Causal assessments of multilevel social determinant factors on meningioma disparities in the United States

Abstract

Background

Prior investigations into meningioma disparities have explored associative relationships of socioeconomic status (SES) and race-ethnicity but face gaps in the range of other social determinants/drivers of health (SDoH) factors considered and sample size. Furthermore, none have explored causal relationships between SDoH-factors and outcomes. Thus, this study aims to utilize a recent, national sampling of meningioma patients incorporating comprehensive inferential and causal-mediation approaches to delineate which SDoH-factors objectively drive care and prognostic disparities.

Methods

This retrospective study of a specialized Surveillance-Epidemiology-End Results 2020 dataset for community-/census tract-level (Yost-Index, a composite SES measure and Rurality-Urbanicity) and individual-level (sex, race-ethnicity) SDoH-factors performed age-adjusted multivariate cox-hazards and logistic regressions, and covariate-adjusted causal-mediation analyses to assess differences in overall survival, treatment receipt, and delay of treatment initiation.

Results

In age-adjusted multivariate analyses of 110,042 meningioma patients from 2010-2018, lower community-level SES significantly increased overall mortality (HR 1.31, 95%CI 1.28-1.34), decreased interventional treatment receipt (Surgery-OR 0.89, 95%CI 0.87-0.91; Radiation 0.83, 0.79-0.87), and increased treatment delay (1.13, 1.09-1.16). Minoritized race/ethnicity featured increased interventional treatment receipt (Surgery 1.18, 1.15-1.22; Radiation 1.18, 1.12-1.24) and decreased treatment delay (0.90, 0.87-0.93). In covariate-adjusted causal analyses, community-level SES showed total mediation effects of race-ethnicity in influencing overall survival and negative partial mediation effects in treatment receipt and delay.

Conclusion

For overall survival, community-level SES primarily drove meningioma disparities even when accounting for other SDoH-factors. For treatment receipt and delay, race-ethnicity caused greater differences that were partially affected by community-level SES. In turn, these comprehensive analyses provide definitive causes of meningioma disparities.

Key Points.

• Community-level SES factors were causal drivers for meningioma survival disparities

• Biological sex and race-ethnicity showed strong associations with patient outcomes

• SES-factors diminished the effects of race-ethnicity on meningioma treatment outcomes

Importance of the Study.

Previous studies evaluating the impact of social determinants of health (SDoH) on meningioma outcomes have primarily been inferential assessments of only a few SDoH factors. As such, current literature has limited generalizability to understanding the real-world intersections of SDoH. This study aims to utilize the SEER database to assess a comprehensive range of SDoH factors and their impact on meningioma outcomes.

Additionally, past studies have been unable to identify causality within SDoH factors, and instead, have predominantly been restricted to assessments of associations. To date, this is the first study to incorporate causal analyses into the analysis of the relationship between SDoH factors and meningioma outcomes. Through assessments of how socioeconomic status alters the influence of other SDoH factors, this study seeks to add further nuance to our understanding of causal factors that drive clinical and treatment disparities in meningioma patients.

Introduction

Meningiomas are primarily benign tumors derived from meningothelial cells that constitute 39.7% of all primary CNS tumors.¹ Due to clinical risk factors such as ionizing radiation, obesity, and hormone receptors, the total case count is expected to surpass 30,000 total cases in 2027.^2,3 In addition, non-clinical risk factors related to one’s social circumstances and community have seen an increasing focus for relaying a disproportionate burden of illness among meningioma patients. These factors, commonly known as social determinants or drivers of health (SDoH) have been observed to impact meningioma patients based on their biological sex and gender, race-ethnicity, socioeconomic status (SES), and other SDoH-aspects.^1–6 Among established literature, female sex, non-Hispanic black race-ethnicity, lower income level, and lack of health insurance have all shown significant associations with more severe symptomatology, longer hospitalization, and higher costs during hospitalization for meningioma patients.^2,5–8

Among more comprehensive efforts to characterize SDoH-impact and meningioma outcomes, several have attempted to tackle issues of sample size, relevant chronology of patients, broader considerations of meningioma subtypes and primary sites, or representation of varied SDoH-factors beyond singular aspects of socioeconomic status and race-ethnicity.^9–14 However, shortcomings remain, as these past studies often lack consideration in one or more of these areas. Furthermore, the expansion of methodologies beyond inferential models towards causal analyses has yet to be performed. Given these limitations, these past findings on socioeconomic status, race-ethnicity, and other SDoH possess limited generalizability for understanding the real-world intersections between varied SDoH and their influence on meningioma outcomes across the US.

The National Cancer Institute-Surveillance, Epidemiology, and End Results (SEER) program features multiple sets of data for assessing tumor outcomes and clinicodemographic characteristics, including specially-authorized datasets featuring individual-level and community/census-level SDoH-measures, such as the Yost-Index/Yost-SES-Index, a composite census-level index comprising socioeconomic, education, housing characteristics, among others. Using more recent iterations of this SDoH-specialized dataset alongside methodologies comprising multilevel inferential and causal-mediation modeling, this study seeks to comprehensively characterize the causal relationships between varied individual- and community-level SDoH-factors with observed disparities of meningioma care and prognosis. We hypothesized that community-level SES as measured by the Yost-SES-Index will be the primary causal driver of care and prognostic disparities in comparison to other individual and community-level SDoH-factors.

Materials and Methods

This retrospective cohort study follows the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline. Per the protocols of Northwestern University’s Institutional Review Board, this study did not require prior IRB/ethics committee approval or waiver of informed consent, as the data consisted of publicly available, deidentified patient data.

Database

This study used the National Cancer Institute’s SEER program, which contains national datasets of patient variables, pathological characteristics, treatment modalities, and prognostic outcomes. The SEER database is comprised of 18 registries across the country that are representative of the demographic makeup of the US allowing for trends to be more generalized. The specific dataset used for this study was the Specially Authorized Head-Neck SEER 2020 Dataset which has several census-level SDoH measurements, including a composite socioeconomic status (SES) measure and two rurality-urbanicity measures (by Rural-Urban Commuting Area (RUCA) codes and Urban Rural Indicator Codes (URIC)). Clinical measures of interests consisted of all-cause mortality, receipt of surgery, receipt of radiotherapy, and delay of treatment initiation (defined as 3 months or more after primary tumor diagnosis). Based on SEER data protocol, receipt or non-receipt of any treatment modality was indicated for the patient at hand.

As a specialized feature of the dataset, the Yost Index consists of an SES measure on the census-level that composites the education level, median household income, percentage living below poverty, median home value, median rent, proportion working blue-collar jobs, and proportion over 16-years-old in the workforce without employment of each patient’s community represented by their designated census tract.¹⁵ Using these measures scores are reported on a scale from 1-100 with 1 being the highest SES and 100 being the lowest. These scores are then divided into quintiles labelled as “Highest,” “Higher,” “Middle,” “Lower,” and “Lowest.” As part of the dataset restrictions, these labels are already pre-assigned to each individual patient without reference to their geographic location in order to remain compliant with data privacy agreements.

Characteristic Definitions

The SEER dataset was searched for adult patients (20 + years) diagnosed with meningioma from 2010-2018 using the International Classification of Diseases for Oncology, 3rd Edition (ICD-O-3) for histology codes 9530-9539. These patients were then categorized based on a total of four clinicodemographic factors, two individual-level and two census tract-level. The two individual-level categories were biological sex and race-ethnicity (Non-Hispanic White, Hispanic, Asian or Pacific Islander, Black, Unknown, and Native American). The two census tract-level categories were Yost Index SES, which was divided into a Low-Middle SES category (“Lowest,” “Lower,” and “Middle” quintiles) and a High SES category (“High” and “Highest” quintiles), and Rurality-Urbanicity, with rurality defined as either having a Rural-Urban Community Area code of “Non-Urban” or a US Census Rurality-Urbanicity designation of either “Mostly Rural” or “All Rural.”

Statistical Analysis

Age-adjusted, multivariate regression models were used to analyze the relationship between the four patient clinicodemographic factors on the individual-level (sex, race-ethnicity) and community-level (Yost Index SES and rurality-urbanicity) and their impacts on patient outcomes. Independent covariates were set as the following: sex (female as comparator, male as reference), race-ethnicity (non-White as comparator, non-Hispanic White as reference), Yost Index SES (Lowest-Middle SES as comparator, Higher-Highest SES as reference), and Rurality-Urbanicity (“Rural”-RUCA designation and “Mostly Rural” and “All Rural”-URIC designations as comparator, “Urban”-RUCA and “Mostly Urban” and “All Urban”-URIC designations as reference).

Using these age-adjusted, multivariate models, logistic regressions were used to analyze treatment outcomes assessed that included dependent outcomes of whether patients underwent primary site surgery, whether they underwent stereotactic radiation therapy, and whether they experienced delays-in-treatment (3 or more months from primary tumor diagnosis to initial treatment). These age-adjusted, multivariate models were also utilized for cox proportional-hazards analyses to evaluate multi-level SDoH impact on all-cause mortality.

Based on the multivariate regression models above, mediation analyses were performed using covariate adjustments of individual-level determinants (age and biological sex) and community-level determinants (rurality-urbanicity) to causally assess the largely reported relationships between race-ethnicity and meningioma outcomes (overall survival/mortality, treatment receipt, and delay of treatment initiation) while mediating for community-level SES/Yost-SES-Index measures.¹⁶ In summary, causal-mediation models revolve around sequential regression models that are then assessed for differences in effect size based on the inclusion/removal of the mediator variable. This allows per-variable quantification of the direct effectual relationship between the independent (e.g., minoritized race/ethnicity) and dependent variates (e.g., outcomes) while elucidating the indirect or confounding effects of the mediator variate (e.g., Yost SES Index) in causing this relationship. Given the heterogenous, nationally-representative study sample, considerations of static/non-temporal confounders outside of the mediator were accounted for in the covariate adjustments (i.e., individual and community-level covariates aforementioned), whereas temporal confounders were accounted for in the hazards models utilized in calculating the percent mediation. A summarized depiction of this mediation modeling and the relationships between them can be found in Figure 1.

Figure 1.

Causal Mediation Analyses of Multilevel Social Determinant Factors. Acycle graph representing covariate adjusted analyses of meningioma patients assessing for the relationship between race-ethnicity and clinical outcomes mediated by community-level SES/Yost-SES-Index.

Chi-squared tests were conducted to investigate differences of study sample representation between “High” (classified from “Higher” and “Highest”-Yost Index score) and Low-Middle Yost Index SES across clinicodemographic characteristics.

Two-sided p-values were used for all analyses. The alpha-criterion was set to 0.05. R-4.2.3 was used to perform all statistical analyses. Patients with any incomplete data within necessitated clinicodemographic variables for certain multivariate models were excluded from their respective

Results

Clinicodemographic Characteristics

110,042 adults diagnosed with meningioma between 2010 and 2018 were obtained from the SEER database. The most common clinicodemographic characteristics were being of 65-84 years of age (n = 46,850, 42.6%), female sex (n = 81,246, 73.8%), and non-Hispanic White race/ethnicity (n = 74,280, 67.5%).

Chi-squared analyses were completed with the sample population divided into high Yost-SES-Index (n = 58,981, 54%) and low Yost-SES-Index (n = 51,061, 46%) cohorts (Table 1). Significant differences in clinicodemographic characteristics between SES cohorts were observed for age groups, race-ethnicity, rural-urban area code, and US census rural-urbanicity (p < 0.001 for all). Significant differences in clinical outcomes were determined for histological type, receipt of primary surgery, receipt of radiation therapy, months from diagnosis to initial treatment, 3-year all-cause mortality, 5-year all-cause mortality, and vital status on last follow-up (p < 0.001 for all) (Table 1)

Table 1.

Clinicodemographic Characteristics

		Yost Socioeconomic Index Score
Characteristic	N	High SES, N = 58,981 (54%)	Low-Middle SES, N = 51,061 (46%)	p-value
Age	110,042			<0.001
20-44 years		5,886 (10.0%)	5,760 (11%)
45-64 years		21,498 (36%)	18,357 (36%)
65-84 years		25,025 (42%)	21,825 (43%)
85 + years		6,572 (11%)	5,119 (10%)
Sex	110,042			0.009
Female		43,355 (74%)	37,891 (74%)
Male		15,626 (26%)	13,170 (26%)
Race	110,042			<0.001
White		43,670 (74%)	30,610 (60%)
Black		3,464 (5.9%)	9,322 (18%)
Hispanic		4,980 (8.4%)	7,696 (15%)
Asian or Pacific Islander		5,973 (10%)	2,869 (5.6%)
Unknown		566 (1.0%)	309 (0.6%)
Native American		328 (0.6%)	255 (0.5%)
Rural-Urban Commuting Area Code	105,662			<0.001
Rural		1,370 (2.5%)	9,649 (19%)
Urban		53,231 (97%)	41,412 (81%)
US Census Rural-Urbanicity	105,660			<0.001
All Rural		937 (1.7%)	4,428 (8.7%)
All Urban		40,814 (75%)	33,259 (65%)
Mostly Rural		2,193 (4.0%)	4,001 (7.8%)
Mostly Urban		10,655 (20%)	9,373 (18%)
Histology Type	110,042			<0.001
Meningioma, NOS		49,820 (84%)	42,782 (84%)
Meningothelial Meningioma		2,729 (4.6%)	2,699 (5.3%)
Atypical Meningioma		2,129 (3.6%)	1,839 (3.6%)
Transitional Meningioma		1,375 (2.3%)	1,047 (2.1%)
Psammomatous Meningioma		973 (1.6%)	805 (1.6%)
Fibrous Meningioma		758 (1.3%)	774 (1.5%)
Meningioma, Malignant		547 (0.9%)	466 (0.9%)
Angiomatous Meningioma		260 (0.4%)	278 (0.5%)
Meningiomatosis, NOS		203 (0.3%)	186 (0.4%)
Clear Cell Meningioma		187 (0.3%)	185 (0.4%)
Primary Surgery Performed	108,762			<0.001
No Surgery		37,123 (64%)	32,646 (65%)
Surgery		21,257 (36%)	17,736 (35%)
Radiation Therapy Performed	110,042			<0.001
No Therapy		54,062 (92%)	47,296 (93%)
Therapy		4,919 (8.3%)	3,765 (7.4%)
Months from Diagnosis to Initial Treatment	92,613			<0.001
< 1 month		11,703 (23%)	10,285 (24%)
1-3 months		9,565 (19%)	7,199 (17%)
3 months or more		28,744 (57%)	25,117 (59%)
3-year All-cause Mortality	80,591			<0.001
Alive, > 3yr		34,473 (80%)	28,617 (76%)
Dead, =<3yr		8,501 (20%)	9,000 (24%)
5-year All-cause Mortality	66,348			<0.001
Alive, > 5yr		24,540 (70%)	19,741 (64%)
Dead, =<5yr		10,743 (30%)	11,324 (36%)
Vital Status on Last Follow-up	110,042			<0.001
Alive		44,915 (76%)	36,399 (71%)
Dead		14,066 (24%)	14,662 (29%)

Survival

Cox-proportional hazard models were performed to examine the individual and census level SDoH factors associated with meningioma overall survival risk (Table 2). Female-Sex demonstrated strong associations as a markedly negative independent effector (HR, 0.69; 95%CI, 0.67-0.71; p < 0.001). Poor Yost-SES, alternatively, was observed to be a markedly positive independent effector increasing risk (1.31; 1.28-1.34; p < 0.001) (Table 2).

Table 2.

Multivariate Cox-Proportional Hazards Model

Disease Classification1	Characteristic	HR	95% CI	p-value
Meningioma, NOS	Sex—Female	0.69	0.67, 0.71	<0.001
	Minority Race/Ethnicity	1.01	0.99, 1.04	0.340
	Yost Index—Lower Socioeconomic Status	1.31	1.28, 1.34	<0.001
	Rurality	1.02	0.98, 1.06	0.384

In causal-mediation analyses across individual- and community-level SDoH characteristics, the significant effects of race-ethnicity on increasing or decreasing overall survival risk were completely mediated by community-level SES as measured by the Yost-SES-Index (Figure 1).

Receipt of Primary Surgery

Multivariate logistic regression models were completed to reveal significant associations between several SDoH factors and receipt of primary surgery for meningioma patients. Of which, being of a minority race/ethnicity was determined as an independent positive predictor (OR, 1.18; 95%CI, 1.15-1.22; p < 0.001). Female-Sex was again a markedly negative predictor (OR, 0.70; 95%CI, 0.68-0.72; p < 0.001) while poor Yost-SES-Index was observed to be a negative predictor for receiving primary surgery (OR, 0.89; 95%CI, 0.87-0.91; p < 0.001) (Table 3).

Table 3.

Multivariate Logistic Regression of Surgery Receipt

Disease Classification1	Characteristic	OR	95% CI	p-value
Meningioma, NOS	Sex—Female	0.70	0.68, 0.72	<0.001
	Minority Race/Ethnicity	1.18	1.15, 1.22	<0.001
	Yost Index—Lower Socioeconomic Status	0.89	0.87, 0.91	<0.001
	Rurality	1.03	0.98, 1.07	0.273

From causal-mediation analyses, the significant effects of minoritized race-ethnicity increasing surgery receipt were partially mediated by community-level SES as measured by the Yost Index. Of note, the mediation effect observed was negative, indicating that community-level SES diminished the impact of minoritized race-ethnicity on increasing surgery receipt (Figure 1).

Receipt of Radiation Therapy

For meningioma patients receiving radiation therapy, both Female-Sex (OR, 0.79; 95%CI, 0.75-0.83; p < 0.001) and poor Yost-SES-Index (OR, 0.83; 95%CI, 0.79-0.87; p < 0.001) were determined to be negative predictors. However, being of minority race-ethnicity was observed to be a positive predictor for radiation therapy receipt (OR, 1.18; 95%CI, 1.12-1.24; p < 0.001) (Table 4).

Table 4.

Multivariate Logistic Regression of Radiation Therapy Receipt

Disease Classification1	Characteristic	OR	95% CI	p-value
Meningioma, NOS	Sex—Female	0.79	0.75, 0.83	<0.001
	Minority Race/Ethnicity	1.18	1.12, 1.24	<0.001
	Yost Index—Lower Socioeconomic Status	0.83	0.79, 0.87	<0.001
	Rurality	1.05	0.97, 1.13	0.218

From causal-mediation analyses, the significant effects of minoritized race-ethnicity increasing radiation therapy receipt were partially mediated by community-level SES as measured by the Yost Index. Similar to prior, the mediation effect observed was negative, indicating that community-level SES diminished the impact of minoritized race-ethnicity on increasing radiation therapy receipt (Figure 1).

Treatment Delay after Preliminary Diagnosis

For treatment delays greater than 3 months, Female-Sex was a markedly positive predictor (OR, 1.38; 95%CI, 1.34-1.43; p < 0.001). Additionally, lower Yost-SES-Index was also observed to be a positive predictor (OR, 1.13; 95%CI, 1.09-1.16; p < 0.001). Conversely, patients of minority race-ethnicity (OR, 0.90; 95%CI, 0.87-0.93; p < 0.001) and rural backgrounds (OR, 0.91; 95%CI, 0.86-0.96; p < 0.001) were both determined to be independent negative predictors for treatment delays (Table 5).

Table 5.

Multivariate Logistic Regression of Delay in Treatment Initiation

Disease Classification1	Characteristic	OR	95% CI	p-value
Meningioma, NOS	Sex—Female	1.38	1.34, 1.43	<0.001
	Minority Race/Ethnicity	0.90	0.87, 0.93	<0.001
	Yost Index—Lower Socioeconomic Status	1.13	1.09, 1.16	<0.001
	Rurality	0.91	0.86, 0.96	<0.001

From causal-mediation analyses, the significant effects of minoritized race-ethnicity were partially mediated by community-level SES as measured by the Yost Index. Of note, the mediation effect observed was negative, indicating that community-level SES diminished the impact of minoritized race-ethnicity on decreasing treatment delays (Figure 1).

Discussion

This multilevel assessment of SDoH factors utilized multivariate regression and mediation approaches to principally characterize the causal relationship behind community-level SES, individual-level race-ethnicity, and other sociodemographic factors with meningioma disparities in the US from 2010 to 2018. Following comprehensive inferential analyses, several SDoH factors were observed to have significant associations with meningioma outcomes, with individual-level biological sex and race-ethnicity and community-level SES factors showing particularly strong associations with multiple measures of patient outcomes. Furthermore, the causal effects of community-level SES on meningioma disparities across race-ethnicity were observed across all measured outcomes. Racial-ethnic differences in overall survival were fully mediated by Yost-SES-Index factors and racial-ethnic differences in treatment receipt while delayed initiation was partially mediated after adjusting for all other sociodemographic factors. In turn, this study both affirms prior observations of the literature with broader SDoH-models and considerations of meningioma subtypes while also elucidating the causal impact of certain sociodemographic factors.

For the receipt of primary surgery and radiation therapy, both Female-Sex and poor Yost-SES-Index presented as substantive negative predictors. Studies from Bhambhvani et al. and Aizer et al. established similar trends, finding that SES was inversely correlated with the likelihood of receiving primary meningioma surgeries such as gross total resections.^12,17,18 Although it is well established that patients of Female-Sex experience higher incidences of meningioma, to our understanding, there are seldom investigations examining the surgical treatment differences across female and male meningioma patients. However, decreased surgical intervention for female patients fall in line with general surgical trends seen with female patients for non-meningioma pathologies and injuries.^19–21 These independent predictors also were consistent among treatment delay trends, with female sex and poor Yost-SES-Index being positive predictors. These results fall in line with prior smaller-scaled studies, such as Carstam et al. showcasing lower income and female sex featuring increased delays in treatment and radiologic diagnosis among a 547 Swedish patient cohort.²² By expanding upon this trend through a US-nationally representative cohort, this present study corroborated previously identified trends while adding nuance through joint modeling of community- and individual-level patient demographics. In turn, our results support previously established associations regarding both a composite SES measure and general trends of biological sex while providing further insight into the treatment outcomes of an existing vulnerable patient population.

Notably, inferential analyses revealed minoritized race-ethnicity as a clinically significant positive predictor of both primary surgery and radiation therapy receipt in the present study. Although prior studies have observed contrasting associations between black race-ethnicity and the receipt of these treatment modalities, more-recent analyses utilizing national patient samplings have showcased similar positive associations as ours regarding Hispanic-specific and Asian/Pacific Islander-specific findings.²³ Consistent with the established literature, these unintuitive trends were also observed with regards to minoritized patients having decreased delays in treatment independent of other individual-level or community-level factors. Given the nuances of meningioma management that allow many cases to be non-interventionally managed for elongated periods after preliminary diagnosis, alongside racial-ethnic minorities showcasing trends of having greater tumor size and symptomatic burden on preliminary presentation and diagnosis,¹³ our findings showcase the downstream effects of these trends with patients of minority race-ethnicity necessitating more immediate interventional treatment compared to non-minority groups. These can include prospective studies assessing adjusted diagnostic protocols for meningioma patients of minority race/ethnicity or supportive social services, such as subsidized transportation or patient education tailored to a patient’s cultural or language characteristics, that would be targeted based on modeling the specific community’s drivers of meningioma disparities.

Given this myriad of trends on the mortality and care-receipt levels, mediation analyses were necessitated in order to causally elucidate whether traditionally observed racial-ethnic or SES-related factors dictated meningioma patient differences in real-world situations. In brief, mediation analyses were statistical techniques originally developed and primary utilized within the clinical psychological and neurology fields for determining causal relationships of numerous factors of interest with a desired (or undesired) outcome.^24–27 Increasingly, they have been utilized in risk stratification models and causal sociodemographic characterizations of disparities in tumor-related pathologies and outcomes.²⁸ With this approach, we showcased that the causes of racial-ethnic disparities of meningioma-related survival hazards were completely mediated/caused by differences in community-level SES factors while adjusting for other sociodemographic covariate factors. Furthermore, we showed that community-level SES partially diminished the effects of race-ethnicity in influencing treatment receipt but not completely. These objective differences definitively showcase how community-level SES largely dictated mortality while race-ethnicity majorly affected care receipt of meningioma patients. With such approaches, alongside others quantifying the differential impact of multidimensional SDoH-factors on other oncologic disparities,^29–35 these large-data foundations bridge the gaps for providing targetable, pertinent SDoH-vulnerabilities that prospective interventions can focus on with limited public health resources. For instance, radiologic diagnostic services have long debated about investing in either low SES or higher minority-race/ethnic populations.^36,37 With these findings showcasing how race-ethnicity largely dictated meningioma care service disparities, investigators and funders can focus on trialing in those specific vulnerable patient communities, hopefully with the goal of no longer observing these differences in the near future.

Despite the strengths of this study, limitations are consistent with other retrospective analyses of large patient datasets. For example, the Yost Index and the specialized SEER dataset do not fully encompass all possible SDoH-factors that would be of interest to investigators. These unaccounted-for cofounders would also potentially affect the adjustor variates in causal-mediator models, which would obscure the observed mediative effects between race/ethnicity and community-level SES on outcome differences. Also inherent to large databases, there are limitations to capturing all clinical variables of interest, such as the specific CPT procedural-diagnostic codes or operative details such as extent of resection beyond treatment receipt values here. As a result, future works should incorporate paid/proprietary datasets that incorporate billing code information on a per-patient level while also retaining census-tract and individual-level characterizations of SDoH-factors.

Funding

This study was not funded.

Disclosures / Conflicts of Interest:

Mr. Fei-Zhang has performed advisory work for Nightingale Open Science outside the submitted work. Dr. Chelius has received a stipend from the American-Academy of Otolaryngology outside the submitted work. No other current or potential conflicts of interest were reported.

Author Contributions:

Concept and design: Fei-Zhang, Sethia, Sheyn, D’Souza, Chelius, Rastatter

Data acquisition and analysis: Fei-Zhang

Data interpretation: All authors.

Drafting of the manuscript: Fei-Zhang, Sethia, Wang

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Fei-Zhang.

Supervision: Sheyn, D’Souza, Chelius, Rastatter

Data Availability

Data analyzed was queried from the publicly available, SEER database, and can be made available upon reasonable request.

Ethics Statement

The Northwestern University IRB/ethics committee has exempted this study due to the data queried consisting of publicly available, de-identified data. No consent to participate was necessitated due to the nature of this study comprising retrospective analyses of a publicly available, de-identified national database.

Consent for publication

The authors of this manuscript consent to the accuracy of the contents of this manuscript and approve for its submission to be published if accepted.

Disclaimer

The content is solely the responsibility of the authors and does not represent the official views of the National Cancer Institute of the United States.