BackgroundEpilepsy is a chronic neurological disorder marked by recurrent and apparently unpredictable seizures and associated with premature death, injury, and diminished quality of life. The unpredictability of seizures is a major concern for people with epilepsy. Thus, developing tools for seizure prediction is a research priority. The Artificial Intelligence to Optimise Seizure Prediction to Empower People With Epilepsy (ATMOSPHERE) project focuses on the development and evaluation of seizure forecasting technology involving mobile technology and machine learning to provide personalized seizure forecasting (risk of seizure in the near future). The project is informed by complex intervention frameworks, which recommend phases of development, feasibility study, clinical evaluation, and implementation. ObjectiveObjective 1 aims to conduct a feasibility study to test and refine the trial methods for a future clinical trial. Objective 2 aims to test and refine the data collection technology, considering usability and technical performance. Objective 3 aims to collect longitudinal data on seizures and their precipitants to refine seizure forecasting. MethodsThis study is a single-arm, mixed methods feasibility study, testing a prototype of the data collection technology, with phase 2 testing a minimum viable product. In total, 60 participants will be recruited via specialist National Health Service epilepsy clinics. Inclusion criteria are adults with epilepsy, experiencing seizures twice per month, able to consent, and engage with technology. Clinicians will screen and gain consent to contact, with researchers obtaining full consent. Participants will be invited to complete the following study procedures: (1) onboarding, (2) use the data collection technology (phase 1 or 2) in their lived context for up to 6 months, (3) complete patient-reported outcome measures and capture clinical-reported outcome measures at baseline and 3 months, and (4) complete a qualitative interview exploring their views of the data collection technology. A study flow diagram will report recruitment rates (outcome 1), diversity of the recruited sample (outcome 2), barriers and facilitators to recruitment (outcome 3), retention rates (outcome 4), and barriers and facilitators to retention (outcome 5). To assess the data collection technology, quantitative technology use data and qualitative interview data will be analyzed to assess usability (outcome 6) and technical performance (outcome 7) of the data collection technology. These outcomes will inform iterative minimum viable product development and testing cycles with stakeholders. ResultsRecruitment is planned to begin in quarter 1 of 2026, with data collection expected to be completed by quarter 2 of 2027. Data analysis will take place during quarter 3, and the results will be published in quarter 4. ConclusionsThis project aims to improve clinical outcomes for people with epilepsy through seizure forecasting technology. To evaluate clinical outcomes, robust trial methodology is critical. This feasibility study will optimize methods for a future full-scale clinical trial as well as refine the seizure forecasting intervention. International Registered Report Identifier (IRRID)PRR1-10.2196/85993

Epilepsy is a chronic neurological condition characterized by recurrent, unprovoked seizures. Epilepsy management often involves a ‘seizure diary’—a patient–generated record of seizure events and accompanying information to inform clinical decisions. However, seizure diaries face accuracy concerns from clinicians and deteriorating engagement over time. Digital Patient–Generated Health Data (PGHD) holds potential to reduce accuracy concerns and increase engagement. As such, PGHD is an area of interest within epilepsy, as smartwatches and mobile apps are increasingly adopted as self–management tools. This study examines the views of people with epilepsy, carers and epilepsy clinicians on a PGHD clinical dashboard within NHS epilepsy clinical pathways. This study explores the opportunities, challenges and preliminary user–requirements for a PGHD clinical dashboard (phase 1). Subsequently, it prioritises those user–requirements from a patient and clinician perspective through a survey (phase 2). A two-phase mixed method approach was used. In Phase 1, six focus groups were held over seven months, from April to October 2025, with 11 people with epilepsy, 2 carers and 9 epilepsy clinicians to explore the opportunities and challenges of presenting PGHD in a clinical dashboard for epilepsy care. The focus groups underwent inductive and deductive thematic analysis; references to specific or potential uses of PGHD were deductively extracted. In Phase 2, a survey was developed using the Kano methodology to classify and prioritise the proposed features according to end–user satisfaction. 11 clinicians, 14 people with epilepsy and 2 carers completed the survey. Participants described the burden of seizure tracking and discussed methods to improve its value, such as predictive and actionable analytics and efficient data communication. People with epilepsy saw potential for PGHD to support autonomous experimentation within self-management whereas clinicians saw it as an opportunity to improve oversight. The focus groups identified 22 features for a clinical dashboard for epilepsy that incorporates holistic and epilepsy–specific PGHD, 19 of which were novel within epilepsy care and beyond the communication of fundamental clinical information. The quantitative survey demonstrated broad alignment between user groups but a surprising disinterest in explainable AI amongst clinicians. The most prioritised features were (i) patient–specific pre– to post–ictal seizure type descriptions and (ii) open identification of AI–generated information. A PGHD clinical dashboard has potential to provide rich, out–of–clinic information to epilepsy clinicians in a standardised format that alleviates time burden. Yet it also can provide a foundation for people with epilepsy, and carers, to exercise self–advocacy and have their holistic needs met. Whilst this study demonstrated alignment between the user groups, the specific needs of each must be balanced to encourage continued data generation by people with epilepsy/carers and use by clinicians in time–poor health care settings. N/A

In the United Kingdom (UK), people experiencing a suspected first seizure should be seen by a clinician with expertise in epilepsy within two weeks. First seizure clinics (FSC) aim to fast-track people outside the standard neurology outpatient system. Despite this, little over one in ten are seen within this timeframe. Most clinics are led by a consultant neurologist. There is a need to increase FSC provision utilising alternative expertise. A consultant nurse for the epilepsies led an innovative FSC, initially alongside a senior clinical fellow. Most patients were offered an electroencephalogram (EEG) (reported contemporaneously) at the FSC and there was provision for phlebotomy, electrocardiogram (ECG) and an urgent (same day) magnetic resonance imaging (MRI). In this prospective cohort study, we report 12-month clinical outcomes of our first 200 adult patients. 183/200 (91.5%) attended. 53 (26.5%) were seen within two weeks. 92% were given a diagnosis at FSC. 51 (25.5%) had epilepsy. 60 (30%) had an isolated seizure, 35% of which were provoked. 29 (14.5%) had functional dissociative seizures (FDS). 85/111 (76.5%) patients who had experienced at least one seizure were seizure-free at 12 months. Of 33 patients with FDS, 18 (55%) were seizure-free at 12 months. We present evidence that alternative methods of providing first seizure review can deliver timely, positive outcomes for patients. However, there must be robust governance arrangements in place involving a consultant neurologist. Many epilepsy specialist nurses will not have the experience, post registration qualifications, renumeration or willingness to undertake FSC. Therefore, there are limitations to the generalisability of these results.

Adverse impacts of epilepsy (e.g., injury, depression, and Sudden Unexpected Death from Epilepsy (SUDEP) can be mitigated by factors that patients may control, such as medication adherence, improved sleep and diet, reduced alcohol and taking care around pregnancy. New guidelines state that risk should be discussed at the time of diagnosis but some clinicians express concern about not wanting to raise anxiety. OBJECTIVE:To explicate practices by which epilepsy expert clinicians broach discussions of risk in specialist epilepsy clinics. METHODS:24 recordings of initial telephone appointments at specialist clinics where epilepsy is diagnosed from two specialist outpatient epilepsy services in England were subjected to Conversation Analysis. Data in British English. A single case study, identified as largely typical of the data set but also highlighting points of interest, is included to illustrate the findings. We also present reflections from analysis of 12 extracts examined in joint-analysis sessions with clinicians, researchers and patients. RESULTS:The analysis revealed that broaching risk was sensitive and challenging. Conversations involved confronting confusion about risk and negotiation between clinician and patient. Clinicians employ questions to establish the patient's knowledge. They were 'repair implicative' that is including lots of changes of sentence direction to achieve mutual understanding (intersubjectivity). Further, the Joint-Analysis highlighted the significance of epistemic matters - who knows what and how. CONCLUSION:Clinicians invite patients to share what they know about risk as a springboard for discussing behaviour change, enabling them to avoid naming specific risks (such as death). However, this often led to interactional trouble, and patients expressed a preference for more direct conversations. PRACTICE IMPLICATIONS:Clinicians can carefully calibrate risk information according to what the patient with epilepsy already knows, sensitively broaching risk of death. However, caution is needed to maximise patient engagement in risk management discussions.

Purpose: Research into epilepsy has experienced decades of chronic underfunding compared to other neurological conditions despite its prevalence and seriousness. To evidence the need for greater investment, the Epilepsy Research Institute (formerly Epilepsy Research UK) funded, led and managed a James Lind Alliance (JLA) Priority Setting Partnership (PSP). This “industry standard” methodology brings together healthcare professionals, patients, carers and patient group representatives to identify and prioritise research uncertainties within a defined area of health or care.Methods: The UK Epilepsy PSP is a once-in-a-generation, national consensus that collated and ranked the research priorities of the UK epilepsy and associated condition community. Following JLA methodology, this 18-month project engaged over 100 patient groups and 5,000 people affected by and working in epilepsy, including medics and allied healthcare professionals, from across the UK.Results: Over 5,400 priorities were received, with anti-seizure medication, sudden unexpected death in epilepsy (SUDEP) and epilepsy in women among the most frequently reported themes. The responses received were categorised and translated into distinct, researchable questions. Questions were excluded if deemed to be “answered” following an evidence check, while research uncertainties (i.e. unanswered and partially answered questions) formed the basis of a second, shortlisting survey. The shortlisted questions were then discussed and debated at the final workshop by participants that broadly represented the UK epilepsy and associated condition community. The final ranking and Top Ten priorities for research into epilepsy were then agreed.Conclusion: The aim of the UK Epilepsy PSP is to encourage and inspire researchers to investigate the research areas prioritised by those most affected by the condition and provide the evidence of need to aid future policy making discussions and support research funding applications.

Background Epilepsy is one of the commonest neurological conditions worldwide and confers a significant mortality risk, partly driven by status epilepticus (SE). Terminating SE is the goal of pharmaceutical rescue therapies. This survey evaluates UK-based healthcare professionals’ clinical practice and experience in community-based rescue therapy prescribing. Methods A cross-sectional, 21 item questionnaire composed of Likert-style and free-text based questions was administered online. It was distributed through a non-discriminative snow-balling methodology to members of the Epilepsy Specialist Nurses’ Association (ESNA) and the British International League Against Epilepsy (ILAE). Quantitative analysis used Chi-squared, Fishers’ exact and Mann-Whitney tests. Qualitative data were analysed through NVivo 14 software, following Braun and Clarke methodology. Results 86 participants comprising of nurses (n=64) and doctors (n=21) responded. Participants’ responses reflected guideline-concordant use of emergency management plans and buccal midazolam (BM) as a first-choice therapy for terminating tonic-clonic seizures in SE. However, significant variation (P<0.05) was found between doctors and nurses in prescribing practices of BM including maximum dose prescribed/day, withdrawal plans and the use in multimorbid patients. Eight themes were identified with some suggestive of concerns of overuse, misuse and abuse of BM by patients/carers. Conclusion This is the first study to give insights to community management of SE using rescue therapies particularly BM. Further evidence-based guidelines are needed for BM use in multimorbid patients and for its deprescribing. Robust safeguarding protocols and vigilance is needed to regulate BM's misuse and abuse potential. Oncoming community-based technology could provide objective assurance for evidencing utility of rescue medications.

PURPOSE:One in five people with autism spectrum disorder have epilepsy and take Anti-Seizure Medications (ASM). However, the impact of ASM on people with autism is under researched. This study evaluates the efficacy and tolerability of Levetiracetam (LEV) for autistic people and epilepsy. METHOD:Data was derived from the English Epilepsy Research Database Register which compares ASM responses in those with neurodevelopmental disorders to those without. Age range was 18-50 years as there were no autistic research participants with autism prescribed LEV over 50. Twelve-month ASM data, including withdrawal rate, seizure frequency and adverse effects were compared. Fisher's exact test was used to assess univariate associations between outcomes and autism with significance accepted as p < 0.05. Logistic regression was used to assess autism group differences after adjustment for potential confounders (age, gender, presence of baseline physical and mental health conditions). RESULTS:Of 175 (aged 18-50) research participants across 18 NHS Trusts, prescribed LEV between 2000 and 2020, 40 were autistic. There was no significant association between withdrawal rate (P = 0.626), or grouped side effects (physical P = 0.165, mental health P = 0.791). Autism was significantly associated with aggression with LEV in univariable analysis but this association was no longer significant after accounting for multiple testing A significant non-linear relationship between efficacy and the autism group (P < 0.001) was found. CONCLUSIONS:This study supports the use of LEV for people with autism and epilepsy as there is no difference in response noted to those without autism. However, they may have less prominent changes in efficacy.

INTRODUCTION:In England, nearly a quarter of people with intellectual disability (PwID) have epilepsy. Though 70 % of PwID have pharmaco-resistant seizures only 10 % are prescribed anti-seizure medication (ASMs) licenced for pharmaco-resistance. Brivaracetam (BRV) licenced in 2016 has had nine post-marketing studies involving PwID. These studies are limited either by lack of controls or not looking at outcomes based on differing levels of ID severity. This study looks at evidence comparing effectiveness and side-effects in PwID to those without ID prescribed Brivaracetam (BRV). METHODS:Pooled case note data for patients prescribed BRV (2016-2022) at 12 UK NHS Trusts were analysed. Demographics, starting and maximum dose, side-effects, dropouts and seizure frequency between ID (mild vs. moderate-profound (M/P)) and general population for a 12-month period were compared. Descriptive analysis, Mann-Whitney, Fisher's exact and logistic regression methods were employed. RESULTS:37 PwID (mild 17 M/P 20) were compared to 102 without ID. Mean start and maximum dose was lower for PwID than non-ID. Mean maximum dose reduced slightly with ID severity. No difference was found between ID and non-ID or between ID groups (Mild vs M/P) in BRV's efficacy i.e. >50 % seizure reduction or tolerability. Mental and behavioural side-effects were more prevalent for PwID (27.0 % ID, 17.6 % no ID) but not significantly higher (P = 0.441) or associated with ID severity (p = 0.255). CONCLUSION:This is the first study on BRV, which compares ID cohorts with differing severity and non-ID. Efficacy, tolerability and side-effects reported are similar across differing ID severity to those with no ID.

Background Epilepsy is a chronic neurological disorder affecting individuals globally, marked by recurrent and apparently unpredictable seizures that pose significant challenges, including increased mortality, injuries, and diminished quality of life. Despite advancements in treatments, a significant proportion of people with epilepsy continue to experience uncontrolled seizures. The apparent unpredictability of these events has been identified as a major concern for people with epilepsy, highlighting the need for innovative seizure forecasting technologies. Objective The ATMOSPHERE study aimed to develop and evaluate a digital intervention, using wearable technology and data science, that provides real-time, individualized seizure forecasting for individuals living with epilepsy. This paper reports the protocol for one of the workstreams focusing on the design and testing of a prototype to capture real-time input data needed for predictive modeling. The first aim was to collaboratively design the prototype (work completed). The second aim is to conduct an “in-the-wild” study to assess usability and refine the prototype (planned research). Methods This study uses a person-based approach to design and test the usability of a prototype for real-time seizure precipitant data capture. Phase 1 (work completed) involved co-design with individuals living with epilepsy and health care professionals. Sessions explored users’ requirements for the prototype, followed by iterative design of low-fidelity, static prototypes. Phase 2 (planned research) will be an “in-the-wild” usability study involving the deployment of a mid-fidelity, functional prototype for 4 weeks, with the collection of mixed methods usability data to assess the prototype’s real-world application, feasibility, acceptability, and engagement. This phase involves primary participants (adults diagnosed with epilepsy) and, optionally, their nominated significant other. The usability study will run in 3 rounds of deployment and data collection, aiming to recruit 5 participants per round, with prototype refinement between rounds. Results The phase-1 co-design study engaged 22 individuals, resulting in the development of a mid-fidelity, functional prototype based on identified requirements, including the tracking of evidence-based and personalized seizure precipitants. The upcoming phase-2 usability study is expected to provide insights into the prototype’s real-world usability, identify areas for improvement, and refine the technology for future development. The estimated completion date of phase 2 is the last quarter of 2024. Conclusions The ATMOSPHERE study aims to make a significant step forward in epilepsy management, focusing on the development of a user-centered, noninvasive wearable device for seizure forecasting. Through a collaborative design process and comprehensive usability testing, this research aims to address the critical need for predictive seizure forecasting technologies, offering a promising approach to improving the lives of individuals with epilepsy. By leveraging predictive analytics and personalized machine learning models, this technology seeks to offer a novel approach to managing epilepsy, potentially improving clinical outcomes, including quality of life, through increased predictability and seizure management. International Registered Report Identifier (IRRID) DERR1-10.2196/60129

IntroductionPeople with Intellectual Disabilities (PwID) are twenty times more likely than general population to have epilepsy. Guidance for prescribing antiseizure medication (ASM) to PwID is driven by trials excluding them. Levetiracetam (LEV) is a first-line ASM in the UK. Concerns exist regarding LEV's behavioural and psychological adverse effects, particularly in PwID. There is no high-quality evidence comparing effectiveness and adverse effects in PwID to those without, prescribed LEV.MethodsPooled casenote data for patients prescribed LEV (2000-2020) at 18 UK NHS Trusts were analysed. Demographics, starting and maximum dose, adverse effects, dropouts and seizure frequency between ID (mild vs. moderate-profound (M/P)) and general population for a 12-month period were compared. Descriptive analysis, Mann-Whitney, Fisher's exact and logistic regression methods were employed.Results173 PwID (mild 53 M/P 120) were compared to 200 without ID. Mean start and maximum dose were similar across all groups. PwID (Mild & M/P) were less likely to withdraw from treatment (P=0.036). No difference was found between ID and non-ID or between ID groups (Mild vs M/P) in LEV's efficacy i.e. >50% seizure reduction. Significant association emerged between ID severity and psychiatric adverse effects (P=0.035). More irritability (14.2%) and aggression (10.8%) were reported in M/P PwID.ConclusionPwID and epilepsy have high rates of premature mortality, comorbidities, treatment resistance and polypharmacy but remain poorly researched for ASM use. This is the largest studied cohort of PwID trialled on LEV compared to general population controls. Findings support prescribing of LEV for PwID as a first-line ASM.

Recent studies using computational and mathematical interrogation of background EEG have revealed eight biomarkers that inform a diagnostic decision-support tool called BioEP. To assess the utility of BioEP for aiding clinical decision making, we conducted a prospective single-site diagnostic belief updating study ([NCT05764252][1]). Eighty-six adults with suspected seizures attended a nurse-led, first-seizure clinic. Using a 7-point scale ranging from 'virtually certain' to 'exceptionally unlikely', two clinicians independently rated the probability of having another epileptic seizure before and after reviewing BioEP scores. Recruitment took place over 1 year. The probability ratings changed (beliefs updated) by at least one category in 35/86 participants for Reviewer 1 (41%; 95% confidence interval: 30-51%) and in 58/86 people for Reviewer 2 (67%; 58 to 77%). The impact of the presentation of new evidence from the BioEP score on reviewer beliefs was substantial and bidirectional. For Reviewer 1 n=20 lower and n=15 higher probability, with n=37 lower and n=21 higher probability for Reviewer 2. Future research will explore the impact of these biomarkers on long-term diagnostic decision making and examine robustness and generalisability in multi-site settings. ### Competing Interest Statement KM and MM are employees of Neuronostics; WW and JRT are co-founders, directors and share-holders of Neuronostics. ### Clinical Trial NCT05764252 ### Funding Statement JRT was supported by EPSRC (EP/N014391/2 & EP/T027703/1). WW was supported by Epilepsy Research UK (F2002). ### Author Declarations I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained. Yes The details of the IRB/oversight body that provided approval or exemption for the research described are given below: This study was approved by the NHS Health Research Authority & Health Care Research Wales (IRAS: 321340) and by the West Midlands - Solihull Research Ethics Committee. I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals. Yes I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance). Yes I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable. Yes The EEG recordings and meta-data are not publicly available due to restrictions by privacy laws. Post-processed data supporting the findings of this study are available upon reasonable request from the corresponding author (WW), and additional meta-data may be made available upon reasonable request from the study sponsor (MM). [1]: /lookup/external-ref?link_type=CLINTRIALGOV&access_num=NCT05764252&atom=%2Fmedrxiv%2Fearly%2F2024%2F12%2F16%2F2024.12.13.24318654.atom

BACKGROUND:People with intellectual disability (PwID) and epilepsy have increased premature and potentially preventable mortality. This is related to a lack of equitable access to appropriate care. The Step Together guidance and toolkit, developed with patient, clinical, charity and commissioning stakeholders, allows evaluation and benchmarking of essential epilepsy service provision for PwID in eight key domains, at a care system level. AIMS:To evaluate care provisions for adult PwID and epilepsy at a system level in the 11 integrated care systems (ICSs) of the Midlands, the largest NHS England region (population: approximately 11 million), using the Step Together toolkit. METHOD:Post training, each ICS undertook its benchmarking with the toolkit and submitted their scores to Epilepsy Action, a national UK epilepsy charity, who oversaw the process. The outcomes were analysed descriptively to provide results, individual and cumulative, at care domain and system levels. RESULTS:The toolkit was completed fully by nine of the 11 ICSs. Across all eight domains, overall score was 44.2% (mean 44.2%, median 43.3%, range 52.4%, interquartile range 23.8-76.2%). The domains of local planning (mean 31.1%, median 27.5%) and care planning (mean 31.4%, median 35.4%) scored the lowest, and sharing information scored the highest (mean 55.2%, median 62.5%). There was significant variability across each domain between the nine ICS. The user/carer participation domain had the widest variation across ICSs (0-100%). CONCLUSIONS:The results demonstrate a significant variance in service provision for PwID and epilepsy across the nine ICSs. The toolkit identifies specific areas for improvement within each ICS and region.

Artificial intelligence (AI) and machine learning (ML) are increasingly being used in medicine. More recently, both AI and ML have been applied in epilepsy research, with the aim of accurately predicting and detecting seizures. Most AL and ML applications need to undergo trials to ensure that accurate data are being captured. The following article discusses such systems which are currently being developed for epilepsy patients, which have/are undergoing patient trials.

There is a demand across the country for highly trained epilepsy specialist nurses (ESNs) due to the increasing number of patients that this long-term condition affects. It is estimated that 65 million people worldwide are affected by epilepsy. If not treated and supported effectively, epilepsy comes with high risk of mortality, comorbidities, stigma and high potential costs to the NHS. Patients who do not have access to a highly trained ESNs can be significantly and negatively affected by wrong diagnosis, gaps in knowledge when facing general care, incorrect and, on occasion, detrimental treatment options, poor epilepsy education and advocacy. Furthermore, patients living in rural areas often do not have access to large tertiary centres or neurologists with an interest in epilepsy. Due to this, access to ESNs running nurse-led clinics in smaller hospitals and community settings is vital to improving patient care. Nonetheless, there is a shortage of highly trained ESNs. The ESN development programme discussed in this article was formed as a concerted action to address recruitment challenges and widen the area and scope of care for people with an epilepsy diagnosis.

OBJECTIVE:This study was undertaken to validate a set of candidate biomarkers of seizure susceptibility in a retrospective, multisite case-control study, and to determine the robustness of these biomarkers derived from routinely collected electroencephalography (EEG) within a large cohort (both epilepsy and common alternative conditions such as nonepileptic attack disorder). METHODS:The database consisted of 814 EEG recordings from 648 subjects, collected from eight National Health Service sites across the UK. Clinically noncontributory EEG recordings were identified by an experienced clinical scientist (N = 281; 152 alternative conditions, 129 epilepsy). Eight computational markers (spectral [n = 2], network-based [n = 4], and model-based [n = 2]) were calculated within each recording. Ensemble-based classifiers were developed using a two-tier cross-validation approach. We used standard regression methods to assess whether potential confounding variables (e.g., age, gender, treatment status, comorbidity) impacted model performance. RESULTS:We found levels of balanced accuracy of 68% across the cohort with clinically noncontributory normal EEGs (sensitivity =61%, specificity =75%, positive predictive value =55%, negative predictive value =79%, diagnostic odds ratio =4.64, area under receiver operated characteristics curve =.72). Group level analysis found no evidence suggesting any of the potential confounding variables significantly impacted the overall performance. SIGNIFICANCE:These results provide evidence that the set of biomarkers could provide additional value to clinical decision-making, providing the foundation for a decision support tool that could reduce diagnostic delay and misdiagnosis rates. Future work should therefore assess the change in diagnostic yield and time to diagnosis when utilizing these biomarkers in carefully designed prospective studies.

Summary Background A retrospective, multi-site case control study was carried out to validate a set of candidate biomarkers of seizure susceptibility. The objective was to determine the robustness of these biomarkers derived from routinely collected EEG within a large cohort (both epilepsy and common alternative conditions which may present with a possible seizure, such as NEAD). Methods The database consisted of 814 EEG recordings from 648 subjects, collected from 8 NHS sites across the UK. Clinically non-contributory EEG recordings were identified by an experienced clinical scientist (N = 281; 152 alternative conditions, 129 epilepsy). Eight computational markers (spectral [N = 2], network-based [N = 4] and model-based [N = 2]) were calculated within each recording. Ensemble-based classifiers were developed using a two-tier cross-validation approach. We used standard regression methods in order to identify whether potential confounding variables (e.g. age, gender, treatment-status, comorbidity) impacted model performance. Findings We found levels of balanced accuracy of 68% across the cohort with clinically non-contributory normal EEGs (sensitivity: 61%, specificity: 75%, positive predictive value: 55%, negative predictive value: 79%, diagnostic odds ratio: 4.64). Group-level analysis found no evidence suggesting any of the potential confounding variables significantly impacted the overall performance. Interpretation These results provide evidence that the set of biomarkers could provide additional value to clinical decision-making, providing the foundation for a decision support tool that could reduce diagnostic delay and misdiagnosis rates. Future work should therefore assess the change in diagnostic yield and time to diagnosis when utilising these biomarkers in carefully designed prospective studies. Research in Context Evidence before this study We searched Google Scholar and Pubmed (March 21, 2022) for the following phrases ((“EEG” OR “electroencephalogram” OR “electroencephalography”) AND (“biomarker”) AND (“epilepsy” OR “seizure”) AND (“resting state” OR “resting-state”) OR (“normal”)). Several of the existing studies developed deep learning approaches for identifying the presence of interictal epileptiform discharges (IED), with the overarching aim to develop an automated stand-alone diagnostic tool. These approaches are particularly sensitive to the potential presence of artefacts in the EEG recordings and typically include spectral rather than network- or model-based features. We found no studies of more than 100 participants that assessed the cross-validated performance of candidate biomarkers on routine EEG recordings that were clinically non-contributory. One study found near-chance performance of a deep-learning based method using spectral features on a smaller cohort of people suspected of epilepsy (N=33 epilepsy; N=30 alternative conditions) with clinically non-contributory EEGs. Another study found overall accuracy of 69% (N=74 epilepsy; N=74 alternative conditions) but this framework did not use any independent cross-validation methods. Estimates of sensitivity of clinical markers of seizure susceptibility in routine EEG recordings vary between 17-56%. To the best of our knowledge no studies have assessed whether computational biomarkers offer sufficient discrimination between people with epilepsy and an alternative diagnosis to provide potential decision support for people with suspected epilepsy. Added value of this study We show that data-driven analysis of routinely collected EEGs that are currently considered clinically non-informative (i.e. absence of apparent epileptiform activity) can be used to distinguish EEGs from people with epilepsy from people with an alternative diagnosis with better-than-chance performance. To the best of our knowledge, this is the largest retrospective study assessing the performance of computational biomarkers derived from clinically non-contributory EEG recordings. The resulting statistical model is interpretable and relies on both spectral and computational (network- and model-based) features. We perform a series of validity and sensitivity analysis to assess the overall robustness of the final statistical model used for classification. We also conduct several statistical tests to analyse any shared characteristics (e.g. site, comorbidity) amongst the primary classes (FP, FN, TP, TN). These findings validate previous biomarker discovery- or development-studies, and provide evidence that they offer better-than-chance performance in a clinically relevant context. Future large-scale studies could consider combining these methods with interictal features for non-specialist settings. Implications of all the available evidence Our study presents evidence that computational analysis of clinically non-contributory EEGs could provide additional decision support for both epilepsy and alternative conditions. Since the statistical model and underlying features are interpretable, they could provide the starting point for further exploring the mechanisms that drive overall seizure-likelihood. Future work should focus on prospective testing and validation (e.g. identification of specific situations or cases in which these methods could be of added value) as well as assessing heterogeneity across different syndromes and diagnoses (e.g. NEAD, focal vs generalised epilepsy).

Background: Nearly a quarter of people with intellectual disability (ID) have epilepsy with large numbers experiencing drug-resistant epilepsy, and premature mortality. To mitigate epilepsy risks the environment and social care needs, particularly in professional care settings, need to be met. Purpose: To compare professional care groups as regards their subjective confidence and perceived responsibility when managing the need of people with ID and epilepsy. Method: A multi-agency expert panel developed a questionnaire with embedded case vignettes with quantitative and qualitative elements to understand training and confidence in the health and social determinants of people with ID and epilepsy. The cross-sectional survey was disseminated amongst health and social care professionals working with people with ID in the UK using an exponential nondiscriminative snow-balling methodology. Group comparisons were undertaken using suitable statistical tests including Fisher's exact, Kruskal-Wallis, and Mann-Whitney. Bonferroni correction was applied to significant (p < 0.05) results. Content analysis was conducted and relevant categories and themes were identified. Results: Social and health professionals (n = 54) rated their confidence to manage the needs of people with ID and epilepsy equally. Health professionals showed better awareness (p < 0.001) of the findings/recommendations of the latest evidence on premature deaths and identifying and managing epilepsy-related risks, including the relevance of nocturnal monitoring. The content analysis highlighted the need for clearer roles, improved care pathways, better epilepsy-specific knowledge, increased resources, and better multi-disciplinary work. Conclusions: A gap exists between health and social care professionals in awareness of epilepsy needs for people with ID, requiring essential training and national pathways. & COPY; 2023 The Author(s). Published by Elsevier Inc. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).