Clinical eHealth最新文献_第4页

Enhancing automatic early arteriosclerosis prediction: an explainable machine learning evidence 增强自动早期动脉硬化预测：一个可解释的机器学习证据

Clinical eHealth

Pub Date : 2024-12-01 DOI: 10.1016/j.ceh.2024.12.003

Eka Miranda , Suko Adiarto

Objective

This paper proposed a machine learning (ML) model to early predict patients with arteriosclerotic heart disease (AHD). We also used model-agnostic ML approaches to find and analyze informative aspects in the prediction model outcomes.

Methods

We employed an Electronic Health Record (EHR) for hematology that contained data on erythrocytes, hematocrit, hemoglobin, mean corpuscular hemoglobin, mean corpuscular hemoglobin concentration, leukocytes, thrombocytes, age, and sex. Our investigation included Decision Tree (DT), Random Forest (RF), Logistic Regression (LR), Bagging Decision Tree (BDT), and Bagging Logistic Regression (BLR) for ML-based AHD detection. To handle imbalanced data and increase classifier accuracy, we used bagging and the Synthetic Minority Oversampling Technique (SMOTE). Following that, we used the Shapley Additive exPlanations (SHAP) framework to explain the ML model and quantify the feature contribution to predictions.

Results

SMOTE-balanced data with RF outperformed on practically all performance measures, including accuracy, precision, recall, f1-score, and ROCAUC, by 82.12 %, 81.31 %, 83.37 %, 82.57 %, and 89 %, respectively. According to the SHAP summary bar plot method for global feature importance, hemoglobin was the most important attribute for detecting and predicting AHD patients. Then, local interpretability in the form of a force plot illustrated the consequences of a single observation’s prediction as well as the magnitude of the SHAP value for each feature. Our findings demonstrated that hemoglobin, erythrocytes, hematocrit, hermch, khermchc, leukocytes, thrombocytes, and age all contributed positively to the prediction of class 1 (AHD patients), however gender had a negative impact on the prediction on a case-by-case basis. For class 0 (patients with no AHD), thrombocytes, hematocrit, and gender contributed positively, but leukocytes, erythrocytes, hemoglobin, and khermchc contributed adversely.

Conclusion

Explainable ML paved the way for early AHD prediction since it examined black-box ML models to determine how each feature contributed to the final prediction.

目的建立动脉硬化性心脏病（AHD）早期预测的机器学习（ML）模型。我们还使用与模型无关的ML方法来查找和分析预测模型结果中的信息方面。方法采用血液学电子健康记录（EHR），包括红细胞、红细胞比容、血红蛋白、平均红细胞血红蛋白、平均红细胞血红蛋白浓度、白细胞、血小板、年龄和性别等数据。我们的研究包括决策树（DT）、随机森林（RF）、逻辑回归（LR）、Bagging决策树（BDT）和Bagging Logistic回归（BLR）用于基于ml的AHD检测。为了处理不平衡数据并提高分类器的准确性，我们使用了装袋和合成少数过采样技术（SMOTE）。接下来，我们使用Shapley加性解释（SHAP）框架来解释ML模型，并量化特征对预测的贡献。结果使用RF的smot -balanced数据在准确率、精密度、召回率、f1-score和ROCAUC等几乎所有性能指标上分别高出82.12%、81.31%、83.37%、82.57%和89%。根据SHAP总体特征重要性汇总条形图方法，血红蛋白是检测和预测AHD患者最重要的属性。然后，以力图形式的局部可解释性说明了单个观测预测的结果以及每个特征的SHAP值的大小。我们的研究结果表明，血红蛋白、红细胞、红细胞压积、hermch、khermchc、白细胞、血小板和年龄都对1级（AHD患者）的预测有积极的影响，而性别对个案预测有负面影响。对于0级（无AHD患者），血小板、红细胞压积和性别有积极作用，但白细胞、红细胞、血红蛋白和血红蛋白有不利作用。可解释ML为早期AHD预测铺平了道路，因为它检查了黑箱ML模型，以确定每个特征如何对最终预测做出贡献。

{"title":"Enhancing automatic early arteriosclerosis prediction: an explainable machine learning evidence","authors":"Eka Miranda , Suko Adiarto","doi":"10.1016/j.ceh.2024.12.003","DOIUrl":"10.1016/j.ceh.2024.12.003","url":null,"abstract":"<div><h3>Objective</h3><div>This paper proposed a machine learning (ML) model to early predict patients with arteriosclerotic heart disease (AHD). We also used model-agnostic ML approaches to find and analyze informative aspects in the prediction model outcomes.</div></div><div><h3>Methods</h3><div>We employed an Electronic Health Record (EHR) for hematology that contained data on erythrocytes, hematocrit, hemoglobin, mean corpuscular hemoglobin, mean corpuscular hemoglobin concentration, leukocytes, thrombocytes, age, and sex. Our investigation included Decision Tree (DT), Random Forest (RF), Logistic Regression (LR), Bagging Decision Tree (BDT), and Bagging Logistic Regression (BLR) for ML-based AHD detection. To handle imbalanced data and increase classifier accuracy, we used bagging and the Synthetic Minority Oversampling Technique (SMOTE). Following that, we used the Shapley Additive exPlanations (SHAP) framework to explain the ML model and quantify the feature contribution to predictions.</div></div><div><h3>Results</h3><div>SMOTE-balanced data with RF outperformed on practically all performance measures, including accuracy, precision, recall, f1-score, and ROCAUC, by 82.12 %, 81.31 %, 83.37 %, 82.57 %, and 89 %, respectively. According to the SHAP summary bar plot method for global feature importance, hemoglobin was the most important attribute for detecting and predicting AHD patients. Then, local interpretability in the form of a force plot illustrated the consequences of a single observation’s prediction as well as the magnitude of the SHAP value for each feature. Our findings demonstrated that hemoglobin, erythrocytes, hematocrit, hermch, khermchc, leukocytes, thrombocytes, and age all contributed positively to the prediction of class 1 (AHD patients), however gender had a negative impact on the prediction on a case-by-case basis. For class 0 (patients with no AHD), thrombocytes, hematocrit, and gender contributed positively, but leukocytes, erythrocytes, hemoglobin, and khermchc contributed adversely.</div></div><div><h3>Conclusion</h3><div>Explainable ML paved the way for early AHD prediction since it examined black-box ML models to determine how each feature contributed to the final prediction.</div></div>","PeriodicalId":100268,"journal":{"name":"Clinical eHealth","volume":"7 ","pages":"Pages 153-163"},"PeriodicalIF":0.0,"publicationDate":"2024-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143160909","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}

引用次数: 0

Leveraging deep edge intelligence for real-time respiratory disease detection 利用深度边缘智能实时检测呼吸系统疾病

Clinical eHealth

Pub Date : 2024-12-01 DOI: 10.1016/j.ceh.2025.01.001

Tahiya Tasneem Oishee, Jareen Anjom, Uzma Mohammed, Md. Ishan Arefin Hossain

Detecting respiratory diseases such as COPD, bronchiolitis, URTI, and pneumonia is crucial for early medical intervention. This study utilizes the ICBHI dataset to train and evaluate deep learning architectures such as CNN-GRU, VGGish, YAMNet, CNN-LSTM, and basic CNN to automate this process. After a detailed analysis of the performance of these models, the CNN-LSTM model achieved an impressive accuracy and F1 score of 96% each. The model is also considerably lightweight, as its weights are further pruned and then quantized using TensorFlow Lite (TFLite), with the model being optimized at a significantly small size of 0.38 MB with only a loss of about 1% in performance. Subsequently, this was deployed to the smartphone application RespiScan. The application uses the prediction capabilities of the disease detection model on patients’ audio recordings. By providing a portable, cost-effective, and efficient, lightweight solution for respiratory health monitoring, this work contributes significantly to timely disease detection. It promotes proactive health management, thereby reducing the burden on healthcare systems. This work can be further validated in real-world conditions, such as for initial preliminary auscultation purposes, to ensure the proposed work’s efficacy across different environmental settings.

检测呼吸道疾病，如慢性阻塞性肺病、细支气管炎、尿路感染和肺炎，对早期医疗干预至关重要。本研究利用ICBHI数据集来训练和评估CNN- gru、VGGish、YAMNet、CNN- lstm和basic CNN等深度学习架构，以实现这一过程的自动化。经过对这些模型性能的详细分析，CNN-LSTM模型取得了令人印象深刻的准确率和96%的F1分数。该模型也是相当轻量级的，因为它的权重被进一步修剪，然后使用TensorFlow Lite （TFLite）进行量化，模型在0.38 MB的小尺寸上进行了优化，性能只损失了大约1%。随后，它被部署到智能手机应用程序RespiScan中。该应用程序利用疾病检测模型对患者录音的预测能力。通过为呼吸系统健康监测提供便携、经济、高效、轻量级的解决方案，这项工作为及时发现疾病做出了重大贡献。它促进积极主动的卫生管理，从而减轻卫生保健系统的负担。这项工作可以在现实条件下进一步验证，例如用于初始初步听诊目的，以确保所提议的工作在不同环境设置中的有效性。

{"title":"Leveraging deep edge intelligence for real-time respiratory disease detection","authors":"Tahiya Tasneem Oishee, Jareen Anjom, Uzma Mohammed, Md. Ishan Arefin Hossain","doi":"10.1016/j.ceh.2025.01.001","DOIUrl":"10.1016/j.ceh.2025.01.001","url":null,"abstract":"<div><div>Detecting respiratory diseases such as COPD, bronchiolitis, URTI, and pneumonia is crucial for early medical intervention. This study utilizes the ICBHI dataset to train and evaluate deep learning architectures such as CNN-GRU, VGGish, YAMNet, CNN-LSTM, and basic CNN to automate this process. After a detailed analysis of the performance of these models, the CNN-LSTM model achieved an impressive accuracy and F1 score of 96% each. The model is also considerably lightweight, as its weights are further pruned and then quantized using TensorFlow Lite (TFLite), with the model being optimized at a significantly small size of 0.38 MB with only a loss of about 1% in performance. Subsequently, this was deployed to the smartphone application RespiScan. The application uses the prediction capabilities of the disease detection model on patients’ audio recordings. By providing a portable, cost-effective, and efficient, lightweight solution for respiratory health monitoring, this work contributes significantly to timely disease detection. It promotes proactive health management, thereby reducing the burden on healthcare systems. This work can be further validated in real-world conditions, such as for initial preliminary auscultation purposes, to ensure the proposed work’s efficacy across different environmental settings.</div></div>","PeriodicalId":100268,"journal":{"name":"Clinical eHealth","volume":"7 ","pages":"Pages 207-220"},"PeriodicalIF":0.0,"publicationDate":"2024-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143160899","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}

引用次数: 0

A survey on define daily dose of watch- and access-category antibiotics in two Indonesian hospitals following the implementation of digital antimicrobial stewardship tool 在实施数字抗微生物药物管理工具后，对印度尼西亚两家医院的监测类和可获得类抗生素的确定日剂量进行了调查

Clinical eHealth

Pub Date : 2024-12-01 DOI: 10.1016/j.ceh.2024.12.004

Ronald Irwanto Natadidjaja , Aziza Ariyani , Hadianti Adlani , Raymond Adianto , Iin Indah Pertiwi , Grace Nerry Legoh , Alvin Lekonardo Rantung , Hadi Sumarsono

Background

In 2023, the World Health Organization (WHO) began targeting a shift in antibiotic prescribing trends from Watch to Access category. The expected target is including 60% of antibiotic prescribing in the Access category.

Method

This survey was a preliminary study, in which our study group designed a digital model of antimicrobial stewardship and the model was known as e-RASPRO. It was an initial review on the implementation of e-RASPRO tool prior to its wider use in future hospitals. The survey on the use of antibiotic Define Daily Dose / 100 patient days (DDD) was carried out in two hospitals in Indonesia at 3 months and 9 months of use, respectively. Hospital 1 as a primary hospital, Hospital 2 as a referral hospital. Data was retrieved retrospectively at the inpatient wards of both hospitals.

Result

Three months before and after the implementation of e-RASPRO in Hospital 1, we found an increase in DDD of prophylactic antibiotic Cefazolin by 167.18 %. In hospital 2, it could not be described because Cefazolin had been used since the hospital applied the manual RASPRO concept. DDD of Watch category antibiotics within 9 months following the implementation of e-RASPRO tool in hospital 1 showed a decrease of 49.01 %. Meanwhile, the implementation of e-RASPRO for 3 months in Hospital 2 still showed an increase in Watch category antibiotics by 20.18 %; however, there was a decrease in DDD of Cephalosporin and Glycopeptide antibiotics by 7.63 % and 49.30 %, respectively. In the meantime, as a way of saving antibiotic use and shifting antibiotic prescribing to the Access category, we found a decrease in DDD of Access category antibiotics in Hospital 1 by 3.64 % and an increase in Hospital 2 by 8.14 %

Conclusion

The survey may indicate that there are savings attempts in antibiotic use as well as an early change in DDD antibiotics from the Watch category to the Access category following the implementation of e-RASPRO tool in both hospitals. The time period of using the digital devices may still affect the results; however, this survey certainly has not illustrated a strong cause-and-effect correlation between the use of e-RASPRO tool and antibiotic DDD.

2023年，世界卫生组织（世卫组织）开始瞄准抗生素处方趋势的转变，从观察到获取类别。预期目标是将60%的抗生素处方纳入可及性类别。方法本研究是一项初步研究，本研究组设计了抗菌药物管理数字模型，该模型称为e-RASPRO。这是在未来医院更广泛使用e-RASPRO工具之前对其实施情况的初步审查。在印度尼西亚的两家医院，分别在使用抗生素3个月和9个月时，对抗生素确定每日剂量/ 100病人日（DDD）的使用情况进行了调查。第一医院是初级医院，第二医院是转诊医院。回顾性检索两家医院住院病房的资料。结果1院实施e-RASPRO前后3个月预防性抗生素头孢唑林的DDD增加了167.18%。在医院2，由于医院采用手动RASPRO概念，头孢唑林一直在使用，因此无法描述。1医院实施e-RASPRO工具后9个月内Watch类抗生素DDD下降49.01%。同时，在第二医院实施e-RASPRO 3个月后，Watch类抗生素仍增加20.18%；头孢菌素类和糖肽类抗生素的DDD分别下降了7.63%和49.30%。与此同时,作为一种节约使用抗生素和抗生素处方转向访问类别,我们发现减少DDD的访问类抗生素在医院1 3.64%,增加医院2 8.14% ConclusionThe调查可能表明有储蓄的尝试在抗生素的使用以及早期改变DDD抗生素从手表类别访问类别后e-RASPRO工具的实现在两个医院。使用数字设备的时间仍可能影响结果；然而，这项调查肯定没有说明使用e-RASPRO工具和抗生素DDD之间存在很强的因果关系。

{"title":"A survey on define daily dose of watch- and access-category antibiotics in two Indonesian hospitals following the implementation of digital antimicrobial stewardship tool","authors":"Ronald Irwanto Natadidjaja , Aziza Ariyani , Hadianti Adlani , Raymond Adianto , Iin Indah Pertiwi , Grace Nerry Legoh , Alvin Lekonardo Rantung , Hadi Sumarsono","doi":"10.1016/j.ceh.2024.12.004","DOIUrl":"10.1016/j.ceh.2024.12.004","url":null,"abstract":"<div><h3>Background</h3><div>In 2023, the World Health Organization (WHO) began targeting a shift in antibiotic prescribing trends from Watch to Access category. The expected target is including 60% of antibiotic prescribing in the Access category.</div></div><div><h3>Method</h3><div>This survey was a preliminary study, in which our study group designed a digital model of antimicrobial stewardship and the model was known as e-RASPRO. It was an initial review on the implementation of e-RASPRO tool prior to its wider use in future hospitals. The survey on the use of antibiotic Define Daily Dose / 100 patient days (DDD) was carried out in two hospitals in Indonesia at 3 months and 9 months of use, respectively. Hospital 1 as a primary hospital, Hospital 2 as a referral hospital. Data was retrieved retrospectively at the inpatient wards of both hospitals.</div></div><div><h3>Result</h3><div>Three months before and after the implementation of e-RASPRO in Hospital 1, we found an increase in DDD of prophylactic antibiotic Cefazolin by 167.18 %. In hospital 2, it could not be described because Cefazolin had been used since the hospital applied the manual RASPRO concept. DDD of Watch category antibiotics within 9 months following the implementation of e-RASPRO tool in hospital 1 showed a decrease of 49.01 %. Meanwhile, the implementation of e-RASPRO for 3 months in Hospital 2 still showed an increase in Watch category antibiotics by 20.18 %; however, there was a decrease in DDD of Cephalosporin and Glycopeptide antibiotics by 7.63 % and 49.30 %, respectively. In the meantime, as a way of saving antibiotic use and shifting antibiotic prescribing to the Access category, we found a decrease in DDD of Access category antibiotics in Hospital 1 by 3.64 % and an increase in Hospital 2 by 8.14 %</div></div><div><h3>Conclusion</h3><div>The survey may indicate that there are savings attempts in antibiotic use as well as an early change in DDD antibiotics from the Watch category to the Access category following the implementation of e-RASPRO tool in both hospitals. The time period of using the digital devices may still affect the results; however, this survey certainly has not illustrated a strong cause-and-effect correlation between the use of e-RASPRO tool and antibiotic DDD.</div></div>","PeriodicalId":100268,"journal":{"name":"Clinical eHealth","volume":"7 ","pages":"Pages 176-189"},"PeriodicalIF":0.0,"publicationDate":"2024-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143159891","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}

引用次数: 0

“AI et al.” The perils of overreliance on Artificial Intelligence by authors in scientific research "人工智能等"科学研究中作者过度依赖人工智能的危害

Clinical eHealth

Pub Date : 2024-09-17 DOI: 10.1016/j.ceh.2024.09.001

Juan S. Izquierdo-Condoy, Jorge Vásconez-González, Esteban Ortiz-Prado

The rapid integration of Artificial Intelligence (AI) into scientific research and publication processes marks a significant shift in knowledge generation. This transition from traditional literature searches to AI-driven algorithms has accelerated tasks such as writing, editing, and summarizing scientific manuscripts. While AI holds promise for improving efficiency and accuracy, concerns have arisen about its potential misuse and the erosion of scientific integrity.

人工智能（AI）与科学研究和出版流程的快速融合标志着知识生成的重大转变。从传统的文献检索到人工智能驱动算法的转变，加速了科学手稿的撰写、编辑和总结等任务。虽然人工智能在提高效率和准确性方面大有可为，但其潜在的滥用和对科学诚信的侵蚀也引起了人们的担忧。

引用次数: 0

A systematic review of eHealth and mHealth interventions for lymphedema patients 针对淋巴水肿患者的电子健康和移动健康干预措施的系统性审查

Clinical eHealth

Pub Date : 2024-08-22 DOI: 10.1016/j.ceh.2024.08.002

Andrea Mangion, Bruno Ivasic, Neil Piller

Lymphedema is a chronic inflammatory disease that causes chronic swelling in the affected area, necessitating daily treatment. Millions of people worldwide are affected. The investigation of strategies to improve the overall health of patients, such as through the utilisation of electronic health (eHealth), is justified considering the ongoing burden of daily self-care. This research aimed to (a) identify current published research in eHealth and mobile health (mHealth) interventions for patients living with lymphedema; (b) assess feasibility and efficacy of the interventions; and (c) understand whether intervention adherence was affected by using eHealth. A systematic review was undertaken. Seven databases including MEDLINE, Scopus, Web of Science, CINAHL, the Cochrane Library, PsycINFO and IEEE Xplore were searched. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses were used. 1857 studies were identified through the database search with 9 meeting the inclusion criteria for a total of 1031 participants. There were 3 types of eHealth, including instructive online content, telehealth, and digital gaming. The efficacy of various eHealth and mHealth modalities was demonstrated in areas such as lymphedema outcomes, self-care, psychosocial outcomes, and disease comprehension. Reports of feasibility demonstrated that eHealth modalities were generally well accepted or preferred over conventional methods. 7 studies reported or discussed adherence and provided insight into the relationship between the design of the eHealth tool and the completion of the intervention. Several distinct categories of eHealth and mHealth interventions were shown to improve disease comprehension, psychosocial and lymphedema outcomes. Findings from this systematic review may have an impact on the design of future studies in this domain, including consideration of early user acceptance testing when developing eHealth tools. With the ongoing progress in eHealth technology, further investigation into eHealth is warranted given the encouraging results observed in a limited number of studies.

淋巴水肿是一种慢性炎症性疾病，会导致患处长期肿胀，需要每天进行治疗。全世界有数百万人受到影响。考虑到日常自我护理的持续负担，有必要研究改善患者整体健康的策略，如通过利用电子健康（eHealth）。本研究旨在：（a）确定目前已发表的针对淋巴水肿患者的电子健康和移动健康（mHealth）干预研究；（b）评估干预的可行性和有效性；以及（c）了解使用电子健康是否会影响干预的坚持性。我们开展了一项系统性研究。检索了七个数据库，包括 MEDLINE、Scopus、Web of Science、CINAHL、Cochrane Library、PsycINFO 和 IEEE Xplore。采用了《系统综述和元分析首选报告项目》（Preferred Reporting Items for Systematic Reviews and Meta-Analyses）。通过数据库搜索确定了 1857 项研究，其中 9 项符合纳入标准，共有 1031 名参与者。电子健康有三种类型，包括指导性在线内容、远程保健和数字游戏。各种电子健康和移动健康模式在淋巴水肿疗效、自我护理、社会心理疗效和疾病理解等方面的疗效得到了证实。有关可行性的报告显示，电子健康模式普遍被广泛接受，或比传统方法更受青睐。有 7 项研究报告或讨论了坚持使用的问题，并深入探讨了电子健康工具的设计与完成干预之间的关系。研究表明，几类不同的电子健康和移动健康干预措施可改善疾病理解、社会心理和淋巴水肿的治疗效果。本系统综述的研究结果可能会对该领域未来研究的设计产生影响，包括在开发电子健康工具时考虑早期用户接受度测试。随着电子健康技术的不断进步，鉴于在有限的几项研究中观察到的令人鼓舞的结果，有必要对电子健康进行进一步的调查。

{"title":"A systematic review of eHealth and mHealth interventions for lymphedema patients","authors":"Andrea Mangion, Bruno Ivasic, Neil Piller","doi":"10.1016/j.ceh.2024.08.002","DOIUrl":"10.1016/j.ceh.2024.08.002","url":null,"abstract":"<div><p>Lymphedema is a chronic inflammatory disease that causes chronic swelling in the affected area, necessitating daily treatment. Millions of people worldwide are affected. The investigation of strategies to improve the overall health of patients, such as through the utilisation of electronic health (eHealth), is justified considering the ongoing burden of daily self-care. This research aimed to (a) identify current published research in eHealth and mobile health (mHealth) interventions for patients living with lymphedema; (b) assess feasibility and efficacy of the interventions; and (c) understand whether intervention adherence was affected by using eHealth. A systematic review was undertaken. Seven databases including MEDLINE, Scopus, Web of Science, CINAHL, the Cochrane Library, PsycINFO and IEEE Xplore were searched. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses were used. 1857 studies were identified through the database search with 9 meeting the inclusion criteria for a total of 1031 participants. There were 3 types of eHealth, including instructive online content, telehealth, and digital gaming. The efficacy of various eHealth and mHealth modalities was demonstrated in areas such as lymphedema outcomes, self-care, psychosocial outcomes, and disease comprehension. Reports of feasibility demonstrated that eHealth modalities were generally well accepted or preferred over conventional methods. 7 studies reported or discussed adherence and provided insight into the relationship between the design of the eHealth tool and the completion of the intervention. Several distinct categories of eHealth and mHealth interventions were shown to improve disease comprehension, psychosocial and lymphedema outcomes. Findings from this systematic review may have an impact on the design of future studies in this domain, including consideration of early user acceptance testing when developing eHealth tools. With the ongoing progress in eHealth technology, further investigation into eHealth is warranted given the encouraging results observed in a limited number of studies.</p></div>","PeriodicalId":100268,"journal":{"name":"Clinical eHealth","volume":"7 ","pages":"Pages 120-132"},"PeriodicalIF":0.0,"publicationDate":"2024-08-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.sciencedirect.com/science/article/pii/S2588914124000108/pdfft?md5=c54858a4e940ccc380deee2df6ac4e8f&pid=1-s2.0-S2588914124000108-main.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142084002","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}

引用次数: 0

Machine learning and transfer learning techniques for accurate brain tumor classification 用于脑肿瘤精确分类的机器学习和迁移学习技术

Clinical eHealth

Pub Date : 2024-08-08 DOI: 10.1016/j.ceh.2024.08.001

Seyed Matin Malakouti, Mohammad Bagher Menhaj, Amir Abolfazl Suratgar

Brain tumors, resulting from uncontrolled and rapid cell growth, pose significant health risks if not treated early. Despite numerous advancements, accurate segmentation and classification remain challenging. This study leverages machine learning (ML) and transfer learning techniques to classify healthy and sick individuals using numerical data and MRI images. We utilized 3762 MRI images alongside Light Gradient Boosting Machine (LightGBM), AdaBoost, gradient boosting, Random Forest, Quadratic Discriminant Analysis, Linear Discriminant Analysis, logistic regression, and transfer learning algorithms. Numerical data was processed with LightGBM, achieving an accuracy of 95.7 %. Transfer learning applied to image data using a modified GoogLeNet model further enhanced classification accuracy to 99.3 %. These results demonstrate the effectiveness of combining ML and transfer learning techniques for accurate brain tumor classification, addressing limitations of prior approaches and offering improved diagnostic reliability. All coding and model implementations were conducted on the Python platform.

脑肿瘤是由不受控制的快速细胞生长引起的，如果不及早治疗，会对健康造成严重威胁。尽管取得了许多进展，但准确的分割和分类仍具有挑战性。本研究利用机器学习（ML）和迁移学习技术，通过数字数据和核磁共振成像图像对健康人和病人进行分类。我们使用了 3762 幅核磁共振图像以及光梯度提升机 (LightGBM)、AdaBoost、梯度提升、随机森林、二次判别分析、线性判别分析、逻辑回归和迁移学习算法。使用 LightGBM 处理了数值数据，准确率达到 95.7%。使用改进的 GoogLeNet 模型对图像数据进行迁移学习，进一步将分类准确率提高到 99.3%。这些结果表明，结合 ML 和迁移学习技术进行准确的脑肿瘤分类非常有效，既解决了以往方法的局限性，又提高了诊断的可靠性。所有编码和模型实现均在 Python 平台上进行。

{"title":"Machine learning and transfer learning techniques for accurate brain tumor classification","authors":"Seyed Matin Malakouti, Mohammad Bagher Menhaj, Amir Abolfazl Suratgar","doi":"10.1016/j.ceh.2024.08.001","DOIUrl":"10.1016/j.ceh.2024.08.001","url":null,"abstract":"<div><p>Brain tumors, resulting from uncontrolled and rapid cell growth, pose significant health risks if not treated early. Despite numerous advancements, accurate segmentation and classification remain challenging. This study leverages machine learning (ML) and transfer learning techniques to classify healthy and sick individuals using numerical data and MRI images. We utilized 3762 MRI images alongside Light Gradient Boosting Machine (LightGBM), AdaBoost, gradient boosting, Random Forest, Quadratic Discriminant Analysis, Linear Discriminant Analysis, logistic regression, and transfer learning algorithms. Numerical data was processed with LightGBM, achieving an accuracy of 95.7 %. Transfer learning applied to image data using a modified GoogLeNet model further enhanced classification accuracy to 99.3 %. These results demonstrate the effectiveness of combining ML and transfer learning techniques for accurate brain tumor classification, addressing limitations of prior approaches and offering improved diagnostic reliability. All coding and model implementations were conducted on the Python platform.</p></div>","PeriodicalId":100268,"journal":{"name":"Clinical eHealth","volume":"7 ","pages":"Pages 106-119"},"PeriodicalIF":0.0,"publicationDate":"2024-08-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.sciencedirect.com/science/article/pii/S2588914124000091/pdfft?md5=802f4d01ec0d0e73328b9da3cc172a18&pid=1-s2.0-S2588914124000091-main.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141963788","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}

引用次数: 0

Internet of Things in healthcare: An adaptive ethical framework for IoT in digital health 医疗保健领域的物联网：数字医疗物联网的适应性伦理框架

Clinical eHealth

Pub Date : 2024-07-14 DOI: 10.1016/j.ceh.2024.07.001

Abubakar Wakili, Sara Bakkali

The emergence of the Internet of Things (IoT) has sparked a profound transformation in the field of digital health, leading to the rise of the Internet of Medical Things (IoMT). These IoT applications, while promising significant enhancements in patient care and health outcomes, simultaneously present a myriad of ethical dilemmas. This paper aims to address these ethical challenges by introducing the Adaptive Ethical Framework for IoT in Digital Health (AEFIDH), a comprehensive evaluation framework designed to examine the ethical implications of IoT technologies within digital health contexts. The AEFIDH is developed using a mixed-methods approach, encompassing expert consultations, surveys, and interviews. This approach was employed to validate and refine the AEFIDH, ensuring it encapsulates critical ethical dimensions, including data privacy, informed consent, user autonomy, algorithmic fairness, regulatory compliance, ethical design, and equitable access to healthcare services. The research reveals pressing issues related to data privacy, security, and user autonomy and highlights the imperative need for an increased focus on algorithmic transparency and the integration of ethical considerations in the design and development of IoT applications. Despite certain limitations, the AEFIDH provides a promising roadmap for guiding the responsible development, deployment, and utilization of IoT technologies in digital health, ensuring its relevance amidst the rapidly evolving digital health landscape. This paper contributes a novel, dynamic framework that encapsulates current ethical considerations and is designed to adapt to future technological evolutions, thereby fostering ethical resilience in the face of ongoing digital health innovation. The framework’s inherent adaptability allows it to evolve in tandem with technological advancements, positioning it as an invaluable tool for stakeholders navigating the ethical terrain of IoT in healthcare.

物联网（IoT）的出现引发了数字医疗领域的深刻变革，导致了医疗物联网（IoMT）的兴起。这些物联网应用在有望显著改善患者护理和医疗效果的同时，也带来了无数伦理难题。本文旨在通过介绍数字医疗物联网自适应伦理框架（AEFIDH）来应对这些伦理挑战，AEFIDH 是一个综合评估框架，旨在研究数字医疗背景下物联网技术的伦理影响。AEFIDH 采用混合方法开发，包括专家咨询、调查和访谈。采用这种方法对 AEFIDH 进行了验证和完善，确保其囊括了关键的伦理维度，包括数据隐私、知情同意、用户自主权、算法公平性、监管合规性、伦理设计以及公平获取医疗保健服务。研究揭示了与数据隐私、安全和用户自主权相关的紧迫问题，并强调了在物联网应用的设计和开发过程中提高算法透明度和整合伦理考虑因素的迫切需要。尽管存在某些局限性，但 AEFIDH 为指导数字健康领域物联网技术的负责任开发、部署和利用提供了一个前景光明的路线图，确保了其在快速发展的数字健康领域中的相关性。本文提出了一个新颖、动态的框架，该框架囊括了当前的伦理考虑因素，旨在适应未来的技术演进，从而在持续的数字健康创新中培养伦理韧性。该框架固有的适应性使其能够与技术进步同步发展，从而成为利益相关者在医疗保健物联网伦理领域导航的宝贵工具。

{"title":"Internet of Things in healthcare: An adaptive ethical framework for IoT in digital health","authors":"Abubakar Wakili, Sara Bakkali","doi":"10.1016/j.ceh.2024.07.001","DOIUrl":"10.1016/j.ceh.2024.07.001","url":null,"abstract":"<div><p>The emergence of the Internet of Things (IoT) has sparked a profound transformation in the field of digital health, leading to the rise of the Internet of Medical Things (IoMT). These IoT applications, while promising significant enhancements in patient care and health outcomes, simultaneously present a myriad of ethical dilemmas. This paper aims to address these ethical challenges by introducing the Adaptive Ethical Framework for IoT in Digital Health (AEFIDH), a comprehensive evaluation framework designed to examine the ethical implications of IoT technologies within digital health contexts. The AEFIDH is developed using a mixed-methods approach, encompassing expert consultations, surveys, and interviews. This approach was employed to validate and refine the AEFIDH, ensuring it encapsulates critical ethical dimensions, including data privacy, informed consent, user autonomy, algorithmic fairness, regulatory compliance, ethical design, and equitable access to healthcare services. The research reveals pressing issues related to data privacy, security, and user autonomy and highlights the imperative need for an increased focus on algorithmic transparency and the integration of ethical considerations in the design and development of IoT applications. Despite certain limitations, the AEFIDH provides a promising roadmap for guiding the responsible development, deployment, and utilization of IoT technologies in digital health, ensuring its relevance amidst the rapidly evolving digital health landscape. This paper contributes a novel, dynamic framework that encapsulates current ethical considerations and is designed to adapt to future technological evolutions, thereby fostering ethical resilience in the face of ongoing digital health innovation. The framework’s inherent adaptability allows it to evolve in tandem with technological advancements, positioning it as an invaluable tool for stakeholders navigating the ethical terrain of IoT in healthcare.</p></div>","PeriodicalId":100268,"journal":{"name":"Clinical eHealth","volume":"7 ","pages":"Pages 92-105"},"PeriodicalIF":0.0,"publicationDate":"2024-07-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.sciencedirect.com/science/article/pii/S258891412400008X/pdfft?md5=218e1f17d86b332d375ee1229471e106&pid=1-s2.0-S258891412400008X-main.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141702840","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}

引用次数: 0

IoMT Tsukamoto Type-2 fuzzy expert system for tuberculosis and Alzheimer’s disease 治疗结核病和阿尔茨海默病的 IoMT Tsukamoto Type-2 模糊专家系统

Clinical eHealth

Pub Date : 2024-05-16 DOI: 10.1016/j.ceh.2024.05.002

M.K. Sharma , Nitesh Dhiman , Ajendra Sharma , Tarun Kumar

Accurate disease monitoring is an extremely time-consuming task for medical experts and technocrats involved, requiring technical support for diagnostic systems. To overcome this situation, we developed an Internet of Medical Things (IoMT) based on Tsukamoto Type 2 Fuzzy Inference System (TT2FIS) that can easily handle diagnostic and predictive aspects in the medical field. In the proposed system, we developed a Tsukamoto type 2 fuzzy inference system that takes the patient’s symptoms as input factors and the medical device as the output factor of the result. The aim of this work is to demonstrate the usefulness of type 2 fuzzy sets in Tuberculosis and Alzheimer’s disease diagnostic system. Numerical calculations are also performed to illustrate the applicability of the proposed method. A validation of the proposed derivation of the proposed IoMT model is also discussed in the results and conclusions section.

对于医学专家和相关技术人员来说，精确的疾病监测是一项极其耗时的任务，需要诊断系统的技术支持。为了克服这种情况，我们开发了一种基于塚本 2 型模糊推理系统（TT2FIS）的医疗物联网（IoMT），可以轻松处理医疗领域的诊断和预测问题。在提议的系统中，我们开发了一个塚本 2 型模糊推理系统，该系统将病人的症状作为输入因素，将医疗设备作为结果的输出因素。这项工作的目的是证明 2 型模糊集在结核病和阿尔茨海默病诊断系统中的实用性。同时还进行了数值计算，以说明所提方法的适用性。结果和结论部分还讨论了对拟议 IoMT 模型推导的验证。

引用次数: 0

Febrile disease modeling and diagnosis system for optimizing medical decisions in resource-scarce settings 发热性疾病建模和诊断系统，用于在资源匮乏的环境中优化医疗决策

Clinical eHealth

Pub Date : 2024-05-10 DOI: 10.1016/j.ceh.2024.05.001

Daniel Asuquo , Kingsley Attai , Okure Obot , Moses Ekpenyong , Christie Akwaowo , Kiirya Arnold , Faith-Michael Uzoka

Febrile diseases are highly prevalent in tropical regions due to elevated humidity and high temperatures. These regions, mainly comprising low- and middle-income countries, often face challenges related to inadequate medical infrastructure and a lack of skilled personnel for accurately diagnosing febrile diseases. Distinguishing one febrile illness from another posed a significant challenge, adding to the complexity of accurate diagnoses. This study developed a multi-symptom multi-disease model to address this challenge, leveraging exploratory data analysis of patient datasets from field studies and the expertise of medical practitioners specializing in tropical diseases. The research investigated the most effective modeling approach for differentiating among 11 febrile illnesses that are prevalent in Nigeria using three intelligent techniques: Extreme Gradient Boost (XGBoost), Fuzzy Cognitive Map (FCM), and Analytic Hierarchy Process (AHP). Comparative analysis demonstrates that AHP surpassed the others, achieving a precision of 84%, recall of 83%, and an F1-score of 84%. Consequently, the AHP technique was integrated into the development of “Febra Diagnostica,” an app aimed at enhancing febrile disease diagnosis in resource-constrained settings. The app was then deployed and utilized in select Nigerian states, offering scalability and empowering frontline health workers in primary health facilities. Febra Diagnostica featured user-friendly interfaces, automated diagnosis and treatment suggestions, streamlined referrals, and provisions for further investigations. Encryption, access control, and multi-factor authentication were some of the security and privacy considerations in the app which gained acceptance from medical experts and adapted to regulatory and ethical policies for smart healthcare systems.

由于湿度大、温度高，发热疾病在热带地区非常普遍。这些地区主要包括中低收入国家，往往面临着医疗基础设施不足和缺乏准确诊断发热疾病的熟练人员等挑战。区分一种发热疾病和另一种发热疾病是一项重大挑战，增加了准确诊断的复杂性。本研究开发了一个多症状多疾病模型来应对这一挑战，该模型利用了对实地研究的患者数据集进行的探索性数据分析以及热带疾病专业医生的专业知识。研究采用三种智能技术，调查了区分尼日利亚流行的 11 种发热疾病的最有效建模方法：极端梯度提升 (XGBoost)、模糊认知图 (FCM) 和层次分析法 (AHP)。比较分析表明，AHP 超越了其他技术，精确度达到 84%，召回率达到 83%，F1 分数达到 84%。因此，AHP 技术被整合到 "Febra Diagnostica "应用程序的开发中，该应用程序的目的是在资源有限的环境中加强发热疾病的诊断。该应用程序随后在尼日利亚部分州进行了部署和使用，提供了可扩展性，并增强了基层医疗机构一线卫生工作者的能力。Febra Diagnostica 具有用户友好的界面、自动诊断和治疗建议、简化的转诊程序以及进一步调查的规定。加密、访问控制和多因素验证是该应用程序在安全和隐私方面的一些考虑因素，它获得了医学专家的认可，并符合智能医疗系统的监管和道德政策。

{"title":"Febrile disease modeling and diagnosis system for optimizing medical decisions in resource-scarce settings","authors":"Daniel Asuquo , Kingsley Attai , Okure Obot , Moses Ekpenyong , Christie Akwaowo , Kiirya Arnold , Faith-Michael Uzoka","doi":"10.1016/j.ceh.2024.05.001","DOIUrl":"https://doi.org/10.1016/j.ceh.2024.05.001","url":null,"abstract":"<div><p>Febrile diseases are highly prevalent in tropical regions due to elevated humidity and high temperatures. These regions, mainly comprising low- and middle-income countries, often face challenges related to inadequate medical infrastructure and a lack of skilled personnel for accurately diagnosing febrile diseases. Distinguishing one febrile illness from another posed a significant challenge, adding to the complexity of accurate diagnoses. This study developed a multi-symptom multi-disease model to address this challenge, leveraging exploratory data analysis of patient datasets from field studies and the expertise of medical practitioners specializing in tropical diseases. The research investigated the most effective modeling approach for differentiating among 11 febrile illnesses that are prevalent in Nigeria using three intelligent techniques: Extreme Gradient Boost (XGBoost), Fuzzy Cognitive Map (FCM), and Analytic Hierarchy Process (AHP). Comparative analysis demonstrates that AHP surpassed the others, achieving a precision of 84%, recall of 83%, and an F1-score of 84%. Consequently, the AHP technique was integrated into the development of “Febra Diagnostica,” an app aimed at enhancing febrile disease diagnosis in resource-constrained settings. The app was then deployed and utilized in select Nigerian states, offering scalability and empowering frontline health workers in primary health facilities. Febra Diagnostica featured user-friendly interfaces, automated diagnosis and treatment suggestions, streamlined referrals, and provisions for further investigations. Encryption, access control, and multi-factor authentication were some of the security and privacy considerations in the app which gained acceptance from medical experts and adapted to regulatory and ethical policies for smart healthcare systems.</p></div>","PeriodicalId":100268,"journal":{"name":"Clinical eHealth","volume":"7 ","pages":"Pages 52-76"},"PeriodicalIF":0.0,"publicationDate":"2024-05-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.sciencedirect.com/science/article/pii/S2588914124000066/pdfft?md5=bc3c9183d27457f4a17f78d6e77caccd&pid=1-s2.0-S2588914124000066-main.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140914083","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}

引用次数: 0

Newsletter: The first Metaverse Medical Digital Human GPT launches 时事通讯：首个 Metaverse 医学数字人 GPT 启动

Clinical eHealth

Pub Date : 2024-04-22 DOI: 10.1016/j.ceh.2024.04.001

引用次数: 0