Key Points
Question
What are the rate and factors associated with neurodiagnostic evaluation among patients with drug-resistant epilepsy (DRE) in the US?
Findings
In this cross-sectional study of 58 779 patients, the proportion of patients receiving 3 categories of neurodiagnostic studies within 2 years of a clinical encounter with DRE was 4.5% in a Medicaid cohort, 8.0% in a commercial insurance cohort, and 14.3% in an academic medical center cohort. Factors independently associated with evaluation included the number of nonemergency visits and focal rather than generalized epilepsy.
Meaning
The findings of this study suggest there is a gap in the use of diagnostic studies to evaluate patients with DRE.
Abstract
Importance
Interdisciplinary practice parameters recommend that patients with drug-resistant epilepsy (DRE) undergo comprehensive neurodiagnostic evaluation, including presurgical assessment. Reporting from specialized centers suggests long delays to referral and underuse of surgery; however, longitudinal data are limited to characterize neurodiagnostic evaluation among patients with DRE in more diverse US settings and populations.
Objective
To examine the rate and factors associated with neurodiagnostic studies and comprehensive evaluation among patients with DRE within 3 US cohorts.
Design, Setting, and Participants
A retrospective cross-sectional study was conducted using the Observational Medical Outcomes Partnership Common Data Model including US multistate Medicaid data, commercial claims data, and Columbia University Medical Center (CUMC) electronic health record data. Patients meeting a validated computable phenotype algorithm for DRE between January 1, 2015, and April 1, 2020, were included. No eligible participants were excluded.
Exposure
Demographic and clinical variables were queried.
Main Outcomes and Measures
The proportion of patients receiving a composite proxy for comprehensive neurodiagnostic evaluation, including (1) magnetic resonance or other advanced brain imaging, (2) video electroencephalography, and (3) neuropsychological evaluation within 2 years of meeting the inclusion criteria.
Results
A total of 33 542 patients with DRE were included in the Medicaid cohort, 22 496 in the commercial insurance cohort, and 2741 in the CUMC database. A total of 31 516 patients (53.6%) were women. The proportion of patients meeting the comprehensive evaluation main outcome in the Medicaid cohort was 4.5% (n = 1520); in the commercial insurance cohort, 8.0% (n = 1796); and in the CUMC cohort, 14.3% (n = 393). Video electroencephalography (24.9% Medicaid, 28.4% commercial, 63.2% CUMC) and magnetic resonance imaging of the brain (35.6% Medicaid, 43.4% commercial, 52.6% CUMC) were performed more regularly than neuropsychological evaluation (13.0% Medicaid, 16.6% commercial, 19.2% CUMC) or advanced imaging (3.2% Medicaid, 5.4% commercial, 13.1% CUMC). Factors independently associated with greater odds of evaluation across all 3 data sets included the number of inpatient and outpatient nonemergency epilepsy visits and focal rather than generalized epilepsy.
Conclusions and Relevance
The findings of this study suggest there is a gap in the use of diagnostic studies to evaluate patients with DRE. Care setting, insurance type, frequency of nonemergency visits, and epilepsy type are all associated with evaluation. A common data model can be used to measure adherence with best practices across a variety of observational data sources.
This cross-sectional study examines the use of neurodiagnostic testing among patients with drug-resistant epilepsy.
Introduction
Epilepsy is a common debilitating disease; an estimated 1% of the US population has active epilepsy, including more than 3 million adults and 400 000 children.1 Patients with uncontrolled seizures are at increased risk of death, injury, cognitive decline, psychiatric illness, and decreased quality of life.2,3,4 Approximately one-third of patients with epilepsy (>1 million) do not respond to 2 or more appropriate medication regimens and are considered to have drug-resistant epilepsy (DRE)5,6; with continued polypharmacy, fewer than 3% of patients with DRE achieve sustained seizure freedom.5,6 Multidisciplinary guidelines recommend that all patients with DRE undergo interdisciplinary comprehensive evaluation, including presurgical assessment, at a specialized center.7,8 Epilepsy surgery offers 50% to 80% seizure freedom and superior cognitive and quality of life outcomes for well-selected patients.9,10,11 Evaluation can also lead to improved nonsurgical diagnosis, therapy, and prognostication.
Previous work suggests that comprehensive evaluation and epilepsy surgery are underused among patients with DRE. However, most findings have been inferred from surveys, cross-sectional inpatient data,12,13 or self-reporting from the most experienced centers.14,15,16 For example, reporting from accredited centers, combined with the prevalence of DRE, suggests that such centers evaluate fewer than 2% of patients with DRE.17 Some centers have also reported delays to referral in excess of 18 years mean disease duration.14,17 According to the Institute of Medicine, across the 100 to 200 000 candidates, only approximately 4000 epilepsy surgeries are performed in the US each year.18 Cross-sectional studies of hospital admissions have found that while the number of DRE admissions has increased in recent decades, surgical volume has remained stable.12,13 Longitudinal data to characterize neurodiagnostic evaluation among patients with DRE in more diverse cohorts that approximate various population level experiences are limited.19
Observational Health Data Sciences and Informatics (OHDSI) is an international open collaborative of more than 220 health care organizations representing the health records of more than 800 million unique patients with a mission to advance large-scale observational health research.20 The OHDSI collaborative maintains the Observational Medical Outcome Partnership (OMOP) Common Data Model, a standard that normalizes the structure and content of observational data across participating sites and databases. In the present study, we aimed to determine the rate and factors associated with comprehensive evaluation among patients with DRE in 3 US OMOP cohorts: a multistate Medicaid-insured cohort, a commercially insured cohort, and the Columbia University Medical Center (CUMC) electronic health record (EHR)–derived database. Our analysis may also be replicated across the OHDSI network of clinical databases.
Methods
Setting and Data Sources
This was a retrospective cross-sectional study of 3 observational databases. First, MarketScan Multi-State Medicaid (MDCD; IBM Watson Health) captures health care service use of individuals covered by Medicaid programs in numerous geographically dispersed states. Compared with Centers for Medicare & Medicaid data, MDCD is a smaller sample but is mapped to OMOP, captures adjudicated claims from both fee-for-service and managed care plans including pharmacy claims data, and maintains unique patient identifiers for improved longitudinal tracking. Second, MarketScan Commercial Claims and Encounters (CCAE) is a health insurance database representing several million individuals enrolled in employer-sponsored private health insurance plans in the US. Third, the CUMC OMOP database is derived from multiple EHR, billing, and pharmacy systems comprising a population of more than 6 million patients with encounters between the 1980s and present. CUMC is an urban academic medical center and home to a National Association of Epilepsy Centers level 4 epilepsy referral center. The database encompasses patient interaction with the entire medical center and is not limited to patients under the care of the epilepsy center. The Columbia University institutional review board approved this study and granted a waiver of informed consent because of the retrospective nature of the data collected. The study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.
The 3 databases were transformed into the OMOP Common Data Model version 5 before all queries. OMOP provides standardized longitudinal patient-level information on demographic characteristics, inpatient and outpatient medical encounters, conditions (billing diagnoses and problem lists), drugs (outpatient prescriptions and inpatient orders and administrations), devices, measurements, and other observations. Concepts from multiple coding systems are harmonized through a common terminology.
Inclusion Criteria
We queried each of the 3 databases to identify patients meeting our previously validated criteria for prevalent DRE21 at or before an index encounter between January 1, 2015, and April 1, 2020. The phenotype requires 1 intractable epilepsy diagnosis and 2 or more unique non–gabapentinoid antiseizure medication exposures (each with ≥90-day drug era) (eTable 1 in Supplement 1 provides the relevant diagnosis and medication codes). This phenotype had a specificity of 93.7% in our center’s database against manual medical record review application of the International League Against Epilepsy (ILAE) consensus definition of DRE.6 Patients were also required to have 1 or more years of continuous observation in the database after the index encounter and followed up to April 1, 2022.
Outcomes
The primary outcome was the occurrence of 3 categories of neurodiagnostic studies critical to presurgical evaluation and readily inferred by procedure codes within 2 years of the index date: 1 or more magnetic resonance imaging (MRI) study (or other advanced imaging) of the brain, 1 or more video electroencephalography (EEG) study, and 1 or more neuropsychological evaluation (eTable 1 in Supplement 1 includes the relevant codes). We secondarily assessed the completion of all 3 studies at any time in the available data.
Variables
We queried the databases for demographic and clinical factors. Demographic variables included age at the index date, gender, observation time before the index date, and observation time after the index date. Race and ethnicity have been associated with the use of epilepsy surgery and are therefore reported to the extent available in each data source. Epilepsy was classified as focal if there was 1 or more focal epilepsy diagnosis and 0 generalized epilepsy diagnoses in the year preceding the index date and generalized if there was 1 or more generalized epilepsy diagnosis and 0 focal epilepsy diagnoses in the year preceding the index date, consistent with prior work.21 We also queried for comorbidities present within 1 year preceding the index date, including brain tumor, anxiety, and depression. We queried the number of each visit type (emergency department, inpatient, and outpatient) with a diagnosis of epilepsy within 1 year preceding the index date (categorized as 0, 1, or ≥2). Procedures included a computed tomography scan of the head within 1 year before the index date. Drug variables included the number of unique non–gabapentinoid antiseizure medication exposure eras of at least 90 days’ duration any time preceding the index date (categorized as 2, 3, 4, or ≥5).
Statistical Analysis
We performed univariate and multivariable logistic regression using the glm function in R, version 4.3.0 (R Foundation for Statistical Computing) to identify factors associated with the primary outcome (ie, completion of 3 categories of neurodiagnostic studies) between 2 years before and after the index date. Variables meeting a 2-sided unpaired significance threshold of P < .05 in univariate analysis were included in multivariable models.
Results
A total of 33 542 patients with DRE in the MDCD cohort, 22 496 in the CCAE cohort, and 2741 in the CUMC database were included. Across the 3 cohorts, a total of 31 516 patients (53.6%) were women and 27 253 (46.4%) were men. Other characteristics of the cohorts are presented in Table 1.
Table 1. Characteristics of Patients With Probable Drug-Resistant Epilepsy in 3 Data Sources.
| Characteristic | Patients, No. (%) | ||
|---|---|---|---|
| MarketScan Multi-State Medicaid Database | MarketScan Commercial Claims and Encounters | Columbia University Medical Center | |
| Total cohort | 33 542 | 22 496 | 2741 |
| Age, y | |||
| 0-17 | 10 211 (30.5) | 5609 (24.9) | 839 (30.9) |
| 18-39 | 12 349 (36.8) | 7995 (35.5) | 834 (30.7) |
| 40-64 | 9841 (29.4) | 8892 (39.5) | 739 (27.2) |
| 65-84 | 1118 (3.3) | NR | 306 (11.3) |
| Gender | |||
| Female | 17 655 (52.6) | 12 406 (55.1) | 1455 (53.1) |
| Male | 15 877 (47.4) | 10 090 (44.85) | 1286 (46.9) |
| Ethnicity | |||
| Not Hispanic or Latino | NR | NR | 1357 (49.5) |
| Hispanic or Latino | 856 (2.6) | NR | 810 (29.6) |
| Not reported | 32 686 (97.4) | NR | 574 (20.9) |
| Race | |||
| Asian | NR | NR | 105 (3.8) |
| Black | 7323 (21.8) | NR | 389 (14.2) |
| White | 19 916 (59.4) | NR | 1252 (45.7) |
| Not reported | 6303 (18.8) | NR | 995 (36.3) |
| Epilepsy type | |||
| Focal | 9882 (29.5) | 8737 (38.8) | 1123 (41.0) |
| Generalized | 6810 (20.3) | 4116 (18.3) | 575 (21.0) |
| Indeterminate/mixed | 16 850 (50.2) | 9643 (42.9) | 1043 (38.1) |
| Brain tumor | 1237 (3.7) | 1554 (6.9) | 373 (13.6) |
| Depression | 12 617 (37.6) | 5555 (24.7) | 651 (23.8) |
| Anxiety | 7958 (23.7) | 6148 (27.3) | 336 (12.3) |
| Drug eras | |||
| 2 | 20 302 (60.5) | 14 619 (65.0) | 1750 (63.8) |
| 3 | 8000 (23.9) | 4880 (21.7) | 598 (21.8) |
| 4 | 3366 (10.0) | 1859 (8.3) | 223 (8.1) |
| ≥5 | 1874 (5.6) | 1138 (5.1) | 170 (6.2) |
| Outpatient visitsa | |||
| 0 | 3350 (10.0) | 2107 (9.4) | 416 (15.2) |
| 1 | 3229 (9.6) | 2945 (13.1) | 445 (16.2) |
| ≥2 | 26 963 (80.4) | 17 444 (77.5) | 1880 (68.6) |
| Inpatient visitsa | |||
| 0 | 24 909 (74.3) | 17 920 (79.7) | 1769 (64.5) |
| 1 | 5601 (16.7) | 3225 (14.3) | 668 (24.4) |
| ≥2 | 3032 (9.0) | 1351 (6.0) | 304 (11.1) |
| Emergency visitsa | |||
| 0 | 19 557 (58.3) | 16 740 (74.4) | 1910 (69.7) |
| 1 | 6292 (18.8) | 3404 (15.1) | 454 (16.6) |
| ≥2 | 7693 (22.9) | 2352 (10.5) | 377 (13.8) |
| CT heada | 8853 (26.4) | 4138 (18.4) | 515 (18.8) |
| Prior observation time, mean (SD), y | 4.35 (2.44) | 4.37 (2.55) | 10.95 (0.22) |
| Post observation time, mean (SD), y | 2.52 (0.62) | 2.31 (0.64) | 4.67 (0.184) |
Abbreviations: CT, computed tomography; NR, not reported.
Count of events in the 1 year preceding the index encounter.
Use of neurodiagnostic studies among patients with DRE in the 3 cohorts is shown in the Figure. eTable 2 in Supplement 1 provides corresponding precise counts and percentages. The proportion of patients meeting the comprehensive evaluation main outcome was 4.5% (n = 1520) in the MDCD cohort, 8.0% (n = 1796) in the CCAE cohort, and 14.3% (n = 393) in the CUMC cohort. Patients with DRE at CUMC had greater use of all neurodiagnostic studies than those in the CCAE cohort, while those in the MDCD cohort had the lowest use. This result held true for each category of neurodiagnostic study and the composite presurgical evaluation outcome both within 2 years of the index date and all time in the database. Video EEG (24.9% MDCD, 28.4% CCAE, 63.2% CUMC) and brain MRI (35.6% MDCD, 43.4% CCAE, 52.6% CUMC) were performed more regularly than neuropsychological evaluation (13.0% MDCD, 16.6% CCAE, 19.2% CUMC) or advanced imaging (3.2% MDCD, 5.4% CCAE, 13.1% CUMC).
Figure. Use of Neurodiagnostic Studies Among Patients With Drug-Resistant Epilepsy (DRE).
Precise counts and percentages are provided in eTable 2 in Supplement 1. Composite presurgical evaluation requires magnetic resonance imaging (MRI) of the brain or advanced imaging study, inpatient video electroencephalography (EEG), and neuropsychological evaluation. CCAE indicates MarketScan Commercial Claims and Encounters; CUMC, Columbia University Medical Center; MDCD, MarketScan Multi-State Medicaid Database.
Results of the multivariable logistic regression testing association with the composite comprehensive presurgical evaluation outcome are reported in Table 2 (eTables 3-5 in the Supplement present the results of univariate analysis). The numbers of outpatient and nonemergency inpatient visits in the year preceding the index date were consistently associated with comprehensive evaluation. For example, patients with 2 or more inpatient visits had greater odds of evaluation than those with no visits in MDCD (odds ratio [OR], 3.51; 95% CI, 3.00-4.10; P < .001), CCAE (OR, 5.08; 95% CI, 4.30-5.99; P < .001), and CUMC (OR, 4.35; 95% CI, 3.02-6.27; P < .001). Patients with 2 or more outpatient visits had greater odds of evaluation than those with no visits in MDCD (OR, 4.18; 95% CI, 2.82-6.49; P < .001), CCAE (OR, 5.05; 95% CI, 3.46-7.71; P < .001), and CUMC (OR, 4.95; 95% CI, 2.76-9.87; P < .001). An incrementally greater number of antiseizure medications was associated with greater odds of presurgical evaluation in the CCAE database only (≥5 drugs: OR, 1.51; 95% CI, 1.24-1.83; P < .001).
Table 2. Association of Demographic and Clinical Variables With Comprehensive Presurgical Evaluation Within 2 Years of the Index Date in Multivariable Logistic Regression.
| Characteristic | MarketScan Multi-State Medicaid Database | MarketScan Commercial Claims and Encounters | Columbia University Medical Center | |||
|---|---|---|---|---|---|---|
| OR (95% CI) | P value | OR (95% CI) | P value | OR (95% CI) | P value | |
| No. | 33 542 | 22 496 | 2741 | |||
| Age, y | ||||||
| 0-17 | 1 [Reference] | NR | 1 [Reference] | NR | 1 [Reference] | NR |
| 18-39 | 0.24 (0.20-0.27) | <.001 | 0.46 (0.41-0.52) | <.001 | 2.54 (1.86-3.49) | <.001 |
| 40-64 | 0.16 (0.13-0.19) | <.001 | 0.37 (0.32-0.42) | <.001 | 2.04 (1.45-2.89) | <.001 |
| 65-84 | 0.04 (0.02-0.09) | <.001 | NR | NR | 1.28 (0.78-2.06) | .32 |
| Gender | ||||||
| Female | 1 [Reference] | NR | NIa | NIa | NIa | NIa |
| Male | 0.98 (0.88-1.09) | .71 | NIa | NIa | NIa | NIa |
| Ethnicity | ||||||
| Not Hispanic or Latino | NR | NR | NR | NR | 1 [Reference] | NR |
| Hispanic or Latino | 1 [Reference] | NR | NR | NR | 0.80 (0.58-1.09) | .22 |
| Not reported | 0.71 (0.54-0.93) | .01 | NR | NR | 0.79 (0.56-1.11) | .22 |
| Race | ||||||
| Asian | NR | NR | NR | NR | 1.20 (0.85-1.67) | .32 |
| Black | 0.92 (0.80-1.06) | .29 | NR | NR | 1.02 (0.57-1.74) | >.94 |
| White | 1 [Reference] | NR | NR | NR | 1 [Reference] | NR |
| Not reported | 0.98 (0.85-1.13) | .81 | NR | NR | 0.84 (0.61-1.14) | .32 |
| Epilepsy type | NI | |||||
| Focal | 1 [Reference] | NR | 1 [Reference] | NR | 1 [Reference] | NR |
| Generalized | 0.42 (0.35-0.49) | <.001 | 0.41 (0.34-0.48) | <.001 | 0.21 (0.14-0.32) | <.001 |
| Indeterminate/mixed | 0.73 (0.65-0.82) | <.001 | 0.76 (0.67-0.85) | <.001 | 0.42 (0.32-0.55) | <.001 |
| Brain tumor | 1.72 (1.35-2.17) | <.001 | 1.61 (1.36-1.91) | <.001 | NIa | NI a |
| Depression | 1.12 (0.96-1.30) | .14 | NIa | NIa | NIa | NIa |
| Anxiety | 1.08 (0.93-1.25) | .31 | 1.27 (1.13-1.42) | <.001 | NIa | NIa |
| Drug eras | NIa | NIa | ||||
| 2 | 1 [Reference] | NR | 1 [Reference] | NR | NIa | NIa |
| 3 | 0.82 (0.71-0.94) | .005 | 1.23 (1.09-1.39) | .001 | NIa | NIa |
| 4 | 0.90 (0.75-1.07) | .22 | 1.37 (1.15-1.62) | <.001 | NIa | NIa |
| ≥5 | 0.86 (0.69-1.07) | .21 | 1.51 (1.24-1.83) | <.001 | NIa | NIa |
| Inpatient visitsb | ||||||
| 0 | 1 [Reference] | NR | 1 [Reference] | NR | 1 [Reference] | NR |
| 1 | 2.74 (2.41-3.13) | <.001 | 3.85 (3.41-4.36) | <.001 | 3.16 (2.41-4.14) | <.001 |
| ≥2 | 3.51 (3.00-4.10) | <.001 | 5.08 (4.30-5.99) | <.001 | 4.35 (3.02-6.27) | <.001 |
| Outpatient visitsb | ||||||
| 0 | 1 [Reference] | NR | 1 [Reference] | NR | 1 [Reference] | NR |
| 1 | 1.90 (1.17-3.16) | .01 | 1.81 (1.15-2.93) | .01 | 3.55 (1.85-7.41) | <.001 |
| ≥2 | 4.18 (2.82-6.49) | <.001 | 5.05 (3.46-7.71) | <.001 | 4.95 (2.76-9.87) | <.001 |
| Emergency visitsb | ||||||
| 0 | 1 [Reference] | NR | 1 [Reference] | NR | NIa | NIa |
| 1 | 1.25 (1.07-1.45) | .004 | 1.06 (0.92-1.22) | .43 | NIa | NIa |
| ≥2 | 1.66 (1.44-1.92) | <.001 | 1.09 (0.93-1.28) | .27 | NIa | NIa |
| CT headb | 0.98 (0.86-1.13) | .78 | 1.03 (0.90-1.18) | .73 | 1.18 (0.88-1.58) | .29 |
Abbreviations: CT, computed tomography; NI, not indicated; NR, not reported; OR, odds ratio; Reference, reference category for binary logistic regression.
Excluded from the multivariable model due to lack of significance (P < .05) in univariate analysis.
Count of events in the 1 year preceding the index date.
Generalized epilepsy was less likely to be evaluated than focal epilepsy in all databases (MDCD: OR, 0.42; 95% CI, 0.35-0.49; P < .001; CCAE: OR, 0.41; 95% CI, 0.34-0.48; P < .001; CUMC: OR, 0.21; 95% CI, 0.14-0.32; P < .001). A history of brain tumor was associated with increased the odds of evaluation in MDCD (OR, 1.72; 95% CI, 1.35-2.17; P < .001) and CCAE (OR, 1.61; 95% CI, 1.36-1.91; P < .001).
Adults (aged ≥18 years) had lower odds of evaluation in MDCD (aged 18-39 years: OR, 0.24; 95% CI, 0.20-0.27; P < .001; aged 40-64 years: OR, 0.16; 95% CI, 0.13-0.19; P < .001; aged 65-84 years: OR, 0.04; 95% CI, 0.02-0.09; P < .001) and CCAE (aged 18-39 years: OR, 0.46; 95% CI, 0.41-0.52; P < .001; aged 40-64 years: OR, 0.37; 95% CI, 0.32-0.42; P < .001), while they had greater odds in the CUMC database, with the exception of those aged 65 years and older (aged 18-39 years: OR, 2.54; 95% CI, 1.86-3.49; P < .001; aged 40-64 years: OR, 2.04; 95% CI, 1.45-2.89; P < .001). Adults in the MDCD cohort had the lowest proportion of evaluation of any demographic group (aged 18-39 years: 2.5%; aged 40-64 years: 2.0%; aged 65-84 years: 0.5%).
Evaluating race and ethnicity was limited by missing data. The MDCD and CUMC data had high rates of unreported race (MDCD: 18.8%; CUMC: 36.3%) and ethnicity (MDCD: 97.4%; CUMC: 20.9%), while CCAE does not report race or ethnicity. In univariate analysis, Hispanic or unreported ethnicity and unreported race were associated with lower odds of comprehensive evaluation at CUMC, and unreported ethnicity was associated with lower odds in MDCD. However, no associations were present after adjustment for other variables in multivariable logistic regression.
Discussion
Key Findings
Using OMOP data standards across 3 US databases, we identified a notable gap in the use of neurodiagnostic evaluation for patients with DRE. Patients in the MDCD cohort were less likely to undergo each neurodiagnostic study and our composite proxy for comprehensive evaluation than those in the CCAE cohort, while those in the CUMC cohort had the highest use. Magnetic resonance imaging and video EEG were performed more frequently than neuropsychological testing and advanced imaging studies. The number of nonemergency inpatient and outpatient epilepsy visits was consistently associated with greater use of evaluation. Patients with focal epilepsy underwent evaluation more frequently than those with generalized epilepsy. Adults were less likely to be evaluated than children in either insurance cohort but more likely at CUMC.
Treatment Gap
Our study captures clinical practice patterns of care for patients with DRE in longitudinal US data, providing a more complete view of the evaluation process in various populations. Previous work suggests that referral to comprehensive centers and epilepsy surgery are underused, but findings have been inferred from cross-sectional inpatient data12,13 and self-reporting from experienced centers.14,15,16 Moreover, reasons for not having surgery are complex and depend on a variety of nuanced clinical factors and shared decision-making between patients and clinicians. However, an interdisciplinary US practice parameter,7 American Academy of Neurology quality measure,22 and International League Against Epilepsy expert consensus statement8 recommend that patients with DRE undergo comprehensive presurgical evaluation. An interdisciplinary evaluation also offers more than just a pathway to surgery. The team may, for example, recognize nonepileptic seizures, diagnose underlying causes, tailor medication regimens, and offer various complementary approaches.17 Our findings suggest that the surgery treatment gap begins earlier in the care pathway than previously reported, with underuse of presurgical neurodiagnostic studies among patients with DRE.
By using the most widely available international common data model, we also establish reproducible methods for quality measurement and future research across the OHDSI network.20,23 Beyond epilepsy, we show that OHDSI data standards can be used to measure adherence with best practices and identify care gaps in multiple populations and settings.
Insurance, Care Setting, and Frequency of Visits
We identified disparities in evaluation based on insurance type and care setting. Patients in the MDCD cohort had lower use of each neurodiagnostic study than those in the CCAE cohort, while those in the CUMC cohort had the highest use. The discrepancy was particularly stark for use of video EEG, a resource-intensive inpatient test. Numerous studies have reported disparities in appropriate care for patients with Medicaid.12,24,25 Patients in the MDCD and CCAE cohorts had similar rates of nonemergency outpatient and inpatient visits and antiseizure medication trials, suggesting disparities may be more pronounced for expensive and logistically complex neurodiagnostic studies not routinely available outside specialized centers.
Patients in the CUMC database had the greatest use of neurodiagnostic studies despite a diverse payer mix, suggesting that care setting (ie, receiving care in a health system with a comprehensive epilepsy center) may be a more important determinant of neurodiagnostic evaluation than insurance type. This finding is preliminary and only reflects the experience of a single center compared with national insurance databases. However, it is consistent with prior hospital discharge data findings that growth in epilepsy hospitalizations has been accompanied by a decline in epilepsy surgery rates as patients are increasingly admitted to low surgical volume hospitals.12
Our study also found that more elective patient contact (outpatient and nonemergency inpatient epilepsy visits) was associated with comprehensive evaluation. It is not clear whether more patient contact leads to evaluation; however, visits are logical prerequisites to diagnostic studies. Moreover, surveys suggest that patient trust in their epilepsy clinician is an important factor in considering surgery.26,27,28 It is plausible that interventions aimed at increasing elective epilepsy patient encounters would lead to increased downstream use of diagnostic studies and surgery.
Epilepsy Type
Patients with generalized epilepsy were less likely to be evaluated than those with focal epilepsy. Focal epilepsy is more amenable to surgical intervention and so it may be clinically appropriate that it is more often evaluated with neurodiagnostic studies. However, the aforementioned guidelines recommend that all patients with DRE, regardless of generalized or focal type, undergo evaluation.7,8,22 Distinguishing between focal and generalized epilepsy can be challenging,29 and diagnostic studies may be needed to classify epilepsy type, diagnose epilepsy syndromes, and guide nonoperative management. Neuromodulation is also increasing surgical options for multifocal and generalized epilepsy.30,31 Lower rates of evaluation among patients with generalized and indeterminate epilepsy may therefore represent an important care gap.
Age and Race and Ethnicity
Younger patients had greater odds of evaluation in both the MDCD and CCAE cohorts. This finding is notable given that early seizure control portends more favorable clinical outcomes and historically long delays in referral for surgery. It may also reflect legitimate concerns about frailty and surgical risk in older individuals.14 However, it also points to a potential bias against evaluating adults, particularly older patients19 and those in the Medicaid population. It is not clear why we observed higher rates of evaluation among adults compared with pediatric patients at our center. One possible explanation is that this finding reflects a higher rate of external referrals for comprehensive evaluation among pediatric patients. Such patients often undergo diagnostic studies before referral that would not be captured in our database.
Assessment of race and ethnicity was limited by underreporting in all databases. Previous work has reported disparities in epilepsy care based on race and ethnicity.32,33 Reporting was highest in our medical center’s database, pointing to one advantage of EHR-derived data compared with insurance claims data, although it was still missing ethnicity in 20.9% and race in 36.3% of patients. We observed lower odds of evaluation among patients with Hispanic or Latino ethnicity and those not reporting race or ethnicity in the CUMC cohort in univariate analysis. No association was found after adjustment for other variables.
Limitations
The study has several important limitations. First, all inclusion criteria, outcomes, and variables were derived from coded data that are primarily used for billing purposes and have variable sensitivity and specificity. There was no access to the medical records, patients, or clinicians. Several authors of the present study found that our DRE inclusion criteria have high specificity in our EHR data, but the positive predictive value in various cohorts is subject to prevalence and not known in the insurance claims databases.12 Our composite outcome is an imperfect proxy for comprehensive evaluation, which typically includes additional diagnostic studies based on clinical indication and an interdisciplinary conference. Moreover, the US practice parameter specifically recommends referral to a National Association of Epilepsy Centers–certified center,7 which were not identifiable in OMOP data. Provider and payer taxonomy were not currently available.
Second, there are limitations in comparing insurance claims data and EHR-derived data. Our EHR-derived CUMC data may not be representative of all academic medical centers with a comprehensive epilepsy center. The CUMC database has an open cohort and may therefore underestimate outcomes for patients who received care at multiple centers during the observation period. The CUMC database also had longer observation times than the claims databases before and after the index date, which may affect clinical characteristics and the likelihood of meeting the outcome. However, given our use of OHDSI data standards, our methods can be readily translated to numerous additional OHDSI-participating sites in the US and elsewhere for further validation and comparison of our findings.
Conclusions
The findings of this cross-sectional study suggest there is a gap in the use of diagnostic studies to evaluate patients with DRE. Care setting, insurance type, frequency of nonemergency visits, and epilepsy type are all associated with evaluation. A common data model can be used to measure adherence to best practices across a variety of observational data sources.
eTable 1. Observational Health Data Sciences and Informatics (OHDSI) Standardized Vocabularies Concept Codes (available at athena.ohsi.org).
eTable 2. Use of Neurodiagnostic Studies Among Patients With Drug-Resistant Epilepsy
eTable 3. Association of Demographic and Clinical Variables With Comprehensive Presurgical Evaluation Within 2 Years of the Index Date – Univariate Analysis – MarketScan Multi-State Medicaid Database (MDCD)
eTable 4. Association of Demographic and Clinical Variables With Comprehensive Presurgical Evaluation Within 2 Years of the Index Date – Univariate Analysis – MarketScan Commercial Claims and Encounters (CCAE)
eTable 5. Association of Demographic and Clinical Variables With Comprehensive Presurgical Evaluation Within 2 Years of the Index Date – Univariate Analysis – Academic Medical Center (CUMC) Database
Data Sharing Statement
References
- 1.Zack MM, Kobau R. National and state estimates of the numbers of adults and children with active epilepsy—United States, 2015. MMWR Morb Mortal Wkly Rep. 2017;66(31):821-825. doi: 10.15585/mmwr.mm6631a1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Sperling MR, Barshow S, Nei M, Asadi-Pooya AA. A reappraisal of mortality after epilepsy surgery. Neurology. 2016;86(21):1938-1944. doi: 10.1212/WNL.0000000000002700 [DOI] [PubMed] [Google Scholar]
- 3.Jokeit H, Ebner A. Long term effects of refractory temporal lobe epilepsy on cognitive abilities: a cross sectional study. J Neurol Neurosurg Psychiatry. 1999;67(1):44-50. doi: 10.1136/jnnp.67.1.44 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Spencer SS, Berg AT, Vickrey BG, et al. ; Multicenter Study of Epilepsy Surgery . Health-related quality of life over time since resective epilepsy surgery. Ann Neurol. 2007;62(4):327-334. doi: 10.1002/ana.21131 [DOI] [PubMed] [Google Scholar]
- 5.Kwan P, Brodie MJ. Early identification of refractory epilepsy. N Engl J Med. 2000;342(5):314-319. doi: 10.1056/NEJM200002033420503 [DOI] [PubMed] [Google Scholar]
- 6.Kwan P, Arzimanoglou A, Berg AT, et al. Definition of drug resistant epilepsy: consensus proposal by the ad hoc task force of the ILAE Commission on Therapeutic Strategies. Epilepsia. 2010;51(6):1069-1077. doi: 10.1111/j.1528-1167.2009.02397.x [DOI] [PubMed] [Google Scholar]
- 7.Engel J Jr, Wiebe S, French J, et al. ; Quality Standards Subcommittee of the American Academy of Neurology; American Epilepsy Society; American Association of Neurological Surgeons . Practice parameter: temporal lobe and localized neocortical resections for epilepsy: report of the Quality Standards Subcommittee of the American Academy of Neurology, in association with the American Epilepsy Society and the American Association of Neurological Surgeons. Neurology. 2003;60(4):538-547. doi: 10.1212/01.WNL.0000055086.35806.2D [DOI] [PubMed] [Google Scholar]
- 8.Jehi L, Jette N, Kwon CS, et al. Timing of referral to evaluate for epilepsy surgery: expert consensus recommendations from the Surgical Therapies Commission of the International League Against Epilepsy. Epilepsia. 2022;63(10):2491-2506. doi: 10.1111/epi.17350 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Wiebe S, Blume WT, Girvin JP, Eliasziw M; Effectiveness and Efficiency of Surgery for Temporal Lobe Epilepsy Study Group . A randomized, controlled trial of surgery for temporal-lobe epilepsy. N Engl J Med. 2001;345(5):311-318. doi: 10.1056/NEJM200108023450501 [DOI] [PubMed] [Google Scholar]
- 10.Engel J Jr, McDermott MP, Wiebe S, et al. ; Early Randomized Surgical Epilepsy Trial (ERSET) Study Group . Early surgical therapy for drug-resistant temporal lobe epilepsy: a randomized trial. JAMA. 2012;307(9):922-930. doi: 10.1001/jama.2012.220 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Dwivedi R, Ramanujam B, Chandra PS, et al. Surgery for drug-resistant epilepsy in children. N Engl J Med. 2017;377(17):1639-1647. doi: 10.1056/NEJMoa1615335 [DOI] [PubMed] [Google Scholar]
- 12.Englot DJ, Ouyang D, Garcia PA, Barbaro NM, Chang EF. Epilepsy surgery trends in the United States, 1990-2008. Neurology. 2012;78(16):1200-1206. doi: 10.1212/WNL.0b013e318250d7ea [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Schiltz NK, Koroukian SM, Lhatoo SD, Kaiboriboon K. Temporal trends in pre-surgical evaluations and epilepsy surgery in the US from 1998 to 2009. Epilepsy Res. 2013;103(2-3):270-278. doi: 10.1016/j.eplepsyres.2012.07.016 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Choi H, Carlino R, Heiman G, Hauser WA, Gilliam FG. Evaluation of duration of epilepsy prior to temporal lobe epilepsy surgery during the past two decades. Epilepsy Res. 2009;86(2-3):224-227. doi: 10.1016/j.eplepsyres.2009.05.014 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Kaiboriboon K, Malkhachroum AM, Zrik A, et al. Epilepsy surgery in the United States: analysis of data from the National Association of Epilepsy Centers. Epilepsy Res. 2015;116:105-109. doi: 10.1016/j.eplepsyres.2015.07.007 [DOI] [PubMed] [Google Scholar]
- 16.Jehi L, Friedman D, Carlson C, et al. The evolution of epilepsy surgery between 1991 and 2011 in nine major epilepsy centers across the United States, Germany, and Australia. Epilepsia. 2015;56(10):1526-1533. doi: 10.1111/epi.13116 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Engel J Jr. What can we do for people with drug-resistant epilepsy? the 2016 Wartenberg Lecture. Neurology. 2016;87(23):2483-2489. doi: 10.1212/WNL.0000000000003407 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Institute of Medicine (US) Committee on the Public Health Dimensions of the Epilepsies . In: England MJ, Liverman CT, Schultz AM, Strawbridge LM, eds. Epilepsy Across the Spectrum: Promoting Health and Understanding. National Academies Press; 2012. Accessed October 5, 2020. https://www.ncbi.nlm.nih.gov/books/NBK91506/ [PubMed] [Google Scholar]
- 19.Hill CE, Lin CC, Terman SW, et al. Predictors of referral for long-term EEG monitoring for Medicare beneficiaries with drug-resistant epilepsy. Epilepsia Open. 2023;8(3):1096-1110. doi: 10.1002/epi4.12789 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Hripcsak G, Duke JD, Shah NH, et al. Observational Health Data Sciences and Informatics (OHDSI): opportunities for observational researchers. Stud Health Technol Inform. 2015;216:574-578. [PMC free article] [PubMed] [Google Scholar]
- 21.Castano VG, Spotnitz M, Waldman GJ, et al. Identification of patients with drug-resistant epilepsy in electronic medical record data using the Observational Medical Outcomes Partnership Common Data Model. Epilepsia. 2022;63(11):2981-2993. doi: 10.1111/epi.17409 [DOI] [PubMed] [Google Scholar]
- 22.Fountain NB, Van Ness PC, Bennett A, et al. Quality improvement in neurology: epilepsy update quality measurement set. Neurology. 2015;84(14):1483-1487. doi: 10.1212/WNL.0000000000001448 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Hripcsak G, Ryan PB, Duke JD, et al. Characterizing treatment pathways at scale using the OHDSI network. Proc Natl Acad Sci U S A. 2016;113(27):7329-7336. doi: 10.1073/pnas.1510502113 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Ermer T, Walters SL, Canavan ME, et al. Understanding the implications of Medicaid expansion for cancer care in the US: a review. JAMA Oncol. 2022;8(1):139-148. doi: 10.1001/jamaoncol.2021.4323 [DOI] [PubMed] [Google Scholar]
- 25.Saldanha IJ, Adam GP, Kanaan G, et al. Health insurance coverage and postpartum outcomes in the US: a systematic review. JAMA Netw Open. 2023;6(6):e2316536. doi: 10.1001/jamanetworkopen.2023.16536 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Solli E, Colwell NA, Say I, et al. Deciphering the surgical treatment gap for drug-resistant epilepsy (DRE): a literature review. Epilepsia. 2020;61(7):1352-1364. doi: 10.1111/epi.16572 [DOI] [PubMed] [Google Scholar]
- 27.Khoo A, de Tisi J, Mannan S, O’Keeffe AG, Sander JW, Duncan JS. Reasons for not having epilepsy surgery. Epilepsia. 2021;62(12):2909-2919. doi: 10.1111/epi.17083 [DOI] [PubMed] [Google Scholar]
- 28.Samanta D, Ostendorf AP, Willis E, et al. Underutilization of epilepsy surgery: part I, a scoping review of barriers. Epilepsy Behav. 2021;117:107837. doi: 10.1016/j.yebeh.2021.107837 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Vulpius SA, Werge S, Jørgensen IF, et al. Text mining of electronic health records can validate a register-based diagnosis of epilepsy and subgroup into focal and generalized epilepsy. Epilepsia. 2023;64(10):2750-2760. doi: 10.1111/epi.17734 [DOI] [PubMed] [Google Scholar]
- 30.Haneef Z, Skrehot HC. Neurostimulation in generalized epilepsy: a systematic review and meta-analysis. Epilepsia. 2023;64(4):811-820. doi: 10.1111/epi.17524 [DOI] [PubMed] [Google Scholar]
- 31.Touma L, Dansereau B, Chan AY, et al. Neurostimulation in people with drug-resistant epilepsy: systematic review and meta-analysis from the ILAE Surgical Therapies Commission. Epilepsia. 2022;63(6):1314-1329. doi: 10.1111/epi.17243 [DOI] [PubMed] [Google Scholar]
- 32.Burneo JG, Shariff SZ, Liu K, Leonard S, Saposnik G, Garg AX. Disparities in surgery among patients with intractable epilepsy in a universal health system. Neurology. 2016;86(1):72-78. doi: 10.1212/WNL.0000000000002249 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Burneo JG, Jette N, Theodore W, et al. ; Task Force on Disparities in Epilepsy Care; North American Commission of the International League Against Epilepsy . Disparities in epilepsy: report of a systematic review by the North American Commission of the International League Against Epilepsy. Epilepsia. 2009;50(10):2285-2295. doi: 10.1111/j.1528-1167.2009.02282.x [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eTable 1. Observational Health Data Sciences and Informatics (OHDSI) Standardized Vocabularies Concept Codes (available at athena.ohsi.org).
eTable 2. Use of Neurodiagnostic Studies Among Patients With Drug-Resistant Epilepsy
eTable 3. Association of Demographic and Clinical Variables With Comprehensive Presurgical Evaluation Within 2 Years of the Index Date – Univariate Analysis – MarketScan Multi-State Medicaid Database (MDCD)
eTable 4. Association of Demographic and Clinical Variables With Comprehensive Presurgical Evaluation Within 2 Years of the Index Date – Univariate Analysis – MarketScan Commercial Claims and Encounters (CCAE)
eTable 5. Association of Demographic and Clinical Variables With Comprehensive Presurgical Evaluation Within 2 Years of the Index Date – Univariate Analysis – Academic Medical Center (CUMC) Database
Data Sharing Statement
