Lancet Digital Health最新文献

{"title":"The urgent need to accelerate synthetic data privacy frameworks for medical research","authors":"Anmol Arora MBBChir MA ,&nbsp;Siegfried Karl Wagner PhD FRCOphth ,&nbsp;Robin Carpenter BSc ,&nbsp;Rajesh Jena MD ,&nbsp;Pearse A Keane MD FRCOphth","doi":"10.1016/S2589-7500(24)00196-1","DOIUrl":"10.1016/S2589-7500(24)00196-1","url":null,"abstract":"<div><div>Synthetic data, generated through artificial intelligence technologies such as generative adversarial networks and latent diffusion models, maintain aggregate patterns and relationships present in the real data the technologies were trained on without exposing individual identities, thereby mitigating re-identification risks. This approach has been gaining traction in biomedical research because of its ability to preserve privacy and enable dataset sharing between organisations. Although the use of synthetic data has become widespread in other domains, such as finance and high-energy physics, use in medical research raises novel issues. The use of synthetic data as a method of preserving the privacy of data used to train models requires that the data are high fidelity with the original data to preserve utility, but must be sufficiently different as to protect against adversarial or accidental re-identification. There is a need for the development of standards for synthetic data generation and consensus standards for its evaluation. As synthetic data applications expand, ongoing legal and ethical evaluations are crucial to ensure that they remain a secure and effective tool for advancing medical research without compromising individual privacy.</div></div>","PeriodicalId":48534,"journal":{"name":"Lancet Digital Health","volume":"7 2","pages":"Pages e157-e160"},"PeriodicalIF":23.8,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142740999","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Artificial intelligence-guided detection of under-recognised cardiomyopathies on point-of-care cardiac ultrasonography: a multicentre study","authors":"Evangelos K Oikonomou MD ,&nbsp;Akhil Vaid MD ,&nbsp;Gregory Holste BA ,&nbsp;Andreas Coppi PhD ,&nbsp;Robert L McNamara MD ,&nbsp;Cristiana Baloescu MD ,&nbsp;Harlan M Krumholz MD ,&nbsp;Zhangyang Wang PhD ,&nbsp;Donald J Apakama MD ,&nbsp;Girish N Nadkarni MD ,&nbsp;Rohan Khera MD","doi":"10.1016/S2589-7500(24)00249-8","DOIUrl":"10.1016/S2589-7500(24)00249-8","url":null,"abstract":"<div><h3>Background</h3><div>Point-of-care ultrasonography (POCUS) enables cardiac imaging at the bedside and in communities but is limited by abbreviated protocols and variation in quality. We aimed to develop and test artificial intelligence (AI) models to screen for under-diagnosed cardiomyopathies from cardiac POCUS.</div></div><div><h3>Methods</h3><div>In a development set of 290 245 transthoracic echocardiographic videos across the Yale–New Haven Health System (YNHHS), we used augmentation approaches, and a customised loss function weighted for view quality to derive a POCUS-adapted, multi-label, video-based convolutional neural network that discriminates hypertrophic cardiomyopathy and transthyretin amyloid cardiomyopathy from controls without known disease. We evaluated the model across independent, internal, and external, retrospective cohorts of individuals undergoing cardiac POCUS across YNHHS and the Mount Sinai Health System (MSHS) emergency departments (between 2012 and 2024) to prioritise key views and validate the diagnostic and prognostic performance of single-view screening protocols.</div></div><div><h3>Findings</h3><div>Between Nov 1, 2023, and March 28, 2024, we identified 33 127 patients (mean age 58·9 [SD 20·5] years, 17 276 [52·2%] were female, 14 923 [45·0%] were male, and for 928 [2·8%] sex was recorded as unknown) at YNHHS and 5624 patients (mean age 56·0 [20·5] years, 1953 [34·7%] were female, 2470 [43·9%] were male, and for 1201 [21·4%] sex was recorded as unknown) at MSHS with 78 054 and 13 796 eligible cardiac POCUS videos, respectively. AI deployed to single-view POCUS videos successfully discriminated hypertrophic cardiomyopathy (eg, area under the receiver operating characteristic curve 0·903 [95% CI 0·795–0·981] in YNHHS; 0·890 [0·839–0·938] in MSHS for apical-4-chamber acquisitions) and transthyretin amyloid cardiomyopathy (0·907 [0·874–0·932] in YNHHS; 0·972 [0·959–0·983] in MSHS for parasternal acquisitions). In YNHHS, 40 (58%) of 69 hypertrophic cardiomyopathy cases and 22 (46%) of 48 transthyretin amyloid cardiomyopathy cases would have had a positive screen by AI-POCUS at a median of 2·1 (IQR 0·9–4·5) years and 1·9 (0·6–3·5) years before diagnosis. Moreover, among 25 261 participants without known cardiomyopathy followed up over a median of 2·8 (1·2–6·4) years, AI-POCUS probabilities in the highest (<em>vs</em> lowest) quintile for hypertrophic cardiomyopathy and transthyretin amyloid cardiomyopathy conferred a 17% (adjusted hazard ratio 1·17, 95% CI 1·06–1·29; p=0·0022) and 32% (1·39, 1·19–1·46; p&lt;0·0001) higher adjusted mortality risk, respectively.</div></div><div><h3>Interpretation</h3><div>We developed and validated an AI framework that enables scalable, opportunistic screening of under-recognised cardiomyopathies through simple POCUS acquisitions.</div></div><div><h3>Funding</h3><div>National Heart, Lung, and Blood Institute, Doris Duke Charitable Foundation, and BridgeBio.</div></div>","PeriodicalId":48534,"journal":{"name":"Lancet Digital Health","volume":"7 2","pages":"Pages e113-e123"},"PeriodicalIF":23.8,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143075873","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Exploring electronic health records to study rare diseases","authors":"The Lancet Digital Health","doi":"10.1016/j.landig.2025.01.008","DOIUrl":"10.1016/j.landig.2025.01.008","url":null,"abstract":"","PeriodicalId":48534,"journal":{"name":"Lancet Digital Health","volume":"7 2","pages":"Page e103"},"PeriodicalIF":23.8,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143075878","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"AI for medical diagnosis: does a single negative trial mean it is ineffective?","authors":"Olga Kostopoulou ,&nbsp;Brendan Delaney","doi":"10.1016/j.landig.2025.01.005","DOIUrl":"10.1016/j.landig.2025.01.005","url":null,"abstract":"","PeriodicalId":48534,"journal":{"name":"Lancet Digital Health","volume":"7 2","pages":"Pages e108-e109"},"PeriodicalIF":23.8,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143075869","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Prediction of emergency admissions: trade-offs between model simplicity and performance","authors":"Shishir Rao ,&nbsp;Kazem Rahimi","doi":"10.1016/j.landig.2024.12.008","DOIUrl":"10.1016/j.landig.2024.12.008","url":null,"abstract":"","PeriodicalId":48534,"journal":{"name":"Lancet Digital Health","volume":"7 2","pages":"Pages e106-e107"},"PeriodicalIF":23.8,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143075845","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"From the 100 Day Mission to 100 lines of software development: how to improve early outbreak analytics","authors":"Carmen Tamayo Cuartero DVM PhD ,&nbsp;Anna C Carnegie MPP ,&nbsp;Zulma M Cucunuba MD PhD ,&nbsp;Anne Cori PhD ,&nbsp;Sara M Hollis MSc ,&nbsp;Rolina D Van Gaalen PhD ,&nbsp;Amrish Y Baidjoe ,&nbsp;Alexander F Spina MPH ,&nbsp;John A Lees PhD ,&nbsp;Simon Cauchemez PhD ,&nbsp;Mauricio Santos PhD ,&nbsp;Juan D Umaña MSc ,&nbsp;Chaoran Chen PhD ,&nbsp;Hugo Gruson PhD ,&nbsp;Pratik Gupte PhD ,&nbsp;Joseph Tsui MSc ,&nbsp;Anita A Shah MPH ,&nbsp;Geraldine Gomez Millan SEP ,&nbsp;David Santiago Quevedo MSc ,&nbsp;Neale Batra MSc ,&nbsp;Prof Adam J Kucharski","doi":"10.1016/S2589-7500(24)00218-8","DOIUrl":"10.1016/S2589-7500(24)00218-8","url":null,"abstract":"<div><div>Since the COVID-19 pandemic, considerable advances have been made to improve epidemic preparedness by accelerating diagnostics, therapeutics, and vaccine development. However, we argue that it is crucial to make equivalent efforts in the field of outbreak analytics to help ensure reliable, evidence-based decision making. To explore the challenges and key priorities in the field of outbreak analytics, the Epiverse-TRACE initiative brought together a multidisciplinary group of experts, including field epidemiologists, data scientists, academics, and software engineers from public health institutions across multiple countries. During a 3-day workshop, 40 participants discussed what the first 100 lines of code written during an outbreak should look like. The main findings from this workshop are summarised in this Viewpoint. We provide an overview of the current outbreak analytic landscape by highlighting current key challenges that should be addressed to improve the response to future public health crises. Furthermore, we propose actionable solutions to these challenges that are achievable in the short term, and longer-term strategic recommendations. This Viewpoint constitutes a call to action for experts involved in epidemic response to develop modern and robust data analytic approaches at the heart of epidemic preparedness and response.</div></div>","PeriodicalId":48534,"journal":{"name":"Lancet Digital Health","volume":"7 2","pages":"Pages e161-e166"},"PeriodicalIF":23.8,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142873186","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Diagnoses supported by a computerised diagnostic decision support system versus conventional diagnoses in emergency patients (DDX-BRO): a multicentre, multiple-period, double-blind, cluster-randomised, crossover superiority trial","authors":"Wolf E Hautz MD ,&nbsp;Thimo Marcin PhD ,&nbsp;Stefanie C Hautz PhD ,&nbsp;Stefan K Schauber PhD ,&nbsp;Prof Gert Krummrey MD ,&nbsp;Martin Müller MD ,&nbsp;Thomas C Sauter MD ,&nbsp;Cornelia Lambrigger RN ,&nbsp;David Schwappach PhD ,&nbsp;Prof Mathieu Nendaz MD ,&nbsp;Gregor Lindner MD ,&nbsp;Simon Bosbach MD ,&nbsp;Ines Griesshammer MD ,&nbsp;Philipp Schönberg MD ,&nbsp;Emanuel Plüss MD ,&nbsp;Valerie Romann MD ,&nbsp;Svenja Ravioli MD ,&nbsp;Nadine Werthmüller MD ,&nbsp;Fabian Kölbener MD ,&nbsp;Prof Aristomenis K Exadaktylos MD ,&nbsp;Laura Zwaan PhD","doi":"10.1016/S2589-7500(24)00250-4","DOIUrl":"10.1016/S2589-7500(24)00250-4","url":null,"abstract":"&lt;div&gt;&lt;h3&gt;Background&lt;/h3&gt;&lt;div&gt;Diagnostic error is a frequent and clinically relevant health-care problem. Whether computerised diagnostic decision support systems (CDDSSs) improve diagnoses is controversial, and prospective randomised trials investigating their effectiveness in routine clinical practice are scarce. We hypothesised that diagnoses made with a CDDSS in the emergency department setting would be superior to unsupported diagnoses.&lt;/div&gt;&lt;/div&gt;&lt;div&gt;&lt;h3&gt;Methods&lt;/h3&gt;&lt;div&gt;This multicentre, multiple-period, double-blind, cluster-randomised, crossover superiority trial was done in four emergency departments in Switzerland. Eligible patients were adults (aged ≥18 years) presenting with abdominal pain, fever of unknown origin, syncope, or non-specific symptoms. Emergency departments were randomly assigned (1:1) to one of two predefined sequences of six alternating periods of intervention or control. Patients presenting during an intervention period were diagnosed with the aid of a CDDSS, whereas patients presenting during a control period were diagnosed without a CDDSS (usual care). Patients and personnel assessing outcomes were masked to group allocation; treating physicians were not. The primary binary outcome (false or true) was a composite score indicating a risk of reduced diagnostic quality, which was deemed to be present if any of the following occurred within 14 days: unscheduled medical care, a change in diagnosis, an unexpected intensive care unit admission within 24 h if initially admitted to hospital, or death. We assessed superiority of supported versus unsupported diagnoses in all consenting patients using a generalised linear mixed effects model. All participants who received any study treatment (including control) and completed the study were included in the safety analysis. This trial is registered with &lt;span&gt;&lt;span&gt;ClinicalTrials.gov&lt;/span&gt;&lt;svg&gt;&lt;path&gt;&lt;/path&gt;&lt;/svg&gt;&lt;/span&gt; (&lt;span&gt;&lt;span&gt;NCT05346523&lt;/span&gt;&lt;svg&gt;&lt;path&gt;&lt;/path&gt;&lt;/svg&gt;&lt;/span&gt;) and is closed to accrual.&lt;/div&gt;&lt;/div&gt;&lt;div&gt;&lt;h3&gt;Findings&lt;/h3&gt;&lt;div&gt;Between June 9, 2022, and June 23, 2023, 15 845 patients were screened and 1204 (591 [49·1%] female and 613 [50·9%] male) were included in the primary efficacy analysis. The median age of participants was 53 years (IQR 34–69). Diagnostic quality risk was observed in 100 (18%) of 559 patients with CDDSS-supported diagnoses and 119 (18%) of 645 with unsupported diagnoses (adjusted odds ratio 0·96 [95% CI 0·71–1·3]). 94 (7·8%) patients suffered a serious adverse event, none related to the study.&lt;/div&gt;&lt;/div&gt;&lt;div&gt;&lt;h3&gt;Interpretation&lt;/h3&gt;&lt;div&gt;Use of a CDDSS did not reduce the occurrence of diagnostic quality risk compared with the usual diagnostic process in adults presenting to emergency departments. Future research should aim to identify specific contexts in which CDDSSs are effective and how existing CDDSSs can be adapted to improve patient outcomes.&lt;/div&gt;&lt;/div&gt;&lt;div&gt;&lt;h3&gt;Funding&lt;/h3&gt;&lt;div&gt;Swiss National Science Foundation and University Hospi","PeriodicalId":48534,"journal":{"name":"Lancet Digital Health","volume":"7 2","pages":"Pages e136-e144"},"PeriodicalIF":23.8,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143075876","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Thank you to The Lancet Digital Health's statistical and peer reviewers in 2024","authors":"The Lancet Digital Health Editors","doi":"10.1016/j.landig.2025.01.009","DOIUrl":"10.1016/j.landig.2025.01.009","url":null,"abstract":"","PeriodicalId":48534,"journal":{"name":"Lancet Digital Health","volume":"7 2","pages":"Pages e110-e112"},"PeriodicalIF":23.8,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143075967","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Using artificial intelligence to switch from accident to sagacity in the serendipitous detection of uncommon diseases","authors":"Roberto Sciagrà","doi":"10.1016/j.landig.2024.12.006","DOIUrl":"10.1016/j.landig.2024.12.006","url":null,"abstract":"","PeriodicalId":48534,"journal":{"name":"Lancet Digital Health","volume":"7 2","pages":"Pages e104-e105"},"PeriodicalIF":23.8,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143075969","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Harnessing temporal patterns in administrative patient data to predict risk of emergency hospital admission","authors":"Benjamin Post MBChB ,&nbsp;Roman Klapaukh PhD ,&nbsp;Prof Stephen J Brett MD ,&nbsp;Prof A Aldo Faisal PhD","doi":"10.1016/S2589-7500(24)00254-1","DOIUrl":"10.1016/S2589-7500(24)00254-1","url":null,"abstract":"<div><h3>Background</h3><div>Unplanned hospital admissions are associated with worse patient outcomes and cause strain on health systems worldwide. Primary care electronic health records (EHRs) have successfully been used to create prediction models for emergency hospitalisation, but these approaches require a broad range of diagnostic, physiological, and laboratory values. In this study, we aimed to capture temporal patterns of patient activity from EHR data and evaluate their effectiveness in predicting emergency hospital admissions compared with conventional methods.</div></div><div><h3>Methods</h3><div>In this retrospective observational study, we used the Secure Anonymised Information Linkage databank to extract temporal patterns of primary care activity from undifferentiated electronic health record timestamp data for 1·37 million patients in Wales aged 18–80 years with at least one recorded Read code between the years 2016 and 2018. Using Gaussian mixture modelling we grouped patients into distinct temporal clusters, performed a three-stage validation of our approach and calculated the risk of emergency hospital admission for each temporal cluster group. Finally, these temporal clusters were combined with five administrative variables and incorporated into four emergency hospital admission prediction models (logistic regression, naive Bayes, XGBoost, and multilayer perceptron [MLP]) and compared with a more traditional, but data-intensive, modelling technique. The primary outcome was emergency hospital admission as the next health-care event.</div></div><div><h3>Findings</h3><div>Six distinct temporal cluster patterns of primary care EHR activity were identified, associated with varying risks of future emergency hospital admission risk. These patterns were visually interpretable, repeatable at a population-level, and clinically plausible. The best emergency hospital admission prediction model (MLP) achieved an area under the receiver operating characteristic (AUROC) of 0·82 and precision of 0·94 in regional cohorts. In external validation in regional cohorts, similar model performance was observed (AUROC 0·82 and precision 0·92). This model also matched the performance of a more complex model (extended feature model) requiring 33 clinical parameters (AUROC 0·82 <em>vs</em> 0·83; precision 0·94 <em>vs</em> 0·90) for the same task on the same dataset.</div></div><div><h3>Interpretation</h3><div>We developed a novel machine learning pipeline that extracts interpretable temporal patterns from simple representations of EHR data and can be incorporated into emergency hospital admission predictors. This framework might enable more rapid development of parsimonious clinical prediction models.</div></div><div><h3>Funding</h3><div>UKRI CDT in AI for Healthcare, UKRI Turing AI Fellowship, NIHR Imperial Biomedical Research Centre, and Research Capability Funding.</div></div>","PeriodicalId":48534,"journal":{"name":"Lancet Digital Health","volume":"7 2","pages":"Pages e124-e135"},"PeriodicalIF":23.8,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143075905","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}