Address correspondence to Dr. Karen Parko, 4150 Clement Street (127), San Francisco, CA 94121, U.S.A. E-mail: Karen.Parko@ucsf.edu
Purpose: To determine the prevalence of epilepsy and seizures in the Navajo.
Methods: We studied 226,496 Navajo residing in the Navajo Reservation who had at least one medical encounter between October 1, 1998 and September 30, 2002. We ascertained and confirmed cases in two phases. First, we identified patients with International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes signifying epilepsy or seizures using Indian Health Service (IHS) administrative data. Second, we reviewed medical charts of a geographic subpopulation of identified patients to confirm diagnoses and assess the positive predictive value of the ICD-9-CM codes in identifying patients with active epilepsy.
Results: Two percent of Navajo receiving IHS care were found to have an ICD-9-CM code consistent with epilepsy or seizures. Based on confirmed cases, the crude prevalence for the occurrence of any seizure (including febrile seizures and recurrent seizures that may have been provoked) in the geographic subpopulation was 13.5 per 1,000 and the crude prevalence of active epilepsy was 9.2 per 1,000. Prevalence was higher among males, children under 5 years of age, and older adults.
Discussion: The estimated prevalence of active epilepsy in the Navajo Nation is above the upper limit of the range of reported estimates from other comparable studies of U.S. communities.
Epilepsy is characterized by recurrent unprovoked seizures and is among the most common serious disorders of the brain occurring worldwide. Among studies of the occurrence of active epilepsy in developed countries, most yield prevalence estimates between 5 and 10 per 1,000 population (Annegers, 2004; Hirtz et al., 2007). Comparable studies from developing countries yield a broader range of values, with a median reported prevalence of 12.4 per 1,000 in Latin America (Burneo et al., 2005) and 15 per 1,000 in sub-Saharan Africa (Preux & Druet-Cabanac, 2005). Studies of specific U.S. communities have yielded prevalence estimates between 6.7 and 8.8 per 1,000 (Haerer et al., 1986; Hauser et al., 1991; Begley et al., 2000; Holden et al., 2005). Some data suggest that the prevalence of epilepsy may be higher among blacks than among whites (Haerer et al., 1986).
The purpose of this study is to examine the prevalence of epilepsy among the Navajo, an American Indian tribe with a population of about 300,000 living in the Southwestern United States. The Navajo Nation is somewhat isolated geographically, and the Navajo have maintained a large measure of their traditional culture. Many continue to rely on traditional Navajo medicine for part of their health care. The Navajo population differs from that of the rest of the United States by several sociodemographic factors. According to 1990 U.S. Census data, compared with the general U.S. population, the Navajo population has a higher percentage of young people (43% younger than age 20), a higher percentage of persons (56%) living below the poverty level, a much higher unemployment rate (about 30%), lower educational levels (40% of adults with high school degrees), and lower annual per capita health care expenditures ($1,397 compared to U.S. average of $3,261) (U.S. Census Bureau, 1990). The principal source of health care is the Indian Health Service (IHS).
Only one other epidemiologic survey of epilepsy in the Navajo Nation has been published (Levy et al., 1995), yielding an age-adjusted prevalence of 8.2 per 1,000 population in one district of the reservation during the years 1971–1978.
We performed a cross-sectional study of all Navajo who received medical or dental care from the IHS within the Navajo Nation between October 1, 1998 and September 30, 2002 [federal fiscal years (FY) 1999–2002]. The study protocol was approved by the National IHS Institutional Review Board, the Navajo Nation Human Research Review Board (NNHRRB), the local community chapter house in Shiprock, New Mexico, and the Shiprock Service Unit Health Board. The publication of this manuscript was approved by the NNHRRB.
IHS administrative data with coded information on all inpatient and outpatient medical encounters were used to estimate the prevalence of active seizures or epilepsy. The medical charts of a sample of about one-third of the identified patients were then reviewed to determine the accuracy of coded diagnostic data indicating epilepsy.
Health care system background
The Navajo Nation encompasses approximately 26,000 square miles (67,000 square kilometers) of trust lands, spanning parts of three western U.S. states, with the largest sector in northwestern Arizona, followed by northeastern New Mexico and southern Utah.
The IHS is an agency within the Department of Health and Human Services that delivers health care to the approximately 2.4 million American Indians and Alaskan Natives in the United States. The IHS is organized into 12 geographical area offices, one of which is Navajo Area Office headquartered in Window Rock, Arizona.
The Navajo Area office provides health care to Native Americans who live on or near the Navajo reservation. There are eight geographic service units in the Navajo Area (see map—Fig. 1). Six of the eight service units have hospitals, with a total of 360 hospital beds and 225 physicians in the Navajo area. During the first 3 years under study, one full-time neurologist (KP) provided neurologic care, working primarily at the Shiprock service unit, but also attending periodic neurology clinics at other service unit hospitals: Tuba City, Kayenta, Gallup, and Chinle (each with four clinics per year) and Crownpoint and Ft. Defiance (each with approximately one clinic per year). In the fourth year of study this neurologist was joined by a second full-time neurologist who provided care at the Gallup service unit. Throughout the study, electroencephalography (EEG) and computerized tomography (CT) were available only at Shiprock and Gallup.
Data sources for case ascertainment
The IHS maintains a standard computerized health care summary for each registered patient, collecting data from each medical or dental visit, including ICD-9-CM diagnostic codes (U.S. Department of Health and Human Services, 2003). IHS maintains a national computerized health summary data set (IHS data set) for all patients seen.
The study population comprised all Navajo tribe members who had at least one medical or dental encounter within any of the eight Navajo Area IHS service units during the 4-year prevalence period, that is, the IHS user population.
The sample for chart review was drawn from three service units (Crownpoint, Kayenta, Shiprock) and consisted of all patients identified from a computerized search for records including ICD-9-CM diagnostic codes indicating seizures or epilepsy. These three geographic units—including more than a third of the total population of Navajo Nation—represented the range in availability of health care and specialty services.
Definition of cases from computerized health summaries
Cases of seizure and epilepsy were identified from patient records containing any of the ICD-9-CM diagnostic codes 345.0–345.9, 333.2, 779.0–779.1, 780.31, and 780.39.
Definition of cases from review of medical charts
Chart reviews were performed by neurologists, clinical pharmacists trained in epilepsy care, and medical students supervised by neurologists. Complete medical records were sought from all patients in the three sampled service units who had been identified with an ICD-9-CM code of interest. Based on documentation in the medical chart, each patient was placed into one of five categories: (1) Epilepsy with documentation of two or more unprovoked seizures, (2) Seizure if they had a single seizure (provoked or not) or two or more provoked seizures (e.g., alcohol withdrawal), (3) Febrile seizure in a child younger than 5 years of age with seizures in the setting of fever that were labeled by caregiver as “febrile seizure,” (4) No seizure if there was no evidence of any seizures or epilepsy, and (5) Missing if the chart was unavailable.
Inter-reviewer reliability was tested for each chart reviewer on each day of chart review by independent coding of three charts by two separate reviewers. There was 100% concordance on the charts that were compared.
From the national IHS data set, we ascertained the prevalence of epilepsy and seizure occurrence per 1,000 population, stratified by demographic characteristics (age, sex, location of home) of identified patients. The population denominators were derived from the Navajo Area IHS user population for FY 2001. Prevalence estimates were adjusted to the 2000 U.S. population by sex and age.
Within the service units of Crownpoint, Kayenta, and Shiprock, we calculated age-specific positive predictive values (PPVs) of epilepsy- and seizure-related ICD-9-CM diagnostic codes in the IHS dataset by comparing these with diagnoses obtained from medical chart reviews. Within these three service units, we calculated the prevalence of epilepsy and seizures by including in the numerator: (i) all confirmed cases (from medical chart reviews) and (ii) putative cases [from the number of missing charts with ICD-9-CM codes in the corresponding IHS data set records that indicate epilepsy (345.x) or seizures (780.3x), multiplied by age-specific PPVs determined for these codes]. These estimates were also age- and sex-adjusted to the U.S. 2000 population.
In all, 226,496 Navajo (117,255 women and 107,714 men in FY 2001) received health care within the IHS in FY 99-02, with a median age of 23 years. The analysis of the IHS data set revealed that 4,181 Navajo received an ICD-9-CM code indicating epilepsy or seizures. The crude prevalence of epilepsy or seizures was 18.6 per 1,000, with an age- and sex-adjusted prevalence of 20 per 1,000. The prevalence was higher among males, very young children, and older adults (Table 1). The crude prevalence varied among the eight geographic service units from 16.3 per 1,000 in Crownpoint to 22.4 per 1,000 in Chinle (Table 2).
Table 1. Prevalence of epilepsy or seizuresa by age and sex—Navajo Nation, 1998–2002
Age group (years)
Total Navajo Area
Subset of three service units Shiprock, Kayenta, Crownpoint
aPresumed cases based on ICD-9-CM codes contained in IHS data set, not verified by medical records review. Shiprock, Kayenta, and Crownpoint are a subset of the total Navajo Area.
bAge adjusted to U.S. 2000 population.
Table 2. Prevalence of epilepsy or seizurea by IHS service unit—Navajo Area, 1998–2002
User population FY 2001
Crude prevalence per 1,000
aPresumed cases based on ICD-9-CM codes contained in IHS data set over 4-year prevalence period, not verified by medical records review.
Restriction of the analysis to the three service units of Crownpoint, Kayenta, and Shiprock produced a slightly lower prevalence: a crude prevalence of epilepsy or seizures of 16.7 per 1,000, and an age- and sex-adjusted prevalence of 18.1 per 1,000. Crude prevalence for males and females was 19.8 and 13.8 per 1,000, respectively (21.9 and 14.5 per 1,000, age adjusted), (Table 1, Fig. 2).
Within the service units of Crownpoint, Kayenta, and Shiprock, the service unit data set yielded 1,367 records indicating possible seizures or epilepsy. Medical charts were located and reviewed for 1,277 of these, enabling a calculation of PPVs for relevant ICD-9-CM codes in the data set. Across all ages, coding with either 345.x or 780.3x yielded PPVs of 0.90 for a clinical diagnosis of epilepsy or seizures and 0.62 for a clinical diagnosis of epilepsy per se. The most common errors in coding were in patients with migraine (ICD-9-CM code 346 miscoded as 345).
On the basis of an analysis taking into account medical chart reviews and applying PPVs to putative cases with missing charts, the crude period prevalence of active epilepsy in the three service units was 9.2 per 1,000 with a corresponding age-adjusted prevalence of 10.2 per 1,000. (Table 3) The crude prevalence for the overall occurrence of seizures (including epilepsy, febrile seizures, and nonrecurring and provoked seizures) was 13.5 per 1,000 with an age-adjusted prevalence of 14.2 per 1,000. (Table 3).
Table 3. Prevalence of epilepsy and seizures—three Navajo service units, 1998–2002
Age group (years)
Epilepsy or seizure prevalenceb
aIHS active user population of Kayenta, Shiprock, and Crownpoint.
bPrevalence per 1,000 population calculated by including in the numerator: (i) all confirmed cases (from medical chart reviews) and (ii) putative cases (from the number of missing charts with ICD-9-CM codes in IHS data set records indicating epilepsy or seizures, multiplied by age-specific positive predictive values determined for these codes.)
cAge- and sex-adjusted to U.S. 2000 census population.
The stratification of data from the three service units by age and sex shows similar estimates of epilepsy for both sexes in childhood, but a divergence in young adulthood with substantially higher prevalence among males in all older age categories (Fig. 2). Including all ages, the age-adjusted prevalence of epilepsy was 12.2 per 1,000 for males and 8.3 per 1,000 for females.
This cross-sectional study emphasizes the importance of employing a population-based study using direct chart review to accurately obtain the prevalence of epilepsy and seizures. When we relied on the ICD-9-CM-coded IHS data alone, we obtained a substantially higher estimate of prevalence, as we included some patients without seizures. We also found that although ICD-9-CM-coded data have a high PPV for seizures, the corresponding predictive value for epilepsy (recurrent unprovoked seizures) is much lower. Therefore, a corrective factor is necessary when relying on ICD-9-CM-coded data to estimate the prevalence of epilepsy.
Our estimates of the prevalence of active epilepsy in the Navajo Nation exceed reported estimates from other comparable studies of U.S. communities (Haerer et al., 1986; Hauser et al., 1991; Begley et al., 2000; Holden et al., 2005). Furthermore, it is likely that the prevalence we found is an underestimate, as our methods failed to capture Navajo who sought medical care from sources outside the IHS, including traditional medical care.
Although slightly increased prevalence among males is commonly reported in studies of other populations, in this study the male-to-female ratio is nearly 1.5 overall, with the disparity principally among people aged 25 years or older. This suggests that Navajo men may have unusual adulthood risks, possibly including head trauma and alcoholism, which are additive to other risk factors.
The different prevalence of epilepsy and seizures that we found within the service units we examined (Table 2) may be understood better by considering differences in access to medical care outside of the IHS. To the extent that off-reservation care is received, cases of epilepsy and seizures may go unrecognized in IHS records. The highest prevalence estimates found in Chinle and Fort Defiance (Table 2)—which are relatively distant from off-reservation sources of medical care (Fig. 1) —might reflect a higher proportion of people with epilepsy in these communities who obtain care from the IHS. In contrast, the lowest prevalence at Crownpoint (Table 2)—which is relatively nearer to Albuquerque, New Mexico (Fig. 1)—might be attributable to a higher proportion who seek specialist care for epilepsy in Albuquerque. However, there are other potential risk factors that may vary by region and cause local variation in prevalence, for example, genetic or sociodemographic factors.
The sample population is fully representative only of Navajo who receive their health care through the IHS, which provides health care free of charge to any registered Navajo tribe member. The population sampled did not include Navajo who live within the Navajo Nation but who receive and self pay for health care from off-reservation, non-IHS providers. Nor did it include those who participate only in traditional medical care, those without health problems who did not obtain IHS medical or dental care, or those Navajo who live outside of the Navajo Nation. It is difficult to estimate the number of Navajo who would fall into any one of these categories. A prior study on Navajo use of native healers found that 62% had used a native healer at least once in their lifetime and 39% had used a native healer during the last year (Kim & Kwok, 1998). However, all of the patients in that study were concomitantly using IHS medical care.
Choice of denominator is an important consideration when calculating prevalence for the Navajo Nation, as different population estimates during the prevalence period are available, for example, the Year 2000 Census (218,455) and the active-user population derived from Navajo Area IHS data for FY 2001 (224,969). Other epidemiologic studies of the Navajo area most often use the user population as the denominator, a mid-interval actual population number considered most representative for cases likely to be ascertained in a study. In our study, receiving a diagnosis of seizures or epilepsy requires receiving care in the hospital or clinic, which is the same requirement as being included in the user population.
Further study is needed to describe the current incidence of epilepsy and seizures among this population, to assess etiologies, to explain disparities in prevalence between men and women, and to identify potential preventive measures in this at-risk population. The prevalence we found is consistent with the reported prevalence in some developing countries (Burneo et al., 2005; Preux & Druet-Cabanac, 2005). In the Navajo Nation, medical care and medication are available, although there may still be significant issues of access to care, quality of care, and health care utilization that are unique to this culture and geography. These deserve further exploration to help address these health disparities.
Given its limited resources, the current emphasis of IHS health care is on preventing disease. However, for diseases such as epilepsy for which preventive strategies may be insufficient, an emphasis is also needed on the delivery of appropriate symptomatic care and chronic disease management.
We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.
Disclosure: The authors have no reported conflicts of interest.
Funding sources: CDC National Center for Chronic Disease Prevention and Health Promotion.
The authors thank the following for their direct assistance: Indian Health Service Information Technology Support Center; Anne Butman, management analyst, DataCom Sciences; Yolinda Cadman; Dr. Nathanial Cobb; Navajo Area medical records department; Northern Navajo Medical Center in Shiprock: Gary Russell-King, in Kayenta: Lorraine Dohi, in Crownpoint: Cynthia Begaye; Clinical pharmacists in the NNMC seizure clinic: Tom Duran, Lauren Dolence, Melissa Stahlecker, Dr. David Labiner, Karla Lindquist, Peter Taylor, and Elena Cherkasova. This work could not have been undertaken without the support from: Duane H. Yazzie and the Shiprock Chapter House; Manuel Morgan and the Shiprock Service Unit Health Board; Richard Champany DDS, CEO, Northern Navajo Medical Center; Phillip Smith, MD, MPH and the National IHS IRB; and Beverly Pigman and the Navajo Nation Human Research Review Board.
Disclaimer: The findings and conclusions in the report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention or the Indian Health Service.