Search

Background/Aims Acute coronary syndromes include unstable angina and myocardial infarction, of which occlusive myocardial infarction is a high-risk subset that is often missed because of a lack of ST elevation. Pre-hospital electrocardiograms may be able to identify myocardial infarction early and reduce mortality. However, it is unclear whether pre-hospital electrocardiograms can accurately detect occlusive myocardial infarction and how this affects outcomes. This study will analyse the outcomes of patients with occlusive myocardial infarction who had a pre-hospital electrocardiogram. Method Electrocardiograms transmitted to the coronary care unit triage service were identified, along with data regarding patient demographics, referrals and mortality. Data were analysed for correlations between demographic and clinical factors and type of myocardial infarction. Results A total of 838 electrocardiograms were identified; 69 (8.2%) showed myocardial infarction and eight (1.0%) showed occlusive myocardial infarction, of which 50% had ST elevation. Patients with occlusive myocardial infarction were more likely to be triaged to the coronary care unit than patients with non-occlusive myocardial infarction (P=0.04). However, 38 (55.1%) of the patients with myocardial infarction and four (50.0%) of the patients with occlusive myocardial infarction were not directed to a coronary care unit or the catheterisation laboratory. Patients with occlusive myocardial infarction had higher index episode mortality rates (P=0.03) and 30-day mortality rates (PConclusions Despite pre-hospital electrocardiogram transmission and adherence to the guidelines, triage of myocardial infarction and occlusive myocardial infarction is imperfect. Refinement of risk scores and adaptation of new technology is required to help identify patients with occlusive myocardial infarction so they can be prioritised for immediate reperfusion therapy.

Treatment decisions in myocardial infarction (MI) are currently stratified by ST elevation (ST-elevation myocardial infarction [STEMI]) or lack of ST elevation (non-ST elevation myocardial infarction [NSTEMI]) on the electrocardiogram. This arose from the assumption that ST elevation indicated acute coronary artery occlusion (OMI). However, one-quarter of all NSTEMI cases are an OMI, and have a higher mortality. The purpose of this study was to identify features that could help identify OMI.Prospectively collected data from patients undergoing percutaneous coronary intervention (PCI) was analyzed. Data included presentation characteristics, comorbidities, treatments, and outcomes. Latent class analysis was undertaken, to determine patterns of presentation and history associated with OMI.A total of 1412 patients underwent PCI for acute MI, and 263 were diagnosed as OMI. Compared to nonocclusive MI, OMI patients are more likely to have fewer comorbidities but no difference in cerebrovascular disease and increased acute mortality (4.2% vs. 1.1%; p .001). Of OMI, 29.5% had delays to their treatment such as immediate reperfusion therapy. With latent class analysis, while clusters of similar patients are observed in the data set, the data available did not usefully identify patients with OMI compared to non-OMI.Features between OMI and STEMI are broadly very similar. However, there was no difference in age and risk of cerebrovascular disease in the OMI/non-OMI group. There are no reliable characteristics therefore for identifying OMI versus non-OMI. Delays to treatment also suggest that OMI patients are still missing out on optimal treatment. An alternative strategy is required to improve the identification of OMI patients.

BackgroundSelf-reported symptom studies rapidly increased our understanding of SARS-CoV-2 during the pandemic and enabled the monitoring of long-term effects of COVID-19 outside the hospital setting. It is now evident that post-COVID syndrome presents with heterogeneous profiles, which need characterisation to enable personalised care among the most affected survivors. This study describes post-COVID profiles, and how they relate to different viral variants and vaccination status.MethodsIn this prospective longitudinal cohort study, we analysed data from 336,652 subjects, with regular health reports through the Covid Symptom Study (CSS) smartphone application. These subjects had reported feeling physically normal for at least 30 days before testing positive for SARS-CoV-2. 9,323 individuals subsequently developed Long-COVID, defined as symptoms lasting longer than 28 days. 1,459 had post-COVID syndrome, defined as more than 12 weeks of symptoms. Clustering analysis of the time-series data was performed to identify distinct symptom profiles for post-COVID patients, across variants of SARS-CoV-2 and vaccination status at the time of infection. Clusters were then characterised based on symptom prevalence, duration, demography, and prior conditions (comorbidities).Using an independent testing sample with additional data (n=140), we investigated the impact of post-COVID symptom clusters on the lives of affected individuals.FindingsWe identified distinct profiles of symptoms for post-COVID syndrome within and across variants: four endotypes were identified for infections due to the wild-type variant; seven for the alpha variant; and five for delta. Across all variants, a cardiorespiratory cluster of symptoms was identified. A second cluster related to central neurological, and a third to cases with the most severe and debilitating multi-organ symptoms. Gastrointestinal symptoms clustered in no more than two specific phenotypes per viral variant. The three main clusters were confirmed in an independent testing sample, and their functional impact was assessed.InterpretationUnsupervised analysis identified different post-COVID profiles, characterised by differing symptom combinations, durations, and functional outcomes. Phenotypes were at least partially concordant with individuals’ reported experiences.Our classification may be useful to understand distinct mechanisms of the post-COVID syndrome, as well as subgroups of individuals at risk of prolonged debilitation.FundingUK Government Department of Health and Social Care, Chronic Disease Research Foundation, The Wellcome Trust, UK Engineering and Physical Sciences Research Council, UK Research and Innovation London Medical Imaging & Artificial Intelligence Centre for Value-Based Healthcare, UK National Institute for Health Research, UK Medical Research Council, British Heart Foundation and Alzheimer’s Society, and ZOE Limited, UK.Research in contextEvidence before this studyWe conducted a search in the PubMed Central database, with keywords: (“Long-COVID*” OR “post?covid*” OR “post?COVID*” OR postCOVID* OR postCovid*) AND (cluster* OR endotype* OR phenotype* OR sub?type* OR subtype).On 15 June 2022, 161 documents were identified, of which 24 either provided descriptions of sub-types or proposed phenotypes of Long-COVID or post-COVID syndrome(s). These included 16 studies attempting manual sub-grouping of phenotypes, 6 deployments of unsupervised methods for patient clustering and automatic semantic phenotyping (unsupervised k-means=2; random forest classification=1; other=2), and two reports of uncommon presentations of Long-COVID/post-COVID syndrome. Overall, two to eight symptom profiles (clusters) were identified, with three recurring clusters. A cardiopulmonary syndrome was the predominant observation, manifesting with exertional intolerance and dyspnoea (n=10), fatigue (n=8), autonomic dysfunction, tachycardia or palpitations (n=5), lung radiological abnormalities including fibrosis (n=2), and chest pain (n=1). A second common presentation consisted in persistent general autoimmune activation and proinflammatory state (n=2), comprising multi-organ mild sequelae (n=2), gastrointestinal symptoms (n=2), dermatological symptoms (n=2), and/or fever (n=1). A third syndrome was reported, with neurological or neuropsychiatric symptoms: brain fog or dizziness (n=2), poor memory or cognition (n=2), and other mental health issues including mood disorders (n=5), headache (n=2), central sensitization (n=1), paresthesia (n=1), autonomic dysfunction (n=1), fibromyalgia (n=2), and chronic pain or myalgias (n=6). Unsupervised clustering methods identified two to six different post-COVID phenotypes, mapping to the ones described above.14 further documents focused on possible causes and/or mechanisms of disease underlying one or more manifestations of Long-COVID or post-COVID and identifying immune response dysregulation as a potential common element. All the other documents were beyond the scope of this work.To our knowledge, there are no studies examining the symptom profile of post-COVID syndrome between different variants and vaccination status. Also, no studies reported the modelling of longitudinally collected symptoms, as time-series data, aiming at the characterisation of post-COVID syndrome.Added-value of this studyOur study aimed to identify symptom profiles for post-COVID syndrome across the dominant variants in 2020 and 2021, and across vaccination status at the time of infection, using a large sample with prospectively collected longitudinal self-reports of symptoms. For individuals developing 12 weeks or more of symptoms, we identified three main symptom profiles which were consistent across variants and by vaccination status, differing only in the ratio of individuals affected by each profile and symptom duration overall.Implications of all the available evidenceWe demonstrate the existence of different post-COVID syndromes, which share commonalities across SARS-CoV-2 variant types in both symptoms themselves and how they evolved through the illness. We describe subgroups of patients with specific post-COVID presentations which might reflect different underlying pathophysiological mechanisms. Given the time-series component, our study is relevant for post-COVID prognostication, indicating how long certain symptoms last. These insights could aid in the development of personalised diagnosis and treatment, as well as helping policymakers plan for the delivery of care for people living with post-COVID syndrome.

Introduction Electrode misplacement and interchange errors are known problems when recording the 12‑lead electrocardiogram (ECG). Automatic detection of these errors could play an important role for improving clinical decision making and outcomes in cardiac care. The objectives of this systematic review and meta-analysis is to 1) study the impact of electrode misplacement on ECG signals and ECG interpretation, 2) to determine the most challenging electrode misplacements to detect using machine learning (ML), 3) to analyse the ML performance of algorithms that detect electrode misplacement or interchange according to sensitivity and specificity and 4) to identify the most commonly used ML technique for detecting electrode misplacement/interchange. This review analysed the current literature regarding electrode misplacement/interchange recognition accuracy using machine learning techniques. Method A search of three online databases including IEEE, PubMed and ScienceDirect identified 228 articles, while 3 articles were included from additional sources from co-authors. According to the eligibility criteria, 14 articles were selected. The selected articles were considered for qualitative analysis and meta-analysis. Results The articles showed the effect of lead interchange on ECG morphology and as a consequence on patient diagnoses. Statistical analysis of the included articles found that machine learning performance is high in detecting electrode misplacement/interchange except left arm/left leg interchange. Conclusion This review emphasises the importance of detecting electrode misplacement detection in ECG diagnosis and the effects on decision making. Machine learning shows promise in detecting lead misplacement/interchange and highlights an opportunity for developing and operationalising deep learning algorithms such as convolutional neural network (CNN) to detect electrode misplacement/interchange.

Supplemental Digital Content is available in the text., Objectives: Timely prehospital diagnosis and treatment of acute coronary syndrome (ACS) are required to achieve optimal outcomes. Clinical decision support systems (CDSS) are platforms designed to integrate multiple data and can aid with management decisions in the prehospital environment. The review aim was to describe the accuracy of CDSS and individual components in the prehospital ACS management. Methods: This systematic review examined the current literature regarding the accuracy of CDSS for ACS in the prehospital setting, the influence of computer-aided decision-making and of 4 components: electrocardiogram, biomarkers, patient history, and examination findings. The impact of these components on sensitivity, specificity, and positive and negative predictive values was assessed. Results: A total of 11,439 articles were identified from a search of databases, of which 199 were screened against the eligibility criteria. Eight studies were found to meet the eligibility and quality criteria. There was marked heterogeneity between studies which precluded formal meta-analysis. However, individual components analysis found that patient history led to significant improvement in the sensitivity and negative predictive values. CDSS which incorporated all 4 components tended to show higher sensitivities and negative predictive values. CDSS incorporating computer-aided electrocardiogram diagnosis showed higher specificities and positive predictive values. Conclusions: Although heterogeneity precluded meta-analysis, this review emphasizes the potential of ACS CDSS in prehospital environments that incorporate patient history in addition to integration of multiple components. The higher sensitivity of certain components, along with higher specificity of computer-aided decision-making, highlights the opportunity for developing an integrated algorithm with computer-aided decision support.

Introduction Treatment urgency of myocardial infarction (MI) is currently stratified by ST elevation (STEMI) or lack of ST elevation (NSTEMI) on the electrocardiogram. This categorisation arose from the theory that ST elevation indicated total acute coronary artery occlusion (1). However, one-quarter of all NSTEMI have acute occlusion, have a higher mortality and are potentially missed by this strategy (2). Purpose The study purpose is to identify clinical features that could help distinguish an occlusive MI from non-occlusive MI. Methods The prospectively British Cardiovascular Intervention Society (BCIS) data set was collected from patients undergoing primary cutaneous intervention (dataset) was analysed from the years 2015 to 2019. Data collected included presentation characteristics, co-morbidities, treatments and outcomes. We further undertook latent class analysis, to determine whether distinct patterns of presentation and history are associated with acutely occluded NSTEMI. This methodology proceeds through segmentation of a dataset into clusters (“classes”) based on case similarities for the set of variables or “indicators” of interest; the classes are then statistically analysed to evaluate the differences between them. Results A total of 2025 patients underwent percutaneous coronary intervention (PCI) between 2015 and 2019. Of these, 1412 underwent PCI for MI, 510 had a STEMI, and 263 had occlusive MI on angiogram. STEMI patients tended to be younger (66 vs 67; p=0.045), in shock (6.1% vs 1.9%; p Conclusion Features between occlusive MI and STEMI are broadly very similar. However, unlike STEMI/NSTEMI, there was no difference in age and risk of cerebrovascular disease in the occlusive MI/non-occlusive MI group. This could indicate occlusion MI occurs in higher numbers in elderly and frail patients and may be frequently missed, possibly from greater collateralisation. There is no reliable characteristic for identifying occlusive between non-occlusive MI. Improved methods are needed for the identification of occlusive MI patients. Funding Acknowledgement Type of funding sources: Public grant(s) – EU funding. Main funding source(s): A project supported by the European Union's INTERREG VA Programme, managed by the Special EU Programmes Body (SEUPB). The funders of this project had no input in designing, implementation or writing of this review. Step 1 of latent class analysis

Electrocardiographic imaging is an imaging modality that has been introduced recently to help in visualizing the electrical activity of the heart and consequently guide the ablation therapy for ventricular arrhythmias. One of the main challenges of this modality is that the electrocardiographic signals recorded at the torso surface are contaminated with noise from different sources. Low amplitude leads are more affected by noise due to their low peak-to-peak amplitude. In this paper, we have studied 6 datasets from two torso tank experiments (Bordeaux and Utah experiments) to investigate the impact of removing or interpolating these low amplitude leads on the inverse reconstruction of cardiac electrical activity. Body surface potential maps used were calculated by using the full set of recorded leads, removing 1, 6, 11, 16, or 21 low amplitude leads, or interpolating 1, 6, 11, 16, or 21 low amplitude leads using one of the three interpolation methods - Laplacian interpolation, hybrid interpolation, or the inverse-forward interpolation. The epicardial potential maps and activation time maps were computed from these body surface potential maps and compared with those recorded directly from the heart surface in the torso tank experiments. There was no significant change in the potential maps and activation time maps after the removal of up to 11 low amplitude leads. Laplacian interpolation and hybrid interpolation improved the inverse reconstruction in some datasets and worsened it in the rest. The inverse forward interpolation of low amplitude leads improved it in two out of 6 datasets and at least remained the same in the other datasets. It was noticed that after doing the inverse-forward interpolation, the selected lambda value was closer to the optimum lambda value that gives the inverse solution best correlated with the recorded one.

Background Even in the era of digital technology, several hospitals still rely on paper-based forms for data entry for patient admission, triage, drug prescriptions, and procedures. Paper-based forms can be quick and convenient to complete but often at the expense of data quality, completeness, sustainability, and automated data analytics. Digital forms can improve data quality by assisting the user when deciding on the appropriate response to certain data inputs (eg, classifying symptoms). Greater data quality via digital form completion not only helps with auditing, service improvement, and patient record keeping but also helps with novel data science and machine learning research. Although digital forms are becoming more prevalent in health care, there is a lack of empirical best practices and guidelines for their design. The study-based hospital had a definite plan to abolish the paper form; hence, it was not necessary to compare the digital forms with the paper form. Objective This study aims to assess the usability of three different interactive forms: a single-page digital form (in which all data input is required on one web page), a multipage digital form, and a conversational digital form (a chatbot). Methods The three digital forms were developed as candidates to replace the current paper-based form used to record patient referrals to an interventional cardiology department (Cath-Lab) at Altnagelvin Hospital. We recorded usability data in a counterbalanced usability test (60 usability tests: 20 subjects×3 form usability tests). The usability data included task completion times, System Usability Scale (SUS) scores, User Experience Questionnaire data, and data from a postexperiment questionnaire. Results We found that the single-page form outperformed the other two digital forms in almost all usability metrics. The mean SUS score for the single-page form was 76 (SD 15.8; P=.01) when compared with the multipage form, which had a mean score of 67 (SD 17), and the conversational form attained the lowest scores in usability testing and was the least preferred choice of users, with a mean score of 57 (SD 24). An SUS score of >68 was considered above average. The single-page form achieved the least task completion time compared with the other two digital form styles. Conclusions In conclusion, the digital single-page form outperformed the other two forms in almost all usability metrics; it had the least task completion time compared with those of the other two digital forms. Moreover, on answering the open-ended question from the final customized postexperiment questionnaire, the single-page form was the preferred choice.

Background A 12-lead electrocardiogram (ECG) is the most commonly used method to diagnose patients with cardiovascular diseases. However, there are a number of possible misinterpretations of the ECG that can be caused by several different factors, such as the misplacement of chest electrodes. Objective The aim of this study is to build advanced algorithms to detect precordial (chest) electrode misplacement. Methods In this study, we used traditional machine learning (ML) and deep learning (DL) to autodetect the misplacement of electrodes V1 and V2 using features from the resultant ECG. The algorithms were trained using data extracted from high-resolution body surface potential maps of patients who were diagnosed with myocardial infarction, diagnosed with left ventricular hypertrophy, or a normal ECG. Results DL achieved the highest accuracy in this study for detecting V1 and V2 electrode misplacement, with an accuracy of 93.0% (95% CI 91.46-94.53) for misplacement in the second intercostal space. The performance of DL in the second intercostal space was benchmarked with physicians (n=11 and age 47.3 years, SD 15.5) who were experienced in reading ECGs (mean number of ECGs read in the past year 436.54, SD 397.9). Physicians were poor at recognizing chest electrode misplacement on the ECG and achieved a mean accuracy of 60% (95% CI 56.09-63.90), which was significantly poorer than that of DL (P Conclusions DL provides the best performance for detecting chest electrode misplacement when compared with the ability of experienced physicians. DL and ML could be used to help flag ECGs that have been incorrectly recorded and flag that the data may be flawed, which could reduce the number of erroneous diagnoses.

Background When a patient is suspected of having an acute myocardial infarction, they are accepted or declined for primary percutaneous coronary intervention partly based on clinical assessment of their 12-lead electrocardiogram (ECG) and ST-elevation myocardial infarction criteria. Objective We retrospectively determined the agreement rate between human (specialists called activator nurses) and computer interpretations of ECGs of patients who were declined for primary percutaneous coronary intervention. Methods Various features of patients who were referred for primary percutaneous coronary intervention were analyzed. Both the human and computer ECG interpretations were simplified to either “suggesting” or “not suggesting” acute myocardial infarction to avoid analysis of complex heterogeneous and synonymous diagnostic terms. Analyses, to measure agreement, and logistic regression, to determine if these ECG interpretations (and other variables such as patient age, chest pain) could predict patient mortality, were carried out. Results Of a total of 1464 patients referred to and declined for primary percutaneous coronary intervention, 722 (49.3%) computer diagnoses suggested acute myocardial infarction, whereas 634 (43.3%) of the human interpretations suggested acute myocardial infarction (P Conclusions Mortality in patients who were declined for primary percutaneous coronary intervention was higher than the reported mortality for ST-elevation myocardial infarction patients at 1 year. Agreement between computerized and human ECG interpretation is poor, perhaps leading to a high rate of inappropriate referrals. Work is needed to improve computer and human decision making when reading ECGs to ensure that patients are referred to the correct treatment facility for time-critical therapy.

This an exploratory paper that discusses the use of artificial intelligence (AI) in ECG interpretation and opportunities for improving the explainability of the AI (XAI) when reading 12-lead ECGs. To develop AI systems, many principles (human rights, well-being, data agency, effectiveness, transparency, accountability, awareness of misuse and competence) must be considered to ensure that the AI is trustworthy and applicable. The current computerised ECG interpretation algorithms can detect different types of heart diseases. However, there are some challenges and shortcomings that need to be addressed, such as the explainability issue and the interaction between the human and the AI for clinical decision making. These challenges create opportunities to develop a trustworthy XAI for automated ECG interpretation with a high performance and a high confidence level. This study reports a proposed XAI interface design in automatic ECG interpretation based on suggestions from previous studies and based on standard guidelines that were developed by the human computer interaction (HCI) community. New XAI interfaces should be developed in the future that facilitate more transparency of the decision logic of the algorithm which may allow users to calibrate their trust and use of the AI system.

Primary percutaneous coronary intervention (PPCI) is a minimally invasive procedure to unblock the arteries which carry blood to the heart. Referred patients are accepted or turned down for PPCI mainly based on the presence of ST segment elevation on the surface electrocardiogram. We explored the features which predict 30 days and 1-year mortality in accepted and turndown patients and report the performance of machine learning (ML) algorithms. Different ML algorithms, namely multiple logistic regression (MLR), decision tree (DT), and a support vector machine (SVM) were used for the prediction of 30 days and 1-year mortality. Upon significance of various features to predict the 30 days and 1-year mortality, the accuracy, sensitivity, and specificity were compared between algorithms. DT outperformed the other algorithms (SVM and MLR) to predict mortality of patients referred to the PPCI service. Greater sensitivity is achieved in predicting 30 days mortality in the accepted group compared to the turndown group, however, the former model included more features.

Linear ECG-lead transformations estimate or derive unrecorded target leads by applying a number of recorded basis leads to a so-called linear ECG-lead transformation matrix. The inverse transform of such a linear ECG-lead transformation performs a transformation in the opposite direction (from the target leads to the basis leads). The pseudo-inverse of a given transformation matrix can be used to perform such an inverse transformation. Linear regression based inverse transformation matrices are, provided that sufficient training data for their development is available, an alternative to pseudo-inverse matrices. The aim of this research was to compare the estimation performance of pseudo-inverse and linear regression based inverse transformations. This comparison was performed for two example inverse transformations. The performance of the different transformations was assessed using root-mean-squared-error (RMSE) values between the QRS-T complexes of recorded and derived leads. Typical mean RMSE values associated with the regression based approach were found to be approximately two thirds to half of the mean RMSE values achieved by the approach based upon the pseudo-inverse. Provided that sufficient data are available, linear regression should be used for the development of inverse ECG-lead transformation matrices.

Background: Acute coronary syndrome (ACS) is one of the main causes of death worldwide. The 12-lead electrocardiogram (ECG) is used to help diagnose ACS, along with clinical risk factors (smoking, diabetes mellitus, hypertension, hscTn and positive family history of ACS. These methods however are associated with many limitations resulting in variable sensitivity/specificity. The aim of this study was to use a machine learning approach to develop an optimum panel of blood protein biomarkers capable of independently diagnosing ACS. Methods: A hybrid feature selection and ML prediction algorithms including two classifiers: 1) decision tree (DT) and 2) logistic regression were applied to protein biomarkers (327 proteins) collected from patients with ACS n=91 or non-ACS n=97. Results: Using this approach, 20 proteins out of 327 proteins were able to accurately distinguish between ACS and non-ACS (logistic regression ROC-AUC=0.8 and accuracy=82.5% and DT ROC-A UC=0.6 and accuracy=64.9%. Conclusion: Logistic regression obtained a higher performance compared to DT and showed promising results uncovering a panel of 20 protein biomarkers which included those associated with progressive atherosclerotic plaques, myocardial injury and inflammation. This approach was able to accurately discriminate between patients with ACS and non-ACS.

BACKGROUND Even in the era of digital technology, several hospitals still rely on paper-based forms for data entry for patient admission, triage, drug prescriptions, and procedures. Paper-based forms can be quick and convenient to complete but often at the expense of data quality, completeness, sustainability, and automated data analytics. Digital forms can improve data quality by assisting the user when deciding on the appropriate response to certain data inputs (eg, classifying symptoms). Greater data quality via digital form completion not only helps with auditing, service improvement, and patient record keeping but also helps with novel data science and machine learning research. Although digital forms are becoming more prevalent in health care, there is a lack of empirical best practices and guidelines for their design. The study-based hospital had a definite plan to abolish the paper form; hence, it was not necessary to compare the digital forms with the paper form. OBJECTIVE This study aims to assess the usability of three different interactive forms: a single-page digital form (in which all data input is required on one web page), a multipage digital form, and a conversational digital form (a chatbot). METHODS The three digital forms were developed as candidates to replace the current paper-based form used to record patient referrals to an interventional cardiology department (Cath-Lab) at Altnagelvin Hospital. We recorded usability data in a counterbalanced usability test (60 usability tests: 20 subjects×3 form usability tests). The usability data included task completion times, System Usability Scale (SUS) scores, User Experience Questionnaire data, and data from a postexperiment questionnaire. RESULTS We found that the single-page form outperformed the other two digital forms in almost all usability metrics. The mean SUS score for the single-page form was 76 (SD 15.8; P=.01) when compared with the multipage form, which had a mean score of 67 (SD 17), and the conversational form attained the lowest scores in usability testing and was the least preferred choice of users, with a mean score of 57 (SD 24). An SUS score of >68 was considered above average. The single-page form achieved the least task completion time compared with the other two digital form styles. CONCLUSIONS In conclusion, the digital single-page form outperformed the other two forms in almost all usability metrics; it had the least task completion time compared with those of the other two digital forms. Moreover, on answering the open-ended question from the final customized postexperiment questionnaire, the single-page form was the preferred choice.

The application of artificial intelligence to interpret the electrocardiogram (ECG) has predominantly included the use of knowledge engineered rule-based algorithms which have become widely used today in clinical practice. However, over recent decades, there has been a steady increase in the number of research studies that are using machine learning (ML) to read or interrogate ECG data.The aim of this study is to review the use of ML with ECG data using a time series approach.Papers that address the subject of ML and the ECG were identified by systematically searching databases that archive papers from January 1995 to October 2019. Time series analysis was used to study the changing popularity of the different types of ML algorithms that have been used with ECG data over the past two decades. Finally, a meta-analysis of how various ML techniques performed for various diagnostic classifications was also undertaken.A total of 757 papers was identified. Based on results, the use of ML with ECG data started to increase sharply (p lt; 0.001) from 2012. Healthcare applications, especially in heart abnormality classification, were the most common application of ML when using ECG data (p lt; 0.001). However, many new emerging applications include using ML and the ECG for biometrics and driver drowsiness. The support vector machine was the technique of choice for a decade. However, since 2018, deep learning has been trending upwards and is likely to be the leading technique in the coming few years. Despite the accuracy paradox, accuracy was the most frequently used metric in the studies reviewed, followed by sensitivity, specificity, F1 score and then AUC.Applying ML using ECG data has shown promise. Data scientists and physicians should collaborate to ensure that clinical knowledge is being applied appropriately and is informing the design of ML algorithms. Data scientists also need to consider knowledge guided feature engineering and the explicability of the ML algorithm as well as being transparent in the algorithm's performance to appropriately calibrate human-AI trust. Future work is required to enhance ML performance in ECG classification.

BACKGROUND A 12-lead electrocardiogram (ECG) is the most commonly used method to diagnose patients with cardiovascular diseases. However, there are a number of possible misinterpretations of the ECG that can be caused by several different factors, such as the misplacement of chest electrodes. OBJECTIVE The aim of this study is to build advanced algorithms to detect precordial (chest) electrode misplacement. METHODS In this study, we used traditional machine learning (ML) and deep learning (DL) to autodetect the misplacement of electrodes V1 and V2 using features from the resultant ECG. The algorithms were trained using data extracted from high-resolution body surface potential maps of patients who were diagnosed with myocardial infarction, diagnosed with left ventricular hypertrophy, or a normal ECG. RESULTS DL achieved the highest accuracy in this study for detecting V1 and V2 electrode misplacement, with an accuracy of 93.0% (95% CI 91.46-94.53) for misplacement in the second intercostal space. The performance of DL in the second intercostal space was benchmarked with physicians (n=11 and age 47.3 years, SD 15.5) who were experienced in reading ECGs (mean number of ECGs read in the past year 436.54, SD 397.9). Physicians were poor at recognizing chest electrode misplacement on the ECG and achieved a mean accuracy of 60% (95% CI 56.09-63.90), which was significantly poorer than that of DL (P CONCLUSIONS DL provides the best performance for detecting chest electrode misplacement when compared with the ability of experienced physicians. DL and ML could be used to help flag ECGs that have been incorrectly recorded and flag that the data may be flawed, which could reduce the number of erroneous diagnoses. CLINICALTRIAL

BACKGROUND When a patient is suspected of having an acute myocardial infarction, they are accepted or declined for primary percutaneous coronary intervention partly based on clinical assessment of their 12-lead electrocardiogram (ECG) and ST-elevation myocardial infarction criteria. OBJECTIVE We retrospectively determined the agreement rate between human (specialists called activator nurses) and computer interpretations of ECGs of patients who were declined for primary percutaneous coronary intervention. METHODS Various features of patients who were referred for primary percutaneous coronary intervention were analyzed. Both the human and computer ECG interpretations were simplified to either “suggesting” or “not suggesting” acute myocardial infarction to avoid analysis of complex heterogeneous and synonymous diagnostic terms. Analyses, to measure agreement, and logistic regression, to determine if these ECG interpretations (and other variables such as patient age, chest pain) could predict patient mortality, were carried out. RESULTS Of a total of 1464 patients referred to and declined for primary percutaneous coronary intervention, 722 (49.3%) computer diagnoses suggested acute myocardial infarction, whereas 634 (43.3%) of the human interpretations suggested acute myocardial infarction (PP=.001). Only the age (odds ratio [OR] 1.07, 95% CI 1.05-1.09) and chest pain (OR 0.59, 95% CI 0.39-0.89) independent variables were statistically significant (P=.008) in predicting mortality after 30 days and 1 year. The odds for mortality within 1 year of referral were lower in patients with chest pain compared to those patients without chest pain. A referral being out of hours was a trending variable (OR 1.41, 95% CI 0.95-2.11, P=.09) for predicting the odds of 1-year mortality. CONCLUSIONS Mortality in patients who were declined for primary percutaneous coronary intervention was higher than the reported mortality for ST-elevation myocardial infarction patients at 1 year. Agreement between computerized and human ECG interpretation is poor, perhaps leading to a high rate of inappropriate referrals. Work is needed to improve computer and human decision making when reading ECGs to ensure that patients are referred to the correct treatment facility for time-critical therapy.

Introduction Acute coronary syndromes (ACS) are associated with high mortality. ST elevation Myocardial Infarction (STEMI) is the highest risk ACS, requiring timely diagnosis and management. A pre-hospital ECG has been shown to improve door to balloon time and therefore improve STEMI outcomes. However, ECG interpretation can be difficult, especially in the pre-hospital ?environment due to issues with experience and training. There are risks associated with under and over diagnoses. To help improve diagnostic accuracy and triage decisions, NHS Highland operates an ECG interpretation service whereupon all ECGs in suspected ACS patients are emailed from the community by the paramedic or primary care practitioner to Coronary Care Unit (CCU). At the CCU either a specialist cardiac nurse or consultant cardiologist provides ECG interpretation and advice to the pre-hospital team. This study reports outcomes from a consecutive cohort of patients. Methodology Consecutive suspected ACS patients referred to the pre-hospital ECG support service in May to July 2017 were included. Data collected included demographics, triage location, final diagnosis, troponin and 1 year outcome. Data were analysed to compare outcomes with provisional pre-hospital diagnosis and triage. Chi-squared test was used to determine if differences between groups were statistically significant (alpha=0.05). Results 428 ECGs were emailed to CCU over a 3 month time period, of which 337 were suitable for further analysis (91 excluded as duplicates or missing patient data). 27 patients from the transmitted ECGs had an ACS, of which 30% were STEMI and 70% were NSTEMI. The majority of patients with ACS (96.3%) had their ECG transmitted by ambulance staff. The ECG computerised diagnosis correctly identified only 30% of the ACS patients. Chest pain was the most common presenting complaint but 19% of ACS had a non chest pain symptom (breathless/collapse). 100% of ACS were admitted with 70% going to A&E (no free bed in cardiology) and 4% going directly to the cardiac catheterisation laboratory. Death rate was high, with 22% mortality during the index admission and 37% 1 year mortality. Conclusions The pre-hospital ECG transmission to CCU provides an additional support and clinical tool to optimise pre-hospital ACS care. However, the computer diagnosis alone misdiagnosed most ACS patients. Implementation of a clinical decision support system with integration of ECG, patient symptoms and clinical signs could improve pre-hospital ACS management.

Introduction Electrode misplacement can adversely affect Electrocardiogram (ECG) morphology and as a consequence, it can affect interpretation and patient diagnosis. Lead misplacement in ECG acquisition can occur in 40% to 60% of the time; which can result. in false positives (Type-I errors) and false negatives (Type-II errors). V1 and V2 electrodes are commonly misplaced high and wide of their correct positions. Vertically misplaced leads V1 and V2 can conceal real STEMI (ST Elevation Myocardial Infarction) events or emulate a false positive. Lead misplacement detection plays an important role to improve ECG data quality and clinical decision making in cardiac care. The aim of the current work was to develop a computer model using deep learning to detect misplacement of leads V1 and V2 for three different types of patients: 1) old Myocardial Infarction (MI) patients, 2) Left Ventricular Hypertrophy (LVH) Patients and 3) normal sinus rhythm patients, and to find in which patient type the ‘general’ patient-type-agnostic computer model achieves the highest performance. Methods 12-lead ECGs for 453 subjects, (normal n=151, LVH n=151, MI n=151) were extracted from body surface potential maps. V1 and V2 leads acquired from the second intercostal space (ICS) are considered as incorrect placement and leads from the fourth (IC) are considered as correct. The computer model was developed using deep learning algorithm (bidirectional long short-term memory network (BLSTM)) to detect V1 and V2 misplacement. The computer model was trained on three datasets together and tested on each dataset separately. In each dataset, data were divided into 67% for training and 33% for testing. The prevalence was 50% for correct and incorrect leads. Results Computer model accuracy using BLSTM of V1 and V2 misplacement detection was 93.0% in normal dataset, 93.0% in MI dataset while the accuracy was 91.0% in LVH dataset. Conclusion Computer model got a high performance to detect V1 and V2 misplacement in normal and MI ECGs datasets, while the performance decreased in LVH dataset to 91.0%. In future we need to consider MI infarct sites such as anterior, inferior etc to increase accuracy.

Background Electrocardiogram (ECG) lead misplacement can adversely affect ECG diagnosis and subsequent clinical decisions. V1 and V2 are commonly placed superior of their correct position. The aim of the current study was to use machine learning approaches to detect V1 and V2 lead misplacement to enhance ECG data quality. Method ECGs for 453 patients, (normal n = 151, Left Ventricular Hypertrophy (LVH) n = 151, Myocardial Infarction n = 151) were extracted from body surface potential maps. These were used to extract both the correct and incorrectly placed V1 and V2 leads. The prevalence for correct and incorrect leads were 50%. Sixteen features were extracted in three different domains: time-based, statistical and time-frequency features using a wavelet transform. A hybrid feature selection approach was applied to select an optimal set of features. To ensure optimal model selection, five classifiers were used and compared. The aforementioned feature selection approach and classifiers were applied for V1 and V2 misplacement in three different positions: first, second and third intercostal spaces (ICS). Results The accuracy for V1 misplacement detection was 93.9%, 89.3%, 72.8% in the first, second and third ICS respectively. In V2, the accuracy was 93.6%, 86.6% and 68.1% in the first, second and third ICS respectively. There is a noticeable decline in accuracy when detecting misplacement in the third ICS which is expected.

Computerised electrocardiogram (ECG) interpretation diagnostic algorithms have been developed to guide clinical decisions like with ST segment elevation myocardial infarction (STEMI) where time in decision making is critical. These computer-generated diagnoses have been proven to strongly influence the final ECG diagnosis by the clinician; often called automation bias. However, the computerised diagnosis may be inaccurate and could result in a wrong or delayed treatment harm to the patient. We hypothesise that an algorithmic certainty index alongside a computer-generated diagnosis might mitigate automation bias. The impact of reporting a certainty index on the final diagnosis is not known.To ascertain whether knowledge of the computer-generated ECG algorithmic certainty index influences operator diagnostic accuracy.Clinicians who regularly analyse ECGs such as cardiology or acute care doctors, cardiac nurses and ambulance staff were invited to complete an online anonymous survey between March and April 2019. The survey had 36 ECGs with a clinical vignette of a typical chest pain and which were either a STEMI, normal, or borderline (but do not fit the STEMI criteria) along with an artificially created certainty index that was either high, medium, low or none. Participants were asked whether the ECG showed a STEMI and their confidence in the diagnosis. The primary outcomes were whether a computer-generated certainty index influenced interpreter's diagnostic decisions and improved their diagnostic accuracy. Secondary outcomes were influence of certainty index between different types of clinicians and influence of certainty index on user's own-diagnostic confidence.A total of 91 participants undertook the survey and submitted 3262 ECG interpretations of which 75% of ECG interpretations were correct. Presence of a certainty index significantly increased the odds ratio of a correct ECG interpretation (OR 1.063, 95% CI 1.022-1.106, p = 0.004) but there was no significant difference between correct certainty index and incorrect certainty index (OR 1.028, 95% CI 0.923-1.145, p = 0.615). There was a trend for low certainty index to increase odds ratio compared to no certainty index (OR 1.153, 95% CI 0.898-1.482, p = 0.264) but a high certainty index significantly decreased the odds ratio of a correct ECG interpretation (OR 0.492, 95% CI 0.391-0.619, p 0.001). There was no impact of presence of a certainty index (p = 0.528) or correct certainty index (p = 0.812) on interpreters' confidence in their ECG interpretation.Our results show that the presence of an ECG certainty index improves the users ECG interpretation accuracy. This effect is not seen with differing levels of confidence within a certainty index, with reduced ECG interpretation success with a high certainty index compared with a trend for increased success with a low certainty index. This suggests that a certainty index improves interpretation when there is an increased element of doubt, possibly forcing the ECG user to spend more time and effort analysing the ECG. Further research is needed looking at time spent analysing differing certainty indices with alternate ECG diagnoses.

Background Acute Coronary Syndrome (ACS) is currently diagnosed using a 12‑lead Electrocardiogram (ECG). Our recent work however has shown that interpretation of the 12‑lead ECG is complex and that clinicians can be sub-optimal in their interpretation. Additionally, ECG does not always identify acute total occlusions in certain patients. Purpose The aim of the present study was to undertake an exploratory analysis to compare protein expression profiles of ACS patients that may in the future augment ECG diagnosis. Methods Patients were recruited consecutively at the cardiac catheterization laboratory at Altnagelvin Hospital over a period of 6 months. A low risk control group was recruited by advertisement. Blood samples were analysed using the multiplex proximity extension assays by OLINK proteomics. Support vector machine (SVM) learning was used as a classifier to distinguish between patient groups on training data. The ST segment elevation level was extracted from each ECG for a subset of patients and combined with the protein markers. Quadratic SVM (QSVM) learning was then used as a classifier to distinguish STEMI from NSTEMI on training and test data. Results Of the 344 participants recruited, 77 were initially diagnosed with NSTEMI, 7 with STEMI, and 81 were low risk controls. The other participants were those diagnosed with angina (stable and unstable) or non-cardiac patients. Of the 368 proteins analysed, 20 proteins together could significantly differentiate between patients with ACS and patients with stable angina (ROC-AUC = 0.96). Six proteins discriminated significantly between the stable angina and the low risk control groups (ROC-AUC = 1.0). Additionally, 16 proteins together perfectly discriminated between the STEMI and NSTEMI patients (ROC-AUC = 1). ECG comparisons with the protein biomarker data for a subset of patients (STEMI n = 6 and NSTEMI n = 6), demonstrated that 21 features (20 proteins + ST elevation) resulted in the highest classification accuracy 91.7% (ROC-AUC = 0.94). The 20 proteins without the ST elevation feature gave an accuracy of 80.6% (ROC-AUC 0.91), while the ST elevation feature without the protein biomarkers resulted in an accuracy of 69.3% (ROC-AUC = 0.81). Conclusions This preliminary study identifies panels of proteins that should be further explored in prospective studies as potential biomarkers that may augment ECG diagnosis and stratification of ACS. This work also highlights the importance for future studies to be designed to allow a comparison of blood biomarkers not only with ECG's but also with cardio angiograms.

Body surface potential mapping (BSPM) provides additional electrophysiological information that can be useful for the detection of cardiac diseases. Moreover, BSPMs are currently utilized in electrocardiographic imaging (ECGI) systems within clinical practice. Missing information due to noisy recordings, poor electrode contact is inevitable. In this study, we present an interpolation method that combines Laplacian minimization and principal component analysis (PCA) techniques for interpolating this missing information.The dataset used consisted of 117 lead BSPMs recorded from 744 subjects (a training set of 384 subjects, and a test set of 360). This dataset is a mixture of normal, old myocardial infarction, and left ventricular hypertrophy subjects. The missing data was simulated by ignoring data recorded from 7 regions: the first region represents three rows of five electrodes on the anterior torso surface (high potential gradient region), and the other six regions were realistic patterns that have been drawn from clinical data and represent the most likely regions of broken electrodes. Three interpolation methods including PCA based interpolation, Laplacian interpolation, and hybrid Laplacian-PCA interpolation methods were used to interpolate the missing data from the remaining electrodes. In the simulated region of missing data, the calculated potentials from each interpolation method were compared with the measured potentials using relative error (RE) and correlation coefficient (CC) over time. In the hybrid Laplacian-PCA interpolation method, the missing data are firstly interpolated using Laplacian interpolation, then the resulting BSPM of 117 potentials was multiplied by the (117 × 117) coefficient matrix calculated using the training set to get the principal components. Out of 117 principal components (PCs), the first 15 PCs were utilized for the second stage of interpolation. The best performance of interpolation was the reason for choosing the first 15 PCs.The differences in the median of relative error (RE) between Laplacian and Hybrid method ranged from 0.01 to 0.35 (p 0.001), while the differences in the median of correlation between them ranged from 0.0006 to 0.034 (p 0.001). PCA-interpolation method performed badly especially in some scenarios where the number of missing electrodes was up to 12 or higher causing a high region of missing data. The figures of median of RE for PCA-method were between 0.05 and 0.6 lower than that for Hybrid method (p 0.001). However, the median of correlation was between 0.0002 and 0.26 lower than the figure for the Hybrid method (p 0.001).Comparison between the three methods of interpolation (Laplacian, PCA, Hybrid) in reconstructing missing data in BSPM showed that the Hybrid method was always better than the other methods in all scenarios; whether the number of missed electrodes is high or low, and irrespective of the location of these missed electrodes.

In this paper, the effect of reducing noise from ECG signals is investigated by applying filters in compliance with the standards for ECG devices, removing and interpolating low amplitude signals. Torso-tank experiment data was used with electrical activity recorded simultaneously from 128 tank electrodes and 108 epicardial sock electrodes. Subsequently, 10 representative beats were selected for analysis. Tikhonov zero-order regularization method was used to solve the inverse problem for the following groups; raw fullset, filtered fullset, raw low amplitude removed, filtered low amplitude removed, raw low amplitude interpolated, filtered low amplitude interpolated torso signals. Pearson’s correlation was used for comparison between measured and computed electrograms and between activation maps derived from them. Filtering the signal according to the standards improved the reconstructed electrograms. In addition, removal of low amplitude signals and replacing them with interpolated signals combined with filtering according to the standard significantly improved the reconstructed electrograms and derived activation maps.