and HNC [3, 4, 5, 6, 20, 21, 22, 23, 24, 25, 26, 27, 28].

This study aims to highlight geographic variances in HNS density in the U.S. and to study the association between HNS distribution and neighborhood‐level SDoH factors as measured with SVI. We hypothesize that HNSs tend to practice in metropolitan areas and that increases in neighborhood‐level SVI scores (increased vulnerability) and their four subthemes will be associated with greater distances to the closest HNS.

2. Methods

2.1. Fellowship‐Trained HNSs

The dataset of HNSs included in this study was obtained from the American Academy of Otolaryngology—Head and Neck Surgery (AAO‐HNS) “Find an ENT” directory website and the American Head & Neck Society (AHNS) fellowship graduate list provided on the AHNS website [31, 32]. The AAO‐HNS “Find an ENT” directory is a public reference service that allows for the identification of otolaryngologists who are members of the Academy. Utilizing this webpage, all otolaryngologists with the specialty area of “Head and Neck Surgery” in the U.S. were queried. After accessing each physician's directory page, the “Education” section of their directory page was reviewed to confirm whether they had completed either a “Head and Neck Oncology” or “Microvascular and Head and Neck Reconstruction” fellowship [18]. In instances where a fellowship was not listed on the physician's directory page, the physician's institutional or personal website was reviewed by searching their name and “head and neck surgery” through Google's search engine.

Similarly, the publicly available AHNS fellowship graduate list was used to identify otolaryngologists who had completed a Head and Neck Oncology fellowship from 1997 to 2022, throughout North America. HNSs listed under the AHNS graduate list were included in the study if they were not already listed in the AAO‐HNS “Find an ENT” directory. We confirmed that each HNS had completed a “Head and Neck Oncology” or “Microvascular and Head and Neck Reconstruction” fellowship by reviewing their institutional or personal website. Similarly, each physician's name and “head and neck surgery” was queried through Google's search engine. HNSs practicing outside of the U.S. were excluded from this study. HNS practice type and practice address were recorded based on institutional or personal website information. While the aforementioned methodology for developing a dataset of U.S. HNSs has been utilized in prior publications [18], it is acknowledged that all surgeons offering care for HNC may not be captured by our methodology.

The genders of all HNSs were determined using NamSor Applied Onomastics, an online name recognition algorithm that infers gender based on names and their cultural, ethnic, and linguistic backgrounds [33, 34]. NamSor provides a probability for each assigned gender ranging from zero (0% accuracy) to one (100% accuracy). Any HNS with a gender identification below 80% accuracy was manually reviewed through its institutional or personal website.

2.2. Social Vulnerability Index and Geographic Data

The list of head and neck oncology surgeons (HNS) was compiled into a CSV file, which was uploaded to the GeoCodio website that converts addresses into specific coordinates and Federal Information Processing Standards codes (FIPS) [35]. Federal Information Processing Standard (FIPS) codes are 11‐digit codes assigned to census tracts, which serve as a neighborhood‐level geographic regions as determined by the U.S. census. The 2020 census tract SVI scores were obtained from the CDC website and were paired to the census tract FIPS code for each surgeon as well as for each census tract in the U.S. (Supplemental Figures 1 and 2) [36]. SVI provides five different scores: Overall SVI for overall vulnerability and four subtheme scores for the aforementioned SVI subthemes (Table 1) [36]. SVI scores are arranged on a continuous scale from 0 to 1, with 0 being the least vulnerable relative to all other U.S. census tracts and 1 as the most vulnerable relative to all other census tracts [36].

TABLE 1.

Overall social Vulnerability Index characterized by its four subthemes: socioeconomic status, household composition‐disability status, racial‐ethnic minority status, and housing‐transportation status.

Overall Social Vulnerability Index
Socioeconomic status
Below 150% poverty
Unemployed
Housing cost burden
No high school diploma
No health insurance
Household Composition‐Disability Status
Aged 65 and older
Aged 17 & younger
Single‐Parent Households
English Language Proficiency
Disability Status
Racial‐Ethnic Minority Status
Hispanic or Latino (of any race)
Black of African American, Not Hispanic or Latino
Asian, Not Hispanic or Latino
American Indian or Alaska Native, Not Hispanic or Latino
Two or More Races, Not Hispanic of Latino
Other Races, Not Hispanic or Latino
Housing‐Transportation Status
Multi‐Unit Structures
Mobile homes
Crowding
No vehicle
Group Quarters

The 2023 U.S. census data was used to obtain populations across the U.S., states, and counties [37]. Furthermore, the U.S. Office of Management and Budget was referenced when determining HNS density across metropolitan statistical areas (MSAs), defined as the major metropolitan areas and their affiliated counties [38].

2.3. Statistical Analysis

Univariable linear regression analyses were performed to characterize the association between each of the five SVI themes and the distance between census tracts and the nearest HNS. SVI scores were maintained as continuous variables for the analysis. Statistical significance was set at p < 0.05. All analyses and map generation were performed using R Studio Version 4.1.2 (RStudio Team [2020]. RStudio: Integrated Development for R. RStudio, PBC, Boston, MA, http://www.rstudio.com/) utilizing the “tidyverse” package [39].

3. Results

A total of 609 HNSs from the AAO‐HNS and AHNS directories were included in the study (Figure 1). The majority of HNSs were male (n = 458, 75.2%), with females comprising 24.8% (n = 151) of the cohort. HNS practice setting was balanced between academic setting (n = 306) and other practice types (n = 303). The Southern region contained the most HNSs (n = 212, 34.8%), followed by a similar distribution of HNSs throughout the Northeast (n = 143, 23.5%), Midwest (n = 142, 23.3%), and Western (n = 112, 18.3%) regions. Based on census tract SVI scores, most HNSs were in the second and third quartiles (n = 363) (Table 2).

FIGURE 1.

Choropleth heatmap visualizing the distance of the closest HNS by county in the United States. Blue dots indicate the practice address of HNS.

TABLE 2.

HNS characteristics and distribution by region and SVI (n = 609).

Characteristic	Frequency (n, %)
Gender
Male	458 (75.2)
Female	151 (24.8)
Practice type
Academic	306 (50.3)
Other	303 (49.7)
Region
Northeast	143 (23.5)
Midwest	142 (23.3)
South	212 (34.8)
West	112 (18.3)
Overall SVI
1st Quartile	96 (6.7)
2nd Quartile	170 (15.4)
3rd Quartile	193 (20.7)
4th Quartile	147 (24.1)
Socioeconomic Status SVI
1st Quartile	135 (22.2)
2nd Quartile	130 (21.3)
3rd Quartile	162 (26.6)
4th Quartile	182 (29.9)
Racial‐Ethnic Minority Status SVI
1st Quartile	44 (7.2)
2nd Quartile	167 (27.4)
3rd Quartile	245 (40.2)
4th Quartile	153 (25.1)
Household Composition‐Disability Status SVI
1st Quartile	313 (51.4)
2nd Quartile	107 (17.6)
3rd Quartile	101 (16.6)
4th Quartile	88 (14.4)
Housing‐Transportation Status SVI
1st Quartile	41 (6.7)
2nd Quartile	94 (15.4)
3rd Quartile	126 (20.7)
4th Quartile	345 (56.7)

The densities of HNSs from the AAO‐HNS and AHNS databases were analyzed by state, county, and MSA. The national average density was 0.182 surgeons for a population of 100,000. State‐level HNS density was greatest in the District of Columbia, Nebraska, and New York, respectively. Conversely, Alaska, Delaware, North Dakota, Rhode Island, Vermont, and Wyoming did not have any HNSs registered with the AAO‐HNS or AHNS databases (Supplemental Table 1). County‐level HNS density was greatest in select Northeastern and Midwestern counties, with the number of HNSs surpassing three standard deviations above the mean (Table 3). When analyzed by MSA, the HNS workforce in the Boston–Cambridge–Newton, Houston–The Woodlands–Sugar Land, Chicago–Naperville–Elgin, and New York–Newark–Jersey City MSAs surpassed 1.5 times the national HNS density (Table 4). The Los Angeles–Long Beach–Anaheim, Atlanta–Sandy Springs–Roswell, Miami–Fort Lauderdale–Pompano Beach, and Dallas–Fort Worth–Arlington MSA regions were found to have a lower density of HNSs compared to the national average.

TABLE 3.

Counties with the greatest number of practicing HNS.

County	State	No. of HNS
Harris County	Texas	26 a
New York County	New York	20 a
Cook County	Illinois	18 a
Suffolk County	Massachusetts	14 a
Cuyahoga County	Ohio	14 a
Philadelphia County	Pennsylvania	13 a
Los Angeles County	California	13 a
Wayne County	Michigan	9
Fulton County	Georgia	9
Baltimore City	Maryland	8
	Total	144

^a HNS amounts are > 3 SDs over the mean.

TABLE 4.

HNS density across the 10 most populated metropolitan statistical areas (MSA).

MSA	No. of HNS in MSA	Population	HNS density per 100,000	Relative density to U.S.
United States	609	334,914,895	0.18184	—
New York–Newark–Jersey City, NY–NJ–PA	58	19,498,249	0.297462608	1.635848042
Los Angeles–Long Beach–Anaheim, CA	16	12,799,100	0.12500879	0.420250434
Chicago–Naperville–Elgin, IL–IN–WI	20	9,262,825	0.21591685	1.727213342
Houston–The Woodlands–Sugar Land, TX	29	7,510,253	0.386138789	1.788368018
Atlanta–Sandy Springs–Roswell, GA	11	6,307,261	0.174402169	0.451656694
Washington–Arlington–Alexandria, DC–VA–MD–WV	13	6,304,975	0.206186385	1.182246678
Philadelphia–Camden–Wilmington, PA–NJ–DE–MD	19	6,246,160	0.304186892	1.475300573
Miami–Fort Lauderdale–Pompano Beach, FL	11	6,183,199	0.177901439	0.584842555
Dallas–Fort Worth–Arlington, TX	3	5,462,593	0.054918973	0.30870449
Boston–Cambridge–Newton, MA–NH	16	4,919,179	0.325257528	1.78870176

Abbreviation: MSA, metropolitan statistical area.

Univariable linear regression analyses were performed to characterize the association between each of the five SVI scores and distance in miles between census tracts and the nearest HNS (Table 5). As Overall SVI increased from 0 to 1 (increasing vulnerability), there was a statistically significant increase in distance to the nearest HNS (β = 12.9 [95% CI 11.05–14.69], p < 0.001). Among the four SVI subthemes, increasing vulnerability among Housing‐Transportation Status led to the greatest increase in distance, extending the distance to an HNS by an average of 18.2 miles (β = 18.2 [95% CI 16.42–20.06], p < 0.001). Socioeconomic Status and Household Composition‐Disability Status subthemes also showed similar trends, with increases in vulnerability resulting in an average increase of 11.5 miles (β = 11.5 [95% CI 9.67–13.32], p < 0.001) and 10.4 miles (β = 10.4 [95% CI 8.61–12.27], p < 0.001) to the nearest HNS, respectively. Conversely, increases in Racial‐Ethnic Minority Status vulnerability were associated with an average 25.8‐mile decrease in the distance to the nearest HNS (β = −25.8 [95% CI −27.64 to −24.01], p < 0.001).

TABLE 5.

Changes in distance (miles) to the closest HNS for each one‐unit increase in SVI.

Characteristic	Intercept	Univariable
		Coefficient	95% CI	p
Overall SVI	30.5	12.9	11.05, 14.69	< 0.001
Socioeconomic Status SVI	31.2	11.5	9.67, 13.32	< 0.001
Household Composition‐Disability Status SVI	31.7	10.4	8.61, 12.27	< 0.001
Racial‐Ethnic Minority Status SVI	49.8	−25.8	−27.64, −24.01	< 0.001
Housing‐Transportation Status SVI	27.8	18.2	16.42, 20.06	< 0.001

4. Discussion

To date, this is the first study that describes the HNS workforce with population density at the county and MSA level while also evaluating HNS distribution across neighborhood‐level SDoH factors. This analysis showed that a disproportionate amount of HNSs practices in select metropolitan areas, as reflected by HNS densities surpassing the national averages in six of the top 10 metropolitan areas in the U.S. Regarding SVI, increased vulnerability in Overall SVI and the subthemes, Housing‐Transportation Status, Socioeconomic Status, and Household Composition‐Disability Status, were associated with greater distances to HNSs. Conversely, increased vulnerability in Racial‐Ethnic Minority Status was associated with closer proximity to HNSs.

The analysis of AAO‐HNS‐ and AHNS‐registered HNS density across MSAs and counties provides insight into the geographic distribution of HNSs in the U.S. Our findings show that the density of registered HNSs is notably higher in certain MSAs, such as Boston‐Cambridge–Newton, Houston–The Woodlands–Sugar Land, Chicago–Naperville–Elgin, and New York–Newark–Jersey City, where it surpasses the national average by 1.5 times. These areas are characterized by substantial healthcare infrastructure and resources, likely attracting specialized physicians, including HNSs. These findings were supported by the top county‐level densities of registered HNSs, which tended to be those with urban centers with significant healthcare infrastructure. Conversely, regions like Los Angeles–Long Beach–Anaheim, Atlanta–Sandy Springs–Roswell, Miami–Fort Lauderdale–Pompano Beach, and Dallas–Fort Worth–Arlington exhibit lower HNS density compared to that of the national average despite being highly populated metropolitan areas. This uneven distribution highlights potential gaps in access to specialized care even across the main metropolitan areas of the U.S. Additionally, the absence of registered HNSs in states like Alaska, Delaware, North Dakota, Rhode Island, Vermont, and Wyoming highlights potential areas of significant regional disparities in regard to HNC care. However, it is acknowledged that the AAO‐HNS and AHNS databases likely do not encompass all surgeons providing HNC care. This emphasizes the need for strategic planning and possibly policy interventions to address geographic disparities in HNS distribution in order to promote more equitable access to specialized care across different regions in the U.S.

Distance was utilized to quantify HNS proximity across census tracts with varying SVI scores to highlight associations between neighborhood‐level SDoH factors and HNC care proximity. The results of this analysis found that increasing vulnerability in Housing‐Transportation Status was associated with the greatest distance to the nearest HNS. In the context of HNC outcomes, Farquhar et al. associated prolonged commutes with advanced HNC presentation in low‐income patients [40]. The results of our study support these findings, as increases in Housing‐Transportation Status vulnerability corresponded to a 1.5‐fold increase in distance to an HNS. Census tracts with fewer transportation resources appear to be further from HNSs, which may contribute to the adverse HNC outcomes described by Farquhar et al.

Furthermore, increased vulnerability in Socioeconomic Status and Household Composition‐Disability Status was associated with greater distances to HNSs. These findings reflect those reported in existing literature stating that patients located further away from healthcare services were more likely to be of lower socioeconomic status [41]. In relation to HNC, previous studies have shown that low‐income patients and those living in economically disadvantaged areas were associated with greater mortality [42, 43]. A cross‐sectional analysis investigating SVI and post‐operative complications showed a similar trend, with high Socioeconomic Status and Household Composition‐Disability Status being most strongly associated with post‐operative complications [44]. Distance to HNC care may contribute to these disparities and further compound adverse outcomes post‐operatively.

The Racial‐Ethnic Minority Status subtheme showed an inverse relationship with distance to the nearest HNS; increased Racial‐Ethnic Minority Status vulnerability was associated with decreased distance to the nearest HNS. This may be due to the uneven distribution of racial and ethnic groups across urban, suburban, and rural counties in the U.S. [45] According to the Pew Research Center, 64% of the residents of metropolitan areas are of Hispanic, Black, Asian, and other non‐White ethnicities [45]. Through the use of a diversity index, the U.S. Census Bureau demonstrated that states such as New York and the District of Columbia contained populations with the greatest variety in racial and ethnic characteristics [46]. These two states happen to contain the highest number of HNSs per 100,000 people, as described by Talwar et al. [18] Similarly, our analysis supports these previous findings as HNS density was greatest in counties within a number of major metropolitan cities such as New York, which is reflected by the number of HNSs surpassing three standard deviations above the mean in that particular county. Given this distribution, increases in Racial‐Ethnic Minority Status vulnerability may correspond to census tracts with greater minority population density and HNS supply, explaining the decrease in distance within our analysis. Although minority and ethnically diverse communities may be geographically closer to HNC care, disparities in outcomes seen among these communities suggest that there are other contributing variables that lead to said disparities [3, 4, 5, 6, 27]. Although these communities may be poised closer to these resources, this does not necessarily mean they are accessing them. Future studies should further explore the contributors to these disparities.

The distribution of HNSs was examined across Overall SVI scores converted into quartiles, showing an abundance of HNSs in the second and third quartiles. An analysis associating SVI with surgically underserved census tracts in Southern California, found that “per capita income estimates” were positively associated with surgeon density [47]. Furthermore, The AAMC has suggested that surgeons are more likely to practice in densely populated urban centers with greater hospital resources. These trends support the findings of this study, as most HNSs are not based in SVI quartiles with greater social vulnerability and presumably fewer resources.

Overall, the findings of this study identify and quantify themes of neighborhood‐level SDoH associated with increased distance to HNSs. The varying increases in distance to HNSs in response to rising Overall SVI, Housing‐Transportation Status, Socioeconomic Status, and Household Composition‐Disability Status suggest that equitable HNS access is a multifactorial issue among census tracts throughout the U.S. This analysis emphasizes the importance of utilizing tools such as SVI to not only identify these disparities but also to strategize solutions to address them. Given that SDoH likely vary within states and counties, SVI is advantageous as it can assess these disparities with more granularity through census tract‐level analysis. Improving equitable access to such healthcare resources is valuable and may potentially contribute to improved outcomes in HNC care. Potential strategies to mitigate this may include telemedicine expansion, incentivizing HNS practice in underserved areas through loan‐forgiveness programs, or increasing community outreach and education. Additionally, policy interventions aimed at improving transportation infrastructure and socioeconomic support in vulnerable areas may enhance access to specialized care. Future research should aim to further understand the most effective interventions for improving access to care.

This study is not without its limitations. Only HNSs affiliated with AAO‐HNS and AHNS were included in our analysis; these databases may not encompass the entire HNS workforce in the U.S., and it is possible that patients in more vulnerable areas are treated by physicians not registered in the aforementioned databases. Additionally, distances to HNSs were measured from the center of each U.S. census tract. This is not an exact measure of distance to HNSs and may not perfectly represent the provider to which patients within a census tract travel As a result, the distances to HNSs may be marginally overestimated or underestimated.

5. Conclusion

The use of SVI in describing the geographic distribution of HNSs demonstrates that neighborhoods experiencing increased social vulnerability are further away from HNC care resources. Although Housing‐Transportation Status seemed to have the greatest influence, the SVI subthemes emphasize that access to medical care is a complex issue associated with multiple neighborhood‐level social factors. Given these associations, SVI may be utilized as a tool to identify areas at risk for HNS shortages while characterizing the factors that may be associated with this disparity.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supplemental Figure 1. Choropleth heatmaps visualizing Overall SVI scores by county in the United States.

Supplemental Figure 2. Choropleth heatmaps visualizing Socioeconomic Status, Household Composition‐Disability Status, Racial‐Ethnic Minority Status, and Housing‐Transportation Status social vulnerability scores by U.S. county.

Supplemental Table 1. HNS density by region and state.

Park A. C., Fehrenbach M. P., Davis R. J., et al., “Social Vulnerability Index as a Tool to Evaluate the Distribution of Head and Neck Oncology Surgeons,” The Laryngoscope 135, no. 9 (2025): 3178–3185, 10.1002/lary.32136.