Poverty, Insurance, and Region as Predictors of Epilepsy Treatment among US Adults

Abstract

Disparities in epilepsy treatment have previously been reported. In the current study, we examine the role of socioeconomic status, health insurance, place of residence, and sociodemographic characteristics in past-year visit to a neurology or epilepsy provider and current use of antiseizure medications. Multiple years of data were compiled from the National Health Interview Surveys, Sample Adult Epilepsy Modules. The sample (n = 1,655) included individuals 18 years and older who have been told by a doctor to have epilepsy or seizures. Independent variables included number of seizures in the past year, health insurance, poverty status, education, region, race/ethnicity, foreign-born status, age, and sex/gender. Two sets of weighted hierarchical logistic regression models were estimated predicting past-year epilepsy visit and current medication use. Accounting for recent seizure activity and other factors, uninsured and people residing outside of the Northeast were less likely to see an epilepsy provider and people living in poverty were less likely to use medications, relative to their comparison groups. However, no racial/ethnic and nativity-based differences in specialty service or medication use were observed. Further research, including longitudinal studies of care trajectories and outcomes, are warranted to better understand health care needs of people with epilepsy, in particular treatment-resistant seizures, and to develop appropriate interventions at the policy, public health, and health system levels.

1. Introduction

According to the U.S. Institute of Medicine (IOM), there are significant social barriers to optimal care and health outcomes for people with epilepsy (PWE) [1]. The IOM cites a range of social factors at play - from gaps in access and provision of care to stigma and discrimination against PWE. Social factors in epilepsy are sometimes examined from the “burden of disease” perspective; the impact of epilepsy has been described as having epidemiological, socioeconomic, and clinical dimensions [2]. Other literature describes specific disparities in epilepsy, including gaps in medical and surgical treatment, disability, incidence and prevalence, and knowledge and attitudes [3]. These disparities have been linked to racial and ethnic background, economic resources, and access to health care [3–5], but the extent to which each of these factors currently shapes epilepsy care and treatment among US adults has not been fully established. This study attempts to address this gap. This knowledge is essential and identifies potential points of intervention at the policy, public health, and health care system levels. Social factors, not only clinical, need to be addressed in order to improve care and outcomes in PWE [1].

1.1. Conceptual framework

Our research was guided by the Social Determinants of Health in Epilepsy (SDH-Epilepsy) conceptual model, derived from the Social Determinants of Health (SDH) Framework [4, 6] and US-based health care disparities frameworks [7]. Per SDH-Epilepsy model, the social determinants of health in epilepsy fall into two categories: structural and intermediary. The structural determinants include the socioeconomic and political context of epilepsy (epilepsy-related public policy and culture/values) and sociodemographic characteristics of PWE, such as social class or socioeconomic status (SES), gender, race, and ethnicity. The structural factors shape epilepsy care and care disparities directly and through mediating factors, including living conditions, behaviors, and individual psychosocial characteristics (e.g., self-perceptions, internalized stigma, or social support).

There are several aspects of health care disparities that are unique to and deserve attention in the US context. First, access to health insurance is not universal and not a right in the US, as it is in other advanced industrial societies [8, 9]. Despite the passing of the Patient Protection and Affordable Care Act (ACA) in 2010, access to health care for adults is still largely tied to employer-based insurance, and out-of-pocket health care expenses are substantial, especially among the elderly and disabled [10]. Thus, having insurance is key to access and quality of care received. Second, availability, pricing, and quality of health care services varies widely by geographic area and is more restricted in rural versus urban locations, making a place an important factor in health care [11, 12]. Third, racial and ethnic minorities have more limited access to health care and receive lower quality of care compared to non-Hispanic whites [7]. Physician groups stress differences in geography, lack of access to adequate health coverage, communication difficulties between patient and provider, cultural barriers, provider stereotyping, and lack of access to providers as explanations for racial and ethnic disparities in health care [13]. Other scholars point directly to existing unconscious discrimination, so-called implicit bias [14–16], and continuing social inequality due to structural racism [17–19] as overarching causes of disparities.

1.2. Current literature

1.2.1. Economic resources and health insurance

Studies have documented that socioeconomic deprivation increases the incidence and prevalence of epilepsy; that PWE have lower education, household income, and health status than the healthy population; and, that finding employment is difficult for PWE [20, 21]. Further research has indicated that these relationships are fairly complex. A study using the SDH framework [21] has found no difference in the prevalence of epilepsy based on poverty status (<200% Federal Poverty Level). However, controlling for material resources (annual income, housing situation), poverty was associated with antiseizure medication nonadherence. This association was explained after accounting for insurance, underscoring the importance of access to epilepsy care and medication coverage for PWE. Other research has also shown a significant association between poor compliance and lower economic resources and insufficient insurance [20]. Another large study [22] using multiple data sources documented that uninsured individuals and those with public insurance programs had significant gaps in access to specialized epilepsy services. In that study, age, race/ethnicity, and the presence of comorbid conditions were also associated with disparities in access to specialized care in PWE.

1.2.2. Race, ethnicity, and immigrant status

Research suggests that epilepsy care and outcomes differ by race and ethnicity [5, 20, 23, 24]. For example, African American PWE have been shown to have higher rates of hospitalizations, ER visits, lower rates of surgery, and more deaths after surgery than their white counterparts. However, to what extent economic resources and health insurance explain this difference remains unclear. In a national study using the Medical Expenditure Panel Survey (MEPS), Blacks were nearly 30% less likely than whites to see an outpatient neurologist for a neurologic condition, even after adjustment for demographic, insurance, and health status differences [25]. However, seizure control did not differ by race, gender, or insurance in analyses of national survey data [26, 27] and in an inner-city sample of PWE [28]. Additional research has shown that another minority group, older Native Americans, are less likely than whites to see neurology providers for their seizures [29]. Some cultural groups (e.g., Hispanic/Latino) may hold traditional beliefs about epilepsy, which may prevent them from pursuing Western medical treatments [5, 30]. Furthermore, immigrants to the United States face unique challenges, above knowledge and attitudes about epilepsy, such as a language barrier or lack of understanding of Western medicine/health systems which may limit their access to epilepsy care or treatments, resulting in poorer outcomes [30].

1.2.3. “Place” and contextual factors

There is an emerging emphasis on the structural and sociocultural SDH because social and health policies and programs can modify certain SDH at the country and local levels [6]. Neighborhood conditions (e.g., community SES, safety, availability of health services) have been linked to health and disparities, but there is a dearth of such data in epilepsy. One study [20] found that PWE were less likely than others to report their neighborhood as safe. This may reflect lower-quality material/living conditions as well as social stigma and isolation experienced by PWE in their communities. A broader contextual factor is geographic place. There are regional differences in epilepsy epidemiology and services based on geography within countries [2]. Notably, the US South has high burden of disease including epilepsy and has recently been referred to as “the Epilepsy Belt” (analogy to the “Stroke Belt”) [29]. US-based national survey data also show the Northeast advantage vis-à-vis other regions in terms of both epilepsy specialist use and seizure control [27]. Some research has also indicated urban-rural differences in epilepsy/epilepsy care, but the evidence is inconclusive [20]. Others have examined the role of area-level income and deprivation in pediatric epilepsy populations and found deprivation to be particularly salient in health resource utilization, even within a universal health insurance context (Canada) [31].

The macro-structural domains of importance include health policies, budgets, health care and treatment regulation, and insurance systems. Economic burden of epilepsy is significant and includes health care system and individual costs [2]. The incidence and prevalence costs in the United States have been estimated at over $11 billion and $12.5 billion, respectively [32]. Effective prevention and treatment of epilepsies are needed to reduce these costs. Proposed solutions include closing gaps in epilepsy care, strengthening policies (e.g., anti-discrimination laws), and supporting epilepsy advocacy [1].

1.2.4. Hypotheses

Based on the past theory and research, we hypothesized that, among US adults with epilepsy, people with fewer economic means (e.g., living in poverty), uninsured, racial/ethnic minorities, and immigrants would be less likely to regularly visit an epilepsy provider and use antiseizure medications than PWE who have more economic resources and health insurance and who are white and US-native. In addition, based on regional variations in the supply of neurologists [33], we hypothesized that PWE living in the South (using region as a proxy/umbrella for “place”) would be less likely to regularly visit an epilepsy provider and use antiseizure medications than PWE living in the Northeast.

2. Material and methods

2.1. Data

The National Health Interview Survey (NHIS) is an annual and nationally representative sample of the U.S. population conducted by the National Center for Health Statistics (NCHS). The main purpose of NHIS is to collect information about individual and household level indicators such as health-related indicators, behaviors, care, access, utilization, and special information regarding non-communicable diseases affecting the nation [34]. The NHIS data used here were downloaded through the Integrated Public Use Microdata Series (IPUMS) of the Minnesota Population Center. The IPUMS NHIS data are based on the Center for Disease Control and Prevention/CDC’s original data collected but is recoded for the purposes of studying characteristics of people within the context of families and co-residents [35].

2.2. Sample: Persons with epilepsy (PWE)

For this project, we limited the analysis sample to the subsample of the population who had experienced a seizure at some point in the past (n = 1,783). The data were extracted from the 2013, 2015, and 2017 NHIS Sample Adult (18 years and over) components. The NHIS epilepsy module that we used to create our sample is administered every other year. The following questions determined whether or not someone is considered a PWE: 1) “Have you ever been told by a doctor or other health professional that you have a seizure disorder or epilepsy?”; 2) “Are you currently taking any medicine to control your seizure disorder or epilepsy?”; 3) “Think back to last year about the same time. About how many seizures of any type have you had in the past year?” Several reports have been published using these data [26, 27, 36–38]. We extend the past research by using a hypothesis-driven approach, multi-year data, and predictive multivariable modeling (see section 2.4). After dropping 128 respondents due to missing information, our final sample size included 1,655 PWE. The study was deemed “Exempt” by the University of Alabama at Birmingham Institutional Review Board.

2.3. Measurement

The dependent variables reflecting epilepsy treatment and care included: currently taking antiseizure medications (yes/no) and having seen a neurologist or an epilepsy specialist in the past year (yes/no). The independent variables included number of seizures in the past year (none, one, two to three, four to ten, and more than ten) and the following social determinants of health care access and utilization: health insurance (private, public, other, and no insurance), poverty status (<200% Federal Poverty Level), education (no high school, high school, some college, college, and graduate school), geographic region (Northeast, Midwest, South, and West), race/ethnicity (non-Hispanic white, black, Hispanic, and other), foreign-born status (foreign-born vs. US-native), age (years), and sex/gender (female vs. male).

2.4. Analysis

Unweighted descriptive statistics for respondents who have epilepsy and who reported the number of seizures in the past were computed and reported in Table 1. The multivariable analyses were weighted using the NHIS-provided weights. Because our two outcomes were binary, we used logistic regression models for the analysis. Best practices for nonlinear models such as these include an examination of multiple transformations of model coefficients in order to confirm the statistical significance and relative magnitude of associations [39]. Thus, we present both odds ratios and predicted probabilities. Predicted probabilities have the advantage of showing the magnitude of differences between the probabilities of epilepsy care across values of covariates, whereas exponentiated coefficients (i.e., odds ratios) offer a more standard accounting of statistical significance. We used Stata 16 [40] for all data management and statistical modeling.

Table 1.

Unweighted descriptive statistics for respondents who have epilepsy or have experienced a seizure (n = 1,655)

	Mean or Proportion	Min	Max
Saw epilepsy specialist in the last year	0.40			0	1
Takes medication	0.52	0	1
Number of seizures in the last year		0	5
None	0.63
One	0.08
Two to three	0.10
Four to ten	0.09
Eleven or more	0.10
Insurance		0	3
Private insurance	0.40
Public insurance	0.34
Other insurance	0.17
No Insurance	0.09
In poverty	0.30	0	1
Educational degree		0	4
No high school	0.20
High school	0.28
Some college	0.32
College	0.13
Graduate school	0.07
Geographic region		1	4
Northeast	0.14
Midwest	0.23
South	0.37
West	0.25
Race/ethnicity		1	4
White, non-Hispanic	0.71
Black	0.15
Hispanic	0.10
Other	0.05
Age	49.12	18	85
Female	0.57	0	1
Foreign-born	0.06	0	1

Notes: Values reflect information pooled from the 2013, 2015, and 2017 surveys.

3. Results

Table 1 includes basic descriptive information for analysis variables. More than half of the sample (52%) was taking antiseizure medications, and 40% saw a neurologist or an epilepsy specialist in the past year. A total of 63% had no seizures while 10% had more than ten seizures in the past year. The majority (74 %) had either private or public health insurance, and 17 % had some other kind of insurance. Nine percent reported having no insurance. Thirty percent of the participants with epilepsy were living in poverty. A fifth had no high school education, 28% completed high school, 32% completed some college, 13% completed college, and 7% completed graduate school. The largest share of this sample (37%) lived in the South while the smallest share (14%) lived in the Northeast. The majority (71%) of the sample was white, 15% was black, and 10% was Hispanic (5% reported other race/ethnicity). The sample was 57% female, and the average age of the participants was 49 years (range 18 to 85 years).

Table 2 contains odds ratios from weighted logistic regression models predicting whether respondents saw a neurologist or epilepsy specialist in the past year. Results showed that, among the social determinants of epilepsy care, health insurance and geographic residence were both significantly associated with seeing a specialist. Even in the full model that included controls for number of seizures and sociodemographic characteristics (Model 5), the odds of seeing a specialist were significantly lower among respondents without insurance compared to respondents with private insurance (OR = 0.30). Likewise, compared to uninsured respondents, having either public or “other” insurance was significantly related to higher odds of seeing a specialist (2.6 and 3.7, respectively). Finally, people residing in the Midwest, the South, and the West were less likely (ORs of 0.54, 0.60, and 0.49, respectively) to visit an epilepsy specialist than people residing in the Northeast. Differences between the Midwest, the South, and the West were not statistically significant.

Table 2.

Odds ratios from weighted logistic regression models predicting whether respondent has seen a neurologist or epilepsy specialist in the past year

	Model 1	Model 2	Model 3	Model 4	Model 5
Number of seizures in past year
None (reference)
One	3.77***	3.69***	3.79***	4.14***	4.19***
	(1.11)	(1.09)	(1.11)	(1.22)	(1.23)
Two to three	4.76***	4.79***	4.89***	6.18***	6.22***
	(1.18)	(1.19)	(1.23)	(1.62)	(1.67)
Four to ten	4.06***	4.09***	4.14***	4.94***	5.02***
	(1.14)	(1.17)	(1.20)	(1.40)	(1.40)
Eleven or more	7.93***	8.01***	8.08***	10.57***	10.90***
	(2.25)	(2.31)	(2.30)	(3.21)	(3.31)
Age		1.00	1.00	0.99	0.99
		(0.00)	(0.00)	(0.01)	(0.01)
Female		0.72*	0.72	0.71*	0.70*
		(0.12)	(0.12)	(0.12)	(0.12)
Race/ethnicity
White (reference)
Black			0.76	0.84	0.85
			(0.19)	(0.22)	(0.22)
Hispanic			1.11	1.32	1.38
			(0.34)	(0.42)	(0.46)
Other			0.72	0.77	0.82
			(0.24)	(0.27)	(0.30)
Foreign-born			1.41	1.40	1.33
			(0.42)	(0.40)	(0.39)
In poverty				0.74	0.72
				(0.16)	(0.16)
Educational degree
None (reference)
High school				1.22	1.21
				(0.27)	(0.27)
College				1.65	1.68
				(0.46)	(0.47)
Graduate/professional degree				1.14	1.15
				(0.42)	(0.42)
Insurance type
Private insurance (reference)
Public insurance				0.79	0.78
				(0.19)	(0.19)
Other insurance				1.14	1.09
				(0.27)	(0.25)
No insurance				0.29***	0.30***
				(0.09)	(0.10)
Geographic region
Northeast (reference)
Midwest					0.54**
					(0.12)
South					0.60*
					(0.13)
West					0.49**
					(0.13)

Notes: Standard errors in parentheses

^*p < 0.05

^**p < 0.01

^***p < 0.001.

In Table 3, the patterns of associations were somewhat similar for medication use. First, number of seizures in the past year and seeing an epilepsy specialist were both associated with medication use in every statistical model. Second, people living in poverty were less likely to use medications (ORs of 0.59 and 0.61 in Model 5) than people not living in poverty. Third, having public insurance appeared to increase the odds of medication use when compared to private insurance (OR = 1.86); however, this association was not significant in supplemental models that excluded controls for seeing an epilepsy specialist.

Table 3.

Odds ratios from weighted logistic regression models predicting whether respondent takes medication for epilepsy

	Model 1	Model 2	Model 3	Model 4	Model 5
Number of seizures in past year
None (reference)
One	3.29***	3.35***	3.54***	3.45***	3.44***
	(1.02)	(1.10)	(1.19)	(1.26)	(1.25)
Two to three	4.41***	5.26***	5.40***	4.91***	4.88***
	(1.75)	(2.13)	(2.17)	(1.92)	(1.92)
Four to ten	4.34***	5.50***	5.50***	5.56***	5.54***
	(1.70)	(2.27)	(2.24)	(2.40)	(2.39)
Eleven or more	3.29***	3.35***	3.54***	3.45***	3.44***
	(1.02)	(1.10)	(1.19)	(1.26)	(1.25)
Saw epilepsy specialist	9.83***	10.72***	10.85***	11.37***	11.29***
	(1.91)	(2.18)	(2.24)	(2.41)	(2.42)
Age		1.02***	1.02***	1.02***	1.02***
		(0.01)	(0.01)	(0.01)	(0.01)
Female		0.55***	0.56***	0.59**	0.59**
		(0.10)	(0.10)	(0.11)	(0.11)
Race/ethnicity
White (reference)
Black			0.95	0.87	0.85
			(0.25)	(0.23)	(0.23)
Hispanic			0.60	0.56	0.55
			(0.20)	(0.19)	(0.19)
Other			0.74	0.72	0.72
			(0.36)	(0.36)	(0.36)
Foreign-born			1.01	1.04	1.04
			(0.36)	(0.37)	(0.37)
In poverty				0.61*	0.61*
				(0.14)	(0.14)
Educational degree
None (reference)
High school				0.62	0.63
				(0.16)	(0.16)
College				0.87	0.88
				(0.26)	(0.26)
Graduate/professional degree				0.54	0.55
				(0.22)	(0.23)
Insurance type
Private insurance (reference)
Public insurance				1.85*	1.86*
				(0.45)	(0.45)
Other insurance				1.64	1.62
				(0.46)	(0.45)
No insurance				1.27	1.27
				(0.47)	(0.47)
Geographic region
Northeast (reference)
Midwest					0.88
					(0.26)
South					1.02
					(0.28)
West					0.92
					(0.28)

Notes: Standard errors in parentheses

^*p < 0.05

^**p < 0.01

^***p < 0.001.

Notes: Values reflect information pooled from the 2013, 2015, and 2017 surveys.

Figure 1 presents differences in the predicted probabilities of epilepsy care between select values of model covariates. These probabilities reflect average marginal effects (AMEs) because all other covariates were held at their observed values during postestimation [39, 41]. Predicted probabilities were calculated in postestimation following the logistic regression models for taking medicine or seeing a specialist (see Model 5 in Tables 1 and 2 for model coefficients). If the difference in probability is negative, then the reference group has a larger probability of the outcome. A difference in probabilities is not significant at 0.05 if its confidence interval is overlapping 0. In the first panel, the largest significant differences in the probability of seeing a specialist are related to health insurance and geographic residence, whereas the largest differences in the probability of taking medicine are related to biological sex and poverty. Although statistically significant, the difference between those with some college and those with no degree in the probability of taking medicine has especially large confidence intervals that nearly intersect with 0.

Figure 1.

Differences in the predicted probabilities of epilepsy care between selected values of model covariates

Notes: Predicted probabilities were calculated in postestimation following the logistic regression models for taking medicine or seeing a specialist (see Model 5 in Tables 1 and 2 for model coefficients). If the difference in probability is negative, then the reference group has a larger probability of the outcome. A difference in probabilities is not significant at .05 if its confidence interval is overlapping 0.

4. Discussion

This study has documented several disparities in visits to an epilepsy provider and antiseizure medication use in the US sample of adult PWE. First, it is alarming that a large proportion of PWE continue to experience reoccurring seizures, considering a broad array of advanced treatment options currently available. Social factors are likely responsible for some of this gap in treatment. In our study, women, uninsured, and people residing outside of the Northeast were less likely to visit an epilepsy provider in the past year compared to their male, insured, and Northeast-based counterparts, after accounting for recent seizure activity. In addition, poverty was the key social predictor of antiseizure medication use among adult PWE. However, racial/ethnic and nativity based differences were not observed in this sample before or after adjustments for socioeconomic factors and health insurance.

Our findings largely collaborate and extend findings from other studies, including prior analyses of the NHIS data [26, 27, 38]. While the 2010 NHIS estimates indicated no differences in epilepsy visits in the past 12 months by age, sex, or race/ethnicity [26], the data for 2013 and 2015 showed some differences in a specialist visit by age group and education, and in seizure control by age, poverty level, marital status, and employment. Our analysis, which used several years of the NHIS data and predictive multivariable modeling, confirmed an independent association of age with medication use but not with a specialist visit, after adjustments for other factors. We also confirmed the lack of association for race/ethnicity and demonstrated the same for foreign-born status. In addition, our findings provide further information about the role of insurance and poverty in the use of epilepsy specialists and antiseizure medications. Past research has documented type of insurance being a barrier to specialized care in PWE [22]. In another study [42] poverty was associated with antiseizure medication nonadherence, but health insurance status explained this relationship. In our analysis, once both factors were taken into account, insurance was a key social predictor of seeing a specialist while poverty was a key barrier in medication use. Thus, in our study the association between antiseizure treatment and poverty extended over and beyond insurance status, indicating that not only access to care but poverty effects more broadly (e.g., distance and transportation barriers) restrict opportunities for quality care and treatment among PWE. Our study was hypothesis-driven and our data and modeling approach were different from those used in the past research, which may account for the differences in findings between our study and the earlier reports.

It is somewhat surprising not to see gaps in epilepsy visits and antiseizure medication use by racial/ethnic minority or immigrant status in the NHIS data. Epilepsy is more prevalent among minorities [5, 43], and disparities in health care by race/ethnicity and immigrant status are pervasive [7, 44]. Therefore, such disparities in epilepsy care can also be expected. Socioeconomic status could explain some of the racial/ethnic disparities in health care utilization [45], but in our study no racial/ethnic or nativity-based differences were observed even before adjustment for poverty, education, and health insurance. However, insurance rates have significantly increased for racial/ethnic groups after the passage of the ACA, reducing disparities in coverage [46], and possibly bringing overall improvements in access to epilepsy treatment. While gaps in access and health care use might have been reduced or closed, there are lingering questions about quality of care experienced by minorities, especially in the light of recent work on the role of structural (institutional) racism (processes of racism embedded in laws, policies, and practices) and implicit bias in health care [19, 47]. Further research is needed to understand the role of race/ethnicity and immigrant status in epilepsy and epilepsy care.

As expected, we observed a higher proportion of epilepsy cases in the South than in the other US regions, similar to the 2018 NHIS report [27]. Another study based on Medicare data [29] has also documented a relatively high burden of epilepsy in the Deep South and labeled the region as “the Epilepsy Belt.” Our study complements the past research by showing that the regional differences persist after all adjustments for seizure activity and sociodemographic characteristics. These findings are consistent with a previous analysis of supply and demand for neurologists nationally and state-by-state [33]. At the national level, over 1,800 more neurologists are needed to meet the demand, and this is reflected in state-by-state estimates: the demand for neurologists in the majority of the states was estimated at 20% or higher than supply, with additional states showing a 6–19% gap. Few states, all but one in the Northeast region (MA, MD, MN, NH, NY, PA, and RI), and the District of Columbia had a supply of neurologists greater than the demand. The Northeast is known for a high concentration of medical centers in urban areas and high quality of care while other regions experience a greater shortfall of facilities and specialists, including specialized epilepsy services. Some Northeastern states (e.g., Massachusetts) also have more robust health insurance systems compared to states in other parts of the country, offering PWE residing in those states better access to epilepsy services compared with some other states [48, 49]. In addition, the Northeast ranks highly on quality of care indicators [50]. Thus, where PWE reside (“place”) makes a difference for their epilepsy care, over and beyond other factors, including type of individual health insurance. Further research is recommended to study contextual social factors in more detail (e.g., distribution of facilities/providers, transportation resources, neighborhood conditions, etc.).

This study has several limitations. The NHIS is a cross-sectional survey and does not allow to track people’s care and treatment over time, and over time in relation to other factors (e.g., SES or place of residence). The survey also relies on self-reported data, which might be subject to recall and social desirability bias. However, findings about similar data in other surveys suggest that reporting bias is small [26]. Further, some cases of epilepsy in the NHIS could be missed or misspecified (e.g., non-epileptic seizures or convulsive syncope reported as epilepsy). In addition, assessment of epilepsy care and treatment is limited in the NHIS. Finally, our study focused on a subset of social determinants of epilepsy treatment; other social statuses not included in our study (e.g., marital status and employment) may provide further explanation for variations in epilepsy specialty service and medication use [27].

Despite these limitations, the study contributes to better describing socially-based variations in two aspects of epilepsy treatement: use of epilepsy specialized services and antiseizure medication use. Information from this study can guide health and disability policies, public health programs, and health care delivery systems to strengthen resources and access to care/treatment for PWE, especially for people with treatment-resistant seizures. Engaging patients/families in policy and program development, as well as research, is also essential for further understanding of the needs of this population and opportunities for improvements.

Highlights.

• Treatment varies by income, insurance, and region, over and beyond other factors

• Poverty is associated with a lower likelihood of antiseizure medication use

• Uninsured are less likely to visit a neurology provider

• People residing in the Northeast are more likely to visit a neurologist

• Epilepsy treatment does not vary by race/ethnicity or immigrant status

Abbreviations:

IOM

Institute of Medicine

PWE

people with epilepsy

SDH

Social Determinants of Health

SES

socioeconomic status

ACA

Affordable Care Act

MEPS

Medical Expenditure Panel Survey

NHIS

National Health Interview Survey

NCHS

National Center for Health Statistics

IPUMS

Integrated Public Use Microdata Series