Pub Date : 2024-11-01DOI: 10.1016/j.pcad.2024.10.002
Aditi Ujjawal , Tabitha Lobo , Henry K. Yaggi , Ian J. Neeland
The growing burden of coronary artery disease (CAD) has led to a deeper exploration of the pathophysiologic mechanisms underlying the disease process with the hope of finding novel treatments to reduce CAD morbidity and mortality. Sleep is a normal physiologic phenomenon essential for maintaining homeostasis. Disruption in sleep physiology has been linked to the activation of pro-inflammatory cytokines that may predispose to a greater risk of CAD. Several studies have evaluated the etiologic relationship between sleep deficiency and CAD. In this review, we attempt to highlight the key mechanisms proposed to play a role in the association of sleep with the pathophysiology of CAD, the findings and limitations of the pertinent studies, and possible future direction for evaluating and leveraging the relationship between sleep and CAD to develop new therapeutics.
{"title":"The connection between sleep deficiency and coronary artery disease: Complexities and controversies","authors":"Aditi Ujjawal , Tabitha Lobo , Henry K. Yaggi , Ian J. Neeland","doi":"10.1016/j.pcad.2024.10.002","DOIUrl":"10.1016/j.pcad.2024.10.002","url":null,"abstract":"<div><div>The growing burden of coronary artery disease (CAD) has led to a deeper exploration of the pathophysiologic mechanisms underlying the disease process with the hope of finding novel treatments to reduce CAD morbidity and mortality. Sleep is a normal physiologic phenomenon essential for maintaining homeostasis. Disruption in sleep physiology has been linked to the activation of pro-inflammatory cytokines that may predispose to a greater risk of CAD. Several studies have evaluated the etiologic relationship between sleep deficiency and CAD. In this review, we attempt to highlight the key mechanisms proposed to play a role in the association of sleep with the pathophysiology of CAD, the findings and limitations of the pertinent studies, and possible future direction for evaluating and leveraging the relationship between sleep and CAD to develop new therapeutics.</div></div>","PeriodicalId":21156,"journal":{"name":"Progress in cardiovascular diseases","volume":"87 ","pages":"Pages 83-89"},"PeriodicalIF":5.6,"publicationDate":"2024-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142396492","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-01DOI: 10.1016/j.pcad.2024.10.009
Mohammad Al Zein , Alicia Khazzeka , Alessandro El Khoury , Jana Al Zein , Dima Zoghaib , Ali H. Eid
Cardiovascular disease (CVD) continues to be a leading cause of global mortality and morbidity. Various established risk factors are linked to CVD, and modifying these risk factors is fundamental in CVD management. Clinical studies underscore the association between dyslipidemia and CVD, and therapeutic interventions that target low-density lipoprotein cholesterol elicit clear benefits. Despite the correlation between low high-density lipoprotein cholesterol (HDLC) and heightened CVD risk, HDL-raising therapies have yet to showcase significant clinical benefits. Furthermore, evidence from epidemiological and genetic studies reveals that not only low HDL-C levels, but also very high levels of HDL-C are linked to increased risk of CVD. In this review, we focus on HDL metabolism and delve into the relationship between HDL and CVD, exploring HDL functions and the observed alterations in its roles in disease. Altogether, the results discussed herein support the conventional wisdom that “too much of a good thing is not always a good thing”. Thus, our recommendation is that a careful reconsideration of the impact of high HDL-C levels is warranted, and shall be revisited in future research.
{"title":"Revisiting high-density lipoprotein cholesterol in cardiovascular disease: Is too much of a good thing always a good thing?","authors":"Mohammad Al Zein , Alicia Khazzeka , Alessandro El Khoury , Jana Al Zein , Dima Zoghaib , Ali H. Eid","doi":"10.1016/j.pcad.2024.10.009","DOIUrl":"10.1016/j.pcad.2024.10.009","url":null,"abstract":"<div><div>Cardiovascular disease (CVD) continues to be a leading cause of global mortality and morbidity. Various established risk factors are linked to CVD, and modifying these risk factors is fundamental in CVD management. Clinical studies underscore the association between dyslipidemia and CVD, and therapeutic interventions that target low-density lipoprotein cholesterol elicit clear benefits. Despite the correlation between low high-density lipoprotein cholesterol (HDL<img>C) and heightened CVD risk, HDL-raising therapies have yet to showcase significant clinical benefits. Furthermore, evidence from epidemiological and genetic studies reveals that not only low HDL-C levels, but also very high levels of HDL-C are linked to increased risk of CVD. In this review, we focus on HDL metabolism and delve into the relationship between HDL and CVD, exploring HDL functions and the observed alterations in its roles in disease. Altogether, the results discussed herein support the conventional wisdom that “too much of a good thing is not always a good thing”. Thus, our recommendation is that a careful reconsideration of the impact of high HDL-C levels is warranted, and shall be revisited in future research.</div></div>","PeriodicalId":21156,"journal":{"name":"Progress in cardiovascular diseases","volume":"87 ","pages":"Pages 50-59"},"PeriodicalIF":5.6,"publicationDate":"2024-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142515603","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-01DOI: 10.1016/j.pcad.2024.10.004
Ana Polo-López , Joaquín Calatayud , Laura López-Bueno , Rodrigo Núñez-Cortés , Lars Louis Andersen , Rubén López-Bueno
Objective
To investigate the prospective dose-response association of accelerometer-measured moderate-to-vigorous physical activity (PA;MVPA) with all-cause mortality and cardiovascular disease (CVD) incidence.
Methods
This prospective cohort of 76,074 participants from the UK Biobank study contained one week of individual accelerometer-based PA data collected between June 1, 2013 and December 23, 2015. Using restricted cubic splines to allow for potential non-linearity, we examined dose-response associations of MVPA with all-cause mortality and incident CVD, respectively.
Results
The median follow-up time was 8.0 years (IQR 7.5–8.5). The dose-response association of MVPA with all-cause mortality and CVD showed a similar L-shaped association, with significant risk reductions already from 10 min of MVPA per week for all-cause mortality (hazard ratio [HR], 0.98 [95 % CI,0.98–0.99]) and 15 min per week for CVD incidence (HR, 0.99 [95 % CI,0.98–0.99]). Doing more MVPA was associated with further risk reduction, but beyond around 500 min per week the benefits levelled off at HR's around 0.6 to 0.7. The highest additional benefit of adding more minutes per week for all-cause mortality and CVD incidence were observed between 100 and 250 weekly minutes of MVPA. From this point forward, the mean risk reduction rates decreased and were close to 0 beyond 500 weekly minutes.
Conclusions
Significant, but small, risk reductions in all-cause mortality and CVD incidence can be achieved with as little as 10 and 15 min of MVPA per week, respectively. However, public health organizations should promote the attainment of 250 min of MVPA per week (with 100 min as a possible first target for inactive individuals), as these thresholds are associated with the greatest efficiency. Beyond that, less pronounced risk reductions can be achieved by accumulating additional MVPA, with hardly any additional benefits beyond 500 weekly minutes.
{"title":"Dose-response association of an accelerometer-measured physical activity with all-cause mortality and cardiovascular disease incidence: Prospective cohort with 76,074 participants","authors":"Ana Polo-López , Joaquín Calatayud , Laura López-Bueno , Rodrigo Núñez-Cortés , Lars Louis Andersen , Rubén López-Bueno","doi":"10.1016/j.pcad.2024.10.004","DOIUrl":"10.1016/j.pcad.2024.10.004","url":null,"abstract":"<div><h3>Objective</h3><div>To investigate the prospective dose-response association of accelerometer-measured moderate-to-vigorous physical activity (PA;MVPA) with all-cause mortality and cardiovascular disease (CVD) incidence.</div></div><div><h3>Methods</h3><div>This prospective cohort of 76,074 participants from the UK Biobank study contained one week of individual accelerometer-based PA data collected between June 1, 2013 and December 23, 2015. Using restricted cubic splines to allow for potential non-linearity, we examined dose-response associations of MVPA with all-cause mortality and incident CVD, respectively.</div></div><div><h3>Results</h3><div>The median follow-up time was 8.0 years (IQR 7.5–8.5). The dose-response association of MVPA with all-cause mortality and CVD showed a similar L-shaped association, with significant risk reductions already from 10 min of MVPA per week for all-cause mortality (hazard ratio [HR], 0.98 [95 % CI,0.98–0.99]) and 15 min per week for CVD incidence (HR, 0.99 [95 % CI,0.98–0.99]). Doing more MVPA was associated with further risk reduction, but beyond around 500 min per week the benefits levelled off at HR's around 0.6 to 0.7. The highest additional benefit of adding more minutes per week for all-cause mortality and CVD incidence were observed between 100 and 250 weekly minutes of MVPA. From this point forward, the mean risk reduction rates decreased and were close to 0 beyond 500 weekly minutes.</div></div><div><h3>Conclusions</h3><div>Significant, but small, risk reductions in all-cause mortality and CVD incidence can be achieved with as little as 10 and 15 min of MVPA per week, respectively. However, public health organizations should promote the attainment of 250 min of MVPA per week (with 100 min as a possible first target for inactive individuals), as these thresholds are associated with the greatest efficiency. Beyond that, less pronounced risk reductions can be achieved by accumulating additional MVPA, with hardly any additional benefits beyond 500 weekly minutes.</div></div>","PeriodicalId":21156,"journal":{"name":"Progress in cardiovascular diseases","volume":"87 ","pages":"Pages 2-7"},"PeriodicalIF":5.6,"publicationDate":"2024-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142402518","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-01DOI: 10.1016/j.pcad.2024.10.006
Marta Belmonte , Alberto Foà , Pasquale Paolisso , Luca Bergamaschi , Emanuele Gallinoro , Alberto Polimeni , Roberto Scarsini , Saverio Muscoli , Sara Amicone , Antonio De Vita , Angelo Villano , Francesco Angeli , Matteo Armillotta , Vincenzo Sucato , Saverio Tremamunno , Doralisa Morrone , Ciro Indolfi , Pasquale Perrone Filardi , Flavio Ribichini , Gaetano Antonio Lanza , Carmine Pizzi
The prevalence of coronary microvascular dysfunction (CMD) beyond the spectrum of chronic coronary syndromes (CCS) is non-negligible, pertaining to pathophysiological and therapeutical implications. Thanks to the availability of accurate and safe non-invasive technique, CMD can be identified as a key player in heart failure, cardiomyopathies, Takotsubo syndrome, aortic stenosis. While CMD is widely recognized as a cause of myocardial ischemia leading to a worse prognosis even in the absence of obstructive coronary artery disease, the characterization of CMD patterns beyond CCS might provide valuable insights on the underlying disease progression, being potentially a “red flag” of adverse cardiac remodeling and a major determinant of response to therapy and outcomes. In this review, we aimed to provide an overview of the latest evidence on the prevalence, mechanistic and prognostic implications of CMD beyond the spectrum of CCS (i.e. heart failure, cardiomyopathies, Takotsubo syndrome, aortic stenosis).
{"title":"Coronary microvascular dysfunction beyond the spectrum of chronic coronary syndromes","authors":"Marta Belmonte , Alberto Foà , Pasquale Paolisso , Luca Bergamaschi , Emanuele Gallinoro , Alberto Polimeni , Roberto Scarsini , Saverio Muscoli , Sara Amicone , Antonio De Vita , Angelo Villano , Francesco Angeli , Matteo Armillotta , Vincenzo Sucato , Saverio Tremamunno , Doralisa Morrone , Ciro Indolfi , Pasquale Perrone Filardi , Flavio Ribichini , Gaetano Antonio Lanza , Carmine Pizzi","doi":"10.1016/j.pcad.2024.10.006","DOIUrl":"10.1016/j.pcad.2024.10.006","url":null,"abstract":"<div><div>The prevalence of coronary microvascular dysfunction (CMD) beyond the spectrum of chronic coronary syndromes (CCS) is non-negligible, pertaining to pathophysiological and therapeutical implications. Thanks to the availability of accurate and safe non-invasive technique, CMD can be identified as a key player in heart failure, cardiomyopathies, Takotsubo syndrome, aortic stenosis. While CMD is widely recognized as a cause of myocardial ischemia leading to a worse prognosis even in the absence of obstructive coronary artery disease, the characterization of CMD patterns beyond CCS might provide valuable insights on the underlying disease progression, being potentially a “red flag” of adverse cardiac remodeling and a major determinant of response to therapy and outcomes. In this review, we aimed to provide an overview of the latest evidence on the prevalence, mechanistic and prognostic implications of CMD beyond the spectrum of CCS (i.e. heart failure, cardiomyopathies, Takotsubo syndrome, aortic stenosis).</div></div>","PeriodicalId":21156,"journal":{"name":"Progress in cardiovascular diseases","volume":"87 ","pages":"Pages 73-82"},"PeriodicalIF":5.6,"publicationDate":"2024-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142515600","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The American Heart Association recently defined the complex interactions among the cardiovascular, renal, and metabolic systems as CKM syndrome. To promote better patient outcomes, having a more profound understanding of CKM pathophysiology and pursuing holistic preventative and therapy strategies is critical. Despite many gaps in understanding CKM syndrome, this study attempts to elucidate two of these gaps: the new emerging biomarkers for screening and the role of inflammation in its pathophysiology. For this review, an extensive search for specific terms was conducted in the following databases: PubMed, Scopus, Web of Science, and Google Scholar. Studies were first assessed by title, abstract, keywords, and selected for portfolio according to eligibility criteria, which led to 38 studies. They provided background information about CKM syndrome; data suggested that serum uric acid, leptin, aldosterone, bilirubin, soluble neprilysin, lipocalin-type-prostaglandin-D-synthase, and endocan could be valuable biomarkers for CKM screening; and finally, the inflammation role in CKM.
美国心脏协会最近将心血管、肾脏和代谢系统之间复杂的相互作用定义为 CKM 综合征。为了改善患者的预后，更深入地了解 CKM 病理生理学并采取全面的预防和治疗策略至关重要。尽管对 CKM 综合征的认识还存在许多空白，但本研究试图阐明其中的两个空白：用于筛查的新兴生物标志物以及炎症在其病理生理学中的作用。为了撰写这篇综述，我们在以下数据库中对特定术语进行了广泛搜索：PubMed、Scopus、Web of Science 和 Google Scholar。首先根据标题、摘要和关键词对研究进行评估，然后根据资格标准筛选出 38 篇研究。这些研究提供了有关 CKM 综合征的背景信息；数据表明，血清尿酸、瘦素、醛固酮、胆红素、可溶性肾酶、脂钙素型-前列腺素-D-合成酶和内切酶可能是筛查 CKM 的有价值的生物标志物；最后，研究了炎症在 CKM 中的作用。
{"title":"Cardiovascular-kidney-metabolic syndrome – An integrative review","authors":"Katiana Simões Kittelson , Arquimedes Gasparotto Junior , Natasha Fillmore , Roberto da Silva Gomes","doi":"10.1016/j.pcad.2024.10.012","DOIUrl":"10.1016/j.pcad.2024.10.012","url":null,"abstract":"<div><div>The American Heart Association recently defined the complex interactions among the cardiovascular, renal, and metabolic systems as CKM syndrome. To promote better patient outcomes, having a more profound understanding of CKM pathophysiology and pursuing holistic preventative and therapy strategies is critical. Despite many gaps in understanding CKM syndrome, this study attempts to elucidate two of these gaps: the new emerging biomarkers for screening and the role of inflammation in its pathophysiology. For this review, an extensive search for specific terms was conducted in the following databases: PubMed, Scopus, Web of Science, and Google Scholar. Studies were first assessed by title, abstract, keywords, and selected for portfolio according to eligibility criteria, which led to 38 studies. They provided background information about CKM syndrome; data suggested that serum uric acid, leptin, aldosterone, bilirubin, soluble neprilysin, lipocalin-type-prostaglandin-D-synthase, and endocan could be valuable biomarkers for CKM screening; and finally, the inflammation role in CKM.</div></div>","PeriodicalId":21156,"journal":{"name":"Progress in cardiovascular diseases","volume":"87 ","pages":"Pages 26-36"},"PeriodicalIF":5.6,"publicationDate":"2024-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142564836","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-01DOI: 10.1016/j.pcad.2024.10.003
Andy Smith , Ross Arena , Simon L. Bacon , Mark A. Faghy , Giovanni Grazzi , Andrea Raisi , Amber L. Vermeesch , Martin Ong'wen , Dejana Popovic , Nicolaas P. Pronk
The purpose of this perspective is to provide recommendations on the use of Artificial Intelligence (AI) in health promotion. To arrive at these recommendations, we followed a 6-step process. The first step was to recruit an international authorship team from the Healthy Living for Pandemic Event Protection (HL- PIVOT) network. This enabled us to achieve an international perspective with insights from Canada, Great Britain, Kenya, Italy, and the US. A philosophical inquiry was conducted addressing 5 questions. What should the relationship be between humans and AI in health promotion? How can the public and professionals trust AI? How can we ensure AI is aligned with our values? How can we ensure the ethical use of data by AI? How can we control AI? 4 hypothetical scenarios were also developed to provide perspectives on: i) Artificial ‘Versus’ Human Intelligence; ii) AI Empowerment in Self-Care; iii) Could AI Improve Patient Provider Relationship; and iii) The Kenyan Cancer Patient at the Height of a Pandemic. Based on the philosophical inquiry and the scenarios 11 recommendations are made by the HL-PIVOT on the use of AI in health promotion. The golden thread running through these recommendations is a human centric approach. The recommendations begin by suggesting that workforce planning should take account of AI. They conclude with the statement that any serious incidents involving an AI in Health Promotion should be reported to the relevant regulatory authority.
{"title":"Recommendations on the use of artificial intelligence in health promotion","authors":"Andy Smith , Ross Arena , Simon L. Bacon , Mark A. Faghy , Giovanni Grazzi , Andrea Raisi , Amber L. Vermeesch , Martin Ong'wen , Dejana Popovic , Nicolaas P. Pronk","doi":"10.1016/j.pcad.2024.10.003","DOIUrl":"10.1016/j.pcad.2024.10.003","url":null,"abstract":"<div><div>The purpose of this perspective is to provide recommendations on the use of Artificial Intelligence (AI) in health promotion. To arrive at these recommendations, we followed a 6-step process. The first step was to recruit an international authorship team from the Healthy Living for Pandemic Event Protection (HL- PIVOT) network. This enabled us to achieve an international perspective with insights from Canada, Great Britain, Kenya, Italy, and the US. A philosophical inquiry was conducted addressing 5 questions. What should the relationship be between humans and AI in health promotion? How can the public and professionals trust AI? How can we ensure AI is aligned with our values? How can we ensure the ethical use of data by AI? How can we control AI? 4 hypothetical scenarios were also developed to provide perspectives on: i) Artificial ‘Versus’ Human Intelligence; ii) AI Empowerment in Self-Care; iii) Could AI Improve Patient Provider Relationship; and iii) The Kenyan Cancer Patient at the Height of a Pandemic. Based on the philosophical inquiry and the scenarios 11 recommendations are made by the HL-PIVOT on the use of AI in health promotion. The golden thread running through these recommendations is a human centric approach. The recommendations begin by suggesting that workforce planning should take account of AI. They conclude with the statement that any serious incidents involving an AI in Health Promotion should be reported to the relevant regulatory authority.</div></div>","PeriodicalId":21156,"journal":{"name":"Progress in cardiovascular diseases","volume":"87 ","pages":"Pages 37-43"},"PeriodicalIF":5.6,"publicationDate":"2024-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142402519","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-01DOI: 10.1016/j.pcad.2024.10.010
T. Elizabeth Workman , Ali Ahmed , Helen M. Sheriff , Venkatesh K. Raman , Sijian Zhang , Yijun Shao , Charles Faselis , Gregg C. Fonarow , Qing Zeng-Treitler
Background
Natural language processing (NLP) can facilitate research utilizing data from electronic health records (EHRs). Large language models can potentially improve NLP applications leveraging EHR notes. The objective of this study was to assess the performance of zero-shot learning using Chat Generative Pre-trained Transformer 4 (ChatGPT-4) for extraction of symptoms and signs, and compare its performance to baseline machine learning and rule-based methods developed using annotated data.
Methods and results
From unstructured clinical notes of the national EHR data of the Veterans healthcare system, we extracted 1999 text snippets containing relevant keywords for heart failure symptoms and signs, which were then annotated by two clinicians. We also created 102 synthetic snippets that were semantically similar to snippets randomly selected from the original 1999 snippets. The authors applied zero-shot learning, using two different forms of prompt engineering in a symptom and sign extraction task with ChatGPT-4, utilizing the synthetic snippets. For comparison, baseline models using machine learning and rule-based methods were trained using the original 1999 annotated text snippets, and then used to classify the 102 synthetic snippets.
The best zero-shot learning application achieved 90.6 % precision, 100 % recall, and 95 % F1 score, outperforming the best baseline method, which achieved 54.9 % precision, 82.4 % recall, and 65.5 % F1 score. Prompt style and temperature settings influenced zero-shot learning performance.
Conclusions
Zero-shot learning utilizing ChatGPT-4 significantly outperformed traditional machine learning and rule-based NLP. Prompt type and temperature settings affected zero-shot learning performance. These findings suggest a more efficient means of symptoms and signs extraction than traditional machine learning and rule-based methods.
{"title":"ChatGPT-4 extraction of heart failure symptoms and signs from electronic health records","authors":"T. Elizabeth Workman , Ali Ahmed , Helen M. Sheriff , Venkatesh K. Raman , Sijian Zhang , Yijun Shao , Charles Faselis , Gregg C. Fonarow , Qing Zeng-Treitler","doi":"10.1016/j.pcad.2024.10.010","DOIUrl":"10.1016/j.pcad.2024.10.010","url":null,"abstract":"<div><h3>Background</h3><div>Natural language processing (NLP) can facilitate research utilizing data from electronic health records (EHRs). Large language models can potentially improve NLP applications leveraging EHR notes. The objective of this study was to assess the performance of zero-shot learning using Chat Generative Pre-trained Transformer 4 (ChatGPT-4) for extraction of symptoms and signs, and compare its performance to baseline machine learning and rule-based methods developed using annotated data.</div></div><div><h3>Methods and results</h3><div>From unstructured clinical notes of the national EHR data of the Veterans healthcare system, we extracted 1999 text snippets containing relevant keywords for heart failure symptoms and signs, which were then annotated by two clinicians. We also created 102 synthetic snippets that were semantically similar to snippets randomly selected from the original 1999 snippets. The authors applied zero-shot learning, using two different forms of prompt engineering in a symptom and sign extraction task with ChatGPT-4, utilizing the synthetic snippets. For comparison, baseline models using machine learning and rule-based methods were trained using the original 1999 annotated text snippets, and then used to classify the 102 synthetic snippets.</div><div>The best zero-shot learning application achieved 90.6 % precision, 100 % recall, and 95 % F1 score, outperforming the best baseline method, which achieved 54.9 % precision, 82.4 % recall, and 65.5 % F1 score. Prompt style and temperature settings influenced zero-shot learning performance.</div></div><div><h3>Conclusions</h3><div>Zero-shot learning utilizing ChatGPT-4 significantly outperformed traditional machine learning and rule-based NLP. Prompt type and temperature settings affected zero-shot learning performance. These findings suggest a more efficient means of symptoms and signs extraction than traditional machine learning and rule-based methods.</div></div>","PeriodicalId":21156,"journal":{"name":"Progress in cardiovascular diseases","volume":"87 ","pages":"Pages 44-49"},"PeriodicalIF":5.6,"publicationDate":"2024-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142515599","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-01DOI: 10.1016/j.pcad.2024.10.014
Ahmed Abdelaziz , Karim Atta , Ahmed Farid Gadelmawla , Mohamed Abdelaziz , Muhammad Desouky , Yasmin Negida , Ahmed A. Ibrahim , Dua Abdelraouf Eldosoky , Ahmed Helmi , Shrouk Ramadan , Emad Singer , Jose Tafur-Soto
{"title":"Safety outcomes of oral anticoagulants in patients with an indication of anti-coagulants after TAVR: A meta-analysis","authors":"Ahmed Abdelaziz , Karim Atta , Ahmed Farid Gadelmawla , Mohamed Abdelaziz , Muhammad Desouky , Yasmin Negida , Ahmed A. Ibrahim , Dua Abdelraouf Eldosoky , Ahmed Helmi , Shrouk Ramadan , Emad Singer , Jose Tafur-Soto","doi":"10.1016/j.pcad.2024.10.014","DOIUrl":"10.1016/j.pcad.2024.10.014","url":null,"abstract":"","PeriodicalId":21156,"journal":{"name":"Progress in cardiovascular diseases","volume":"87 ","pages":"Pages 97-99"},"PeriodicalIF":5.6,"publicationDate":"2024-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142564933","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号