Hybrid architecture based intelligent diagnosis assistant for GP.

As the first point of contact for patients, General Practitioners (GPs) play a crucial role in the National Health Service (NHS). An accurate primary diagnosis from the GP can alleviate the burden on specialists and reduce the time needed to re-confirm the patient's condition, allowing for more efficient further examinations. However, GPs have broad but less specialized knowledge, which limits the accuracy of their diagnosis. Therefore, it is imperative to introduce an intelligent system to assist GPs in making decisions. This paper introduces two data augmentation methods, the Complaint Symptoms Integration Method and Symptom Dot Separating Method, to integrate essential information into the Integration dataset. Additionally, it proposes a hybrid architecture that fuses the features of words from different representation spaces. Experiments demonstrate that, compared to commonly used pre-trained attention-based models, our hybrid architecture delivers the best classification performance for four common neurological diseases on the enhanced Integration dataset. For example, the classification accuracy of the BERT+CNN hybrid architecture is 0.897, which is a 5.1% improvement over both BERT and CNN with 0.846. Finally, this paper develops an AI diagnosis assistant web application that leverages the superior performance of this architecture to help GPs complete primary diagnosis efficiently and accurately.

Technology-enabled seizure detection and reporting: The epilepsy network project.

The traditional models of epilepsy care provision have not changed substantially in more than a century, despite rapid advances in computing, technology and materials science. One consequence of these advances has been the near universal prevalence of smartphones. Wearable devices with a complex sensor arrays are an emerging technology. These devices provide a coalescence of digital computing and communication tools that offer a new way to detect, report and communicate about seizures. The pilot of a smartphone-based application for patients and cares allowing real-time reporting of seizures securely to the relevant epilepsy care team is described. Wrist-worn devices were evaluated for their ability to detect epileptic seizures. Relevant information, such as seizure notifications and live alerts for notification of emergency attendance or admission to the hospital, are sent securely to the epilepsy care team in real time. Tailored specialist advice following notification is provided along traditional lines. Compared to the year preceding the pilot, the interval between seizure occurrence in the community and notification of the specialist team reduced, with faster response times in terms of advice. There was a 30% reduction in admissions for patients with epilepsy and 10% reduction in length of stay. Patients using the technology report an increased feeling of empowerment. This model of care has several challenges and requires modification of existing working practices if benefits for patients are to be fully realised. The benefits and challenges of technology-enabled care in are discussed from the perspective of the experience from development to clinical deployment.

Digital Care in Epilepsy: A Conceptual Framework for Technological Therapies.

Epilepsy is associated with a significant increase in morbidity and mortality. The likelihood is significantly greater for those patients with specific risk factors. Identifying those at greatest risk of injury and providing expert management from the earliest opportunity is made more challenging by the circumstances in which many such patients present. Despite increasing recognition of the importance of earlier identification of those at risk, there is little or no improvement in outcomes over more than 30 years. Despite ever increasing sophistication of drug development and delivery, there has been no meaningful improvement in 1-year seizure freedom rates over this time. However, in the last few years, there has been an increase in patient-triggered interventions based on automated monitoring of indicators and risk factors facilitated by technological advances. The opportunities such approaches provide will only be realized if accompanied by current working practice changes. Replacing traditional follow-up appointments at arbitrary intervals with dynamic interventions, remotely and at the point and place of need provides a better chance of a substantial reduction in seizures for people with epilepsy. Properly implemented, electronic platforms can offer new opportunities to provide expert advice and management from first presentation thus improving outcomes. This perspective paper provides and proposes an informed critical opinion built on current evidence base of an outline techno-therapeutic approach to harnesses these technologies. This conceptual framework is generic, rather than tied to a specific product or solution, and the same generalized approach could be beneficially applied to other long-term conditions.

Genetic dysfunction of MT-ATP6 causes axonal Charcot-Marie-Tooth disease.

OBJECTIVE: Charcot-Marie-Tooth (CMT) disease is the most common inherited neuromuscular disorder, affecting 1 in 2,500 individuals. Mitochondrial DNA (mtDNA) mutations are not generally considered within the differential diagnosis of patients with uncomplicated inherited neuropathy, despite the essential requirement of ATP for axonal function. We identified the mtDNA mutation m.9185T>C in MT-ATP6, encoding the ATP6 subunit of the mitochondrial ATP synthase (OXPHOS complex V), at homoplasmic levels in a family with mitochondrial disease in whom a severe motor axonal neuropathy was a striking feature. This led us to hypothesize that mutations in the 2 mtDNA complex V subunit encoding genes, MT-ATP6 and MT-ATP8, might be an unrecognized cause of isolated axonal CMT and distal hereditary motor neuropathy (dHMN). METHODS: A total of 442 probands with CMT type 2 (CMT2) (270) and dHMN (172) were screened for MT-ATP6/8 mutations after exclusion of mutations in known CMT2/dHMN genes. Mutation load was quantified using restriction endonuclease analysis. Blue-native gel electrophoresis (BN-PAGE) was performed to analyze the effects of m.9185T>C on complex V structure and function. RESULTS: Three further probands with CMT2 harbored the m.9185T>C mutation. Some relatives had been classified as having dHMN. Patients could be separated into 4 groups according to their mutant m.9185T>C levels. BN-PAGE demonstrated both impaired assembly and reduced activity of the complex V holoenzyme. CONCLUSIONS: We have shown that m.9185T>C in MT-ATP6 causes CMT2 in 1.1% of genetically undefined cases. This has important implications for diagnosis and genetic counseling. Recognition that mutations in MT-ATP6 cause CMT2 enhances current understanding of the pathogenic basis of axonal neuropathy.