Mortality in people with epilepsy: A statewide retrospective cohort study

Epilepsy Research 122 (2016) 7–14

Contents lists available at www.sciencedirect.com

Epilepsy Research journal homepage: www.elsevier.com/locate/epilepsyres

Mortality in people with epilepsy: A statewide retrospective cohort study Dulaney A. Wilson a,∗ , Angela M. Malek a , Janelle L. Wagner b , Braxton B. Wannamaker c , Anbesaw W. Selassie a a

Medical University of South Carolina, Department of Public Health Sciences, Charleston, SC, United States Medical University of South Carolina, College of Nursing, Charleston, SC, United States c Medical University of South Carolina, Department of Neurology, Charleston, SC, United States b

a r t i c l e

i n f o

Article history: Received 25 June 2015 Received in revised form 18 January 2016 Accepted 28 January 2016 Keywords: Epilepsy Mortality Disparities

a b s t r a c t Rationale: People with epilepsy (PWE) have a higher risk of mortality than the general population, because of disparities in the receipt of appropriate epilepsy care, which may be affected by socioeconomic status, race/ethnicity and insurance coverage. Increased epilepsy prevalence has been associated with black race, low educational attainment, unemployment, and low income levels. Rural/urban residence may affect health through individual or environmental factors. Health disparities seen in rural residents are likely amplified in rural PWE because of limited access to specialized care. This analysis aims to examine the risk of mortality attributable to rural residence in the statewide population of South Carolina (SC) after adjusting for potential confounders. Methods: This statewide retrospective cohort study of PWE seen in SC non-federal hospitals and emergency departments from 2000 to 2013 describes the hazard of mortality by rural/urban residential status in addition to other demographic and clinical characteristics. Differences in proportions were assessed by comparison of 95% confidence intervals. The association of rural/urban residence with mortality was further evaluated with Cox proportional hazard regression controlling for demographic and clinical covariables. Results: 62,794 PWE were identified, of whom 21,451 (25.7%) had died. Deceased PWE were more likely to be rural residents, black, older than age 45, Medicare insured, in the middle income group, and have 5 or more comorbid conditions compared with living PWE. After adjustment for all other covariables, the risk of mortality did not differ by rural/urban residence. Blacks had a weak but significantly higher risk than whites (hazard ratio (HR) = 1.14; 95% confidence interval (CI) = 1.11, 1.18) while PWE of other races had a slightly lower risk of mortality (HR = 0.79; 95% CI = 0.67, 0.93). Male PWE had higher hazard as did Medicare, Medicaid or commercially insured PWE, those living in zip codes with annual median incomes less than $36,000, and those with 2 or more comorbid conditions. Conclusions: While other covariables were more strongly associated with mortality after adjustment (older age, insurance coverage, income level of zip code, and number of comorbidities), the finding of a higher hazard in black PWE than white PWE after adjustment for rural/urban residence and other demographic and clinical covariables is a concern. Further, the increased risk of mortality with higher numbers of comorbid conditions warrants regular management of these conditions. © 2016 Elsevier B.V. All rights reserved.

1. Introduction

∗ Corresponding author at: Medical University of South Carolina, Department of Public Health Sciences, 135 Cannon Street, Suite 303, MSC 835, Charleston, SC 29425, United States. E-mail address: [email protected] (D.A. Wilson). http://dx.doi.org/10.1016/j.eplepsyres.2016.01.008 0920-1211/© 2016 Elsevier B.V. All rights reserved.

The quality of and access to health care has been the focus of rural health research; however, health disparities can involve a complex interaction of factors including race/ethnicity, age, sex, educational level, income level, insurance status and place of residence (Burneo et al., 2009; Hartley, 2004; Johnson et al., 2008; Marmot, 2005; Scott and Wilson, 2011; Smedley et al., 2001; Szaflarski, 2014) Healthy People 2020 lists geographic region

8

D.A. Wilson et al. / Epilepsy Research 122 (2016) 7–14

(rural/urban residence) as an important social determinant of health (Healthy People 2020, 2015) Rural/urban residence may affect health through individual (use of self-care, no routine source of care, lifestyle and behavior) or environmental (poverty, income inequality, access to care, health care shortages) factors (James, 2014). Rural residents tend to be older, report poorer health, take limited physical activity, be overweight/obese, have higher rates of smoking, fewer health care visits and are more likely to see a general practice physician rather than a specialist (Bethea et al., 2012; Chan et al., 2006; Meit et al., 2014) Rural counties have the lowest numbers of active specialty physicians or dentists per 100,000 population; 65% of rural counties in the U.S. are considered health professional shortage areas (HPSA) (Meit et al., 2014; Probst et al., 2004) In the Southern U.S., 22% of rural residents live below the poverty line (Meit et al., 2014). An analysis of mortality from 1969 to 2009 found increasing disparities in mortality by rural/urban residence with higher mortality seen in rural areas after adjusting for poverty (Singh and Siahpush, 2014a) Further, rural residents with chronic diseases requiring long-term care have shown increased mortality (Eberhardt and Pamuk, 2004; Hill et al., 2014; Kulshreshtha et al., 2014; Terashima et al., 2014). Health disparities, both in healthcare access and outcomes have been documented in people with epilepsy (PWE). Incidence and prevalence of epilepsy are higher in blacks (Faught et al., 2012; Kroner et al., 2013; Theodore et al., 2006) and higher epilepsy prevalence is associated with indicators of low socioeconomic status (SES) such as low educational attainment, unemployment, and low income levels (Elliott et al., 2009, 2008; Ferguson et al., 2008; Geerts et al., 2011; Heaney et al., 2002; Hesdorffer et al., 2005; Kobau et al., 2004, 2006, 2007; Konda et al., 2009; Kroner et al., 2013; Li et al., 2008; Ottman et al., 2011; Pickrell et al., 2015; Sillanpaa, 2004; Steer et al., 2014; Wiebe et al., 2009). Worldwide, epilepsy prevalence appears to be higher in rural areas, especially in underdeveloped countries (Camfield and Camfield, 2015; Gourie-Devi et al., 2004). PWE, especially those with remote symptomatic or intractable epilepsy, are acknowledged to have a higher risk of premature mortality than the general population, especially in the first years after diagnosis (Forsgren et al., 2005; Gaitatzis et al., 2004; Hitiris et al., 2007; Lhatoo et al., 2001; Morgan et al., 2000; Neligan et al., 2011; Nevalainen et al., 2014; Selassie et al., 2014; Theodore et al., 2006; Trinka et al., 2013; Wiebe et al., 2009). Disproportionate mortality in subpopulations of PWE may result from disparities in the receipt of appropriate epilepsy care, which has shown to be affected by income status (Begley et al., 2011; Gresenz et al., 2000), racial/ethnic group (Kelvin et al., 2007; McClelland et al., 2010; Schiltz et al., 2013), and insurance coverage (Baca et al., 2013; Halpern et al., 2011; Hauptman et al., 2013a; Schiltz et al., 2013). Health disparities associated with rural residency are likely amplified in PWE because of barriers to education and employment (Steer et al., 2014). The few studies of geographic disparities in PWE have focused on healthcare factors including access to and use of epilepsy providers, diagnostic tools (e.g., EEG), and treatments (e.g., AEDs and surgery). In a study of pediatric epilepsy surgery patients in California, the time from seizure onset to epilepsy surgery evaluation was shorter for those living closer to the treating facility (Hauptman et al., 2013b). In a Canadian study, rural PWE used emergency department services more frequently and dental services less frequently than those in more heavily populated regions (Wiebe et al., 2009). The risk of mortality in PWE is likely to be influenced by a cumulative effect of SES over the lifetime (Hesdorffer et al., 2005); adjustment for one or two indicators of SES may not account for the complex relationship of these determinants of

health (Lawlor et al., 2005). This analysis aims to examine the risk of mortality attributable to rural residence in the statewide population of South Carolina (SC) after adjusting for potential confounders. 2. Methods 2.1. Data sources Data analyzed in this study derive from hospital discharge and emergency department (ED) visit datasets (including hospitalbased outpatient department (OPD) visits) from all non-federal hospitals in SC. SC health care providers are required to submit these data to the SC Revenue and Fiscal Affairs Office, Health and Demographics (H&D) section which assigns a unique ID to each individual, allowing linkage across various datasets (Weis et al., 2006). Since these data are required for billing, data completeness is over 98% (Varma et al., 2010). Additional data on mortality was obtained from death certificate information maintained by the SC Public Health Statistics and Information Services. The Medical University of South Carolina Institutional Review Board approved this study. 2.1.1. Study design and case ascertainment The design is a retrospective cohort study of mortality in PWE in SC. All individuals with hospital, ED, or OPD visits from January 1, 2000 to December 31, 2013 with International Classification of Diseases, 9th revision, Clinical Modification (ICD-9-CM) diagnosis codes for epilepsy (ICD-9-CM 345.0, 345.1, 345.4–345.6, 345.8, or 345.9) or “seizure, not otherwise specified” (ICD-9-CM 780.39) were identified (American Medical Association, 2009). Diagnoses assigned by clinical providers are coded by certified health information specialists operating under the standard coding guideline of National Center for Health Statistics and the Center for Medicare and Medicaid Services. Case ascertainment was based on the aforementioned ICD-9-CM coded diagnoses of epilepsy or unspecified seizures using an operational version (Fig. 1) of a decision algorithm (Appendix A) developed for the South Carolina Epilepsy Surveillance System that uses clinical and diagnostic criteria of epilepsy to increase the probability of identifying “true” epilepsy (Selassie, 2012). Briefly, a case of epilepsy had to satisfy one of the following requirements: (1) at least two visits with a diagnosis code for epilepsy during the study period; (2) a single visit coded for epilepsy PLUS a prior visit coded for unspecified seizure; (3) two or more visits coded for unspecified seizure within one year or (4)

All individuals with either ICD-9-CM codes 345.x or 780.39 At least two visits coded with 345.x Yes

No Yes

1 visits for 345.x PLUS prior visit coded 780.39 No

Yes

2 or more visits coded 780.39 within 1 year

Yes

Single visit coded 780.39 PLUS code for indicators of epilepsy* No

Epilepsy

No

Not Epilepsy

*Indicators include: codes for vagal nerve stimulator implantation, epilepsy surgery, ketogenic diet in children, video EEG.

Fig. 1. Operational algorithm for classifying epilepsy based on diagnosis codes.

D.A. Wilson et al. / Epilepsy Research 122 (2016) 7–14

a single visit coded for unspecified seizure plus other procedures suggestive of epilepsy such vagal nerve stimulator implantation, epilepsy surgery, ketogenic diet in children and adolescents age 18 and younger or video EEG. 2.1.2. Inclusion and exclusion Of the 148,705 individuals with an ICD-9-CM diagnosis code for epilepsy or seizure, 72,705 met our operational definition of epilepsy. Out of state residents and individuals without residency information were excluded (7753). Also excluded were 2158 observations with missing or incorrect data elements required to calculate follow-up time (such as a date of death before the first epilepsy diagnosis). After exclusions, the cohort consisted of 62,794 SC PWE. 2.1.3. Definitions The outcome of interest was mortality from any cause, based on SC death certificate data through December 31, 2013. The primary independent variable of interest, rural/urban residence, was defined using Rural Urban Commuting Area (RUCA) codes specific to zip codes, the smallest geographic area available in these data (Morrill et al., 1999; USDA Economic Research Service, 2011; WWAMI Rural Health Research Center, 2015a). RUCA codes range from 1 to 10 and incorporate measures of urbanization, population density and commuting to urban areas. As individuals may have lived in multiple zip codes over the study period, we determined the RUCA designation for each individual based on the most frequent zip code reported for that individual. Individuals were

9

further grouped on the basis of a dichotomous RUCA code aggregate (urban = RUCA 1.0, 1.1, 2.0, 2.1, 2.2, 3.0, 4.1, 5.1, 7.1, 8.1, 10.1; rural = all other RUCA) (WWAMI Rural Health Research Center, 2015b). Characteristics of PWE considered to be social determinants of health were evaluated for inclusion in the model. These include race/ethnicity (grouped as white, black and other, including Hispanic), sex, age at the earliest visit for epilepsy or seizure, median income of the zip code of residence, and insurance status (classified as commercial, Medicare, Medicaid and uninsured). In the U.S., insurance status is linked to socioeconomic status (SES). Individuals using Indigent Care programs and Medicaid HMOs were expected to be of similar SES and were grouped with Medicaid. Uninsured individuals included those listed as self-pay. Person-years at risk were calculated from the date of the earliest visit for epilepsy or seizure to date of death or the censoring data of December 31, 2013 for those known to be alive. Income was assigned based on the median income of zip code of residence. Using guidelines from the U.S. Department of Housing and Urban Development, income was categorized based on a percentage of the SC median household income: (1) under $36,000 (80%), (2) $36,000 to $54,000 (>80–120%), and (3) $54,000 and over (≥120%) (U.S. Department of Housing and Urban Development). Clinical characteristics evaluated include the number of visits with an epilepsy or seizure code as the primary diagnosis, and the occurrence of one or more of 36 common comorbidities of epilepsy noted in the literature (listed in Appendix B) (Rai et al., 2012; Thurman et al., 2011). As a measure of epilepsy severity, the proportion of total visits with a diagnosis

Table 1 Demographic and clinical characteristics of SC PWE by mortality status.

Race White Black Other Sex Female Male Age at first diagnosis 0–5 6–18 19–45 46–64 ≥65 Mean ± std. Median (IQR) Residence Urban Rural Payer Uninsured Medicare Medicaid Commercial Income (thousands of US$) High (≥54) Middle (36–54) Low (<36) Mean ± std. Median (IQR) Proportion of visits with epilepsy or seizure diagnosis <50% ≥50% Common comorbidities of epilepsy 0–1 2–4 ≥5 Mean ± std. Median (IQR)

Deceased PWE N = 21,451 (25.7%) n (%; 95% CI)

Living PWE N = 41,343 (49.6%) n (%; 95% CI)

Total N = 62,794 n (%; 95% CI)

12,264 (57.2; 56.5–57.8) 9037 (42.1; 41.5–42.8) 150 (0.7; 0.6–0.8)

25,110 (60.7; 60.3–61.2) 14,967 (36.2; 35.7–36.7) 1266 (3.1; 2.9–3.2)

37,374 (59.5; 60.2–60.8) 24,004 (38.2; 36.9–37.5) 1416 (2.3; 2.2–2.4)

10,349 (48.2; 47.6–48.9) 11,102 (51.8; 51.1–52.4)

20,301 (49.1; 48.6–49.6) 21,042 (50.9; 50.4–51.4)

30,650 (48.8; 50.1–50.7) 32,144 (51.2; 49.3–49.9)

232 (1.1; 0.9–1.2) 353 (1.6; 1.5–1.8) 3217 (15.0; 14.5–15.5) 7165 (33.4; 32.8–34.0) 10,484 (48.9; 48.2–49.5) 61.8 ± 18.5 64.0 (26.0)

5745 (13.9; 13.6–14.2) 6041 (14.6; 14.3–15.0) 14,992 (36.3; 35.8–36.7) 9882 (23.9; 23.5–24.3) 4683 (11.3; 11.0–11.6) 35.1 ± 23.0 35.0 (36.0)

5977 (9.5; 7.9–8.3) 6394 (10.2; 10.2–10.6) 18,209 (29.0; 31.6–32.2) 17,047 (27.1; 27.2–27.8) 15,167 (24.2; 21.8–22.4) 44.2 ± 25.0 46.0 (40.0)

13,912 (64.9; 64.2–65.5) 7539 (35.1; 34.5–35.8)

28,575 (69.1; 68.7–69.6) 12,768 (30.9; 30.4–31.3)

42,487 (67.7; 67.8–68.4) 20,307 (32.3; 31.6–32.2)

619 (2.9; 2.7–3.1) 12,916 (60.2; 59.6–60.9) 4021 (18.7; 18.2–19.3) 3895 (18.2; 17.6–18.7)

5408 (13.1; 12.8–13.4) 10,179 (24.6; 24.2–25.0) 14,572 (35.2; 34.8–35.7) 11,184 (27.1; 26.6–27.5)

6027 (9.6; 10.3–10.7) 23,095 (36.8; 34.9–35.5) 18,593 (29.6; 29.7–30.3) 15,079 (24.0; 24.0–24.6)

2401 (11.2; 10.8–11.6) 8194 (38.2; 37.5–38.8) 10,856 (50.6; 49.9–51.3) 40.9 ± 11.7 38.5 (15.7)

5059 (12.2; 11.9–12.6) 13,355 (32.3; 31.9–32.8) 22,929 (55.5; 55.0–55.9) 42.9 ± 12.1 41.1 (16.3)

7460 (11.9; 53.5–54.1) 21,549 (34.3; 34.0–34.6) 33,785 (53.8; 11.7–12.1) 42.2 ± 12.0 41.0 (16.0)

p-Value

<0.001

0.0411

<0.001

<0.001

<0.001

0.2179 13,182 (61.5; 60.8–62.1) 8269 (38.5; 37.9–39.2)

25,197 (60.9; 60.5–61.4) 16,146 (39.1; 38.6–39.5)

38,379 (61.1; 64.7–65.3) 24,415 (38.9; 34.7–35.3)

928 (4.3; 4.1–4.6) 6066 (28.3; 27.7–28.9) 14,457 (67.4; 66.8–68.0) 6.2 ± 3.1 6.0 (4.0)

11,283 (27.3; 26.9–27.7) 14,108 (34.1; 33.7–34.6) 15,952 (38.6; 38.1–39.1) 4.0 ± 3.3 3.0 (5.0)

12,211 (19.4; 20.1–20.7) 20,174 (32.1; 31.5–32.1) 30,409 (48.4; 47.5–48.2) 4.8 ± 3.4 4.0 (5.0)

<0.001

10

D.A. Wilson et al. / Epilepsy Research 122 (2016) 7–14

of epilepsy or seizure was calculated and grouped into (1) less than 50th percentile and (2) 50th percentile or more. The number of comorbidities for each individual were enumerated and grouped as 0–1, 2–4, and 5 or more. 2.1.4. Statistical analysis Data were analyzed with the SAS software package, V9.4 (SAS Institute, 2012). The associations of demographic and clinical characteristics with all-cause mortality were evaluated with chi square tests of homogeneity and independence (Table 1). Stratumspecific 95% confidence intervals (CI) were constructed for the independent variables under the assumption of independence and normal approximation. Overlapping CIs likely suggest no significant difference in stratum-specific proportions. Distributional assumptions of continuous variables were explored with numerical and graphical methods. The association of all-cause mortality with demographic and clinical variables was further evaluated with Cox Proportional Hazard (CPH) regression modeling. Variables considered to be potential confounders were entered simultaneously in the model. Covariables were assessed for potential confounding or interaction in the relationship between rural/urban residence and mortality. There was no evidence of interaction or multicollinearity, which was assessed with variance inflation factors. Unadjusted and adjusted hazard ratios (HR) with their corresponding 95% CIs are reported (Table 2). 3. Results The analytic cohort consisted of 62,794 PWE, of whom 21,451 (25.7%) had died by December 31, 2013 (Table 1). Deceased PWE showed significantly higher proportions of rural residents, black PWE, older than age 45, Medicare insured, in the middle income

Fig. 2. Unadjusted survival for rural/urban residency.

group, and had 5 or more comorbid conditions compared with living PWE. Fig. 2 shows the unadjusted Kaplan–Meier survival estimates by rural/urban residence. Rural PWE have significantly lower unadjusted survival (log-rank test for equality of strata p < 0.001). Table 2 shows the unadjusted and adjusted HRs for all-cause mortality by rural/urban residency in addition to other demographic and clinical characteristics. After controlling for other demographic and clinical covariables, rural residency was no longer a significant predictor of mortality (HR = 1.02; 95% CI = 0.99, 1.05). Blacks had a weak but significantly higher risk than whites (HR = 1.14; 95% CI = 1.11, 1.18) and Other races (HR = 0.79; 95% CI = 0.67, 0.93). The adjusted cumulative hazard of mortality by race is shown in Fig. 3. Males had ∼13% higher risk than females. After adjustment for other covariables, compared with the uninsured, the risk in Medicare and Medicaid insured increased by 96% and 74%, respectively, while commercially insured PWE had a 2-fold

Table 2 Unadjusted and adjusted hazard ratios for all-cause mortality in SC residents with epilepsy, 2000–2013. Unadjusted

Adjusted

HR Race White Black 1.12 0.34 Other Sex Female 1.01 Male Age at first diagnosis 0–5 1.41 6–18 4.44 19–45 12.70 46–64 28.51 ≥65 Residence Urban 1.14 Rural Payer Uninsured 6.56 Medicare 2.25 Medicaid 2.84 Commercial Income (thousands of US$) High (≥54) 1.16 Middle (36–54) 0.94 Low (<36) Proportion of visits with epilepsy or seizure diagnosis <50% 0.90 ≥50% Common comorbidities of epilepsy 0–1 4.27 2–4 6.66 ≥5

95% CI

HR

Referent

95% CI Referent

(1.09–1.15) (0.29–0.39)

1.14 0.79

(0.98–1.04)

1.13

Referent

(1.11–1.18) (0.67–0.93) Referent

Referent

(1.10–1.16) Referent

(1.19–1.66) (3.89–5.08) (11.14–14.47) (25.03–32.48)

1.37 3.87 9.16 19.80

Referent

(1.16–1.62) (3.38–4.43) (8.01–10.48) (17.29–22.67) Referent

(1.11–1.18)

1.02

(6.05–7.11) (2.07–2.45) (2.61–3.09)

1.96 1.74 2.13

(1.11–1.21) (0.90–0.98)

1.05 1.11

(0.88–0.93)

0.95

(3.98–4.57) (6.23–7.12)

1.94 2.12

Referent

(0.99–1.05) Referent

Referent

(1.80–2.14) (1.59–1.90) (1.95–2.32) Referent

Referent

(1.00–1.09) (1.06–1.17) Referent

Referent

(0.92–0.98) Referent (1.80–2.08) (1.98–2.28)

D.A. Wilson et al. / Epilepsy Research 122 (2016) 7–14

Fig. 3. Cumulative hazard of death by race/ethnicity.

higher hazard of mortality. The risk was higher in PWE residing in a zip code with low median income (less than $36,000) than PWE living in zip codes with higher median income. Compared to PWE with zero or 1 comorbidity, PWE with 2–4 comorbid conditions had 2-fold higher risk and PWE with 5 or more comorbidities had 2-fold higher risk. 4. Discussion This analysis provides a comprehensive statewide view of differences in mortality in PWE who visited an ED or hospital by rural/urban residence adjusted for median income of residential zip code, race, sex, age, and other demographic and clinical covariables. After adjustment, the mortality in SC PWE did not differ by rural/urban residence; as in other studies, demographic and clinical factors appear to explain geographic variability (Bethea et al., 2012; Steer et al., 2014). We did find that the adjusted risk of mortality was slightly but significantly elevated in black PWE and was decreased in PWE of other races compared with white PWE (Fig. 3). The increased risk in black PWE mirrors the higher age-adjusted mortality rate seen in black SC residents (1055 per 100,000) compared with white SC residents (847 per 100,000) (Centers for Disease Control and Prevention National Center for Health Statistics, 2014). This higher risk is potentially attributable to demographic factors as lower life expectancy has been seen in blacks living in poverty, residing in rural areas, and having low educational attainment (Probst et al., 2004; Singh and Siahpush, 2014a,b; Wong et al., 2002). For rural PWE, the proximity or availability of health care in general may not explain health disparities seen in SC; indeed, as a small state, even rural areas are not prohibitively distant from healthcare providers and facilities. However, the availability of specialized care for PWE may differ based on the economic characteristics of the areas. In SC most epileptologists and neurologists practice in an urban setting limiting the availability of specialized care for PWE; data from the SC Epilepsy Surveillance System from 2006 to 2010 show that only 22.8% of PWE received their care from these specialists (Selassie, 2012). In the U.S. 85% of rural counties with a black majority are HPSAs compared with 65% of all rural counties in the U.S. (Probst et al., 2004). A study of cancer screening rates found that significantly lower screening rates in blacks were attributable to the racial composition of counties; majority black counties had much lower rates of screening (Probst et al., 2004). Thus, the increased risk of mortality among black PWE is unlikely to be a direct result of overt differences in epilepsy care. Rather, socioeconomic factors may indirectly influence appropriate epilepsy care. The decreased risk seen in Hispanic/other PWE and uninsured PWE are likely due to the younger ages in those populations and the lack of death information

11

available on individuals not legally residing in the US. PWE living in areas with low median incomes (less than $36,000 or 80% of the SC median income) had a significant albeit weakly increased risk of mortality (HR = 1.11; 95% CI = 1.06, 1.17) than those in areas with higher median incomes. This is consistent with other studies that found low SES associated with higher mortality in the general population and in PWE (Kaiboriboon et al., 2014; Koroukian et al., 2012; Wong et al., 2002). The incidence and prevalence of epilepsy in the U.S. population was recently estimated at 0.79 per 1000 people and 8.5 per 1000 people, respectively (Helmers et al., 2015). In SC, the estimated incidence and prevalence of epilepsy was 0.95 per 1000 people and 8 per 1000 people, respectively (Selassie, 2012; Selassie et al., 2005). The estimated incidence from SC statewide data is higher than recent U.S. estimates but the difference may be attributable to socioeconomic factors known to be associated with higher epilepsy incidence (such as low income, low educational attainment, and unemployment) that are prevalent in SC. Identifying and ameliorating factors detrimental to health is important for improving health and reducing disparities in health and outcomes (Lawlor et al., 2005). Rural PWE encounter all of the risk factors common to rural residents compounded by factors associated with epilepsy. The relationship between rural residence and epilepsy is complicated by other factors such as race, education, occupation, and income level. For example, rural residents are more likely to smoke and PWE have higher smoking rates (Kobau et al., 2007; Meit et al., 2014). If rural PWE are more likely to smoke, health effects from smoking in PWE may obscure the relationship between epilepsy and mortality as will comorbid disease. For example, in these data, 27% of rural PWE have a diagnosis of diabetes compared with 24% of urban residents. When categorizing by rural residence and race, 21% of urban white PWE, 22% of rural white PWE, 31% of urban black PWE and 33% of rural black PWE had a diagnosis of diabetes. While we did not evaluate specific comorbidities in these analyses, we did see that PWE with two or more comorbid conditions had a higher risk of mortality than PWE with one or none. More comorbidities are associated with higher age (data not shown) and are expected to correspond with poorer health status, both of which increase mortality risk. 4.1. Strengths Our study has several strengths. First, data come from a legally mandated, multifaceted database of all hospital, emergency department and hospital-based outpatient encounters spanning 14 years with a unique ID allowing identification of individuals over time and avoiding reliance on self-report used in many epidemiological studies of epilepsy. Second, the case definition adhered to guidelines suggested by the ILEA for the use of ICD-9-CM codes in epidemiological studies to identify PWE and is similar to other recent studies of epilepsy (Kaiboriboon et al., 2014; Helmers et al., 2015; Schiltz et al., 2013). Third, the use of ICD-9-CM diagnostic codes in administrative data has been validated for identification of PWE (Jette et al., 2010; Kee et al., 2012; Reid et al., 2012; Selassie et al., 2005; Thurman et al., 2011). SC data from 2001 and 2002 were validated with medical record review. A diagnosis code of epilepsy (ICD-9-CM 345.x) was indicative of epilepsy in 95% of the cases while a diagnosis of unspecified seizure (ICD-9-CM 780.39) was indicative of epilepsy in ∼83% of cases (Selassie et al., 2005). Reimbursement changes in 2006 resulted in a significant increase in the use of ICD-9-CM codes for epilepsy (with a corresponding decrease in codes for unspecified seizure) (Cardenas et al., 2014). Anecdotal evidence from neurology and epileptology colleagues shows that, prior to the 2006 coding mandate, physicians were likely to assign a seizure diagnosis rather than an epilepsy diagnosis in order to avert

12

D.A. Wilson et al. / Epilepsy Research 122 (2016) 7–14

negative consequences for the patient such as loss of driving privileges, health insurance or employment. This coding shift is unlikely to reflect a change in incidence or prevalence but a move toward appropriate coding and thus, we expect that our decision algorithm will perform as well for later data. Third, the use of all non-federal encounters gives an overview of a statewide population. Fourth, SC is an ideal venue for the evaluation of social determinants of health and disparities in vulnerable subgroups of PWE such as minorities and rural residents because significant proportions of the population belong to those groups. The proportion of blacks in SC (29.6%) is more than double that of the US population. The proportion of black PWE in our analytic cohort was higher than in the SC population (38.2%). Of our analytic cohort, 32.3% resided in a rural areas; application of the RUCA designation previously described in the methods section to 2010 population estimates has 18.4% of the SC and 18% of the U.S. population residing in rural areas. 5. Limitations Despite the aforementioned strengths, there are important limitations worth noting. First, validation of diagnosis codes for the reliable identification of epilepsy in SC was conducted prior to the coding shift in 2006. As a result, the estimate provided in this report may have been distorted. However, individuals with a single visit coded for epilepsy (most likely to have been misclassified) were not included in the analysis. The remaining analytic cohort satisfies the most conservative ILAE definition of epilepsy. Second, in a small state like South Carolina, the classification of rural and urban residential areas may not be useful because of the relatively short distance to healthcare facilities. Third, there is residual confounding due to limited information on all potential covariables that may affect the relationship between rural residence, race/ethnicity and mortality in PWE. The administrative data does not provide information on medication, education level, employment, physical characteristics such as obesity or lifestyle factors

such as smoking or physical activity. Lack of information on individual income level was addressed by using the median income for each individual’s zip code as approximation of household income. Fourth, the use of surveillance data offers limited information to make a reasonable assessment of epilepsy type or severity. Further, the demographic distributions seen in the early years may be distorted due to the incremental case capture during the beginning phase of the epilepsy surveillance system. Fifth, the data do not include PWE who receive healthcare in federal hospitals and private physician practices, reducing the generalizability of the study. This limitation, however, applies to all public health data systems in the US. Sixth, statistically significant differences in mortality associated with a specific characteristic does not imply causation (Burneo et al., 2009). 6. Conclusion Using statewide data on all-cause mortality in PWE, we determined that the apparent increased risk of mortality in rural PWE is attenuated by adjustment for race and other demographic and clinical characteristics. While rural/urban residence was no longer significant, black PWE had 14% higher risk of mortality compared with whites. Further, higher numbers of comorbid conditions in PWE increase mortality, warranting the need to manage these conditions on a regular basis. Acknowledgments This research was supported by The Centers for Disease Control and Prevention (CDC), National Center for Chronic Disease Prevention and Health Promotion, Epilepsy Program Office (Grant 01DP003251). Appendix A.

D.A. Wilson et al. / Epilepsy Research 122 (2016) 7–14

Appendix B. ICD-9-CM codes used to define comorbid conditions Comorbidity Somatic Alzheimer’s Dementia Anemia Asthma/pulmonary disease Cardiovascular disease Celiac disease Cerebral palsy Cysticercosis Diabetes Gastric reflux GI bleed Hearing loss HIV/AIDS Intestinal problems Migraine Multiple sclerosis Nutritional deficiencies Onchocerciasis/toxocariasis Osteoporosis Parkinson’s disease Peptic ulcer Stroke Traumatic brain injury Vision loss Psychiatric Alcohol misuse Anxiety Depression Drug misuse Personality disorder Psychoses Schizophrenia Somatoform disorder Suicidal ideation/attempt Neurodevelopmental ADHD Autism spectrum disorder Cognitive dysfunction Intellectual disability

ICD-9-CM code(s) 331.0 280.1–281.9, 285.9 490–496 401–405, 410–417, 420–429 579.0 343 123.1 250.00–250.33, 250.40–250.73, 250.90–250.93 530.81 578 389 042, 044, V08 560–569 346 340 260–269 125.3, 128.0 733.0, V82.81 332 531–535 430–438 800, 801, 803, 804, 850–854, 959.01 369 305.00–305.03 300.0–300.7 300.4, 309.0, 309.1, 311 305.20–305.93 301 293.8, 296.0–298.9, 295, 299.10, 299.11 300.8 300.9, V62.84 314.0 299.0 315, V40.0 317–319

References American Medical Association, 2009. International Classification of Diseases, 9th revision. Clinical Modification Saunders, Chicago, IL. Baca, C.B., Vickrey, B.G., Vassar, S., Hauptman, J.S., Dadour, A., Oh, T., Salamon, N., Vinters, H.V., Sankar, R., Mathern, G.W., 2013. Time to pediatric epilepsy surgery is related to disease severity and nonclinical factors. Neurology 80, 1231–1239. Begley, C., Basu, R., Lairson, D., Reynolds, T., Dubinsky, S., Newmark, M., Barnwell, F., Hauser, A., Hesdorffer, D., 2011. Socioeconomic status, health care use, and outcomes: persistence of disparities over time. Epilepsia 52, 957–964. Bethea, T.N., Lopez, R.P., Cozier, Y.C., White, L.F., McClean, M.D., 2012. The relationship between rural status, individual characteristics, and self-rated health in the Behavioral Risk Factor Surveillance System. J. Rural Health 28, 327–338. Burneo, J.G., Jette, N., Theodore, W., Begley, C., Parko, K., Thurman, D.J., Wiebe, S., Task Force on Disparities in Epilepsy C, North American Commission of the International League Against E, 2009. Disparities in epilepsy: report of a systematic review by the North American Commission of the International League Against Epilepsy. Epilepsia 50, 2285–2295. Camfield, P., Camfield, C., 2015. Incidence, prevalence and aetiology of seizures and epilepsy in children. Epileptic. Disord. 17, 117–123. Cardenas, V.M., Roman, G.C., Perez, A., Hauser, W.A., 2014. Why U.S. epilepsy hospital stays rose in 2006. Epilepsia 55, 1347–1354. Centers for Disease Control Prevention National Center for Health Statistics, 2014. Compressed Mortality File 1999–2013 on CDC WONDER Online Database, released October 2014 Data are from the Compressed Mortality File 1999–2013 Series 20 No. 2S, 2014, as compiled from data provided by the 57 vital statistics jurisdictions through the Vital Statistics Cooperative Program 2014, Available from: http://wonder.cdc.gov/cmf-icd10.html [cited 19.11.15]. Chan, L., Hart, L.G., Goodman, D.C., 2006. Geographic access to health care for rural Medicare beneficiaries. J. Rural Health 22, 140–146.

13

Eberhardt, M.S., Pamuk, E.R., 2004. The importance of place of residence: examining health in rural and nonrural areas. Am. J. Public Health 94, 1682–1686. Elliott, J.O., Charyton, C., Lu, B., Moore, J.L., 2009. Serious psychological distress and health outcomes for persons with epilepsy in poverty. Seizure 18, 332–338. Elliott, J.O., Moore, J.L., Lu, B., 2008. Health status and behavioral risk factors among persons with epilepsy in Ohio based on the 2006 Behavioral Risk Factor Surveillance System. Epilepsy Behav. 12, 434–444. Faught, E., Richman, J., Martin, R., Funkhouser, E., Foushee, R., Kratt, P., Kim, Y., Clements, K., Cohen, N., Adoboe, D., Knowlton, R., Pisu, M., 2012. Incidence and prevalence of epilepsy among older U.S Medicare beneficiaries. Neurology 78, 448–453. Ferguson, P.L., Chiprich, J., Smith, G., Dong, B., Wannamaker, B.B., Kobau, R., Thurman, D.J., Selassie, A.W., 2008. Prevalence of self-reported epilepsy, health care access, and health behaviors among adults in South Carolina. Epilepsy Behav. 13, 529–534. Forsgren, L., Hauser, W.A., Olafsson, E., Sander, J.W., Sillanpaa, M., Tomson, T., 2005. Mortality of epilepsy in developed countries: a review. Epilepsia 46 (Suppl. 11), 18–27. Gaitatzis, A., Johnson, A.L., Chadwick, D.W., Shorvon, S.D., Sander, J.W., 2004. Life expectancy in people with newly diagnosed epilepsy. Brain 127, 2427–2432. Geerts, A., Brouwer, O., van Donselaar, C., Stroink, H., Peters, B., Peeters, E., Arts, W.F., 2011. Health perception and socioeconomic status following childhood-onset epilepsy: the Dutch study of epilepsy in childhood. Epilepsia 52, 2192–2202. Gourie-Devi, M., Gururaj, G., Satishchandra, P., Subbakrishna, D.K., 2004. Prevalence of neurological disorders in Bangalore, India: a community-based study with a comparison between urban and rural areas. Neuroepidemiology 23, 261–268. Gresenz, C.R., Stockdale, S.E., Wells, K.B., 2000. Community effects on access to behavioral health care. Health Serv. Res. 35, 293–306. Halpern, M.T., Renaud, J.M., Vickrey, B.G., 2011. Impact of insurance status on access to care and out-of-pocket costs for U.S. individuals with epilepsy. Epilepsy Behav. 22, 483–489. Hartley, D., 2004. Rural health disparities, population health, and rural culture. Am. J. Public Health 94, 1675–1678. Hauptman, J.S., Dadour, A., Oh, T., Baca, C.B., Vickrey, B.G., Vassar, S., Sankar, R., Salamon, N., Vinters, H.V., Mathern, G.W., 2013a. Time to pediatric epilepsy surgery is longer and developmental outcomes lower for government compared with private insurance. Neurosurgery 73, 152–157. Hauptman, J.S., Dadour, A., Oh, T., Baca, C.B., Vickrey, B.G., Vassar, S.D., Sankar, R., Salamon, N., Vinters, H.V., Mathern, G.W., 2013b. Sociodemographic changes over 25 years of pediatric epilepsy surgery at UCLA. J. Neurosurg. Pediatr. 11, 250–255. Healthy People 2020, 2015. Foundation Health Measures: Disparities 2015, Available from: http://www.healthypeople.gov//about/foundation-healthmeasures/Disparities [cited 03.04.15]. Heaney, D.C., MacDonald, B.K., Everitt, A., Stevenson, S., Leonardi, G.S., Wilkinson, P., Sander, J.W., 2002. Socioeconomic variation in incidence of epilepsy: prospective community based study in south east England. Br Med J 325, 1013–1016. Helmers, S.L., Thurman, D.J., Durgin, T.L., Pai, A.K., Faught, E., 2015. Descriptive epidemiology of epilepsy in the U.S. population: a different approach. Epilepsia 56, 942–948. Hesdorffer, D.C., Tian, H., Anand, K., Hauser, W.A., Ludvigsson, P., Olafsson, E., Kjartansson, O., 2005. Socioeconomic status is a risk factor for epilepsy in Icelandic adults but not in children. Epilepsia 46, 1297–1303. Hill, J.L., You, W., Zoellner, J.M., 2014. Disparities in obesity among rural and urban residents in a health disparate region. BMC Public Health 14, 1051. Hitiris, N., Mohanraj, R., Norrie, J., Brodie, M.J., 2007. Mortality in epilepsy. Epilepsy Behav. 10, 363–376. James, W.L., 2014. All rural places are not created equal: revisiting the rural mortality penalty in the United States. Am. J. Public Health 104, 2122–2129. Jette, N., Reid, A.Y., Quan, H., Hill, M.D., Wiebe, S., 2010. How accurate is ICD coding for epilepsy? Epilepsia 51, 62–69. Johnson, S., Abonyi, S., Jeffery, B., Hackett, P., Hampton, M., McIntosh, T., Martz, D., Muhajarine, N., Petrucka, P., Sari, N., 2008. Recommendations for action on the social determinants of health: a Canadian perspective. Lancet 372, 1690–1693. Kaiboriboon, K., Schiltz, N.K., Bakaki, P.M., Lhatoo, S.D., Koroukian, S.M., 2014. Premature mortality in poor health and low income adults with epilepsy. Epilepsia 55, 1781–1788. Kee, V.R., Gilchrist, B., Granner, M.A., Sarrazin, N.R., Carnahan, R.M., 2012. A systematic review of validated methods for identifying seizures, convulsions, or epilepsy using administrative and claims data. Pharmacoepidemiol. Drug Saf. 21 (Suppl. 1), 183–193. Kelvin, E.A., Hesdorffer, D.C., Bagiella, E., Andrews, H., Pedley, T.A., Shih, T.T., Leary, L., Thurman, D.J., Hauser, W.A., 2007. Prevalence of self-reported epilepsy in a multiracial and multiethnic community in New York City. Epilepsy Res. 77, 141–150. Kobau, R., DiIorio, C.A., Price, P.H., Thurman, D.J., Martin, L.M., Ridings, D.L., Henry, T.R., 2004. Prevalence of epilepsy and health status of adults with epilepsy in Georgia and Tennessee: Behavioral Risk Factor Surveillance System, 2002. Epilepsy Behav. 5, 358–366. Kobau, R., Gilliam, F., Thurman, D.J., 2006. Prevalence of self-reported epilepsy or seizure disorder and its associations with self-reported depression and anxiety: results from the 2004 Health Styles Survey. Epilepsia 47, 1915–1921.

14

D.A. Wilson et al. / Epilepsy Research 122 (2016) 7–14

Kobau, R., Zahran, H., Grant, D., Thurman, D.J., Price, P.H., Zack, M.M., 2007. Prevalence of active epilepsy and health-related quality of life among adults with self-reported epilepsy in California: California Health Interview Survey, 2003. Epilepsia 48, 1904–1913. Konda, K., Ablah, E., Konda, K.S., Liow, K., 2009. Health behaviors and conditions of persons with epilepsy: a bivariate analysis of 2006 BRFSS data. Epilepsy Behav. 16, 120–127. Koroukian, S.M., Bakaki, P.M., Raghavan, D., 2012. Survival disparities by Medicaid status: an analysis of 8 cancers. Cancer 118, 4271–4279. Kroner, B.L., Fahimi, M., Kenyon, A., Thurman, D.J., Gaillard, W.D., 2013. Racial and socioeconomic disparities in epilepsy in the District of Columbia. Epilepsy Res. 103, 279–287. Kulshreshtha, A., Goyal, A., Dabhadkar, K., Veledar, E., Vaccarino, V., 2014. Urban–rural differences in coronary heart disease mortality in the United States: 1999–2009. Public Health Rep. 129, 19–29. Lawlor, D.A., Davey Smith, G., Patel, R., Ebrahim, S., 2005. Life-course socioeconomic position, area deprivation, and coronary heart disease: findings from the British Women’s Heart and Health Study. Am. J. Public Health 95, 91–97. Lhatoo, S.D., Johnson, A.L., Goodridge, D.M., MacDonald, B.K., Sander, J.W., Shorvon, S.D., 2001. Mortality in epilepsy in the first 11 to 14 years after diagnosis: multivariate analysis of a long-term, prospective, population-based cohort. Ann. Neurol. 49, 336–344. Li, X., Sundquist, J., Sundquist, K., 2008. Socioeconomic and occupational risk factors for epilepsy: a nationwide epidemiological study in Sweden. Seizure 17, 254–260. Marmot, M., 2005. Social determinants of health inequalities. Lancet 365, 1099–1104. McClelland 3rd, S., Guo, H., Okuyemi, K.S., 2010. Racial disparities in the surgical management of intractable temporal lobe epilepsy in the United States: a population-based analysis. Arch. Neurol. 67, 577–583. Meit, M., Knudson, A., Gilbert, T., Yu, A.T.-C., Tanenbaum, E., Ormson, E., TenBroeck, S., Bayne, A., Popat, S., 2014, October. NORC Walsh Center for Rural Health Analysis. The 2014 Update of the Rural-Urban Chartbook: Rural Health Research & Policy Centers, Available from: https://ruralhealth.und.edu/ projects/health-reform-policy-research-center/pdf/2014-rural-urbanchartbook-update.pdf [14.05.15]. Morgan, C.L., Ahmed, Z., Kerr, M.P., 2000. Social deprivation and prevalence of epilepsy and associated health usage. J. Neurol. Neurosurg. Psychiatry 69, 13–17. Morrill, R., Cromartie, J., Hart, G., 1999. Metropolitan, urban, and rural commuting areas: toward a better depiction of the United States settlement system. Urban Geogr. 20, 727–748. Neligan, A., Bell, G.S., Johnson, A.L., Goodridge, D.M., Shorvon, S.D., Sander, J.W., 2011. The long-term risk of premature mortality in people with epilepsy. Brain 134, 388–395. Nevalainen, O., Ansakorpi, H., Simola, M., Raitanen, J., Isojarvi, J., Artama, M., Auvinen, A., 2014. Epilepsy-related clinical characteristics and mortality: a systematic review and meta-analysis. Neurology 83, 1968–1977. Ottman, R., Lipton, R.B., Ettinger, A.B., Cramer, J.A., Reed, M.L., Morrison, A., Wan, G.J., 2011. Comorbidities of epilepsy: results from the Epilepsy Comorbidities and Health (EPIC) survey. Epilepsia 52, 308–315. Pickrell, W.O., Lacey, A.S., Bodger, O.G., Demmler, J.C., Thomas, R.H., Lyons, R.A., Smith, P.E., Rees, M.I., Kerr, M.P., 2015. Epilepsy and deprivation, a data linkage study. Epilepsia 56, 585–591. Probst, J.C., Moore, C.G., Glover, S.H., Samuels, M.E., 2004. Person and place: the compounding effects of race/ethnicity and rurality on health. Am. J. Public Health 94, 1695–1703. Rai, D., Kerr, M.P., McManus, S., Jordanova, V., Lewis, G., Brugha, T.S., 2012. Epilepsy and psychiatric comorbidity: a nationally representative population-based study. Epilepsia 53, 1095–1103. Reid, A.Y., St Germaine-Smith, C., Liu, M., Sadiq, S., Quan, H., Wiebe, S., Faris, P., Dean, S., Jette, N., 2012. Development and validation of a case definition for epilepsy for use with administrative health data. Epilepsy Res. 102, 173–179. SAS Institute, 2012. Statistical Analytical Software, Version 9, 4th ed. SAS Institute Inc., Cary, NC. Schiltz, N.K., Koroukian, S.M., Singer, M.E., Love, T.E., Kaiboriboon, K., 2013. Disparities in access to specialized epilepsy care. Epilepsy Res. 107, 172–180. Scott, A.J., Wilson, R.F., 2011. Social determinants of health among African Americans in a rural community in the Deep South: an ecological exploration. Rural Remote Health 11, 1634.

Selassie, A.W., 2012. IOM data gathering efforts: South Carolina. In: England, M.J., Liverman, C.T., Schultz, A.M., Strawbridge, L.M. (Eds.), Epilepsy Across the Spectrum: Promoting Health and Understanding. Institute of Medicine, National Academies Press, Washington, DC, pp. 491–508. Selassie, A.W., Ferguson, P., Wannamaker, B.B., Pickelsimer, E.E., Turner, R., Smith, G.G., Bailey, W., Tyrell, M., Lineberry, L., Gu, J.K., Thurman, D.J., 2005. The South Carolina Epidemiological Studies of Epilepsy and Seizure Disorders. Medical University of South Carolina, Charleston, SC, USA, pp. 1–139. Selassie, A.W., Wilson, D.A., Malek, A.M., Wagner, J.L., Smith, G., Martz, G., Edwards, J., Wannamaker, B., Zack, M.M., Kobau, R., 2014. Premature deaths among children with epilepsy – South Carolina, 2000–2011. Morb. Mortal. Wkly. Rep. 63, 989–994. Sillanpaa, M., 2004. Learning disability: occurrence and long-term consequences in childhood-onset epilepsy. Epilepsy Behav. 5, 937–944. Singh, G.K., Siahpush, M., 2014a. Widening rural-urban disparities in all-cause mortality and mortality from major causes of death in the USA, 1969–2009. J. Urban Health 91, 272–292. Singh, G.K., Siahpush, M., 2014b. Widening rural-urban disparities in life expectancy, U.S., 1969–2009. Am. J. Prev. Med. 46, e19–e29. Smedley, B.D., Syme, S.L., Committee on Capitalizing on Social S, Behavioral Research to Improve the Public’s H, 2001. Promoting health: intervention strategies from social and behavioral research. Am. J. Health Promot. 15, 149–166. Steer, S., Pickrell, W.O., Kerr, M.P., Thomas, R.H., 2014. Epilepsy prevalence and socioeconomic deprivation in England. Epilepsia 55, 1634–1641. Szaflarski, M., 2014. Social determinants of health in epilepsy. Epilepsy Behav. 41, 283–289. Terashima, M., Rainham, D.G., Levy, A.R., 2014. A small-area analysis of inequalities in chronic disease prevalence across urban and non-urban communities in the Province of Nova Scotia, Canada, 2007–2011. BMJ Open 4, e004459. Theodore, W.H., Spencer, S.S., Wiebe, S., Langfitt, J.T., Ali, A., Shafer, P.O., Berg, A.T., Vickrey, B.G., 2006. Epilepsy in North America: a report prepared under the auspices of the global campaign against epilepsy, the International Bureau for Epilepsy, the International League Against Epilepsy, and the World Health Organization. Epilepsia 47, 1700–1722. Thurman, D.J., Beghi, E., Begley, C.E., Berg, A.T., Buchhalter, J.R., Ding, D., Hesdorffer, D.C., Hauser, W.A., Kazis, L., Kobau, R., Kroner, B., Labiner, D., Liow, K., Logroscino, G., Medina, M.T., Newton, C.R., Parko, K., Paschal, A., Preux, P.M., Sander, J.W., Selassie, A., Theodore, W., Tomson, T., Wiebe, S., Epidemiology, I.C.o., 2011. Standards for epidemiologic studies and surveillance of epilepsy. Epilepsia 52 (Suppl. 7), 2–26. Trinka, E., Bauer, G., Oberaigner, W., Ndayisaba, J.P., Seppi, K., Granbichler, C.A., 2013. Cause-specific mortality among patients with epilepsy: results from a 30-year cohort study. Epilepsia 54, 495–501. U.S. Department of Housing and Urban Development, Available from: http://www. hud.gov/offices/cpd/communitydevelopment/programs/neighborhoodspg/ nspeligibleuses.doc [01.06.15] 2015. Guide to Neighborhood Stabilization Program (NSP) Eligible Uses. USDA Economic Research Service, 2011. Rural–Urban Commuting Area Codes, Available from: http://www.ers.usda.gov/data-products/rural-urbancommuting-area-codes.aspx [updated 2011-11-092015-05-14]. Varma, A., Hill, E.G., Nicholas, J., Selassie, A., 2010. Predictors of early mortality after traumatic spinal cord injury: a population-based study. Spine (Phila., PA, 1976) 35, 778–783. Weis, M.A., Bradberry, C., Carter, L.P., Ferguson, J., Kozareva, D., 2006. An exploration of human services system contacts prior to suicide in South Carolina: an expansion of the South Carolina Violent Death Reporting System. Inj. Prev. 12 (Suppl. 2), ii17–ii21. Wiebe, S., Camfield, P., Jette, N., Burneo, J.G., 2009. Epidemiology of epilepsy: prevalence, impact, comorbidity and disparities. Can. J. Neurol. Sci. 36 (Suppl. 2), S7–S16. Wong, M.D., Shapiro, M.F., Boscardin, W.J., Ettner, S.L., 2002. Contribution of major diseases to disparities in mortality. N. Engl. J. Med. 347, 1585–1592. WWAMI Rural Health Research Center, 2015a. Maps; Map Classification, Available from: http://depts.washington.edu/uwruca/ruca-maps.php [cited 2015 2015-04-09]. WWAMI Rural Health Research Center, 2015b. Rural Urban Commuting Area Codes Data; Using RUCA Data, Available from: http://depts.washington.edu/uwruca/ ruca-uses.php [14.05.15].