Machine learning for (non-)epileptic tissue detection from the intraoperative electrocorticogram.

OBJECTIVE: Clinical visual intraoperative electrocorticography (ioECoG) reading intends to localize epileptic tissue and improve epilepsy surgery outcome. We aimed to understand whether machine learning (ML) could complement ioECoG reading, how subgroups affected performance, and which ioECoG features were most important. METHODS: We included 91 ioECoG-guided epilepsy surgery patients with Engel 1A outcome. We allocated 71 training and 20 test set patients. We trained an extra trees classifier (ETC) with 14 spectral features to classify ioECoG channels as covering resected or non-resected tissue. We compared the ETC's performance with clinical ioECoG reading and assessed whether patient subgroups affected performance. Explainable artificial intelligence (xAI) unveiled the most important ioECoG features learnt by the ETC. RESULTS: The ETC outperformed clinical reading in five test set patients, was inferior in six, and both were inconclusive in nine. The ETC performed best in the tumor subgroup (area under ROC curve: 0.84 [95%CI 0.79-0.89]). xAI revealed predictors of resected (relative theta, alpha, and fast ripple power) and non-resected tissue (relative beta and gamma power). CONCLUSIONS: Combinations of subtle spectral ioECoG changes, imperceptible by the human eye, can aid healthy and pathological tissue discrimination. SIGNIFICANCE: ML with spectral ioECoG features can support, rather than replace, clinical ioECoG reading, particularly in tumors.

N-of-1 trials in epilepsy: A systematic review and lessons paving the way forward.

OBJECTIVE: Defined as prospective single-patient crossover studies with repeated paired cycles of active and control intervention, N-of-1 trials have gained attention as an option to obtain high-quality evidence of efficacy, particularly for patients with rare epilepsies in whom conduction of well-powered randomized controlled trials can be challenging. The objective of this systematic review is to provide an appraisal of the literature on N-of-1 trials in individuals with epilepsy. METHODS: We searched PubMed and Embase on January 12, 2024, for studies meeting the following criteria: prospectively planned, within-patient, multiple-crossover design in individuals with epilepsy and outcomes related to comorbidities. Information on design, outcome measurements, intervention, and analyses was retrieved. Risk of bias assessment was performed using the Risk of Bias in N-of-1 Trials (RoBiNT) scale. We highlighted methodological aspects of the N-of-1 trials identified and discuss future recommendations. RESULTS: Five studies met our inclusion criteria. An additional multiple-crossover trial that evaluated treatment effects exclusively at group level was also included because of its relevance to N-of-1 study methodology. The studies enrolled individuals with focal seizures, absences or cognitive impairement and electrographic discharges. Treatments included established or investigational antiseizure medications, off-label medications, neurostimulation or lifestyle intervention. Three of the five N-of-1 trials reported on individual cases. The studies' strengths were the use of individualized treatment dosages and symptom-specific patient-reported outcomes. Limitations were related to minimal reporting of baseline characteristics and seizure burden. SIGNIFICANCE: The trials identified by our search exemplify how the N-of-1 design can be applied to assess interventions in individuals with epilepsy-related disorders. Future N-of-1 trials of antiseizure interventions should take into account baseline seizure frequency, should apply statistical models suited to capture seizure frequency changes reliably and make predefined interim assessments. Non-seizure outcome measures evaluable over short periods should be considered. Tailored N-of-1 methodology could pave the way to evidence-based, treatment selection for patients with rare epilepsies.

Three Prospective Case Studies Examining Mifepristone's Efficacy in Patients with Treatment-Resistant PTSD.

Despite the availability of various treatment approaches for patients with posttraumatic stress disorder (PTSD), some patients do not respond to these therapies, and novel treatment approaches are needed. This study investigated the efficacy of mifepristone, a glucocorticoid receptor antagonist, in treatment-resistant PTSD patients. Three patients with PTSD who were resistant to standard psychological and pharmacological treatments were prescribed mifepristone (600-1,200 mg/day) for 1 week. A baseline-controlled single-case design was used, involving a 2-week baseline phase (no intervention), a 1-week intervention phase (mifepristone), and a 2-week postintervention phase. The primary outcome measure, self-reported PTSD symptom severity (PCL-5), was assessed daily, with participants providing their own control condition. Two of the three patients experienced a significant reduction in PTSD symptom severity after the intervention phase and no longer met the diagnostic criteria for PTSD. These positive results were maintained during long-term follow-up. These findings support the potential effectiveness of mifepristone in the treatment of patients with treatment-resistant PTSD. However, our findings must be interpreted with caution, and further studies with larger sample sizes and more rigorous designs are necessary to confirm the promising results.

Potential merits and flaws of large language models in epilepsy care: A critical review.

The current pace of development and applications of large language models (LLMs) is unprecedented and will impact future medical care significantly. In this critical review, we provide the background to better understand these novel artificial intelligence (AI) models and how LLMs can be of future use in the daily care of people with epilepsy. Considering the importance of clinical history taking in diagnosing and monitoring epilepsy-combined with the established use of electronic health records-a great potential exists to integrate LLMs in epilepsy care. We present the current available LLM studies in epilepsy. Furthermore, we highlight and compare the most commonly used LLMs and elaborate on how these models can be applied in epilepsy. We further discuss important drawbacks and risks of LLMs, and we provide recommendations for overcoming these limitations.

Footprint of publication selection bias on meta-analyses in medicine, environmental sciences, psychology, and economics.

Publication selection bias undermines the systematic accumulation of evidence. To assess the extent of this problem, we survey over 68,000 meta-analyses containing over 700,000 effect size estimates from medicine (67,386/597,699), environmental sciences (199/12,707), psychology (605/23,563), and economics (327/91,421). Our results indicate that meta-analyses in economics are the most severely contaminated by publication selection bias, closely followed by meta-analyses in environmental sciences and psychology, whereas meta-analyses in medicine are contaminated the least. After adjusting for publication selection bias, the median probability of the presence of an effect decreased from 99.9% to 29.7% in economics, from 98.9% to 55.7% in psychology, from 99.8% to 70.7% in environmental sciences, and from 38.0% to 29.7% in medicine. The median absolute effect sizes (in terms of standardized mean differences) decreased from d = 0.20 to d = 0.07 in economics, from d = 0.37 to d = 0.26 in psychology, from d = 0.62 to d = 0.43 in environmental sciences, and from d = 0.24 to d = 0.13 in medicine.

Determinants of Functional Outcome after Pediatric Hemispherotomy.

OBJECTIVE: We aimed to evaluate determinants of functional outcome after pediatric hemispherotomy in a large and recent multicenter cohort. METHODS: We retrospectively investigated the functional outcomes of 455 children who underwent hemispherotomy at 5 epilepsy centers in 2000-2016. We identified determinants of unaided walking, voluntary grasping with the hemiplegic hand, and speaking through Bayesian multivariable regression modeling using missing data imputation. RESULTS: Seventy-five percent of children were seizure-free, and 44% stopped antiseizure medication at a 5.1-year mean follow-up (range = 1-17.1). Seventy-seven percent of children could walk unaided, 8% could grasp voluntarily, and 68% could speak at the last follow-up. Children were unlikely to walk when they had contralateral magnetic resonance imaging (MRI) abnormalities (40/73, p = 0.04), recurrent seizures following hemispherotomy (62/109, p = 0.04), and moderately (50/61, p = 0.03) or severely impaired (127/199, p = 0.001) postsurgical intellectual functioning, but were likely to walk when they were older at outcome determination (p = 0.01). Children were unlikely to grasp voluntarily with the hand contralateral to surgery when they had Rasmussen encephalitis (0/61, p = 0.001) or Sturge-Weber syndrome (0/32, p = 0.007). Children were unlikely to speak when they had contralateral MRI abnormalities (30/69, p = 0.002) and longer epilepsy duration (p = 0.01), but likely to speak when they had Sturge-Weber syndrome (29/35, p = 0.01), were older at surgery (p = 0.04), and were older at outcome determination (p < 0.001). INTERPRETATION: Etiology and bilaterality of structural brain abnormalities were key determinants of functional outcome after hemispherotomy. Longer epilepsy duration affected language outcomes. Not surprisingly, walking and talking ability increased with older age at outcome evaluation. ANN NEUROL 2024;95:377-387.

Variation in seizure risk increases from antiseizure medication withdrawal among patients with well-controlled epilepsy: A pooled analysis.

OBJECTIVE: Guidelines suggest considering antiseizure medication (ASM) discontinuation in seizure-free patients with epilepsy. Past work has poorly explored how discontinuation effects vary between patients. We evaluated (1) what factors modify the influence of discontinuation on seizure risk; and (2) the range of seizure risk increase due to discontinuation across low- versus high-risk patients. METHODS: We pooled three datasets including seizure-free patients who did and did not discontinue ASMs. We conducted time-to-first-seizure analyses. First, we evaluated what individual patient factors modified the relative effect of ASM discontinuation on seizure risk via interaction terms. Then, we assessed the distribution of 2-year risk increase as predicted by our adjusted logistic regressions. RESULTS: We included 1626 patients, of whom 678 (42%) planned to discontinue all ASMs. The mean predicted 2-year seizure risk was 43% [95% confidence interval (CI) 39%-46%] for discontinuation versus 21% (95% CI 19%-24%) for continuation. The mean 2-year absolute seizure risk increase was 21% (95% CI 18%-26%). No individual interaction term was significant after correcting for multiple comparisons. The median [interquartile range (IQR)] risk increase across patients was 19% (IQR 14%-24%; range 7%-37%). Results were unchanged when restricting analyses to only the two RCTs. SIGNIFICANCE: No single patient factor significantly modified the influence of discontinuation on seizure risk, although we captured how absolute risk increases change for patients that are at low versus high risk. Patients should likely continue ASMs if even a 7% 2-year increase in the chance of any more seizures would be too much and should likely discontinue ASMs if even a 37% risk increase would be too little. In between these extremes, individualized risk calculation and a careful understanding of patient preferences are critical. Future work will further develop a two-armed individualized seizure risk calculator and contextualize seizure risk thresholds below which to consider discontinuation. PLAIN LANGUAGE SUMMARY: Understanding how much antiseizure medications (ASMs) decrease seizure risk is an important part of determining which patients with epilepsy should be treated, especially for patients who have not had a seizure in a while. We found that there was a wide range in the amount that ASM discontinuation increases seizure risk-between 7% and 37%. We found that no single patient factor modified that amount. Understanding what a patient's seizure risk might be if they discontinued versus continued ASM treatment is critical to making informed decisions about whether the benefit of treatment outweighs the downsides.

Clinical characteristics and diagnoses of 1213 children referred to a first seizure clinic.

OBJECTIVE: New-onset seizure-like events (SLEs) are common in children, but differentiating between epilepsy and its mimics is challenging. This study provides an overview of the clinical characteristics, diagnoses, and corresponding etiologies of children evaluated at a first seizure clinic (FSC), which will be helpful for all physicians involved in the care of children with SLEs. METHODS: We included 1213 children who were referred to the FSC of a Dutch tertiary children's hospital over a 13-year period and described their clinical characteristics, first routine EEG recording results, and the distribution and specification of their eventual epilepsy and non-epilepsy diagnoses. The time interval to correct diagnosis and the diagnostic accuracy of the FSC were evaluated. RESULTS: "Epilepsy" was eventually diagnosed in 407 children (33.5%), "no epilepsy" in 737 (60.8%), and the diagnosis remained "unclear" in 69 (5.7%). Epileptiform abnormalities were seen in 60.9% of the EEG recordings in the "epilepsy" group, and in 5.7% and 11.6% of the "no epilepsy" and "unclear" group, respectively. Of all children with final "epilepsy" and "no epilepsy" diagnoses, 68.6% already received their diagnosis at FSC consultation, and 2.9% of the children were initially misdiagnosed. The mean time to final diagnosis was 2.0 months, and 91.3% of all children received their final diagnosis within 12 months after the FSC consultation. SIGNIFICANCE: We describe the largest pediatric FSC cohort to date, which can serve as a clinical frame of reference. The experience and expertise built at FSCs will improve and accelerate diagnosis in children with SLEs. PLAIN LANGUAGE SUMMARY: Many children experience events that resemble but not necessarily are seizures. Distinguishing between seizures and seizure mimics is important but challenging. Specialized first-seizure clinics can help with this. Here, we report data from 1213 children who were referred to the first seizure clinic of a Dutch children's hospital. One-third of them were diagnosed with epilepsy. In 68.8% of all children-with and without epilepsy-the diagnosis was made during the first consultation. Less than 3% were misdiagnosed. This study may help physicians in what to expect regarding the diagnoses in children who present with events that resemble seizures.

The importance of discriminative power rather than significance when evaluating potential clinical biomarkers in epilepsy research.

OBJECTIVE: The quest for epilepsy biomarkers is on the rise. Variables with statistically significant group-level differences are often misinterpreted as biomarkers with sufficient discriminative power. This study aimed to demonstrate the relationship between significant group-level differences and a variable's power to discriminate between individuals. METHODS: We simulated normal-distributed datasets from hypothetical populations with varying sample sizes (25-800), effect sizes (Cohen's d: .25-2.50), and variability (standard deviation: 10-35) to assess the impact of these parameters on significance and discriminative power. The simulation data were illustrated by assessing the discriminative power of a potential real-case biomarker-the EEG beta band power-to diagnose generalized epilepsy, using data from 66 children with generalized epilepsy and 385 controls. Additionally, we evaluated recently reported epilepsy biomarkers by comparing their effect sizes to our simulation-derived effect size criterion. RESULTS: Group size affects significance but not discriminative power. Discriminative power is much more related to variability and effect size. Our real data example supported these simulation results by demonstrating that group-level significance does not translate, one to one, into discriminative power. Although we found a significant difference in the beta band power between children with and without epilepsy, the discriminative power was poor due to a small effect size. A Cohen's d of at least 1.25 is required to reach good discriminative power in univariable prediction modeling. Slightly over 60% of the biomarkers in our literature search met this criterion. SIGNIFICANCE: Rather than statistical significance of group-level differences, effect size should be used as an indicator of a variable's biomarker potential. The minimal required effects size for individual biomarkers-a Cohen's d of 1.25-is large. This calls for multivariable approaches, in which combining multiple variables with smaller effect sizes could increase the overall effect size and discriminative power.

Methodology reporting improved over time in 176,469 randomized controlled trials.

OBJECTIVES: To describe randomized controlled trial (RCT) methodology reporting over time. STUDY DESIGN AND SETTING: We used a deep learning-based sentence classification model based on the Consolidated Standards of Reporting Trials (CONSORT) statement, considered minimum requirements for reporting RCTs. We included 176,469 RCT reports published between 1966 and 2018. We analyzed the reporting trends over 5-year time periods, grouping trials from 1966 to 1990 in a single stratum. We also explored the effect of journal impact factor (JIF) and medical discipline. RESULTS: Population, Intervention, Comparator, Outcome (PICO) items were commonly reported during each period, and reporting increased over time (e.g., interventions: 79.1% during 1966-1990 to 87.5% during 2010-2018). Reporting of some methods information has increased, although there is room for improvement (e.g., sequence generation: 10.8-41.8%). Some items are reported infrequently (e.g., allocation concealment: 5.1-19.3%). The number of items reported and JIF are weakly correlated (Pearson's r (162,702) = 0.16, P < 0.001). The differences in the proportion of items reported between disciplines are small (<10%). CONCLUSION: Our analysis provides large-scale quantitative support for the hypothesis that RCT methodology reporting has improved over time. Extending these models to all CONSORT items could facilitate compliance checking during manuscript authoring and peer review, and support metaresearch.

Continuous Theta-Burst Stimulation of the Contralesional Primary Motor Cortex for Promotion of Upper Limb Recovery After Stroke: A Randomized Controlled Trial.

BACKGROUND: Despite improvements in acute stroke therapies and rehabilitation strategies, many stroke patients are left with long-term upper limb motor impairment. We assessed whether an inhibitory repetitive transcranial magnetic stimulation treatment paradigm started within 3 weeks after stroke onset promotes upper limb motor recovery. METHODS: We performed a single-center randomized, sham-controlled clinical trial. Patients with ischemic stroke or intracerebral hemorrhage and unilateral upper limb motor impairment were randomized to 10 daily sessions of active or sham continuous theta-burst stimulation (cTBS) of the contralesional primary motor cortex combined with standard upper limb therapy, started within 3 weeks after stroke onset. The primary outcome was the change in the Action Research Arm Test score from baseline (pretreatment) at 3 months after stroke. Secondary outcomes included the score on the modified Rankin Scale at 3 months and the length of stay at the rehabilitation center. Statistical analyses were performed using mixed models for repeated measures. RESULTS: We enrolled 60 patients between April 2017 and February 2021, of whom 29 were randomized to active cTBS and 31 to sham cTBS. One patient randomized to active cTBS withdrew consent before the intervention and was excluded from the analyses. The mean difference in the change in Action Research Arm Test score from baseline at 3 months poststroke was 9.6 points ([95% CI, 1.2-17.9]; P=0.0244) in favor of active cTBS. Active cTBS was associated with better scores on the modified Rankin Scale at 3 months (OR, 0.2 [95% CI, 0.1-0.8]; P=0.0225) and with an 18 days shorter length of stay at the rehabilitation center than sham cTBS ([95% CI, 0.0-36.4]; P=0.0494). There were no serious adverse events. CONCLUSIONS: Ten daily sessions of cTBS of the contralesional primary motor cortex combined with upper limb training, started within 3 weeks after stroke onset, promote recovery of the upper limb, reduce disability and dependence and leads to earlier discharge from the rehabilitation center. REGISTRATION: URL: https://trialsearch.who.int/; Unique identifier: NTR6133.

Toward responsible clinical n-of-1 strategies for rare diseases.

N-of-1 strategies can provide high-quality evidence of treatment efficacy at the individual level and optimize evidence-based selection of off-label treatments for patients with rare diseases. Given their design characteristics, n-of-1 strategies are considered to lay at the intersection between medical research and clinical care. Therefore, whether n-of-1 strategies should be governed by research or care regulations remains a debated issue. Here, we delineate differences between medical research and optimized clinical care, and distinguish the regulations which apply to either. We also set standards for responsible optimized clinical n-of-1 strategies with (off-label) treatments for rare diseases. Implementing clinical n-of-1 strategies as defined here could aid in optimized treatment selection for such diseases.

Antiseizure Medication Withdrawal Practice Patterns: A Survey Among Members of the American Academy of Neurology and EpiCARE.

Background and Objectives: To describe neurologist practice patterns, challenges, and decision support needs pertaining to withdrawal of antiseizure medications (ASMs) in patients with well-controlled epilepsy. Methods: We sent an electronic survey to (1) US and (2) European physician members of the American Academy of Neurology and (3) members of EpiCARE, a European Reference Network for rare and complex epilepsies. Analyses included frequencies and percentages, and we showed distributions through histograms and violin plots. Results: We sent the survey to 4,923 individuals; 463 consented, 411 passed eligibility questions, and 287 responded to at least 1 of these questions. Most respondents indicated that they might ever consider ASM withdrawal, with respondents treating mostly children being more likely ever to consider withdrawal (e.g., medical monotherapy: children 96% vs adults 81%; p < 0.05). The most important factors when making decisions included seizure probability (83%), consequences of seizures (73%), and driving (74%). The top challenges when making decisions included unclear seizure probability (81%), inadequate guidelines (50%), and difficulty communicating probabilities (45%). Respondents would consider withdrawal after a median of 2-year seizure freedom, but also responded that they would begin withdrawal on average only when the postwithdrawal seizure relapse risk in the coming 2 years was less than 15%-30%. Wide variation existed in the use of words or numbers in respondents' counsel methods, for example, percentages vs frequencies or probability of seizure freedom vs seizure. The most highly rated point-of-care methods to inform providers of calculated risk were Kaplan-Meier curves and showing percentages only, rather than pictographs or text recommendations alone. Discussion: Most surveyed neurologists would consider withdrawing ASMs in seizure-free individuals. Seizure probability was the largest factor driving decisions, yet estimating seizure probabilities was the greatest challenge. Respondents on average indicated that they may withdraw ASM after a minimum seizure-free duration of 2 years, yet also on average were willing to withdraw when seizure risk decreased below 15%-30%, which is lower than most patients' postwithdrawal risk at 2-year seizure freedom and lower than the equivalent even of a first seizure of life. These findings will inform future efforts at developing decision support tools aimed at optimizing ASM withdrawal decisions.

Unrestricted weighted least squares represent medical research better than random effects in 67,308 Cochrane meta-analyses.

OBJECTIVES: To evaluate how well meta-analysis mean estimators represent reported medical research and establish which meta-analysis method is better using widely accepted model selection measures: Akaike information criterion (AIC) and Bayesian information criterion (BIC). STUDY DESIGN AND SETTING: We compiled 67,308 meta-analyses from the Cochrane Database of Systematic Reviews (CDSR) published between 1997 and 2020, collectively encompassing nearly 600,000 medical findings. We compared unrestricted weighted least squares (UWLS) vs. random effects (RE); fixed effect was also secondarily considered. RESULTS: The probability that a randomly selected systematic review from the CDSR would favor UWLS over RE is 79.4% (95% confidence interval [CI95%]: 79.1; 79.7). The odds ratio that a Cochrane systematic review would substantially favor UWLS over RE is 9.33 (CI95%: 8.94; 9.73) using the conventional criterion that a difference in AIC (or BIC) of two or larger represents a 'substantial' improvement. UWLS's advantage over RE is most prominent in the presence of low heterogeneity. However, UWLS also has a notable advantage in high heterogeneity research, across different sizes of meta-analyses and types of outcomes. CONCLUSION: UWLS frequently dominates RE in medical research, often substantially. Thus, the UWLS should be reported routinely in the meta-analysis of clinical trials.

Not surgical technique, but etiology, contralateral MRI, prior surgery, and side of surgery determine seizure outcome after pediatric hemispherotomy.

OBJECTIVE: We aimed to assess determinants of seizure outcome following pediatric hemispherotomy in a contemporary cohort. METHODS: We retrospectively analyzed the seizure outcomes of 457 children who underwent hemispheric surgery in five European epilepsy centers between 2000 and 2016. We identified variables related to seizure outcome through multivariable regression modeling with missing data imputation and optimal group matching, and we further investigated the role of surgical technique by Bayes factor (BF) analysis. RESULTS: One hundred seventy seven children (39%) underwent vertical and 280 children (61%) underwent lateral hemispherotomy. Three hundred forty-four children (75%) achieved seizure freedom at a mean follow-up of 5.1 years (range 1 to 17.1). We identified acquired etiology other than stroke (odds ratio [OR] 4.4, 95% confidence interval (CI) 1.1-18.0), hemimegalencephaly (OR 2.8, 95% CI 1.1-7.3), contralateral magnetic resonance imaging (MRI) findings (OR 5.5, 95% CI 2.7-11.1), prior resective surgery (OR 5.0, 95% CI 1.8-14.0), and left hemispherotomy (OR 2.3, 95% CI 1.3-3.9) as significant determinants of seizure recurrence. We found no evidence of an impact of the hemispherotomy technique on seizure outcome (the BF for a model including the hemispherotomy technique over the null model was 1.1), with comparable overall major complication rates for different approaches. SIGNIFICANCE: Knowledge about the independent determinants of seizure outcome following pediatric hemispherotomy will improve the counseling of patients and families. In contrast to previous reports, we found no statistically relevant difference in seizure-freedom rates between the vertical and horizontal hemispherotomy techniques when accounting for different clinical features between groups.

Antiseizure medication withdrawal risk estimation and recommendations: A survey of American Academy of Neurology and EpiCARE members.

OBJECTIVE: Choosing candidates for antiseizure medication (ASM) withdrawal in well-controlled epilepsy is challenging. We evaluated (a) the correlation between neurologists' seizure risk estimation ("clinician predictions") vs calculated predictions, (b) how viewing calculated predictions influenced recommendations, and (c) barriers to using risk calculation. METHODS: We asked US and European neurologists to predict 2-year seizure risk after ASM withdrawal for hypothetical vignettes. We compared ASM withdrawal recommendations before vs after viewing calculated predictions, using generalized linear models. RESULTS: Three-hundred and forty-six neurologists responded. There was moderate correlation between clinician and calculated predictions (Spearman coefficient 0.42). Clinician predictions varied widely, for example, predictions ranged 5%-100% for a 2-year seizure-free adult without epileptiform abnormalities. Mean clinician predictions exceeded calculated predictions for vignettes with epileptiform abnormalities (eg, childhood absence epilepsy: clinician 65%, 95% confidence interval [CI] 57%-74%; calculated 46%) and surgical vignettes (eg, focal cortical dysplasia 6-month seizure-free mean clinician 56%, 95% CI 52%-60%; calculated 28%). Clinicians overestimated the influence of epileptiform EEG findings on withdrawal risk (26%, 95% CI 24%-28%) compared with calculators (14%, 95% 13%-14%). Viewing calculated predictions slightly reduced willingness to withdraw (-0.8/10 change, 95% CI -1.0 to -0.7), particularly for vignettes without epileptiform abnormalities. The greatest barrier to calculator use was doubting its accuracy (44%). SIGNIFICANCE: Clinicians overestimated the influence of abnormal EEGs particularly for low-risk patients and overestimated risk and the influence of seizure-free duration for surgical patients, compared with calculators. These data may question widespread ordering of EEGs or time-based seizure-free thresholds for surgical patients. Viewing calculated predictions reduced willingness to withdraw particularly without epileptiform abnormalities.

Frequency of and factors associated with antiseizure medication discontinuation discussions and decisions in patients with epilepsy: A multicenter retrospective chart review.

OBJECTIVE: Guidelines suggest considering antiseizure medication (ASM) discontinuation in patients with epilepsy who become seizure-free. Little is known about how discontinuation decisions are being made in practice. We measured the frequency of, and factors associated with, discussions and decisions surrounding ASM discontinuation. METHODS: We performed a multicenter retrospective cohort study at the University of Michigan (UM) and two Dutch centers: Wilhelmina Children's Hospital (WCH) and Stichting Epilepsie Instellingen Nederland (SEIN). We screened all children and adults with outpatient epilepsy visits in January 2015 and included those with at least one visit during the subsequent 2 years where they were seizure-free for at least one year. We recorded whether charts documented (1) a discussion with the patient about possible ASM discontinuation and (2) any planned attempt to discontinue at least one ASM. We conducted multilevel logistic regressions to determine factors associated with each outcome. RESULTS: We included 1058 visits from 463 patients. Of all patients who were seizure-free at least one year, 248/463 (53%) had documentation of any discussion and 98/463 (21%) planned to discontinue at least one ASM. Corresponding frequencies for patients who were seizure-free at least 2 years were 184/285 (65%) and 74/285 (26%). The probability of discussing or discontinuing increased with longer duration of seizure freedom. Still, even for patients who were 10 years seizure-free, our models predicated that in only 49% of visits was a discontinuation discussion documented, and in only 16% of visits was it decided to discontinue all ASMs. Provider-to-provider variation explained 18% of variation in whether patients discontinued any ASM. SIGNIFICANCE: Only approximately half of patients with prolonged seizure freedom had a documented discussion about ASM discontinuation. Discontinuation was fairly rare even among low-risk patients. Future work should further explore barriers to and facilitators of counseling and discontinuation attempts.

Transforming epilepsy research: A systematic review on natural language processing applications.

Despite improved ancillary investigations in epilepsy care, patients' narratives remain indispensable for diagnosing and treatment monitoring. This wealth of information is typically stored in electronic health records and accumulated in medical journals in an unstructured manner, thereby restricting complete utilization in clinical decision-making. To this end, clinical researchers increasing apply natural language processing (NLP)-a branch of artificial intelligence-as it removes ambiguity, derives context, and imbues standardized meaning from free-narrative clinical texts. This systematic review presents an overview of the current NLP applications in epilepsy and discusses the opportunities and drawbacks of NLP alongside its future implications. We searched the PubMed and Embase databases with a "natural language processing" and "epilepsy" query (March 4, 2022) and included original research articles describing the application of NLP techniques for textual analysis in epilepsy. Twenty-six studies were included. Fifty-eight percent of these studies used NLP to classify clinical records into predefined categories, improving patient identification and treatment decisions. Other applications of NLP had structured clinical information retrieval from electronic health records, scientific papers, and online posts of patients. Challenges and opportunities of NLP applications for enhancing epilepsy care and research are discussed. The field could further benefit from NLP by replicating successes in other health care domains, such as NLP-aided quality evaluation for clinical decision-making, outcome prediction, and clinical record summarization.

Predictive models for starting antiseizure medication withdrawal following epilepsy surgery in adults.

More than half of adults with epilepsy undergoing resective epilepsy surgery achieve long-term seizure freedom and might consider withdrawing antiseizure medications. We aimed to identify predictors of seizure recurrence after starting postoperative antiseizure medication withdrawal and develop and validate predictive models. We performed an international multicentre observational cohort study in nine tertiary epilepsy referral centres. We included 850 adults who started antiseizure medication withdrawal following resective epilepsy surgery and were free of seizures other than focal non-motor aware seizures before starting antiseizure medication withdrawal. We developed a model predicting recurrent seizures, other than focal non-motor aware seizures, using Cox proportional hazards regression in a derivation cohort (n = 231). Independent predictors of seizure recurrence, other than focal non-motor aware seizures, following the start of antiseizure medication withdrawal were focal non-motor aware seizures after surgery and before withdrawal [adjusted hazard ratio (aHR) 5.5, 95% confidence interval (CI) 2.7-11.1], history of focal to bilateral tonic-clonic seizures before surgery (aHR 1.6, 95% CI 0.9-2.8), time from surgery to the start of antiseizure medication withdrawal (aHR 0.9, 95% CI 0.8-0.9) and number of antiseizure medications at time of surgery (aHR 1.2, 95% CI 0.9-1.6). Model discrimination showed a concordance statistic of 0.67 (95% CI 0.63-0.71) in the external validation cohorts (n = 500). A secondary model predicting recurrence of any seizures (including focal non-motor aware seizures) was developed and validated in a subgroup that did not have focal non-motor aware seizures before withdrawal (n = 639), showing a concordance statistic of 0.68 (95% CI 0.64-0.72). Calibration plots indicated high agreement of predicted and observed outcomes for both models. We show that simple algorithms, available as graphical nomograms and online tools (predictepilepsy.github.io), can provide probabilities of seizure outcomes after starting postoperative antiseizure medication withdrawal. These multicentre-validated models may assist clinicians when discussing antiseizure medication withdrawal after surgery with their patients.

Indicators of questionable research practices were identified in 163,129 randomized controlled trials.

OBJECTIVES: To explore indicators of the following questionable research practices (QRPs) in randomized controlled trials (RCTs): (1) risk of bias in four domains (random sequence generation, allocation concealment, blinding of participants and personnel, and blinding of outcome assessment); (2) modifications in primary outcomes that were registered in trial registration records (proxy for selective reporting bias); (3) ratio of the achieved to planned sample sizes; and (4) statistical discrepancy. STUDY DESIGN AND SETTING: Full texts of all human RCTs published in PubMed in 1996-2017 were automatically identified and information was collected automatically. Potential indicators of QRPs included author-specific, publication-specific, and journal-specific characteristics. Beta, logistic, and linear regression models were used to identify associations between these potential indicators and QRPs. RESULTS: We included 163,129 RCT publications. The median probability of bias assessed using Robot Reviewer software ranged between 43% and 63% for the four risk of bias domains. A more recent publication year, trial registration, mentioning of CONsolidated Standards Of Reporting Trials-checklist, and a higher journal impact factor were consistently associated with a lower risk of QRPs. CONCLUSION: This comprehensive analysis provides an insight into indicators of QRPs. Researchers should be aware that certain characteristics of the author team and publication are associated with a higher risk of QRPs.