European heart journal. Digital health最新文献

{"title":"The effectiveness of a telemedical program for COVID-19 positive high-risk patients in domestic isolation","authors":"L. Brunelli, L. Poelzl, J. Hirsch, C. Engler, F. Naegele, T. Egelseer-Bruendl, T. Scheffauer, C. Rassel, C. Schmit, F. Nawabi, A. Luckner-Hornischer, A. Bauer, G. Poelzl","doi":"10.1093/ehjdh/ztac076.2802","DOIUrl":"https://doi.org/10.1093/ehjdh/ztac076.2802","url":null,"abstract":"Abstract Background For almost two years, the Covid-19 pandemic has posed an enormous challenge to healthcare systems. Recurrent waves of disease brought the health systems to the limit of their resilience. Purpose The Tele-Covid telemedicine care program was installed in December 2020 to monitor high-risk patients in home isolation. Close monitoring allows early detection of disease deterioration and timely intensification of therapy, ideally avoiding intensive care. Conversely, if the course of the disease is stable, unnecessary hospitalisation can be avoided, thus reducing the burden on the healthcare system. Methods Patient acquisition was performed in collaboration with the local public health service and primary care physicians. Covid-19 positive high-risk patients (age >65 years and/or severe comorbidities) from the greater Innsbruck area were fitted with an ear sensor-based home monitoring system. The ear sensor measures SpO2, respiratory rate, body temperature and heart rate. The monitoring team (25 medical students supervised by 6 physicians) provided continuous monitoring of vital signs (24/7). After validation of the measurements, the collected parameters were evaluated using a specially developed risk score. If a defined risk score was exceeded, the patient was contacted by telephone. The combination of the clinical condition and the risk score determined the further course of action: (a) wait and see, (b) notify the primary care physician, or (c) refer for inpatient admission. The program was active from December 2020 to March 2022. In Summer 2021, the program was temporarily paused due to the epidemiological situation. Results A total of 132 patients (59.8% women) were monitored. The median age was 74 years (IQR: [67.3–80.8]). 91 patients (68.9%) had at least one relevant comorbidity. During the monitoring period, hospitalisation was required in 20 patients (15.2%), 3 of whom were transferred to the intensive care unit. Of the hospitalised patients, 3 (15%) patients died. During the same monitoring period, the Austrian Ministry of Health reported a mortality rate of 20.5% of all hospitalised patients in Austria aged 70–79 years. Subjectively, the patients felt safe due to close monitoring. Conclusion The Tele-Covid program is the successful implementation of a remote monitoring system in a pandemic situation. In the future, a broad application of the program is feasible. Funding Acknowledgement Type of funding sources: Public grant(s) – National budget only. Main funding source(s): Funded by the Region of the Tyrol","PeriodicalId":72965,"journal":{"name":"European heart journal. Digital health","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"88488707","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Importance of external validation and subgroup analysis of artificial intelligence in the detection of low ejection fraction from electrocardiograms.","authors":"Ryuichiro Yagi, Shinichi Goto, Yoshinori Katsumata, Calum A MacRae, Rahul C Deo","doi":"10.1093/ehjdh/ztac065","DOIUrl":"10.1093/ehjdh/ztac065","url":null,"abstract":"<p><strong>Aim: </strong>Left ventricular systolic dysfunction (LVSD) carries an increased risk for overt heart failure and mortality, yet treatable to mitigate disease progression. An artificial intelligence (AI)-enabled 12-lead electrocardiogram (ECG) model demonstrated promise in LVSD screening, but the performance dropped unexpectedly in external validation. We thus sought to train <i>de novo</i> models for LVSD detection and investigated their performance across multiple institutions and across a broader set of patient strata.</p><p><strong>Methods and results: </strong>ECGs taken within 14 days of an echocardiogram were obtained from four academic hospitals (three in the United States and one in Japan). Four AI models were trained to detect patients with ejection fraction (EF) <40% using ECGs from each of the four institutions. All the models were then evaluated on the held-out test data set from the same institution and data from the three external institutions. Subgroup analyses stratified by patient characteristics and common ECG abnormalities were performed. A total of 221 846 ECGs were identified from the 4 institutions. While the Brigham and Women's Hospital (BWH)-trained and Keio-trained models yielded similar accuracy on their internal test data [area under the receiver operating curve (AUROC) 0.913 and 0.914, respectively], external validity was worse for the Keio-trained model (AUROC: 0.905-0.915 for BWH trained and 0.849-0.877 for Keio-trained model). Although ECG abnormalities including atrial fibrillation, left bundle branch block, and paced rhythm-reduced detection, the models performed robustly across patient characteristics and other ECG features.</p><p><strong>Conclusion: </strong>While using the same model architecture, different data sets produced models with different performances for detecting low-EF highlighting the importance of external validation and extensive stratification analysis.</p>","PeriodicalId":72965,"journal":{"name":"European heart journal. Digital health","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-11-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ftp.ncbi.nlm.nih.gov/pub/pmc/oa_pdf/3c/b5/ztac065.PMC9779862.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"10747227","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Electronic health record-based facilitation of familial hypercholesterolaemia detection sensitivity of different algorithms in genetically confirmed patients.","authors":"Niekbachsh Mohammadnia, Ralph K Akyea, Nadeem Qureshi, Willem A Bax, Jan H Cornel","doi":"10.1093/ehjdh/ztac059","DOIUrl":"10.1093/ehjdh/ztac059","url":null,"abstract":"<p><strong>Aims: </strong>Familial hypercholesterolaemia (FH) is a disorder of LDL cholesterol clearance, resulting in increased risk of cardiovascular disease. Recently, we developed a Dutch Lipid Clinic Network (DLCN) criteria-based algorithm to facilitate FH detection in electronic health records (EHRs). In this study, we investigated the sensitivity of this and other algorithms in a genetically confirmed FH population.</p><p><strong>Methods and results: </strong>All patients with a healthcare insurance-related coded diagnosis of 'primary dyslipidaemia' between 2018 and 2020 were assessed for genetically confirmed FH. Data were extracted at the time of genetic confirmation of FH (T1) and during the first visit in 2018-2020 (T2). We assessed the sensitivity of algorithms on T1 and T2 for DLCN ≥ 6 and compared with other algorithms [familial hypercholesterolaemia case ascertainment tool (FAMCAT), Make Early Diagnoses to Prevent Early Death (MEDPED), and Simon Broome (SB)] using EHR-coded data and using all available data (i.e. including non-coded free text). 208 patients with genetically confirmed FH were included. The sensitivity (95% CI) on T1 and T2 with EHR-coded data for DLCN ≥ 6 was 19% (14-25%) and 22% (17-28%), respectively. When using all available data, the sensitivity for DLCN ≥ 6 was 26% (20-32%) on T1 and 28% (22-34%) on T2. For FAMCAT, the sensitivity with EHR-coded data on T1 was 74% (67-79%) and 32% (26-39%) on T2, whilst sensitivity with all available data was 81% on T1 (75-86%) and 45% (39-52%) on T2. For Make Early Diagnoses to Prevent Early Death MEDPED and SB, using all available data, the sensitivity on T1 was 31% (25-37%) and 17% (13-23%), respectively.</p><p><strong>Conclusions: </strong>The FAMCAT algorithm had significantly better sensitivity than DLCN, MEDPED, and SB. FAMCAT has the best potential for FH case-finding using EHRs.</p>","PeriodicalId":72965,"journal":{"name":"European heart journal. Digital health","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-10-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ftp.ncbi.nlm.nih.gov/pub/pmc/oa_pdf/47/b8/ztac059.PMC9779787.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"9153430","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Home monitoring of arterial pulse-wave velocity during COVID-19 total or partial lockdown using connected smart scales<sup />.","authors":"Rosa Maria Bruno, Jean Louis Pépin, Jean Philippe Empana, Rui Yi Yang, Vincent Vercamer, Paul Jouhaud, Pierre Escourrou, Pierre Boutouyrie","doi":"10.1093/ehjdh/ztac027","DOIUrl":"10.1093/ehjdh/ztac027","url":null,"abstract":"<p><strong>Aims: </strong>To investigate the impact of coronavirus disease 2019 lockdown on trajectories of arterial pulse-wave velocity in a large population of users of connected smart scales that provide reliable measurements of pulse-wave velocity.</p><p><strong>Methods and results: </strong>Pulse-wave velocity recordings obtained by Withings Heart Health & Body Composition Wi-Fi Smart Scale users before and during lockdown were analysed. We compared two demonstrative countries: France, where strict lockdown rules were enforced (<i>n</i> = 26 196) and Germany, where lockdown was partial (<i>n</i> = 26 847). Subgroup analysis was conducted in users of activity trackers and home blood pressure monitors. Linear growth curve modelling and trajectory clustering analyses were performed. During lockdown, a significant reduction in vascular stiffness, weight, blood pressure, and physical activity was observed in the overall population. Pulse-wave velocity reduction was greater in France than in Germany, corresponding to 5.2 month reduction in vascular age. In the French population, three clusters of stiffness trajectories were identified: decreasing (21.1%), stable (60.6%), and increasing pulse-wave velocity clusters (18.2%). Decreasing and increasing clusters both had higher pulse-wave velocity and vascular age before lockdown compared with the stable cluster. Only the decreasing cluster showed a significant weight reduction (-400 g), whereas living alone was associated with increasing pulse-wave velocity cluster. No clusters were identified in the German population.</p><p><strong>Conclusions: </strong>During total lockdown in France, a reduction in pulse-wave velocity in a significant proportion of French users of connected smart bathroom scales occurred. The impact on long-term cardiovascular health remains to be established.</p>","PeriodicalId":72965,"journal":{"name":"European heart journal. Digital health","volume":null,"pages":null},"PeriodicalIF":3.9,"publicationDate":"2022-10-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ftp.ncbi.nlm.nih.gov/pub/pmc/oa_pdf/ee/8f/ztac027.PMC9384477.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"10639204","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Atrial fibrillation virtual ward: reshaping the future of AF care","authors":"A. Kotb, S. Armstrong, I. Antoun, I. Koev, A. Mavilakandy, J. Barker, Z. Vali, G. Panchal, X. Li, M. Lazdam, M. Ibrahim, A. Sandilands, S. Chin, R. Somani, G. André Ng","doi":"10.1093/eurheartj/ehac544.2804","DOIUrl":"https://doi.org/10.1093/eurheartj/ehac544.2804","url":null,"abstract":"Abstract Background Atrial fibrillation (AF) hospital admissions represent significant AF related treatment costs nationally. In the year 2019–2020 our hospital reported 1,333 admissions with a primary diagnosis of AF, with a 10% annual increase. A virtual ambulatory AF ward providing multidisciplinary care with remote hospital-level monitoring could reshape the future model of AF management. Methods An AF virtual ward was implemented at our UK tertiary centre, as a proof-of-concept model of care. Patients admitted with a primary diagnosis of AF satisfying the AF virtual ward (AFVW) entry criteria (i.e., haemodynamically stable, HR <140 bpm with other acute conditions excluded) were given access to a single lead ECG recording device, a Bluetooth integrated blood pressure machine and pulse oximeter with instruction to record daily ECGs, blood pressure readings, oxygen saturations and fill an online AF symptom questionnaire via a smart phone or electronic tablet. Data were uploaded to an integrated digital platform for review by the clinical team who undertook twice daily virtual ward rounds. Medication adjustment was arranged through the hospital pharmacy. Data was collected prospectively for patients admitted to the AF virtual ward between 31 January and 11 March 2022. Outcomes included length of hospital stay, admission avoidance and re-admissions. Re-admission avoidance was assessed using the index admission criteria as a parameter for re-admission likelihood. Patients' satisfaction was assessed using the NHS family and friends' test (FFT). Results Over the 6-week period a total of 14 patients were enrolled. One patient was unable to be onboarded because of technology related anxiety with 13 patients onboarded to the virtual ward, 30.7% (n=4) did not have smart phones and were provided with electronic tablets. The age on admission was 64±10 years (mean±SD) with the oldest at 78 years of age. All patients were in AF with a mean heart rate of 122±24 bpm, and 38.5% (n=5) were discharged from the virtual ward in sinus rhythm. One patient was onboarded directly from pacemaker clinic and hence hospital admission was completely avoided, and 5 re-admissions were avoided for 3 patients. One patient required brief readmission due to persistent tachycardia requiring acute cardioversion. The FFT yielded 100% positive responses among patients. Conclusion This proof-of-concept is a first real world experience of a virtual ward for hospital patients with fast AF. It demonstrates a promising new telemedicine-based care model and with clear appetite among both patients and health professionals. This model of care has the potential to reduce the financial and backlog pressures caused by AF admissions without compromising patients' care or safety. Work is ongoing to further confirm the safety and cost-effectiveness upon further progress in a larger patient cohort. Funding Acknowledgement Type of funding sources: None.","PeriodicalId":72965,"journal":{"name":"European heart journal. Digital health","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"74121504","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A new mobile smartphone application, AF-EduApp, for atrial fibrillation patients: what do they use most?","authors":"L. Knaepen, R. Theunis, M. Delesie, J. Vijgen, P. Dendale, L. Desteghe, H. Heidbuchel","doi":"10.1093/ehjdh/ztac076.2822","DOIUrl":"https://doi.org/10.1093/ehjdh/ztac076.2822","url":null,"abstract":"Abstract Background The management of atrial fibrillation (AF) is complex and based on three main pillars: avoid stroke, better symptom control and cardiovascular risk factor management. Therefore, a holistic, multidisciplinary approach is needed in which the patient has a central role. Smartphone ownership increases strongly in the elderly population (in Belgian 65+ years old: 52% in 2018 to 82% in 2020). This digital growth creates opportunities for a closer patient follow-up. An in-house developed application, AF-EduApp, focused on delivering targeted education and guiding self-care, has been validated and is currently being studied in an ongoing clinical trial. Purpose Intermediate analysis of the user data of AF-EduApp. Methods At two Belgian hospitals, an open, prospective, randomized trial is currently performed. A total of 153 AF patients hospitalized or seen at an out-patient visit were included. Patients could use the application during a follow-up of 12 months. The AF-EduApp consists of six different modules: education, questionnaires with immediate patient feedback, medication overview with reminders, measurements (e.g. blood pressure, heart rate), appointments, and the possibility to ask questions to the caregivers. Knowledge about AF and its treatment was tested through the Jessa Atrial fibrillation Knowledge Questionnaire (JAKQ) with feedback on incorrectly answered questions. The main aim of the AF-EduApp is to improve patients' medication adherence through improved education and medication reminders. Results Currently, a total of 132 patients have completed a follow-up of 12 months (follow-up days: mean 357.3±60.7 and median: 365.5 [350.3–382.0]). The app was used on average 122.5±126.6 days (median: 55.0 [23.3–241.0]), or 34.3% of the available days. As shown in Fig. 1, the measurements and medication modules were the most used module (on 66.1% resp. 55.2% of the days). The education module was the least used module (3.5% of the days); the average education time was 17.0±27.7 min (median: 6.1 [1.4–20.6]). Within the measurement module (mean: 80.9±109.4 days used), the most frequently entered parameter was blood pressure, with on average 208.3±351.3 entries (median: 53.5 [7.0–296.3]) (Fig. 2). AF episodes was the least entered data (average 37.0±185.0 times; median 8.0 [4.0–19.0.3]). Conclusion Patients actively engaged with an educational smartphone AF application on 1/3th of the available days. The measurement module was the most used (to enter health data) together with the medication module (to confirm intake after reminder). It shows that many patients appreciate the mHealth tool to “connect” with their condition. The clinical trial tries to answer whether such increasing interaction leads to improved self-management and outcomes. Funding Acknowledgement Type of funding sources: Private company. Main funding source(s): The AF-EduApp study is supported by an BMS/Pfizer European Thrombosis Investigator Initiated Research Progr","PeriodicalId":72965,"journal":{"name":"European heart journal. Digital health","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"75126930","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A service evaluation of Zio XT: the Liverpool experience","authors":"A. Fawzy, J. Edmonds, A. Shannon, D. Wright","doi":"10.1093/ehjdh/ztac076.2812","DOIUrl":"https://doi.org/10.1093/ehjdh/ztac076.2812","url":null,"abstract":"Abstract Introduction The Zio XT is an adhesive, ambulatory heart rhythm monitoring device that can be worn for up to 14 days. It can be fitted by patients and utilises an Artificial Intelligence-based algorithm for rhythm analysis, offering potential convenience, accuracy and efficiency compared to Holter monitors. However, there is a lack of data regarding its efficacy and long-term impact. Thus, until further evidence ensues, NICE guidelines recommend Zio XT as a potential option for those requiring prolonged rhythm monitoring. Purpose We evaluated the efficacy of Zio XT for heart rhythm monitoring compared to Holter monitors. Methods 200 sequential patients that had Holter monitors and 204 that had Zio XT were included. Zio cases were randomly selected over 6 months to avoid the learning curve effect. Primary outcomes included time to results and the arrhythmia detection rate. Secondary outcomes included the proportion of patients that had heart rhythm monitoring in the 12 months preceding their investigation, those who required further tests as well as rates of outpatient appointments (OPAs) for device fitting and follow-up, and procedures such as device implantation and ablations. Results Data from 22 (10.8%) Zio patches was unavailable due to these being lost/not returned/unwearable, thus post-investigation outcomes were analysed for 182 Zio and 200 Holter cases. Zio XT was associated with a significantly shorter time to results compared to Holter monitors (median time: 21 days (interquartile range (IQR) 18–25) vs. 46 days (IQR 37.3–87.8), p<0.001), and a higher significant arrhythmia detection rate (55.4% vs. 17.5%, p<0.001). 26.5% of Zio patients had heart-rhythm monitoring in the preceding 12 months, compared to the 14.5% in the Holter group, p=0.003, with 55.8% having Holters and 28.8% having Zios previously, in the Zio group. A higher proportion of Zio recipients also required repeat tests (19.4 vs. 8.5%, p=0.002). Reasons for this included post-intervention monitoring (44.1%), lack of results due to devices being lost/faulty/not returned (41.2%) and a lack of diagnosis (14.7%). Zio monitoring was associated with a significant reduction in the need for OPAs for fitting (0.5% vs. 96%, p<0.001) and follow-up (70.1% vs. 87.0, p<0.001), and resulted in a significant increase in ablations (5.9% vs. 1.0%, p=0.005) but not device implantations (5.9% vs. 3.9, p=0.209). Conclusion Our findings indicate that Zio XT is associated with a statistically significant reduction in time to results, higher arrhythmia detection rate and a reduced need for OPAs. We demonstrated a higher rate of both Holter and Zio testing before and Zio testing after these investigations. We postulate that this has partly been due to a learning curve effect with the introduction of a new technology compared to the Holter which has been in use for many decades. Further large-scale evaluation is recommended to yield vital information on management pathways and cost efficacy","PeriodicalId":72965,"journal":{"name":"European heart journal. Digital health","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"85762063","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Inside the “brain” of an artificial neural network: an interpretable deep learning approach to paroxysmal atrial fibrillation diagnosis from electrocardiogram signals during sinus rhythm","authors":"P. Pantelidis, E. Oikonomou, S. Lampsas, N. Souvaliotis, M. Spartalis, M. Vavuranakis, M. Bampa, P. Papapetrou, G. Siasos, M. Vavuranakis","doi":"10.1093/ehjdh/ztac076.2781","DOIUrl":"https://doi.org/10.1093/ehjdh/ztac076.2781","url":null,"abstract":"Abstract Background With the ongoing, rapid advances in Deep Learning (DL), such solutions can now detect medical conditions even invisible to the human eye. In this direction, efforts have been made to develop DL algorithms that diagnose paroxysmal atrial fibrillation (PAF) from electrocardiogram (ECG) signals in sinus rhythm (SR). However, many of the available approaches function as “black boxes”, with physicians unable to understand and trust their predictions. Purpose To train a DL model to detect PAF patients while in SR and apply an algorithm that interprets and visualises its decisions. Methods We obtained ECG samples from PAF and non-PAF patients during SR, from the PAF Prediction Challenge Database. After discarding unannotated samples and augmenting the sample size (by dividing each signal into 30-second segments), we split the whole dataset into a train (68%), a validation (16%) and a test (16%) set. No pair of samples belonging to different sets originated from the same patient. We trained the InceptionTime neural network on the train/validation sets and tested on the “unseen” test set after “hiding” the correct answers. Its performance was evaluated with the following metrics: Accuracy, f1-score, precision and recall (sensitivity). After repeating this process 20 times, we obtained a distribution for each score. Finally, we adjusted the Grad-CAM interpretation algorithm to our data and used it to visualise the areas perceived as important by the model. Results After pre-processing, 4,080, 30-second, two-lead ECG signals were allocated to the train set, 960 to the validation and 960 to the test set. Each subset contained an equal number of PAF and non-PAF samples. After repeated training and testing, we obtained a median accuracy of 0.84 (interquartile range, IQR: 0.66–0.88), an f1-score of 0.82 (IQR: 0.68–0.88) and a median precision and recall equal to 0.93 (IQR: 0.67–0.99) and 0.77 (IQR: 0.68–0.93), respectively. The Grad-CAM technique highlighted the ECG areas of interest that led to each decision. We selected and present both PAF-positive and -negative samples, perceived either correctly or falsely. Interestingly, correct model decisions tend to focus on the P-wave, while false ones fixate on other regions. Conclusions Although a pilot study with considerable limitations (small sample size, disregard of possible confounding due to comorbidities or other factors), this work shows how DL can be employed to distinguish between PAF and non-PAF patients from SR ECG samples, and confirms the potential of DL-enabled approaches to offer novel diagnostic capabilities. Most importantly, our effort provides a comprehensible, visual interpretation of the model's decisions. Demystifying DL behaviour can, not only improve such efforts by explaining false decisions, but also cultivate trust among clinicians and, possibly, point out directions for future research, since we can now see through the magnifying lens of a neural network. Funding Ack","PeriodicalId":72965,"journal":{"name":"European heart journal. Digital health","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81104687","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Right heart catherisation – a virtual reality","authors":"M. Brown, N. Krishnananthan, V. Paul","doi":"10.1093/ehjdh/ztac076.2787","DOIUrl":"https://doi.org/10.1093/ehjdh/ztac076.2787","url":null,"abstract":"Abstract Introduction Right heart catheterisation (RHC) is the gold standard for assessing patients with pulmonary hypertension. Doctors require training in this procedure in a safe and friendly environment with minimal risk to patients. Due to the Covid pandemic, formal RHC teaching workshops were cancelled in our country, so we sought to develop a Virtual Reality Right Heart Catheterisation (VRRHC) training program to fulfil this area of need without the need for face to face contact. The aim was to improve training, competency and confidence in this technique with improved diagnostic skills and reduction of procedural errors. Method We approached a health technology company to design a VRRHC training module based on our current RHC simulation workshops. Phase 1 required virtual insertion of RHC via the right internal jugular vein using micro-puncture, double Seldinger technique under ultrasound guidance, followed by insertion of the RHC to the right atrium, right ventricle and pulmonary artery with pulmonary artery occlusion using real time pressure tracings and fluoroscopy. Thermodilution cardiac outputs and chamber saturations were also performed. The proprietary platform technology was delivered via a laptop and VR headset. Clinicians perform the VRRHC with imaging, monitoring and haptic feedback with the collection of real time performance tracking allowing user data (e.g. failed steps and proficiency scores) to be captured and subsequently visualised in the learning management system. We collected analytics and data on user engagement, experience and retention, targeted learning outcomes and learning curve, reduction in operating costs, reduction in procedure times due to higher proficiency, early diagnosis of pulmonary hypertension, reduced complications, improved interpretation and diagnosis. Results The program was launched in October 2021. Preliminary data shows a learning curve is associated with both using VR (10–15 minutes) and the RHC procedure itself. Initial time to completion of the RHC was 30–40 mins, reducing to 20–30 minutes with experience and 15 minutes in experts. Completion rates increase with experience from 40–50% to 100% and error rates reduce with frequency of completion. Conclusion A Virtual Reality Right Heart Catheter training program is safe, feasible and non-invasive. Increased experience results in increased completion rates, reduced procedure time and reduced errors. Using this program will potentially have beneficial effects on doctor training, outcomes, patient safety and health economics with no risk to a real patient. Funding Acknowledgement Type of funding sources: Private grant(s) and/or Sponsorship. Main funding source(s): Janssen Pharmaceuticals VRRHC images VRRHC hardware and utilisation","PeriodicalId":72965,"journal":{"name":"European heart journal. Digital health","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"79872846","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Predicting cardiovascular risk factors from facial & full body photography using deep learning","authors":"M. S. Knorr, M. Neyazi, J. P. Bremer, J. Brederecke, F. M. Ojeda, F. Ohm, M. Augustin, S. Blankenberg, N. Kirsten, R. B. Schnabel","doi":"10.1093/eurheartj/ehac544.2780","DOIUrl":"https://doi.org/10.1093/eurheartj/ehac544.2780","url":null,"abstract":"Abstract Introduction The early and easy detection of pathological cardiovascular phenotypes can lead to an early medical intervention and thus slow or limit the development of cardiovascular diseases. As full body photographs are easily obtainable without the need of medical expertise, this modality holds the potential to be viable for screening of populations. Purpose Utilizing data from a population-based study, we examined the possibility to detect cardiovascular risk factors from total body photographs using deep learning. Methods A population-based cohort study was utilized. The first data release provides facial and full body photographs in dermatologic standard poses of 6500 participants (median age 62.0 years, 49.6% men) and corresponding cardiovascular risk factors. Here, we focus on the most prevalent ones: smoking status (prevalence: 19.0%), hypertension (35.3%) and diabetes (8.2%). Here we employ a 2D-Convolutional Resnet-18 Neural Network for predicting the risk factors. It receives the full body image, the facial image and age and sex as input. We compare this to a logistic regression model only including sex and age. Logistic Regression and Neural Network are employed in a 5-fold validation scheme and t-tested for significance. Results Our model provided a good detection of arterial hypertension (AUC 0.711, CI 0.684–0.739), while a logistic regression on age and sex yielded a significantly worse prediction (AUC 0.681, CI 0.679– 0.683, p<0.05). Additionally, it made a good detection of a positive smoking status (AUC 0.733, CI 0.711–0.754), being significantly better than a respective logistic regression on age and sex (AUC 0.598, CI 0.597–0.6, p<0.001). Lastly, it classified diabetes well (AUC 0.744, CI 0.724–0.764, p<0.001) in comparison to the logistic regression (AUC 0.681, CI 0.679–0.683, p<0.001). Conclusion The presence of cardiovascular risk factors can be detected from a total body photograph. As total body photographs can be easily obtained with a majority of digital cameras, including smart phones, this model represents a potentially widely applicable diagnostic tool to easily screen large parts of the population for relevant cardiovascular risk factors, making an early medical intervention possible. Funding Acknowledgement Type of funding sources: None. Figure 1","PeriodicalId":72965,"journal":{"name":"European heart journal. Digital health","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"86635865","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}