ABSTRACT
Rationale
Longer duration of epilepsy before surgery is a predictor of poor outcome. While referral delays of surgical candidates are well documented, factors causing delay during the presurgical evaluation remain unclear and may vary depending on institutional characteristics. By benchmarking the duration of presurgical evaluation across multiple centers and identifying patient and evaluation characteristics contributing to duration, we can ascertain best practices and address modifiable contributors to reduce delays.
Methods
We queried the Pediatric Epilepsy Research Consortium Surgery Database, a prospective, observational multicenter study enrolling children 0–18 years at 27 US pediatric epilepsy centers, for all patients undergoing initial presurgical evaluation for drug‐resistant epilepsy (DRE). We included patients with completed evaluations and data on duration from initiation of presurgical evaluation to final surgical decision. We compared patient characteristics and evaluation components between those with long duration evaluations (> 75% quartile) and those with short evaluations (< 25% quartile). Akaike information criteria selection identified variables associated with longer duration. From these, we developed a logistic prediction model for evaluation duration, using a random 80/20 training/testing split of the entire cohort. The model was tested among institutions with ≥ 10 patients in the cohort to assess its accuracy in predicting long durations. Linear models for each site assessed each variable's impact on duration. Variables with < 10% of the patient population at each site were excluded. Beta values were compared to identify intra‐ and inter‐institution variability and to delineate institutions with the shortest added duration for each variable.
Results
Of 2318 patients undergoing surgical evaluation, 1655 (71%) from 23 sites had complete data. Median evaluation duration was 8 weeks (interquartile range 3–22); 453 (27%) were short‐duration evaluations and 414 (25%) were long‐duration evaluations. Multiple patient and evaluation characteristics were associated with duration (Table 1). Table 6 provides the average duration each variable contributes to evaluation by site, highlighting the shortest durations compared with other groups.
Conclusions
Duration of presurgical evaluation for DRE can be accurately modeled using multiple patient characteristics and testing strategies commonly employed in epilepsy surgery evaluations. This predictive model can not only estimate evaluation duration but also identify opportunities to improve systemic efficiency. Institution‐level modeling identifies specific program strengths, providing an opportunity to learn from successful processes. Subsequent research will focus on institutional process mapping to better understand systemic practices that lead to improved efficiencies, then sharing these processes across the consortium to shorten evaluation durations.
Keywords: duration of epilepsy surgery evaluation, epilepsy surgery, pediatric epilepsy
1. Introduction
Epilepsy surgery is an effective treatment option for patients with drug‐resistant epilepsy (DRE), leading to improvement in seizure frequency, better neurodevelopmental outcomes, and enhanced quality of life [1]. Longer duration of DRE is associated with inferior neurodevelopmental outcomes in children and reduced postsurgical seizure‐freedom rates [2, 3, 4, 5, 6, 7]. Unaddressed DRE is most critical in young children, where even brief delays during sensitive neurodevelopmental windows can lead to lasting neurological impairments [2]. Delays also increase the likelihood of seizure‐related injuries, comorbidities, and mortality, including sudden unexpected death in epilepsy (SUDEP) [8]. Despite its proven efficacy, epilepsy surgery remains underutilized and is often delayed by a variety of barriers that hinder timely referral and efficient presurgical evaluations. Factors such as limited access to specialized epilepsy centers, scheduling inefficiencies, a lack of healthcare resources, and cognitive biases during clinical decision‐making have been shown to significantly contribute to delayed presurgical evaluation of DRE [9].
Most studies examining delays in surgical therapy focus on referral delays, defined as the duration between epilepsy onset and referral to a surgical epilepsy center, or total epilepsy duration, defined as the duration from seizure onset to surgical therapy [10]. Addressing referral delays requires changing clinical practice, often through increased awareness and education of referring providers regarding candidate selection and the benefits of surgical therapy in DRE. Despite the Quality Standards Subcommittee of the American Academy of Neurology [11] supporting prompt surgical referral for patients with DRE, studies demonstrate persistent barriers to timely surgical referral to epilepsy surgery ([12] and [13]).
Given the reliance of surgical evaluation on multiple presurgical testing modalities, it is understandable that the surgical evaluation process can be lengthy and also lengthen the time to surgical decision. Time to referral is only one component in the duration of DRE before surgery. The length of time to complete the presurgical evaluation, which varies depending on the processes, expertise, and resources of any epilepsy center, is also an important and perhaps the most feasibly modifiable component, as epilepsy centers have more control over their own workflows. In this study, we seek to analyze the factors affecting time from initiation of presurgical evaluation to final surgical decisions in pediatric epilepsy patients using a large, multicenter, prospective database of children referred for surgical evaluation in the United States. After identifying the characteristics leading to prolonged evaluations, we aim to identify centers with practices associated with higher efficiencies, characterized by shorter durations of evaluations. This study provides novel insights by specifically examining delays occurring after referral to an epilepsy center, a phase over which institutions have direct control, and by quantifying the impact of various patient and evaluation characteristics on this duration.
2. Methods
2.1. Study Design and Population
For the purpose of this study, we included patients ages birth to 18 years from the Pediatric Epilepsy Research Consortium (PERC) database who (1) underwent initial presurgical evaluation with surgical decision rendered, and (2) had data available on the initiation of phase 1 video electroencephalography (EEG) to the final surgical decision, which may have followed a phase 2, invasive EEG investigation. Detailed methods of the PERC surgery database have been previously described [14].
2.2. Data Collection
We included all patients meeting the inclusion criteria from database initiation in January 2018 to March 2024. Study data were collected and managed using REDCap (Research Electronic Data Capture), an electronic data capture tool hosted at Cook Children's Medical Center. REDCap is a secure, web‐based software platform designed to support data capture for research studies, providing (1) an intuitive interface for validated data capture; (2) audit trails for tracking data manipulation and export procedures; (3) automated export procedures for seamless data downloads to common statistical packages; and (4) procedures for data integration and interoperability with external sources.
2.3. Statistical Analysis
Demographic and clinical characteristics of the cohort were summarized using descriptive statistics. Continuous variables were expressed as medians, means, ranges, and interquartile ranges (IQR). Categorical variables were reported as frequencies and percentages. The distribution of the overall duration period, defined as time in weeks from phase 1 initiation to final surgical decision, was tested for normality with the Shapiro–Wilk normality test, the Kolmogorov–Smirnov test, and a normal Q–Q plot. Comparisons of categorical predictors and the duration period were performed using the Wilcoxon rank‐sum test for predictors with two categories and the Kruskal‐Wallis test for predictors with more than two categories. Numeric predictors and the duration period were tested for correlation using the Spearman rank correlation test. A significance level of alpha = 0.05 was used for all comparisons.
The cohort was then split into patients with “long” duration evaluations, defined by patients in the fourth quartile, and patients with “short” durations, defined by a duration period in the lower first quartile. Patient characteristics and evaluation components were compared between these two groups using the chi‐square test for categorical variables and the t‐test for numeric variables.
All variables showing significant differences in affecting the duration period were added into a linear regression model predicting the duration time. Stepwise variable selection using Akaike information criteria (AIC) identified the best predictive variables associated with total duration. From these, linear models for each institution in the cohort assessed each variable's impact on duration time in weeks. Variables with < 10% of the patient population at each site were excluded. Beta values were compared to identify intra‐ and inter‐institution variability and to delineate institutions with the shortest added duration for each variable. We developed a logistic prediction model for long versus short evaluation durations, using a random 80/20 training/testing split of the entire cohort to test overall accuracy. The model was tested among institutions with ≥ 10 patients in the cohort to assess accuracy in predicting durations longer than the median duration time.
3. Results
3.1. Overview of Pediatric Epilepsy Cohort
At the time of data acquisition from the PERC database, there were 3203 patients enrolled in total, of which 2318 (72%) were undergoing initial surgical evaluation, and 1655 (71%) from 23 institutions had completed duration data. Baseline patient characteristics were initially studied to understand the cohort as a whole (Table 1). The cohort comprised 53% males, with a racial composition of primarily white (75.4%), non‐Hispanic (80.4%) participants. Half (50.3%) of the cohort lived within 50 miles of their surgical center, while a third (34.3%) traveled from more than 100 miles away. Magnetic resonance imaging (MRI) results demonstrated suspected epileptogenic lesions in 72%, and 52% had abnormal neurological examinations. Insurance was similarly split between public (45%) and private (54%). The mean age (and range) in years at seizure onset was 4.90 (0–18), age at phase 1 conference was 10.03 (0.1–18.9), and age at surgery was 10.86 (0.1–20.5). The number of failed antiseizure medications (ASMs) was 3.43 (0–19). Ancillary testing was obtained at the discretion of the surgical center. Within this cohort 71.6% underwent neuropsychological testing, 60.4% received positron emission tomography (PET), 27.4% received a functional MRI (fMRI), 21.9% received a magnetoencephalography (MEG), and 14.7% underwent single photon emission computed tomography (SPECT) injection and scanning. The majority of patients (1389, 83.9%) were offered surgery, 897 (65%) were offered a one‐stage surgery, and 492 (35%) were offered a two‐stage approach, inclusive of invasive EEG monitoring. A total of 1136 (81.8%) completed surgery, with 49.5% undergoing surgical procedures intended to render them seizure‐free and 50.5% undergoing planned palliative procedures. At a follow‐up of at least 6 months (mean of 22.8 months, median of 18.5 months, ranging from 6 to 114 months), 729 (64%) had outcome data, with 345 (47%) achieving seizure freedom. Of the patients that underwent definitive surgical procedures intended to render seizure freedom, 68.8% achieved seizure freedom after 6 months. Of the remainder, 11.7% achieved > 90% reduction, 5.3% had ≥ 50% reduction, 7.4% had < 50% reduction, 6.3% underwent additional surgery, and death occurred in < 1%. Of the patients receiving palliative surgery, 21.3% achieved seizure freedom, 17.1% achieved > 90% reduction, 22.9% had ≥ 50% reduction, 29.9% had < 50% reduction, 7.0% underwent additional surgery, and death occurred in 1.5%.
Table 1.
Patient characteristics.
| Cohort characteristics | N = 1655 |
|---|---|
| Male | 876 (53%) |
| Female | 776 (47%) |
| White | 1248 (75%) |
| Black or African American | 135 (8%) |
| Asian | 62 (4%) |
| American Indian/Alaska Native | 19 (1) |
| More than one race | 48 (3%) |
| Unknown | 143 (9%) |
| Hispanic or Latino | 279 (17%) |
| Not Hispanic or Latino | 1331 (80%) |
| Unknown | 45 (3%) |
| Distance from surgical center | |
| 50 miles or less | 828 (50%) |
| 51–100 miles | 258 (16%) |
| 101–500 miles | 464 (28%) |
| Over 500 miles | 96 (6%) |
| International | 8 (< 1%) |
| Normal MRI | 460 (28%) |
| Abnormal MRI | 1191 (72%) |
| Normal neurological exam | 774 (47%) |
| Abnormal neurological exam | 863 (52%) |
| Type of insurance | |
| Public | 737 (45%) |
| Private | 899 (54%) |
| Self‐pay | 16 (1%) |
| Mean age of onset | 4.90 |
| Mean age at phase 1 | 10.03 |
| Mean age at surgery | 10.86 |
| Mean no. of failed ASMs | 3.43 |
| MEG | 363 (22%) |
| PET | 1001 (60%) |
| SPECT | 244 (15%) |
| fMRI | 453 (27%) |
| Neuropsychology testing | 1185 (72%) |
| Surgery offered | 1389 (84%) |
| 1 stage | 897 (65%) |
| 2 stage | 492 (35%) |
| Surgery performed | 1136 (82%) |
| Definitive | 562 (49%) |
| Palliative | 574 (51%) |
3.2. Total Duration From Phase 1 to Surgical Decision
Median and mean evaluation durations were 8 weeks (IQR 3–22) and 18 weeks, respectively. Short evaluation duration, defined as those lasting ≤ 3 weeks, occurred in 453 patients, while 414 had long durations, defined as those lasting ≥ 22 weeks. The normal Q–Q plot, Shapiro–Wilk normality test (p‐value < 0.001), and Kolmogorov–Smirnov test (p‐value < 0.001) all showed signs of non‐normal distribution of this duration period.
Variables and their effect on the total duration time from initial phase 1 evaluation to the final surgical decision made at the epilepsy conference were analyzed (Tables 2 and 3). Variables that showed significant (p < 0.05) differences between mean and median duration periods were Hispanic or Latino ethnicity; residing closer to the epilepsy center; older age at phase 1 referral; seizures less frequent than daily, seizures less frequent than weekly, and seizures more frequent than monthly; multifocal ictal EEG localization; and a non‐lesional MRI. Ancillary testing including MEG, PET, SPECT, and neuropsychological testing, and an overall higher total number of ancillary tests performed was also associated with a significant increase in mean and median evaluation duration. This held true for the occurrence of a two‐stage surgery and a higher number of conferences required to render a surgical decision as well. Lastly, having a longer duration period of presurgical testing was significantly associated with patients not achieving seizure freedom after 6 months (p value = 0.014).
Table 2.
Patient characteristics (binary) that affect overall duration time.
| Variables | Median duration (weeks) | Mean duration (weeks) | p value |
|---|---|---|---|
| Lesion (n = 1191) | 8 | 17.0 | 0.022 |
| Non‐lesional (n = 460) | 8 | 20.6 | |
| Abnormal neuro exam (n = 863) | 8 | 17.6 | 0.70 |
| Normal neuro exam (n = 774) | 8 | 18.5 | |
| MEG (n = 363) | 19 | 25.9 | < 0.001 |
| No MEG (n = 1292) | 6 | 15.8 | |
| PET (n = 1001) | 9 | 19.2 | < 0.001 |
| No PET (n = 654) | 6 | 16.2 | |
| SPECT (n = 244) | 16 | 29.2 | < 0.001 |
| No SPECT (n = 1411) | 7 | 16.1 | |
| fMRI (n = 453) | 10 | 19.6 | 0.033 |
| No fMRI (n = 1202) | 8 | 17.4 | |
| NP testing (n = 1185) | 9 | 19.3 | < 0.001 |
| No NP testing (n = 470) | 5 | 14.8 | |
| Male (n = 876) | 7.35 | 17.6 | 0.127 |
| Female (n = 776) | 9 | 18.5 | |
| Hispanic/Latino (n = 279) | 10.1 | 21.5 | 0.004 |
| Non‐Hispanic/Latino (n = 1331) | 8 | 17.3 | |
| Definitive (n = 562) | 8 | 17 | 0.075 |
| Palliative (n = 574) | 8 | 20 | |
| 1 stage (n = 897) | 6 | 14.4 | < 0.001 |
| 2 stage (n = 492) | 16 | 27.4 | |
| Daily seizures (n = 793) | 7 | 16.9 | 0.003 |
| Not (n = 862) | 9 | 19 | |
| Weekly seizures (n = 436) | 11 | 21.8 | < 0.001 |
| Not (n = 1219) | 7.3 | 16.6 | |
| Monthly seizures (n = 249) | 9 | 17.5 | 0.18 |
| Not (n = 1406) | 8 | 18.1 | |
| > Monthly seizures (n = 173) | 6 | 14.4 | 0.04 |
| Not (n = 1482) | 8 | 18.4 | |
| Seizure freedom after 6 months (n = 346) | 8 | 14.7 | 0.014 |
| No seizure freedom after 6 months (n = 391) | 9 | 22.0 |
Table 3.
Patient characteristics (multicategorical and quantitative) that affect overall duration time.
| Variables | Median duration (weeks) | Mean duration (weeks) | p value |
|---|---|---|---|
| Race | 0.322 | ||
| White (n = 1248) | 8 | 17.3 | |
| Black (n = 135) | 10 | 21.0 | |
| Asian (n = 62) | 10 | 25.5 | |
| American Indian/Alaska Native (n = 19) | 5 | 19.7 | |
| More than one race (n = 48) | 4 | 11.8 | |
| Etiology | 0.443 | ||
| Acquired (n = 375) | 6 | 15.6 | |
| Congenital (n = 441) | 7 | 15.2 | |
| Genetic (n = 191) | 8 | 17.7 | |
| Infectious (n = 18) | 10 | 12.9 | |
| Inflammatory/autoimmune (n = 27) | 17 | 28 | |
| Metabolic (n = 1) | 22 | 22 | |
| Ictal EEG localization | < 0.001 | ||
| Generalized (n = 99) | 4 | 13.5 | |
| Indeterminate (n = 41) | 9 | 17.3 | |
| Mixed general/focal (n = 65) | 8 | 14.6 | |
| Multifocal (n = 211) | 11 | 23.2 | |
| Single focus (n = 506) | 9.5 | 18.8 | |
| Type of insurance | 0.063 | ||
| Private (n = 899) | 8 | 17.2 | |
| Public (n = 737) | 8 | 19.2 | |
| Self‐pay (n = 16) | 4 | 7.6 | |
| Distance from hospital | < 0.001 | ||
| 50 miles or less (n = 828) | 10 | 20.2 | |
| 51–100 miles (n = 258) | 7.5 | 19.3 | |
| 101–500 miles (n = 464) | 7 | 14.8 | |
| Over 500 miles (n = 96) | 4.45 | 12.2 | |
| International (n = 8) | 2 | 2.51 | |
| Quantitative variables | Spearman correlation rho | ||
| Age at onset | 0.027 | 0.278 | |
| Age at referral | 0.075 | 0.002 | |
| Number of conferences | 0.438 | < 0.001 | |
| Number of ancillary tests | 0.242 | < 0.001 | |
| Number of seizure types | ‐0.037 | 0.136 | |
After removing patients with duration of evaluation periods in the middle 50%, the patients identified as having short and long durations were compared to find what variables contributed to these extreme periods. Variables that were significantly associated with longer durations included having Hispanic or Latino ethnicity, living closer to the hospital, having an older age at referral, having less frequent seizures, multifocal seizure foci, and a non‐lesional MRI. Undergoing a MEG, PET, SPECT, and neuropsychological testing was each associated with a greater duration as well as having undergone a greater number of ancillary tests, having been discussed at more epilepsy surgery conferences, and requiring a two‐stage surgery. Further details about this analysis, including frequencies and the odds ratios for each predictor, are found in Tables 4 and 5.
Table 4.
Patient variables (binary) associated with short versus long duration of the evaluation period.
| Predictors (n = 867) | Short duration (n = 453) | Long duration (n = 414) | p value | Odds ratios and interpretation |
|---|---|---|---|---|
| Non‐lesional (n = 256) | 118 | 138 | 0.02 | 1.42 times more likely for longer evaluation duration if non‐lesional |
| Lesional (n = 611) | 335 | 276 | ||
| Normal neuro exam (n = 428) | 224 | 204 | 1 | NA |
| Abnormal neuro exam (n = 439) | 229 | 210 | ||
| No MEG (n = 687) | 429 | 258 | < 0.001 | 10.72 times more likely for longer duration if MEG completed |
| MEG (n = 180) | 24 | 156 | ||
| No PET (n = 358) | 206 | 152 | 0.01 | 1.43 times more likely for longer duration if PET completed |
| PET (n = 509) | 247 | 262 | ||
| No SPECT (n = 729) | 411 | 318 | < 0.001 | 2.94 times more likely for longer duration if SPECT completed |
| SPECT (n = 138) | 42 | 96 | ||
| No fMRI (n = 615) | 332 | 283 | 0.13 | NA |
| fMRI (n = 252) | 121 | 131 | ||
| No NP testing (n = 250) | 156 | 94 | < 0.001 | 1.79 times more for longer duration if NP testing completed |
| NP testing (n = 617) | 297 | 320 | ||
| Female (n = 421) | 213 | 208 | 0.38 | NA |
| Male (n = 446) | 240 | 206 | ||
| Hispanic/Latino (n = 147) | 57 | 90 | < 0.001 | 1.91 times more likely for longer duration among Hispanics |
| Non‐Hispanic/Latino (n = 695) | 381 | 314 | ||
| Definitive (n = 296) | 167 | 129 | 0.10 | NA |
| Palliative (n = 319) | 158 | 161 | ||
| 1 stage (n = 479) | 308 | 171 | < 0.001 | 4.60 times more likely for longer duration if 2 stage study completed |
| 2 stage (n = 278) | 78 | 200 |
Note: NA, not available.
Table 5.
Patient variables (categorical and quantitative) associated with short versus long duration of the evaluation period.
| Predictors (n = 867) | Short duration (n = 453) | Long duration (n = 414) | p value | Odds ratios and interpretation |
|---|---|---|---|---|
| Race | 0.55 | NA | ||
| White (n = 643) | 342 | 301 | ||
| Black (n = 81) | 41 | 40 | ||
| Asian (n = 29) | 12 | 17 | ||
| American Indian/Alaska Native (n = 13) | 7 | 6 | ||
| More than one race (n = 30) | 19 | 11 | ||
| Seizure frequency | 0.006 | Duration of evaluation is longer if seizures occur less often than daily | ||
| Daily (n = 423) | 240 | 183 | ||
| Weekly (n = 225) | 100 | 125 | ||
| Monthly (n = 124) | 57 | 67 | ||
| > Monthly (n = 95) | 56 | 39 | ||
| Etiology | 0.32 | NA | ||
| Acquired (n = 193) | 119 | 74 | ||
| Congenital (n = 235) | 138 | 97 | ||
| Genetic (n = 97) | 51 | 46 | ||
| Infectious (n = 8) | 4 | 4 | ||
| Inflammatory/autoimmune (n = 19) | 8 | 11 | ||
| Metabolic (n = 1) | 0 | 1 | ||
| Ictal EEG localization | < 0.001 | There are significant differences between the types of EEG localization and having a short or long duration | ||
| Generalized (n = 110) | 79 | 31 | ||
| Indeterminate (n = 38) | 18 | 20 | ||
| Mixed general/focal (n = 45) | 26 | 19 | ||
| Multifocal (n = 178) | 77 | 101 | ||
| Single focus (n = 400) | 187 | 213 | ||
| Type of insurance | 0.19 | NA | ||
| Private (n = 471) | 255 | 216 | ||
| Public (n = 385) | 191 | 194 | ||
| Self‐pay (n = 10) | 7 | 3 | ||
| Distance from hospital | < 0.001 | There are significant differences between the distances from the hospital and having a short or long duration | ||
| 50 miles or less (n = 442) | 201 | 241 | ||
| 51–100 miles (n = 123) | 65 | 58 | ||
| 101–500 miles (n = 241) | 145 | 96 | ||
| Over 500 miles (n = 56) | 37 | 19 | ||
| International (n = 5) | 5 | 0 |
| Quantitative variables | p value | Interpretation | ||
|---|---|---|---|---|
| Age at onset (years) | 4.88 | 4.99 | 0.73 | NA |
| Age at referral (years) | 9.79 | 10.80 | 0.005 | Older age at referral is significantly associated with longer durations |
| Number of conferences | 1.07 | 1.71 | < 0.01 | More conferences are significantly associated with longer durations |
| Number of ancillary tests (defined as PET/SPECT/MEG/fMRI/NP) | 1.94 | 2.55 | < 0.01 | More ancillary testing is significantly associated with longer durations |
| Number of seizure types | 1.18 | 1.14 | 0.14 | NA |
Note: NA, not available.
3.3. Regression Equations on Duration Period
The variables with independent, significant association to long versus short duration (Tables 4 and 5) were included in a multivariate regression model. Stepwise selection using AIC was performed on the full model with 13 variables. Nine were found to have strong predictive power associated with the duration of evaluation (i.e., Hispanic or Latino, living 50 miles or less from the hospital, having a non‐lesional MRI, ancillary testing with MEG or SPECT, the total number of ancillary tests performed, undergoing a two‐stage surgery, having multi‐ or single‐focus ictal localization [rather than generalized, mixed generalized/focal, or indeterminate], and discussed at multiple conferences). These nine variables were then used to create a model against the 20 sites that had 10 or more patients in the data set (Table 6). A logistic model predicted long duration periods with an overall accuracy of 71%. For 12 sites, the logistic model predicted long durations with ≥ 75% accuracy; overall accuracy estimates ranged from 50% to 100%. The accuracy of the logistic model was determined for each institution (Table 6). Beta values highlighted indicate the two shortest durations compared with other institutions. Beta values in red text show variables where < 10% of the site's population had this variable occur; due to the small sample size, they may not reflect how efficiently a site is at evaluating patients with that characteristic.
Table 6.
Linear models detailing individual sites and number of weeks attributed to each variable.
| Site | n | Accuracy predicting long duration | Non‐lesional | MEG | SPECT | Hispanic or Latino | 2 stage | Multi‐ or single‐focus ictal | 50 miles or less from hospital | Each conference | Each ancillary test |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Institution 1 | 123 | 50% | 11.11 (33%) | 21.29 (15%) | −2.23 (7%) | 2.86 (23%) | −9.89 (21%) | −1.50 (54%) | 5.37 (59%) | 11.90 (1.19) | −1.48 (1.69) |
| Institution 2 | 39 | 51% | −14.46 (23%) | 45.55 (10%) | 6.09 (46%) | 88.67 (10%) | 4.98 (49%) | 27.15 (87%) | −18.91 (64%) | 21.19 (1.56) | −23.53 (2.97) |
| Institution 3 | 67 | 59% | −2.25 (36%) | 36.89 (4%) | −8.66 (6%) | 7.48 (42%) | −8.78 (34%) | 3.77 (70%) | 3.41 (88%) | 20.88 (1.19) | −1.52 (2.12) |
| Institution 4 | 31 | 66% | 2.96 (35%) | NA | NA | −11.77 (6%) | −0.26 (26%) | 0.57 (55%) | −5.64 (19%) | 17.8 (1.10) | −1.55 (1.52) |
| Institution 5 | 79 | 67% | −20.28 (18%) | 204.31 (3%) | NA | −1.75 (15%) | 30.01 (16%) | −3.88 (67%) | −15.14 (78%) | 23.78 (1.41) | −7.04 (1.48) |
| Institution 6 | 170 | 69% | 0.70 (34%) | 4.02 (2%) | 0.65 (7%) | 9.51 (2%) | 2.73 (23%) | 2.47 (55%) | −0.79 (54%) | 20.82 (1.05) | −0.95 (1.96) |
| Institution 7 | 17 | 71% | −8.18 (24%) | 0.82 (71%) | −4.73 (12%) | 11.57 (47%) | 26.89 (65%) | −34.76 (65%) | −14.72 (65%) | NA (1) | 1.34 (2.35) |
| Institution 8 | 64 | 73% | −2.79 (20%) | −0.81 (30%) | −3.40 (55%) | 1.02 (8%) | 11.92 (27%) | −12.23 (75%) | 1.34 (30%) | 7.80 (1.58) | 6.77 (3.16) |
| Institution 9 | 54 | 75% | −2.57 (24%) | 12.54 (4%) | 3.37 (48%) | −1.42 (11%) | 3.39 (20%) | −4.38 (65%) | 0.22 (61%) | 5.86 (1.30) | 0.50 (2.13) |
| Institution 10 | 21 | 76% | 7.63 (38%) | NA | NA | 8.64 (9.5%) | 8.89 (48%) | −23.89 (91%) | 9.27 (33%) | 16.59 (1.29) | −8.89 (2.38) |
| Institution 11 | 189 | 77% | 2.77 (22%) | NA | −5.13 (2%) | −4.21 (13%) | 0.96 (22%) | −0.51 (65%) | 1.54 (42%) | 18.54 (1.17) | −1.84 (2.57) |
| Institution 12 | 302 | 79% | 3.95 (32%) | 2.19 (59%) | 5.48 (12%) | −0.91 (27%) | −2.66 (32%) | 1.44 (56%) | 0.59 (43%) | 21.88 (1.40) | 0.64 (2.44) |
| Institution 13 | 93 | 81% | −1.94 (26%) | 3.91 (19%) | −4.01 (37%) | −1.00 (4%) | 1.63 (46%) | −2.57 (75%) | 1.65 (3%) | 11.01 (1.53) | −0.26 (2.24) |
| Institution 14 | 35 | 81% | 15.23 (20%) | NA | 35.39 (26%) | 25.56 (14%) | −11.28 (46%) | 10.84 (69%) | −2.8 (74%) | 22.96 (1.54) | −12.23 (2.03) |
| Institution 15 | 47 | 82% | 0.98 (30%) | −2.62 (4%) | −14.12 (2%) | −8.20 (13%) | 1.71 (47%) | −2.50 (64%) | 8.79 (68%) | 7.82 (1.26) | −4.26 (2.06) |
| Institution 16 | 77 | 82% | 0.25 (34%) | 14.78 (42%) | 1.70 (18%) | 7.71 (8%) | −4.08 (18%) | 2.91 (53%) | −0.12 (58%) | 20.26 (1.16) | 0.96 (2.34) |
| Institution 17 | 63 | 85% | 24.02 (38%) | 9.48 (21%) | 41.71 (6%) | −0.63 (35%) | −15.89 (21%) | 19.65 (64%) | 21.98 (29%) | 83.29 (1.12) | −11.29 (1.86) |
| Institution 18 | 73 | 93% | 50.34 (8%) | −14.17 (12%) | 31.60 (7%) | 17.23 (37%) | 29.11 (26%) | 12.15 (70%) | 4.57 (74%) | 14.48 (1.49) | 1.40 (2.42) |
| Institution 19 | 68 | 97% | −24.26 (33%) | −29.20 (18%) | 10.26 (47%) | −47.45 (3%) | 42.29 (44%) | 20.61 (88%) | −22.56 (57%) | 21.29 (1.51) | 6.56 (2.44) |
| Institution 20 | 30 | 100% | 18.57 (30%) | −3.81 (27%) | NA | −28.10 (7%) | −5.59 (47%) | 23.34 (73%) | −7.30 (20%) | 12.62 (1.03) | 14.26 (2) |
*Each beta value of the linear model represents the estimated number of weeks a site will add to its duration time if that characteristic is present. The numbers in parentheses for categorical variables represent the frequency at which that characteristic appears in that site's patient sample. For quantitative variables (number of conferences and ancillary tests), numbers in parentheses are the average numbers reported at that site. Variables in RED have frequencies less than 10% of the patient sample and are not considered optimal examples. Values highlighted in GREEN are the two sites with the shortest duration added for the characteristic. NA, not available.
4. Discussion
This study's key findings include the significant variability in presurgical evaluation timelines across different pediatric epilepsy centers, with median durations ranging from 8 to 19 weeks (IQR 3–22). It is well established that a variety of barriers delay referral to epilepsy surgical evaluation and lengthen the time to surgical decision, including a lack of specialized clinics, scheduling delays, and cognitive biases during the decision‐making process [9]. Longer durations of epilepsy are associated with inferior neurodevelopmental outcomes [2] and worse seizure‐freedom rates in youth postsurgery [2, 3, 4, 5, 6, 7]; however, the degree to which these delays are due to lengthy presurgical evaluation timelines has not been explored. In this study, we examined the factors influencing the duration between the initiation of an epilepsy surgery evaluation and the final surgical decisions in children with epilepsy. By identifying variables that predict longer duration for evaluations, we may be able to inform future practice by (1) eliminating testing that delays evaluation without enhancing outcomes or (2) improving processes related to evaluation.
In the current study, the median and average times from the initiation of the presurgical evaluation to final surgical decision across patients were highly variable, with non‐normal distribution patterns. There is substantial heterogeneity in evaluation timelines, which were influenced by factors ranging from patient sociodemographics and clinical features to specific testing modalities included in the evaluations.
Among sociodemographic factors, significantly longer evaluation times were reported for Hispanic or Latino ethnicities. In fact, they were almost two times more likely to have evaluations of long durations than the cohort with short durations, reflecting potentially decreased access to care and other socioeconomic barriers. While our database does not contain information on language preference, existing literature highlights that barriers to epilepsy care for Hispanic or Latino youth include a lack of resources available in Spanish, lower health literacy, and mistrust, which have been statistically identified as major contributors to delays in diagnosis and treatment [15, 16]. Systemic barriers, such as lack of adequate insurance and geographic constraints, further increase these delays and reduce access to specialized epilepsy care [17], with much lower rates of epilepsy surgery in Hispanic youth (15%) compared with non‐Hispanic white patients (60%). Addressing such disparities calls for targeted interventions, including culturally competent care models, more appropriate insurance policies, and community outreach programs. These merit priority for timely evaluation and to help reduce delays in accessing presurgical evaluation among Hispanic/Latino populations.
We also found that patients referred at older ages were more likely to have a surgical evaluation of long duration. Several factors likely contribute to this finding. Many younger patients with DRE present with seizures and structural lesions, which may require less ancillary testing. Likewise, many of the ancillary tests used in presurgical evaluations have limited utility in younger populations, thus adding less complexity to presurgical evaluations in the younger age groups. Older patients may be more likely to participate in testing essential to identifying the critical cortex, thus further adding complexity in these cases. Finally, older children may be more likely to have established routines such as school and extracurricular activities, which complicates scheduling for evaluations and medical procedures.
Other predictors for longer evaluations included proximity to the epilepsy center, consistent with existing literature [18]. This is likely in part due to the effort institutions exert in batching presurgical testing modalities for patients who live further from the epilepsy center to reduce the travel time associated with the evaluation. Some institutions report an increased likelihood of performing the majority, or all, of the evaluation in the inpatient setting for patients who do not live close to the epilepsy center. This suggests that for those living close to the center, coordination of presurgical testing may be used less often or effectively, resulting in an unnecessary delay. In our study, the initial neurological examination and patient gender did not significantly influence the presurgical evaluation timeline. Previous studies suggest that gender and neurological examinations generally have minimal impact on the presurgical process, indicating that the duration of evaluations is more closely related to the complexity of the epilepsy case, required diagnostic evaluations, and other factors [19].
Clinical features contributing to patient complexity were associated with longer durations of evaluation—the most significant of which is a non‐lesional MRI. When no causative lesion is found, ancillary testing including different neuroimaging modalities and possibly a two‐stage evaluation with intracranial electrode implantation is required to provide localizing information, which adds to the duration of the evaluation. Unsurprisingly, presurgical trajectories among patients with non‐lesional epilepsy were longer compared with patients with lesional epilepsy, as previously reported [20].
Ancillary tests such as PET, SPECT, and MEG were associated with a longer presurgical evaluation. Ancillary tests are typically reserved for further localization of seizure onset in non‐lesional or multifocal cases. Some testing, such as MEG and invasive EEG, may be used for the localization of eloquent cortices near seizure onset zones, adding greater complexity to the surgical evaluation. MEG scans are particularly useful in cases with ambiguous localization of the epileptogenic zone or suspected larger seizure onset zones for accurate localization. However, MEG has very limited availability; only 22 centers associated with the American Clinical Magnetoencephalography Society (ACMEGS) exist in the United States, accounting for less than 17% of epilepsy centers accredited by the National Association of Epilepsy Centers. MEG scarcity is driven by institutional factors such as economic priorities, program strength, patient profiles, and systemic challenges, including regulatory barriers and insurance policies that further extend presurgical evaluation timelines [21]. When a MEG was included, the complete presurgical evaluation was more than 10 times more likely to be ≥ 22 weeks in our study. Therefore, while our database does not capture the specific reasons for delays in obtaining ancillary tests, the inherent shortage and logistical complexities of MEG and similar specialized tests likely contribute significantly to the prolonged evaluation durations we observed.
fMRI is a critical tool in presurgical evaluation for epilepsy, especially for mapping language and memory functions. However, fMRI has been associated with increased evaluation timelines that could be due to its complex nature, requiring specialized expertise in testing and image post‐processing. Interestingly, while fMRI appeared to contribute to the evaluation duration, it was not significantly associated with a longer vs. a shorter duration in our analysis, likely because it has become increasingly commonplace in most epilepsy centers. Lastly, neuropsychological testing is mandatory in epilepsy surgery evaluations for assessing cognitive function, localizing seizure onset zones, and determining lateralization of brain functions, especially in cases involving language or memory. However, adequate neuropsychological testing availability varies depending on finances, geographic location, and institutional resources. For example, many epilepsy centers in the United States lack appropriate neuropsychological personnel or funding for all the tests that could be done, which may prolong the presurgical process [22].
Most patients with public health insurance experience long delays for surgical treatment of DRE [18]. There is evidence from various studies regarding the longer time intervals between DRE diagnosis and surgery in patients with public health insurance compared with those with private insurance [9]. Insurance denials and subsequent appeals further delay the presurgical evaluation. Public and private insurers often balk at necessary diagnostic tests and surgical interventions despite substantial data demonstrating the overall cost‐effectiveness of epilepsy surgery [18]. Interestingly, we did not find insurance status as a predictor of long durations. This is likely because insurance denials contribute more significantly to the duration of referral, whereas once insurance approval for surgical evaluation is obtained, delays may be limited.
Invasive monitoring has become an extremely helpful tool in assessing complex epilepsy cases, especially when related to multifocal seizures, deep epileptogenic zones, and seizure onset near eloquent cortices. Our results show that patients with multifocal ictal EEG patterns and those undergoing two‐stage surgeries had longer periods of evaluation. The association between invasive monitoring and longer evaluation times reflects more than clinical complexity. Scheduling and coordinating invasive monitoring procedures, which requires hospital resources, specialized personnel, and patient availability, adds more time to the presurgical evaluation.
When comparing evaluation duration across institutions for significant variables, there is a great amount of variation (Table 6). Institutions have the ability to handle these patients and epilepsy characteristics efficiently, yet that is not consistent across surgical centers. For example, a non‐lesional patient can expect a range of 24 weeks less or 24 weeks more than expected, depending on the institution where their evaluation is completed. The surgical center's experience, processes, and availability to resources may have an impact on this variability across sites.
Our study is a comprehensive evaluation of patient characteristics and evaluation strategies leading to longer presurgical evaluations for pediatric epilepsy surgery. Data included was collected from the PERC surgery database provided by 23 geographically diverse institutions in the United States. Our study demonstrates variability in the presurgical evaluation processes between each participating epilepsy center, leading to differences in the length of assessment for patients with DRE. Such differences likely arise from a combination of institutional resources, availability of specialized personnel, clinical workflows, scheduling practices, and systemic challenges unique to each center. Some centers prefer streamlined assessments combining several investigations into an inpatient stay, while others do these investigations over several outpatient visits, thus prolonging the overall time. Recognizing factors that contribute to long evaluation durations and identifying centers with more efficient processes is a strength of this study. This provides a strong foundation for process improvement strategies, potentially leading to shorter, more efficient presurgical evaluations. Once identified, efficient evaluation workflows can be mapped and replicated, improving efficiencies across epilepsy centers and reducing the duration of epilepsy before surgical therapy for pediatric DRE.
Regarding the association between longer evaluation duration and poorer seizure‐freedom outcomes, we acknowledge the inherent complexity of patients requiring more extensive presurgical testing. These patients often have more challenging epilepsy pathologies, which may reduce their overall likelihood of seizure freedom regardless of evaluation speed. However, our study focused on identifying modifiable factors within the evaluation process. While a direct comparison of similar cases (matched for complexity) between fast and slow institutions was beyond the scope of this particular analysis, the observed association between longer evaluation duration and reduced seizure freedom suggests that even within complex cases, efficiency in evaluation remains a critical factor in improving patient outcomes. Future studies could explore this relationship with more detailed case matching.
4.1. Study Limitations
The PERC surgery database is intended to collect data on all patients referred for epilepsy surgery, yet some centers may prioritize enrollment of patients undergoing surgery, leading to selection bias. The prediction model was validated using centers with > 10 patients in the cohort; thus, the model may not accurately predict duration at centers with smaller volumes. However, the database includes many National Association of Epilepsy Center level IV facilities across the United States, providing a generalizable assessment of current practice in the United States. The study collects limited data reflecting social determinants of health, and the impact of such variables on durations cannot be directly recorded.
Moreover, among the wide set of demographic and clinical factors tested, we did not examine the possible impact of cultural or linguistic barriers that may occur with Hispanic/Latino patients. Other sources of unmeasured variability may include variation in institutional protocol regarding ancillary testing and workflow for preoperative evaluation, although ancillary testing was partially assessed with the site‐specific regression models. The judgment on the evaluation duration, whether long or short, was based on statistical rather than clinical cut‐offs using the 75th and 25th percentiles, which may not be sensitive to clinical nuances and the subjective experiences of families navigating such processes.
5. Conclusion
A systematic review and meta‐analysis on the effect of earlier or later definitive epilepsy surgery on seizure outcome concluded that a shorter duration of epilepsy is the only modifiable factor that is associated with favorable seizure outcome after epilepsy surgery [10]. Identification of the numerous factors that contribute to delays in referring patients for surgical evaluation has prompted the Surgical Therapies Commission of the International League Against Epilepsy to publish recommendations to offer all patients with DRE a presurgical evaluation [13], emphasizing that epilepsy surgery is still grossly underutilized and too often offered only after patients have experienced long durations of epilepsy. In this study, we show that US institutions performing moderate to high volumes of pediatric epilepsy surgeries vary widely in the time each takes to complete an epilepsy surgery evaluation. While the variables discussed contribute differently to this duration and vary between institutions, time to surgery is lengthened simply with prolonged presurgical evaluation, increasing the likelihood of less favorable neurodevelopmental and seizure‐freedom rates. Here, we show that some epilepsy programs are successfully designed to offer streamlined, rapid presurgical evaluations lasting only weeks compared with those that require several months for complete evaluation. Some variability is explained by centers having more factors associated with lengthy evaluation durations, such as older ages at referral, a greater percentage of Hispanic or Latino patients, a heavy reliance on MEG, a requirement for more provider conferences, or a preference for two‐stage surgeries.
However, our data do not clarify why some programs completed separate components of these evaluations faster than others. Some epileptologists or epilepsy surgeons may hold their institutions accountable for providing timely care despite the complex, nonstandardized approach that pediatric epilepsy surgery evaluation demands. Some programs may have dedicated epilepsy surgery care coordinators to navigate the complex scheduling of presurgical testing. Some institutions may offer epilepsy surgery care coordinators or even preferential scheduling practices after seeing higher attrition rates with longer evaluation durations, which decrease the ratio of patients that complete a financially beneficial neurosurgery compared with those who only undergo components of the presurgical evaluation. Alternatively, some programs may consider the consequences of delayed surgery during a lengthy evaluative process with incurred injuries and mortalities, encouraging them to reduce evaluation duration as much as possible. Lastly, the differences may not be explained by specific provider efforts, use of care coordinators, identification of financial benefit, or understanding of the data, but may be due to the innate organization, culture, or explicit goals of the larger institution within which the epilepsy program operates. However, moving forward, if health systems fail to understand that lengthy presurgical evaluation durations are detrimental in epilepsy surgery and do not incorporate programmatic efforts targeting modifiable factors that will shorten the evaluative process, they will remain complicit in producing less‐than‐optimal neurodevelopmental outcomes and seizure‐freedom rates in pediatric epilepsy patients. Further study will be conducted to identify successful strategies employed by institutions that complete presurgical evaluations quickly. While strategies will need to be tailored depending on the resources at each institution, there are likely generalizable practices that would help reduce the time to surgery in many programs.
Current Knowledge of the Topic
Epilepsy surgery is an effective treatment for DRE in children, leading to improved seizure control and neurodevelopmental outcomes. However, delays in surgical evaluation and intervention are common due to various barriers, including limited access to specialized centers and inefficiencies in presurgical processes.
Question Addressed by the Study
This study aimed to analyze the factors affecting the duration from the initiation of presurgical evaluation to the final surgical decision in pediatric epilepsy patients, using data from a large, multicenter prospective database.
What This Study Adds to Our Knowledge
The study identifies specific variables that contribute to prolonged presurgical evaluations, such as non‐lesional MRI findings, Hispanic or Latino ethnicity, and the use of ancillary tests like MEG and SPECT. It also highlights institutions with more efficient evaluation processes, providing insights into practices that can reduce evaluation durations.
Potential Impact on the Practice of Neurology
By understanding the factors that lead to delays in presurgical evaluations, neurology practices can implement targeted strategies to streamline these processes, potentially improving surgical outcomes and reducing the duration of epilepsy in pediatric patients. This can lead to better neurodevelopmental outcomes and higher rates of seizure freedom postsurgery.
Author Contributions
Ruba Al‐Ramadhani: writing – original draft, formal analysis. Ann Hyslop: formal analysis, writing – original draft. Avery R. Caraway: data curation, formal analysis, methodology. Edward J. Novotny: writing – review and editing. Adam P. Ostendorf: writing – review and editing. Krista L. Eschbach: writing – review and editing. Allyson L. Alexander: writing – review and editing. Lily C. Wong‐Kisiel: writing – review and editing. Dewi F. Depositario‐Cabacar: writing – review and editing. Chima O. Oluigbo: writing – review and editing. Cemal Karakas: writing – review and editing. Samir R. Karia: writing – review and editing. Priyamvada Tatachar: writing – review and editing. Jeffrey Bolton: writing – review and editing. Pilar D. Pichon: writing – review and editing. Daniel W. Shrey: writing – review and editing. Erin Fedak Romanowski: writing – review and editing. Nancy A. McNamara: writing – review and editing. Ernesto Gonzalez‐Giraldo: writing – review and editing. Kurtis Auguste: writing – review and editing. Danilo Bernardo: writing – review and editing. Rani K. Singh: writing – review and editing. Pradeep K. Javarayee: writing – review and editing. Jenny J. Lin: writing – review and editing. Jason C. Coryell: writing – review and editing. Shilpa B. Reddy: writing – review and editing. Abhinaya Ganesh: writing – review and editing. Michael A. Ciliberto: writing – review and editing. Debopam Samanta: writing – review and editing. Kristen H. Arredondo: writing – review and editing. Ahmad Marashly: writing – review and editing. Zachary M. Grinspan: writing – review and editing. Dallas Armstrong: writing – review and editing. Taylor J. Abel: writing – review and editing. Janelle Wagner: writing – review and editing. Derryl J. Miller: writing – review and editing. Fernando N. Galan: writing – review and editing.
Ethics Statement
This study was approved by the Cook Children's Health Care System Institutional Review Board (IRB Number: 2017‐059) and conducted in accordance with the 1964 Helsinki Declaration and its later amendments.
Conflicts of Interest
The authors declare no conflicts of interest.
All co‐authors have reviewed and approved the content of this manuscript and Annals of Child Neurology (77:187, 2015) requirements for authorship have been met.
References
- 1. Dwivedi R., Ramanujam B., Chandra P. S., et al., “Surgery for Drug‐Resistant Epilepsy in Children,” New England Journal of Medicine 377, no. 17 (2017): 1639–1647, 10.1056/NEJMoa1615335. [DOI] [PubMed] [Google Scholar]
- 2. Kadish N. E., Bast T., Reuner G., et al., “Epilepsy Surgery in the First 3 Years of Life: Predictors of Seizure Freedom and Cognitive Development,” Neurosurgery 84, no. 6 (2019): e368–e377, 10.1093/neuros/nyy376. [DOI] [PubMed] [Google Scholar]
- 3. Pelliccia V., Deleo F., Gozzo F., et al., “Early and Late Epilepsy Surgery in Focal Epilepsies Associated With Long‐Term Epilepsy‐Associated Tumors,” Journal of Neurosurgery 127, no. 5 (2017): 1147–1152, 10.3171/2016.9.JNS161176. [DOI] [PubMed] [Google Scholar]
- 4. Ramantani G., Kadish N. E., Mayer H., et al., “Frontal Lobe Epilepsy Surgery in Childhood and Adolescence: Predictors of Long‐Term Seizure Freedom, Overall Cognitive and Adaptive Functioning,” Neurosurgery 83, no. 1 (2018): 93–103, 10.1093/neuros/nyx340. [DOI] [PubMed] [Google Scholar]
- 5. Simasathien T., Vadera S., Najm I., Gupta A., Bingaman W., and Jehi L., “Improved Outcomes With Earlier Surgery for Intractable Frontal Lobe Epilepsy,” Annals of Neurology 73, no. 5 (2013): 646–654, 10.1002/ana.23862. [DOI] [PubMed] [Google Scholar]
- 6. Edelvik A., Rydenhag B., Olsson I., et al., “Long‐Term Outcomes of Epilepsy Surgery in Sweden: A National Prospective and Longitudinal Study,” Neurology 81, no. 14 (2013): 1244–1251, 10.1212/WNL.0b013e3182a6ca7b. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Lamberink H. J., Otte W. M., Blümcke I., et al., “Seizure Outcome and Use of Antiepileptic Drugs After Epilepsy Surgery According to Histopathological Diagnosis: A Retrospective Multicentre Cohort Study,” Lancet Neurology 19, no. 9 (2020): 748–757, 10.1016/S1474-4422(20)30220-9. [DOI] [PubMed] [Google Scholar]
- 8. Casadei C. H., Carson K. W., Mendiratta A., et al., “All‐Cause Mortality and SUDEP in a Surgical Epilepsy Population,” Epilepsy & Behavior 108 (2020): 107093, 10.1016/j.yebeh.2020.107093. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Samanta D., Ostendorf A. P., Willis E., et al., “Underutilization of Epilepsy Surgery: Part I: A Scoping Review of Barriers,” Epilepsy & Behavior 117 (2021): 107837, 10.1016/j.yebeh.2021.107837. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Bjellvi J., Olsson I., Malmgren K., and Wilbe Ramsay K., “Epilepsy Duration and Seizure Outcome in Epilepsy Surgery: A Systematic Review and Meta‐Analysis,” Neurology 93, no. 2 (2019): e159–e166, 10.1212/WNL.0000000000007753. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. J. Engel, Jr. , Wiebe S., French J., et al., “Practice Parameter: Temporal Lobe and Localized Neocortical Resections for Epilepsy: Report of the Quality Standards Subcommittee of the American Academy of Neurology,” Neurology 60, no. 10 (2003): 538–547, 10.1212/01.WNL.0000055086.35806.2D. [DOI] [PubMed] [Google Scholar]
- 12. Haneef Z., Stern J., Dewar S., and J. Engel, Jr. , “Referral Pattern for Epilepsy Surgery After Evidence‐Based Recommendations: A Retrospective Study,” Neurology 75, no. 8 (2010): 699–704, 10.1212/WNL.0b013e3181eee457. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Jehi L., Jette N., Kwon C. S., 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 63, no. 10 (2022): 2491–2506, 10.1111/epi.17350. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Perry M. S., Shandley S., Perelman M., et al., “Surgical Evaluation in Children < 3 Years of Age With Drug‐Resistant Epilepsy: Patient Characteristics, Diagnostic Utilization, and Potential for Treatment Delays,” Epilepsia 63, no. 1 (2022): 96–107, 10.1111/epi.17124. [DOI] [PubMed] [Google Scholar]
- 15. Nathan C. L. and Gutierrez C., “FACETS of Health Disparities in Epilepsy Surgery and Gaps That Need to Be Addressed,” Neurology: Clinical Practice 8, no. 4 (2018): 340–345, 10.1212/CPJ.0000000000000490. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Bautista R. E. D. and Jain D., “Detecting Health Disparities Among Caucasians and African‐Americans With Epilepsy,” Epilepsy & Behavior 20, no. 1 (2011): 52–56, 10.1016/j.yebeh.2010.10.016. [DOI] [PubMed] [Google Scholar]
- 17. Sánchez Fernández I., Stephen C., and Loddenkemper T., “Disparities in Epilepsy Surgery in the United States of America,” Journal of Neurology 264, no. 8 (2017): 1735–1745, 10.1007/s00415-017-8560-6. [DOI] [PubMed] [Google Scholar]
- 18. Campbell J. M., Yost S., Gautam D., et al., “Delays in the Diagnosis and Surgical Treatment of Drug‐Resistant Epilepsy: A Cohort Study,” Epilepsia 65 (2024): 1314–1321, 10.1111/epi.17944. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Bien C. G., Raabe A. L., Schramm J., Becker A., Urbach H., and Elger C. E., “Trends in Presurgical Evaluation and Surgical Treatment of Epilepsy at One Centre From 1988–2009,” Journal of Neurology, Neurosurgery & Psychiatry 84, no. 1 (2013): 54–61, 10.1136/jnnp-2011-301763. [DOI] [PubMed] [Google Scholar]
- 20. Sanders M. W., Van der Wolf I., Jansen F. E., et al., “Outcome of Epilepsy Surgery in MRI‐Negative Patients Without Histopathologic Abnormalities in the Resected Tissue,” Neurology 102, no. 4 (2024): e456–e468, 10.1212/WNL.0000000000208007. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Bagić A. I. and Burgess R. C., “Utilization of MEG Among the US Epilepsy Centers: A Survey‐Based Appraisal,” Journal of Clinical Neurophysiology 37, no. 6 (2020): 599–605, 10.1097/WNP.0000000000000716. [DOI] [PubMed] [Google Scholar]
- 22. Morrison C. E., MacAllister W. S., and Barr W. B., “Neuropsychology Within a Tertiary Care Epilepsy Center,” Archives of Clinical Neuropsychology 33, no. 3 (2018): 354–364, 10.1093/arclin/acx134. [DOI] [PubMed] [Google Scholar]