Artificial intelligence (AI) for neurologists: do digital neurones dream of electric sheep?

Artificial intelligence (AI) is routinely mentioned in journals and newspapers, and non-technical outsiders may have difficulty in distinguishing hyperbole from reality. We present a practical guide to help non-technical neurologists to understand healthcare AI. AI is being used to support clinical decisions in treating neurological disorders. We introduce basic concepts of AI, such as machine learning and natural language processing, and explain how AI is being used in healthcare, giving examples its benefits and challenges. We also cover how AI performance is measured, and its regulatory aspects in healthcare. An important theme is that AI is a general-purpose technology like medical statistics, with broad utility applicable in various scenarios, such that niche approaches are outpaced by approaches that are broadly applicable in many disease areas and specialties. By understanding AI basics and its potential applications, neurologists can make informed decisions when evaluating AI used in their clinical practice. This article was written by four humans, with generative AI helping with formatting and image generation.

Diagnostic signature for heart failure with preserved ejection fraction (HFpEF): a machine learning approach using multi-modality electronic health record data.

BACKGROUND: Heart failure with preserved ejection fraction (HFpEF) is thought to be highly prevalent yet remains underdiagnosed. Evidence-based treatments are available that increase quality of life and decrease hospitalization. We sought to develop a data-driven diagnostic model to predict from electronic health records (EHR) the likelihood of HFpEF among patients with unexplained dyspnea and preserved left ventricular EF. METHODS AND RESULTS: The derivation cohort comprised patients with dyspnea and echocardiography results. Structured and unstructured data were extracted using an automated informatics pipeline. Patients were retrospectively diagnosed as HFpEF (cases), non-HF (control cohort I), or HF with reduced EF (HFrEF; control cohort II). The ability of clinical parameters and investigations to discriminate cases from controls was evaluated by extreme gradient boosting. A likelihood scoring system was developed and validated in a separate test cohort. The derivation cohort included 1585 consecutive patients: 133 cases of HFpEF (9%), 194 non-HF cases (Control cohort I) and 1258 HFrEF cases (Control cohort II). Two HFpEF diagnostic signatures were derived, comprising symptoms, diagnoses and investigation results. A final prediction model was generated based on the averaged likelihood scores from these two models. In a validation cohort consisting of 269 consecutive patients [with 66 HFpEF cases (24.5%)], the diagnostic power of detecting HFpEF had an AUROC of 90% (P < 0.001) and average precision of 74%. CONCLUSION: This diagnostic signature enables discrimination of HFpEF from non-cardiac dyspnea or HFrEF from EHR and can assist in the diagnostic evaluation in patients with unexplained dyspnea. This approach will enable identification of HFpEF patients who may then benefit from new evidence-based therapies.

Evaluation and improvement of the National Early Warning Score (NEWS2) for COVID-19: a multi-hospital study.

BACKGROUND: The National Early Warning Score (NEWS2) is currently recommended in the UK for the risk stratification of COVID-19 patients, but little is known about its ability to detect severe cases. We aimed to evaluate NEWS2 for the prediction of severe COVID-19 outcome and identify and validate a set of blood and physiological parameters routinely collected at hospital admission to improve upon the use of NEWS2 alone for medium-term risk stratification. METHODS: Training cohorts comprised 1276 patients admitted to King's College Hospital National Health Service (NHS) Foundation Trust with COVID-19 disease from 1 March to 30 April 2020. External validation cohorts included 6237 patients from five UK NHS Trusts (Guy's and St Thomas' Hospitals, University Hospitals Southampton, University Hospitals Bristol and Weston NHS Foundation Trust, University College London Hospitals, University Hospitals Birmingham), one hospital in Norway (Oslo University Hospital), and two hospitals in Wuhan, China (Wuhan Sixth Hospital and Taikang Tongji Hospital). The outcome was severe COVID-19 disease (transfer to intensive care unit (ICU) or death) at 14 days after hospital admission. Age, physiological measures, blood biomarkers, sex, ethnicity, and comorbidities (hypertension, diabetes, cardiovascular, respiratory and kidney diseases) measured at hospital admission were considered in the models. RESULTS: A baseline model of 'NEWS2 + age' had poor-to-moderate discrimination for severe COVID-19 infection at 14 days (area under receiver operating characteristic curve (AUC) in training cohort = 0.700, 95% confidence interval (CI) 0.680, 0.722; Brier score = 0.192, 95% CI 0.186, 0.197). A supplemented model adding eight routinely collected blood and physiological parameters (supplemental oxygen flow rate, urea, age, oxygen saturation, C-reactive protein, estimated glomerular filtration rate, neutrophil count, neutrophil/lymphocyte ratio) improved discrimination (AUC = 0.735; 95% CI 0.715, 0.757), and these improvements were replicated across seven UK and non-UK sites. However, there was evidence of miscalibration with the model tending to underestimate risks in most sites. CONCLUSIONS: NEWS2 score had poor-to-moderate discrimination for medium-term COVID-19 outcome which raises questions about its use as a screening tool at hospital admission. Risk stratification was improved by including readily available blood and physiological parameters measured at hospital admission, but there was evidence of miscalibration in external sites. This highlights the need for a better understanding of the use of early warning scores for COVID.

Pre-existing cardiovascular disease rather than cardiovascular risk factors drives mortality in COVID-19.

BACKGROUND: The relative association between cardiovascular (CV) risk factors, such as diabetes and hypertension, established CV disease (CVD), and susceptibility to CV complications or mortality in COVID-19 remains unclear. METHODS: We conducted a cohort study of consecutive adults hospitalised for severe COVID-19 between 1st March and 30th June 2020. Pre-existing CVD, CV risk factors and associations with mortality and CV complications were ascertained. RESULTS: Among 1721 patients (median age 71 years, 57% male), 349 (20.3%) had pre-existing CVD (CVD), 888 (51.6%) had CV risk factors without CVD (RF-CVD), 484 (28.1%) had neither. Patients with CVD were older with a higher burden of non-CV comorbidities. During follow-up, 438 (25.5%) patients died: 37% with CVD, 25.7% with RF-CVD and 16.5% with neither. CVD was independently associated with in-hospital mortality among patients < 70 years of age (adjusted HR 2.43 [95% CI 1.16-5.07]), but not in those ≥ 70 years (aHR 1.14 [95% CI 0.77-1.69]). RF-CVD were not independently associated with mortality in either age group (< 70 y aHR 1.21 [95% CI 0.72-2.01], ≥ 70 y aHR 1.07 [95% CI 0.76-1.52]). Most CV complications occurred in patients with CVD (66%) versus RF-CVD (17%) or neither (11%; p < 0.001). 213 [12.4%] patients developed venous thromboembolism (VTE). CVD was not an independent predictor of VTE. CONCLUSIONS: In patients hospitalised with COVID-19, pre-existing established CVD appears to be a more important contributor to mortality than CV risk factors in the absence of CVD. CVD-related hazard may be mediated, in part, by new CV complications. Optimal care and vigilance for destabilised CVD are essential in this patient group. Trial registration n/a.

Bleeding in cardiac patients prescribed antithrombotic drugs: electronic health record phenotyping algorithms, incidence, trends and prognosis.

BACKGROUND: Clinical guidelines and public health authorities lack recommendations on scalable approaches to defining and monitoring the occurrence and severity of bleeding in populations prescribed antithrombotic therapy. METHODS: We examined linked primary care, hospital admission and death registry electronic health records (CALIBER 1998-2010, England) of patients with newly diagnosed atrial fibrillation, acute myocardial infarction, unstable angina or stable angina with the aim to develop algorithms for bleeding events. Using the developed bleeding phenotypes, Kaplan-Meier plots were used to estimate the incidence of bleeding events and we used Cox regression models to assess the prognosis for all-cause mortality, atherothrombotic events and further bleeding. RESULTS: We present electronic health record phenotyping algorithms for bleeding based on bleeding diagnosis in primary or hospital care, symptoms, transfusion, surgical procedures and haemoglobin values. In validation of the phenotype, we estimated a positive predictive value of 0.88 (95% CI 0.64, 0.99) for hospitalised bleeding. Amongst 128,815 patients, 27,259 (21.2%) had at least 1 bleeding event, with 5-year risks of bleeding of 29.1%, 21.9%, 25.3% and 23.4% following diagnoses of atrial fibrillation, acute myocardial infarction, unstable angina and stable angina, respectively. Rates of hospitalised bleeding per 1000 patients more than doubled from 1.02 (95% CI 0.83, 1.22) in January 1998 to 2.68 (95% CI 2.49, 2.88) in December 2009 coinciding with the increased rates of antiplatelet and vitamin K antagonist prescribing. Patients with hospitalised bleeding and primary care bleeding, with or without markers of severity, were at increased risk of all-cause mortality and atherothrombotic events compared to those with no bleeding. For example, the hazard ratio for all-cause mortality was 1.98 (95% CI 1.86, 2.11) for primary care bleeding with markers of severity and 1.99 (95% CI 1.92, 2.05) for hospitalised bleeding without markers of severity, compared to patients with no bleeding. CONCLUSIONS: Electronic health record bleeding phenotyping algorithms offer a scalable approach to monitoring bleeding in the population. Incidence of bleeding has doubled in incidence since 1998, affects one in four cardiovascular disease patients, and is associated with poor prognosis. Efforts are required to tackle this iatrogenic epidemic.

In response to Ballantyne and Schaefer's 'Consent and the ethical duty to participate in health data research'.

We welcome Ballantyne & Schaefer's discussion of the issues concerning consent and use of health data for research. In response to their acknowledgement of the need for public debate and discussion, we provide evidence from our own public consultation on this topic.

Late cortical plasticity in motor and auditory cortex: role of met-allele in BDNF Val66Met polymorphism.

The brain-derived neurotropic factor (BDNF) Val66Met polymorphism has been associated with abnormalities of synaptic plasticity in animal models, and abnormalities in motor cortical plasticity have also been described in humans using transcranial direct current stimulation. No study has yet been done on plasticity in non-motor regions, and the effect of two Met alleles (i.e. 'Met dose') is not well understood. We studied the effect of the BDNF Val66Met polymorphism on the after-effects of transcranial direct current stimulation and tetanic auditory stimulation in 65 subjects (23; Val66Val, 22; Val66Met and 20; Met66Met genotypes). In the first session, motor evoked potentials (MEP) were recorded under stereotaxic guidance for 90 min after 9 min of anodal transcranial direct current stimulation (TDCS). In the second session, auditory-evoked potentials (AEP) were recorded before and after 2 min of auditory 13 Hz tetanic stimulation. There was a difference in MEP facilitation post-TDCS comparing Met carriers with non-Met carriers, with Met carriers having a modest late facilitation at 30-90 min. There was no difference in responses between Val66Met genotype and Met66Met genotype subjects. Tetanic auditory stimulation also produced late facilitation of N1-P2 AEP at 25 min, but there was no apparent effect of genetic status. This study indicates that Met66Met carriers behave like Val66Met carriers for TDCS-induced plasticity, and produce a late facilitation of MEPs. Auditory cortical plasticity was not affected by the BDNF Val66Met polymorphism. This study sheds light on the differences between auditory and motor cortical plasticity and the role of the BDNF Val66Met polymorphism.

Secondary and primary dystonia: pathophysiological differences.

Primary dystonia is thought to be a disorder of the basal ganglia because the symptoms resemble those of patients who have anatomical lesions in the same regions of the brain (secondary dystonia). However, these two groups of patients respond differently to therapy suggesting differences in pathophysiological mechanisms. Pathophysiological deficits in primary dystonia are well characterized and include reduced inhibition at many levels of the motor system and increased plasticity, while emerging evidence suggests additional cerebellar deficits. We compared electrophysiological features of primary and secondary dystonia, using transcranial magnetic stimulation of motor cortex and eye blink classical conditioning paradigm, to test whether dystonia symptoms share the same underlying mechanism. Eleven patients with hemidystonia caused by basal ganglia or thalamic lesions were tested over both hemispheres, corresponding to affected and non-affected side and compared with 10 patients with primary segmental dystonia with arm involvement and 10 healthy participants of similar age. We measured resting motor threshold, active motor threshold, input/output curve, short interval intracortical inhibition and cortical silent period. Plasticity was probed using an excitatory paired associative stimulation protocol. In secondary dystonia cerebellar-dependent conditioning was measured using delayed eye blink classical conditioning paradigm and results were compared with the data of patients with primary dystonia obtained previously. We found no difference in motor thresholds, input/output curves or cortical silent period between patients with secondary and primary dystonia or healthy controls. In secondary dystonia short interval intracortical inhibition was reduced on the affected side, whereas it was normal on the non-affected side. Patients with secondary dystonia had a normal response to the plasticity protocol on both the affected and non-affected side and normal eye blink classical conditioning that was not different from healthy participants. In contrast, patients with primary dystonia showed increased cortical plasticity and reduced eye blink classical conditioning. Normal motor cortex plasticity in secondary dystonia demonstrates that abnormally enhanced cortical plasticity is not required for clinical expression of dystonia, and normal eye blink conditioning suggests an absence of functional cerebellar involvement in this form of dystonia. Reduced short interval intracortical inhibition on the side of the lesion may result from abnormal basal ganglia output or may be a consequence of maintaining an abnormal dystonic posture. Dystonia appears to be a motor symptom that can reflect different pathophysiological states triggered by a variety of insults.

Foresight-a generative pretrained transformer for modelling of patient timelines using electronic health records: a retrospective modelling study.

BACKGROUND: An electronic health record (EHR) holds detailed longitudinal information about a patient's health status and general clinical history, a large portion of which is stored as unstructured, free text. Existing approaches to model a patient's trajectory focus mostly on structured data and a subset of single-domain outcomes. This study aims to evaluate the effectiveness of Foresight, a generative transformer in temporal modelling of patient data, integrating both free text and structured formats, to predict a diverse array of future medical outcomes, such as disorders, substances (eg, to do with medicines, allergies, or poisonings), procedures, and findings (eg, relating to observations, judgements, or assessments). METHODS: Foresight is a novel transformer-based pipeline that uses named entity recognition and linking tools to convert EHR document text into structured, coded concepts, followed by providing probabilistic forecasts for future medical events, such as disorders, substances, procedures, and findings. The Foresight pipeline has four main components: (1) CogStack (data retrieval and preprocessing); (2) the Medical Concept Annotation Toolkit (structuring of the free-text information from EHRs); (3) Foresight Core (deep-learning model for biomedical concept modelling); and (4) the Foresight web application. We processed the entire free-text portion from three different hospital datasets (King's College Hospital [KCH], South London and Maudsley [SLaM], and the US Medical Information Mart for Intensive Care III [MIMIC-III]), resulting in information from 811 336 patients and covering both physical and mental health institutions. We measured the performance of models using custom metrics derived from precision and recall. FINDINGS: Foresight achieved a precision@10 (ie, of 10 forecasted candidates, at least one is correct) of 0·68 (SD 0·0027) for the KCH dataset, 0·76 (0·0032) for the SLaM dataset, and 0·88 (0·0018) for the MIMIC-III dataset, for forecasting the next new disorder in a patient timeline. Foresight also achieved a precision@10 value of 0·80 (0·0013) for the KCH dataset, 0·81 (0·0026) for the SLaM dataset, and 0·91 (0·0011) for the MIMIC-III dataset, for forecasting the next new biomedical concept. In addition, Foresight was validated on 34 synthetic patient timelines by five clinicians and achieved a relevancy of 33 (97% [95% CI 91-100]) of 34 for the top forecasted candidate disorder. As a generative model, Foresight can forecast follow-on biomedical concepts for as many steps as required. INTERPRETATION: Foresight is a general-purpose model for biomedical concept modelling that can be used for real-world risk forecasting, virtual trials, and clinical research to study the progression of disorders, to simulate interventions and counterfactuals, and for educational purposes. FUNDING: National Health Service Artificial Intelligence Laboratory, National Institute for Health and Care Research Biomedical Research Centre, and Health Data Research UK.

AI chatbots not yet ready for clinical use.

As large language models (LLMs) expand and become more advanced, so do the natural language processing capabilities of conversational AI, or "chatbots". OpenAI's recent release, ChatGPT, uses a transformer-based model to enable human-like text generation and question-answering on general domain knowledge, while a healthcare-specific Large Language Model (LLM) such as GatorTron has focused on the real-world healthcare domain knowledge. As LLMs advance to achieve near human-level performances on medical question and answering benchmarks, it is probable that Conversational AI will soon be developed for use in healthcare. In this article we discuss the potential and compare the performance of two different approaches to generative pretrained transformers-ChatGPT, the most widely used general conversational LLM, and Foresight, a GPT (generative pretrained transformer) based model focused on modelling patients and disorders. The comparison is conducted on the task of forecasting relevant diagnoses based on clinical vignettes. We also discuss important considerations and limitations of transformer-based chatbots for clinical use.

Mapping and evaluating national data flows: transparency, privacy, and guiding infrastructural transformation.

The importance of big health data is recognised worldwide. Most UK National Health Service (NHS) care interactions are recorded in electronic health records, resulting in an unmatched potential for population-level datasets. However, policy reviews have highlighted challenges from a complex data-sharing landscape relating to transparency, privacy, and analysis capabilities. In response, we used public information sources to map all electronic patient data flows across England, from providers to more than 460 subsequent academic, commercial, and public data consumers. Although NHS data support a global research ecosystem, we found that multistage data flow chains limit transparency and risk public trust, most data interactions do not fulfil recommended best practices for safe data access, and existing infrastructure produces aggregation of duplicate data assets, thus limiting diversity of data and added value to end users. We provide recommendations to support data infrastructure transformation and have produced a website (https://DataInsights.uk) to promote transparency and showcase NHS data assets.

Federated learning for medical imaging radiology.

Federated learning (FL) is gaining wide acceptance across the medical AI domains. FL promises to provide a fairly acceptable clinical-grade accuracy, privacy, and generalisability of machine learning models across multiple institutions. However, the research on FL for medical imaging AI is still in its early stages. This paper presents a review of recent research to outline the difference between state-of-the-art [SOTA] (published literature) and state-of-the-practice [SOTP] (applied research in realistic clinical environments). Furthermore, the review outlines the future research directions considering various factors such as data, learning models, system design, governance, and human-in-loop to translate the SOTA into SOTP and effectively collaborate across multiple institutions.

Latent Transformer Models for out-of-distribution detection.

Any clinically-deployed image-processing pipeline must be robust to the full range of inputs it may be presented with. One popular approach to this challenge is to develop predictive models that can provide a measure of their uncertainty. Another approach is to use generative modelling to quantify the likelihood of inputs. Inputs with a low enough likelihood are deemed to be out-of-distribution and are not presented to the downstream predictive model. In this work, we evaluate several approaches to segmentation with uncertainty for the task of segmenting bleeds in 3D CT of the head. We show that these models can fail catastrophically when operating in the far out-of-distribution domain, often providing predictions that are both highly confident and wrong. We propose to instead perform out-of-distribution detection using the Latent Transformer Model: a VQ-GAN is used to provide a highly compressed latent representation of the input volume, and a transformer is then used to estimate the likelihood of this compressed representation of the input. We demonstrate this approach can identify images that are both far- and near- out-of-distribution, as well as provide spatial maps that highlight the regions considered to be out-of-distribution. Furthermore, we find a strong relationship between an image's likelihood and the quality of a model's segmentation on it, demonstrating that this approach is viable for filtering out unsuitable images.

Cerebellar theta burst stimulation impairs eyeblink classical conditioning.

Theta burst stimulation (TBS) protocols of repetitive transcranial magnetic stimulation (rTMS) have after-effects on excitability of motor areas thought to be due to LTP- and LTD-like processes at cortical synapses. The present experiments ask whether, despite the low intensities of stimulation used and the anatomy of the posterior fossa, TBS can also influence the cerebellum. Acquisition and retention of eyeblink classical conditioning (EBCC) was examined in 30 healthy volunteers after continuous theta burst stimulation (cTBS) over the right cerebellar hemisphere. In subjects who received cerebellar cTBS, conditioned responses were fewer and their onsets were earlier (in the last half of the acquisition blocks) than those from control subjects. There was, however, no effect of cerebellar cTBS on the re-acquisition of EBCC in another session of EBCC 7­10 days later. There was also no effect of cerebellar cTBS on the re-acquisition of EBCC in subjects not naïve to EBCC when the stimulation was delivered immediately before a re-acquisition session. Control experiments verified that suppressive effects of cTBS on EBCC were not due to changes in motor cortical excitability or sensory disturbance caused by cTBS. Based on previous EBCC studies in various cerebellar pathologies, our data are compatible with the hypothesis that cerebellar cTBS has a focal cerebellar cortical effect, and are broadly in line with data from studies of EBCC in various animal models. These results confirm that cerebellar TBS has measurable effects on the function of the cerebellum, and indicate it is a useful non-invasive technique with which to explore cerebellar physiology and function in humans.

Tardive dyskinesia is caused by maladaptive synaptic plasticity: a hypothesis.

It has been 50 years since the first patients with tardive dyskinesia (TD) were described, but the pathophysiology is only partially understood and effective treatments have remained elusive. Newer atypical antipsychotics with less nonspecific activity at dopamine receptors have not heralded the end of tardive dyskinesia and merely highlight the incomplete understanding of the disorder. We present an overview of the existing pathophysiology of the disorder and incorporate recent developments in genetics and the study of human synaptic plasticity in other hyperkinetic movement disorders. We propose a hypothesis that dopamine-receptor sensitization and altered function of the N-methyl-D-aspartate receptor produces maladaptive synaptic plasticity, which allows the encoding of abnormal motor programs, and propose studies that would falsify or support this hypothesis. In conclusion, a maladaptive synaptic plasticity" hypothesis goes some way toward filling in the gaps of existing theories of TD with the pathophysiology of other hyperkinetic movement disorders. © 2012 Movement Disorder Society.

Human θ burst stimulation enhances subsequent motor learning and increases performance variability.

Intermittent theta burst stimulation (iTBS) transiently increases motor cortex excitability in healthy humans by a process thought to involve synaptic long-term potentiation (LTP), and this is enhanced by nicotine. Acquisition of a ballistic motor task is likewise accompanied by increased excitability and presumed intracortical LTP. Here, we test how iTBS and nicotine influences subsequent motor learning. Ten healthy subjects participated in a double-blinded placebo-controlled trial testing the effects of iTBS and nicotine. iTBS alone increased the rate of learning but this increase was blocked by nicotine. We then investigated factors other than synaptic strengthening that may play a role. Behavioral analysis and modeling suggested that iTBS increased performance variability, which correlated with learning outcome. A control experiment confirmed the increase in motor output variability by showing that iTBS increased the dispersion of involuntary transcranial magnetic stimulation-evoked thumb movements. We suggest that in addition to the effect on synaptic plasticity, iTBS may have facilitated performance by increasing motor output variability; nicotine negated this effect on variability perhaps via increasing the signal-to-noise ratio in cerebral cortex.

COVID-related hospitalization, intensive care treatment, and all-cause mortality in patients with psychosis and treated with clozapine.

Clozapine, an antipsychotic, is associated with increased susceptibility to infection with COVID-19, compared to other antipsychotics. Here, we investigate associations between clozapine treatment and increased risk of adverse outcomes of COVID-19, namely COVID-related hospitalisation, intensive care treatment, and death, amongst patients taking antipsychotics with schizophrenia-spectrum disorders. Using the clinical records of South London and Maudsley NHS Foundation Trust, we identified 157 individuals who had an ICD-10 diagnosis of schizophrenia-spectrum disorders, were taking antipsychotics (clozapine or other antipsychotics) at the time of COVID-19 pandemic in the UK and had a laboratory-confirmed COVID-19 infection. The following health outcomes were measured: COVID-related hospitalisation, COVID-related intensive care treatment and death. We tested associations between clozapine treatment and each outcome using logistic regression models, adjusting for gender, age, ethnicity, neighbourhood deprivation, obesity, smoking status, diabetes, asthma, bronchitis and hypertension using propensity scores. Of the 157 individuals who developed COVID-19 while on antipsychotics (clozapine or other antipsychotics), there were 28% COVID-related hospitalisations, 8% COVID-related intensive care treatments and 8% deaths of any cause during the 28 days follow-up period. amongst those taking clozapine, there were 25% COVID-related hospitalisations, 7% COVID-related intensive care treatments and 7% deaths. In both unadjusted and adjusted analyses, we found no significant association between clozapine and any of the outcomes. Thus, we found no evidence that patients with clozapine treatment at time of COVID-19 infection had increased risk of hospitalisation, intensive care treatment or death, compared to non-clozapine antipsychotic-treated patients. However, further research should be considered in larger samples to confirm this.

Anticoagulation for atrial fibrillation in people with serious mental illness in the general hospital setting.

OBJECTIVE: People with serious mental illnesses (SMI) have an increased risk of stroke compared to the general population. This study aims to evaluate oral anticoagulation prescription trends in atrial fibrillation (AF) patients with and without a comorbid SMI. METHODS: An open-source retrieval system for clinical data (CogStack) was used to identify a cohort of AF patients with SMI who ever had an inpatient admission to King's College Hospital from 2011 to 2020. A Natural Language Processing pipeline was used to calculate CHA2DS2-VASc and HASBLED risk scores from Electronic Health Records free text. Antithrombotic prescriptions of warfarin and Direct acting oral anti-coagulants (DOACs) (apixaban, rivaroxaban, dabigatran, edoxaban) were extracted from discharge summaries. RESULTS: Among patients included in the study (n = 16 916), 2.7% had a recorded co-morbid SMI diagnosis. Compared to non-SMI patients, those with SMI had significantly higher CHA2DS2-VASc (mean (SD): 5.3 (1.96) vs 4.7 (2.08), p < 0.001) and HASBLED scores (mean (SD): 3.2 (1.27) vs 2.5 (1.29), p < 0.001). Among AF patients having a CHA2DS2-VASc ≥2, those with co-morbid SMI were less likely than non-SMI patients to be prescribed an OAC (44% vs 54%, p < 0.001). However, there was no evidence of a significant difference between the two groups since 2019. CONCLUSION: Over recent years, DOAC prescription rates have increased among AF patients with SMI in acute hospitals. More research is needed to confirm whether the introduction of DOACs has reduced OAC treatment gaps in people with serious mental illness and to assess whether the use of DOACs has improved health outcomes in this population.

Moving towards vertically integrated artificial intelligence development.

Substantial interest and investment in clinical artificial intelligence (AI) research has not resulted in widespread translation to deployed AI solutions. Current attention has focused on bias and explainability in AI algorithm development, external validity and model generalisability, and lack of equity and representation in existing data. While of great importance, these considerations also reflect a model-centric approach seen in published clinical AI research, which focuses on optimising architecture and performance of an AI model on best available datasets. However, even robustly built models using state-of-the-art algorithms may fail once tested in realistic environments due to unpredictability of real-world conditions, out-of-dataset scenarios, characteristics of deployment infrastructure, and lack of added value to clinical workflows relative to cost and potential clinical risks. In this perspective, we define a vertically integrated approach to AI development that incorporates early, cross-disciplinary, consideration of impact evaluation, data lifecycles, and AI production, and explore its implementation in two contrasting AI development pipelines: a scalable "AI factory" (Mayo Clinic, Rochester, United States), and an end-to-end cervical cancer screening platform for resource poor settings (Paps AI, Mbarara, Uganda). We provide practical recommendations for implementers, and discuss future challenges and novel approaches (including a decentralised federated architecture being developed in the NHS (AI4VBH, London, UK)). Growth in global clinical AI research continues unabated, and introduction of vertically integrated teams and development practices can increase the translational potential of future clinical AI projects.