Journal of Medical Internet Research最新文献_第9页

Accuracy of Medical Image-Based Deep Learning for Detecting Microvascular Invasion in Hepatocellular Carcinoma: Systematic Review and Meta-Analysis. 基于医学图像的深度学习检测肝细胞癌微血管侵犯的准确性：系统评价和荟萃分析。

IF 6 2区医学 Q1 HEALTH CARE SCIENCES & SERVICES

Journal of Medical Internet Research

Pub Date : 2026-03-02 DOI: 10.2196/82000

Wei Feng, Bo Qu, Shuo Han

Background: Hepatocellular carcinoma (HCC) is a leading cause of cancer-related mortality worldwide. Microvascular invasion (MVI) is a critical pathological indicator of postoperative recurrence and poor prognosis in patients with HCC. Some researchers have explored the diagnostic accuracy of deep learning (DL) based on various imaging modalities for MVI.Objective: This meta-analysis aimed to systematically evaluate the preoperative diagnostic performance of DL models using medical images to predict MVI in HCC, and to investigate the impact of different imaging modalities and validation strategies on model performance and generalizability.Methods: PubMed, Cochrane Library, Embase, and Web of Science were searched up to October 16, 2025. Studies investigating the detection of MVI in HCC using imaging-based DL techniques were eligible. Studies focusing solely on image segmentation were excluded. The Quality Assessment of Diagnostic Accuracy Studies-2 tool was used to assess risk of bias. A bivariate mixed-effects meta-analysis was performed to calculate the pooled sensitivity, specificity, and area under the summary receiver operating characteristic curve (SROC). Subgroup analyses were conducted by imaging modality and validation set generation method.Results: This meta-analysis included 52 studies with 19,531 patients with HCC. The pooled analysis revealed that imaging-based DL models had an overall sensitivity of 0.80 (95% CI 0.78-0.83), a specificity of 0.82 (95% CI 0.80-0.85), and an SROC of 0.88 for MVI prediction. Subgroup analysis showed that models based on preoperative contrast-enhanced computed tomography performed excellently, with a sensitivity of 0.84 (95% CI 0.79-0.88), a specificity of 0.83 (95% CI 0.77-0.88), and an SROC of 0.90. These results suggest that contrast-enhanced computed tomography is the most promising noninvasive method for current clinical applications. Meanwhile, DL models using pathological sections achieved the highest diagnostic performance: a sensitivity of 0.91 (95% CI 0.87-0.94), a specificity of 0.90 (95% CI 0.68-0.97), and an SROC of 0.92. This establishes the ultimate benchmark for performance optimization for all noninvasive models. A key finding was that model performance was less consistent in independent external validation (SROC: 0.85) than in internal validation (SROC: 0.90). This discrepancy indicates that overreliance on internal validation may overestimate model efficacy and underscores the decisive role of rigorous external validation in assessing real-world generalizability.Conclusions: This study is the first to systematically assess the use of imaging-based DL for diagnosing MVI in HCC. The results demonstrate a significant potential for these models in predicting MVI. However, their clinical applicability requires rigorous evaluation, given the scarcity of independent external val

背景：肝细胞癌（HCC）是世界范围内癌症相关死亡的主要原因。微血管侵犯（Microvascular invasion， MVI）是HCC患者术后复发和预后不良的重要病理指标。一些研究人员已经探索了基于各种成像模式的深度学习（DL）对MVI的诊断准确性。目的：本荟萃分析旨在系统评价肝细胞癌（HCC）影像学DL模型的术前诊断性能，并探讨不同影像学方式和验证策略对模型性能和普遍性的影响。方法：检索截止到2025年10月16日的PubMed、Cochrane Library、Embase和Web of Science。使用基于成像的DL技术检测HCC中的MVI的研究是合格的。仅关注图像分割的研究被排除在外。使用诊断准确性研究质量评估-2工具评估偏倚风险。进行双变量混合效应荟萃分析，计算汇总敏感性、特异性和总受试者工作特征曲线（SROC）下的面积。采用成像方式和验证集生成方法进行亚组分析。结果：该荟萃分析包括52项研究，19531例HCC患者。合并分析显示，基于成像的DL模型预测MVI的总体敏感性为0.80 (95% CI 0.78-0.83)，特异性为0.82 (95% CI 0.80-0.85)， SROC为0.88。亚组分析显示，基于术前对比增强计算机断层扫描的模型表现出色，敏感性为0.84 (95% CI 0.79-0.88)，特异性为0.83 (95% CI 0.77-0.88)， SROC为0.90。这些结果表明，对比增强计算机断层扫描是目前临床应用中最有前途的无创方法。同时，使用病理切片的DL模型具有最高的诊断性能：敏感性为0.91 (95% CI 0.87-0.94)，特异性为0.90 (95% CI 0.68-0.97)， SROC为0.92。这为所有非侵入性模型的性能优化建立了最终基准。一个关键的发现是，独立外部验证（SROC: 0.85）的模型性能一致性低于内部验证（SROC: 0.90）。这一差异表明过度依赖内部验证可能会高估模型的有效性，并强调了严格的外部验证在评估现实世界的泛化性方面的决定性作用。结论：本研究首次系统地评估了基于成像的DL在HCC中诊断MVI的应用。结果表明，这些模型在预测MVI方面具有很大的潜力。然而，由于缺乏独立的外部验证队列，它们之间存在显著的异质性，并且观察到模型性能下降，因此它们的临床适用性需要严格的评估。因此，遵循标准化报告准则的前瞻性多中心研究是未来的关键方向。这些研究还应侧重于开发集成算法，将组织病理学见解转化为术前成像数据，以建立强大的临床工具。

{"title":"Accuracy of Medical Image-Based Deep Learning for Detecting Microvascular Invasion in Hepatocellular Carcinoma: Systematic Review and Meta-Analysis.","authors":"Wei Feng, Bo Qu, Shuo Han","doi":"10.2196/82000","DOIUrl":"10.2196/82000","url":null,"abstract":"Background: Hepatocellular carcinoma (HCC) is a leading cause of cancer-related mortality worldwide. Microvascular invasion (MVI) is a critical pathological indicator of postoperative recurrence and poor prognosis in patients with HCC. Some researchers have explored the diagnostic accuracy of deep learning (DL) based on various imaging modalities for MVI.Objective: This meta-analysis aimed to systematically evaluate the preoperative diagnostic performance of DL models using medical images to predict MVI in HCC, and to investigate the impact of different imaging modalities and validation strategies on model performance and generalizability.Methods: PubMed, Cochrane Library, Embase, and Web of Science were searched up to October 16, 2025. Studies investigating the detection of MVI in HCC using imaging-based DL techniques were eligible. Studies focusing solely on image segmentation were excluded. The Quality Assessment of Diagnostic Accuracy Studies-2 tool was used to assess risk of bias. A bivariate mixed-effects meta-analysis was performed to calculate the pooled sensitivity, specificity, and area under the summary receiver operating characteristic curve (SROC). Subgroup analyses were conducted by imaging modality and validation set generation method.Results: This meta-analysis included 52 studies with 19,531 patients with HCC. The pooled analysis revealed that imaging-based DL models had an overall sensitivity of 0.80 (95% CI 0.78-0.83), a specificity of 0.82 (95% CI 0.80-0.85), and an SROC of 0.88 for MVI prediction. Subgroup analysis showed that models based on preoperative contrast-enhanced computed tomography performed excellently, with a sensitivity of 0.84 (95% CI 0.79-0.88), a specificity of 0.83 (95% CI 0.77-0.88), and an SROC of 0.90. These results suggest that contrast-enhanced computed tomography is the most promising noninvasive method for current clinical applications. Meanwhile, DL models using pathological sections achieved the highest diagnostic performance: a sensitivity of 0.91 (95% CI 0.87-0.94), a specificity of 0.90 (95% CI 0.68-0.97), and an SROC of 0.92. This establishes the ultimate benchmark for performance optimization for all noninvasive models. A key finding was that model performance was less consistent in independent external validation (SROC: 0.85) than in internal validation (SROC: 0.90). This discrepancy indicates that overreliance on internal validation may overestimate model efficacy and underscores the decisive role of rigorous external validation in assessing real-world generalizability.Conclusions: This study is the first to systematically assess the use of imaging-based DL for diagnosing MVI in HCC. The results demonstrate a significant potential for these models in predicting MVI. However, their clinical applicability requires rigorous evaluation, given the scarcity of independent external val","PeriodicalId":16337,"journal":{"name":"Journal of Medical Internet Research","volume":"28 ","pages":"e82000"},"PeriodicalIF":6.0,"publicationDate":"2026-03-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12954728/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147433486","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}

引用次数: 0

Context-Aware Sentence Classification of Radiology Reports Using Synthetic Data: Development and Validation Study. 使用合成数据的放射学报告的上下文感知句子分类：开发和验证研究。

IF 6 2区医学 Q1 HEALTH CARE SCIENCES & SERVICES

Journal of Medical Internet Research

Pub Date : 2026-02-28 DOI: 10.2196/86365

Tomohiro Kikuchi, Yosuke Yamagishi, Kohei Yamamoto, Toshiaki Akashi, Harushi Mori, Hisaki Makimoto, Takahide Kohro

Background: Automated structuring of radiology reports is essential for data utilization and the development of medical artificial intelligence models. However, manual annotation by experts is labor-intensive, and processing real clinical data through commercial large language models (LLMs) presents significant privacy risks. These challenges are particularly pronounced for non-English languages like Japanese, where specialized medical corpora are scarce. While synthetic data generation offers a potential privacy-preserving alternative, its effectiveness in capturing complex clinical nuances-such as negation and contextual dependencies-to train robust classification models without any real-world training data has not been fully established.Objective: This study aimed to develop a context-aware sentence classification model for Japanese radiology reports using an entirely synthetic training pipeline, thereby eliminating the reliance on real-world clinical data during the development phase. Furthermore, we sought to evaluate the generalizability of this approach by validating the model performance on diverse, multi-institutional real-world reports.Methods: Japanese radiology reports (n=3,104) were generated using GPT-4.1 and automatically annotated at the sentence level into four categories (background, positive finding, negative finding, and continuation) using GPT-4.1-mini. This Synthetic data was partitioned into training (n=2,670), validation (n=334), and test (n=100) sets. We fine-tuned several models, including lightweight local LLMs (Qwen3 and Llama 3.2 series) using Low-Rank Adaptation (LoRA) and Japanese text classification models (BERT base Japanese v3, JMedRoBERTa-base, and ModernBERT-Ja-130M). External validation was performed using 280 real-world reports (3,477 sentences) from seven institutions in the Japan Medical Image Database (J-MID), with ground-truth labels established by board-certified radiologists. Evaluation metrics included accuracy, macro-averaged F1 (Macro F1) score, and positive predictive value for positive findings (PPV_1).Results: All models achieved high performance on the synthetic test set (accuracy: 0.938-0.951; Macro F1 score: 0.924-0.940). Overall performance declined on the external validation dataset (accuracy: 0.783-0.813; Macro F1 score: 0.761-0.790), reflecting distributional differences between synthetic and real-world reports; however, PPV_1 remained stable and high across datasets (e.g., 0.957 on the synthetic test set vs. 0.952 on the external validation dataset for Qwen3 (4B)). Parsing errors occurred in LLM-based approaches (19-260 sentences, 0.55%-7.48% in the external dataset).Conclusions: This study demonstrates the feasibility of developing context-aware sentence classification models for Japanese radiology reports using a training pipeline based entirely on synthetic data. The stabi

背景：放射学报告的自动化结构对于数据利用和医学人工智能模型的发展至关重要。然而，专家手工标注是一项劳动密集型工作，并且通过商业大型语言模型（llm）处理真实临床数据存在重大隐私风险。这些挑战对于日语等非英语语言尤其明显，因为这些语言缺乏专门的医学语料库。虽然合成数据生成提供了一种潜在的隐私保护替代方案，但其在捕捉复杂的临床细微差别（如否定和上下文依赖性）以训练健壮的分类模型方面的有效性尚未完全建立，而无需任何真实世界的训练数据。目的：本研究旨在使用完全合成的训练管道为日语放射学报告开发一个上下文感知的句子分类模型，从而消除在开发阶段对真实临床数据的依赖。此外，我们试图通过验证模型在不同的、多机构的真实世界报告中的表现来评估这种方法的普遍性。方法：使用GPT-4.1生成日本放射学报告（n= 3104），并使用GPT-4.1-mini在句子级别自动标注为背景、阳性发现、阴性发现和延续四类。该合成数据被划分为训练集（n= 2670）、验证集（n=334）和测试集（n=100）。我们对几个模型进行了微调，包括使用Low-Rank Adaptation （LoRA）和日语文本分类模型（BERT base Japanese v3、JMedRoBERTa-base和ModernBERT-Ja-130M）的轻量级本地llm （Qwen3和Llama 3.2系列）。外部验证使用来自日本医学图像数据库（J-MID）中七个机构的280份真实报告（3,477个句子）进行，并使用由委员会认证的放射科医生建立的基础事实标签。评价指标包括准确性、宏观平均F1 （Macro F1）评分和阳性结果的阳性预测值（PPV_1）。结果：所有模型在综合测试集上均取得了较高的性能（准确率：0.938 ~ 0.951；Macro F1得分：0.924 ~ 0.940）。外部验证数据集的整体性能下降（准确性：0.783-0.813；宏观F1得分：0.761-0.790），反映了合成报告与真实报告之间的分布差异；然而，PPV_1在数据集上保持稳定和高（例如，在合成测试集上为0.957，而在Qwen3 （4B）的外部验证数据集上为0.952）。基于llm的方法中出现了解析错误（19-260个句子，外部数据集中为0.55%-7.48%）。结论：本研究证明了使用完全基于合成数据的训练管道为日语放射学报告开发上下文感知句子分类模型的可行性。PPV_1的稳定性表明，尽管在外部验证中观察到性能下降，但该模型成功捕获了识别真实世界报告中积极发现所需的基本临床术语和语言模式。这种方法大大减少了人工标注需求和隐私风险，为构建结构化放射学数据集提供了可扩展的基础，以支持临床相关医学人工智能模型的开发。临床试验:

{"title":"Context-Aware Sentence Classification of Radiology Reports Using Synthetic Data: Development and Validation Study.","authors":"Tomohiro Kikuchi, Yosuke Yamagishi, Kohei Yamamoto, Toshiaki Akashi, Harushi Mori, Hisaki Makimoto, Takahide Kohro","doi":"10.2196/86365","DOIUrl":"https://doi.org/10.2196/86365","url":null,"abstract":"Background: Automated structuring of radiology reports is essential for data utilization and the development of medical artificial intelligence models. However, manual annotation by experts is labor-intensive, and processing real clinical data through commercial large language models (LLMs) presents significant privacy risks. These challenges are particularly pronounced for non-English languages like Japanese, where specialized medical corpora are scarce. While synthetic data generation offers a potential privacy-preserving alternative, its effectiveness in capturing complex clinical nuances-such as negation and contextual dependencies-to train robust classification models without any real-world training data has not been fully established.Objective: This study aimed to develop a context-aware sentence classification model for Japanese radiology reports using an entirely synthetic training pipeline, thereby eliminating the reliance on real-world clinical data during the development phase. Furthermore, we sought to evaluate the generalizability of this approach by validating the model performance on diverse, multi-institutional real-world reports.Methods: Japanese radiology reports (n=3,104) were generated using GPT-4.1 and automatically annotated at the sentence level into four categories (background, positive finding, negative finding, and continuation) using GPT-4.1-mini. This Synthetic data was partitioned into training (n=2,670), validation (n=334), and test (n=100) sets. We fine-tuned several models, including lightweight local LLMs (Qwen3 and Llama 3.2 series) using Low-Rank Adaptation (LoRA) and Japanese text classification models (BERT base Japanese v3, JMedRoBERTa-base, and ModernBERT-Ja-130M). External validation was performed using 280 real-world reports (3,477 sentences) from seven institutions in the Japan Medical Image Database (J-MID), with ground-truth labels established by board-certified radiologists. Evaluation metrics included accuracy, macro-averaged F1 (Macro F1) score, and positive predictive value for positive findings (PPV_1).Results: All models achieved high performance on the synthetic test set (accuracy: 0.938-0.951; Macro F1 score: 0.924-0.940). Overall performance declined on the external validation dataset (accuracy: 0.783-0.813; Macro F1 score: 0.761-0.790), reflecting distributional differences between synthetic and real-world reports; however, PPV_1 remained stable and high across datasets (e.g., 0.957 on the synthetic test set vs. 0.952 on the external validation dataset for Qwen3 (4B)). Parsing errors occurred in LLM-based approaches (19-260 sentences, 0.55%-7.48% in the external dataset).Conclusions: This study demonstrates the feasibility of developing context-aware sentence classification models for Japanese radiology reports using a training pipeline based entirely on synthetic data. The stabi","PeriodicalId":16337,"journal":{"name":"Journal of Medical Internet Research","volume":" ","pages":""},"PeriodicalIF":6.0,"publicationDate":"2026-02-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147321680","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}

引用次数: 0

Influencing the Influencers: How Health Experts Are Partnering With Content Creators to Fight Misinformation Online. 影响影响者：健康专家如何与内容创作者合作打击在线错误信息。

IF 6 2区医学 Q1 HEALTH CARE SCIENCES & SERVICES

Journal of Medical Internet Research

Pub Date : 2026-02-27 DOI: 10.2196/93450

Wendy Glauser

引用次数: 0

Machine Learning in Left Ventricular Hypertrophy Detection: Systematic Review and Meta-Analysis. 机器学习在左心室肥厚检测中的应用：系统回顾和荟萃分析。

IF 6 2区医学 Q1 HEALTH CARE SCIENCES & SERVICES

Journal of Medical Internet Research

Pub Date : 2026-02-27 DOI: 10.2196/76637

Yilin Li, Ke Zhao, Jing Wu

Background: In recent years, researchers have investigated machine learning (ML)-based approaches for the detection of left ventricular hypertrophy (LVH). However, the accuracy of ML in detecting LVH varies across different modeling variables and models. Systematic evidence is lacking in understanding how different ML approaches affect LVH detection accuracy.

Objective: The aim of this study is to systematically assess the diagnostic accuracy of these ML approaches to inform the development of artificial intelligence tools.

Methods: PubMed, Embase, Cochrane Library, and Web of Science were comprehensively searched up to November 12, 2025. The Prediction Model Risk of Bias Assessment Tool was used to evaluate the risk of bias. Subgroup analyses were performed based on ML model types and modeling variables (electrocardiogram [ECG], clinical features, and echocardiography). Only diagnostic 2×2 tables from validation sets were pooled for meta-analysis, with all statistical analyses performed using Stata.

Results: A total of 25 studies were included in the analysis. The performance of ML models varied with input data types and algorithms. A meta-analysis showed that ECG-based models, in comparison, exhibited a sensitivity of 0.76 (95% CI 0.66-0.84) and a specificity of 0.84 (95% CI 0.78-0.89). Echocardiography-based models had a sensitivity ranging from 0.71 to 0.94 and a specificity ranging from 0.67 to 0.96. The models based on clinical features had a sensitivity of 0.78 (95% CI 0.69-0.85) and a specificity of 0.71 (95% CI 0.65-0.76). A subgroup analysis of the ECG-based models revealed that the deep learning model produced a sensitivity of 0.71 (95% CI 0.60-0.80) and a specificity of 0.79 (95% CI 0.65-0.88).

Conclusions: ML demonstrates reasonably high accuracy in detecting LVH. However, these conclusions are derived from limited evidence. Meanwhile, the extreme heterogeneity reported in the meta-analysis requires more critical interpretation. Current conclusions regarding model accuracy should be interpreted with caution. Therefore, future research should focus on constructing high-performance ML models based on imaging data for LVH diagnosis.

背景：近年来，研究人员研究了基于机器学习（ML）的左心室肥厚（LVH）检测方法。然而，ML检测LVH的准确性在不同的建模变量和模型中有所不同。在了解不同的ML方法如何影响LVH检测准确性方面，缺乏系统的证据。目的：本研究的目的是系统地评估这些机器学习方法的诊断准确性，为人工智能工具的开发提供信息。方法：综合检索PubMed、Embase、Cochrane Library、Web of Science，检索截止日期为2025年11月12日。使用预测模型偏倚风险评估工具评估偏倚风险。根据ML模型类型和建模变量（心电图、临床特征和超声心动图）进行亚组分析。只有来自验证集的诊断2×2表被合并进行meta分析，所有统计分析都使用Stata进行。结果：共纳入25项研究。机器学习模型的性能随输入数据类型和算法而变化。荟萃分析显示，相比之下，基于心电图的模型的敏感性为0.76 (95% CI 0.66-0.84)，特异性为0.84 （95% CI 0.78-0.89）。基于超声心动图的模型灵敏度为0.71 ~ 0.94，特异性为0.67 ~ 0.96。基于临床特征的模型敏感性为0.78 (95% CI 0.69-0.85)，特异性为0.71 （95% CI 0.65-0.76）。基于ecg的模型的亚组分析显示，深度学习模型的敏感性为0.71 (95% CI 0.60-0.80)，特异性为0.79 （95% CI 0.65-0.88）。结论：ML检测LVH具有较高的准确性。然而，这些结论来自有限的证据。同时，meta分析中报告的极端异质性需要更严格的解释。目前关于模型准确性的结论应谨慎解释。因此，未来的研究应侧重于基于影像数据构建用于LVH诊断的高性能ML模型。

{"title":"Machine Learning in Left Ventricular Hypertrophy Detection: Systematic Review and Meta-Analysis.","authors":"Yilin Li, Ke Zhao, Jing Wu","doi":"10.2196/76637","DOIUrl":"10.2196/76637","url":null,"abstract":"Background: In recent years, researchers have investigated machine learning (ML)-based approaches for the detection of left ventricular hypertrophy (LVH). However, the accuracy of ML in detecting LVH varies across different modeling variables and models. Systematic evidence is lacking in understanding how different ML approaches affect LVH detection accuracy.Objective: The aim of this study is to systematically assess the diagnostic accuracy of these ML approaches to inform the development of artificial intelligence tools.Methods: PubMed, Embase, Cochrane Library, and Web of Science were comprehensively searched up to November 12, 2025. The Prediction Model Risk of Bias Assessment Tool was used to evaluate the risk of bias. Subgroup analyses were performed based on ML model types and modeling variables (electrocardiogram [ECG], clinical features, and echocardiography). Only diagnostic 2×2 tables from validation sets were pooled for meta-analysis, with all statistical analyses performed using Stata.Results: A total of 25 studies were included in the analysis. The performance of ML models varied with input data types and algorithms. A meta-analysis showed that ECG-based models, in comparison, exhibited a sensitivity of 0.76 (95% CI 0.66-0.84) and a specificity of 0.84 (95% CI 0.78-0.89). Echocardiography-based models had a sensitivity ranging from 0.71 to 0.94 and a specificity ranging from 0.67 to 0.96. The models based on clinical features had a sensitivity of 0.78 (95% CI 0.69-0.85) and a specificity of 0.71 (95% CI 0.65-0.76). A subgroup analysis of the ECG-based models revealed that the deep learning model produced a sensitivity of 0.71 (95% CI 0.60-0.80) and a specificity of 0.79 (95% CI 0.65-0.88).Conclusions: ML demonstrates reasonably high accuracy in detecting LVH. However, these conclusions are derived from limited evidence. Meanwhile, the extreme heterogeneity reported in the meta-analysis requires more critical interpretation. Current conclusions regarding model accuracy should be interpreted with caution. Therefore, future research should focus on constructing high-performance ML models based on imaging data for LVH diagnosis.","PeriodicalId":16337,"journal":{"name":"Journal of Medical Internet Research","volume":"28 ","pages":"e76637"},"PeriodicalIF":6.0,"publicationDate":"2026-02-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12954682/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147344538","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}

引用次数: 0

Digital Health Interventions to Promote Physical Activity Among Adolescents: Systematic Review. 促进青少年体育活动的数字健康干预：系统回顾。

IF 6 2区医学 Q1 HEALTH CARE SCIENCES & SERVICES

Journal of Medical Internet Research

Pub Date : 2026-02-27 DOI: 10.2196/82395

Rui Shi Fan, Jia Jun Jiang, Qing Yuan Zhou, Xin Yue Zhang, Zhou Hang Wu, Liu Ji

Background: Insufficient physical activity among adolescents is a major global public health concern. Digital health interventions (DHIs) have gained increasing attention as a promising approach to promoting physical activity in adolescents. However, existing systematic reviews predominantly focus on single-intervention formats or specific study designs, while reviews that integrate multiple DHIs and diverse study designs remain scarce.Objective: This systematic review aims to synthesize evidence from diverse DHIs and multiple study designs to assess their effectiveness in promoting physical activity among adolescents.Methods: The review protocol was registered in PROSPERO (International Prospective Register of Systematic Reviews; CRD420251117923). This systematic review searched literature published between January 1, 2014, and June 30, 2025, across Web of Science, PubMed, EBSCO, Scopus, Embase, the Cochrane Library, ProQuest, and Google Scholar. The final search was completed on August 3, 2025. Using the PICOS (population, intervention, comparator, outcomes, and study design) framework, the review included adolescents aged 10-19 years and focused on evidence-based research promoting physical activity through DHIs. The review was limited to peer-reviewed English-language literature and excluded studies solely focused on measurement tools, those not evaluating intervention effectiveness, or those not involving adolescents. Two reviewers independently screened studies and extracted data. Research quality was assessed using the Joanna Briggs Institute tool. Findings were synthesized through narrative synthesis and qualitative content analysis.Results: A total of 24 studies were included, involving approximately 12,183 adolescents. Study designs comprised 10 randomized controlled trials, 4 quasi-experimental studies, 3 quantitative research studies, 3 cross-sectional studies, and 4 mixed methods studies. Overall, 7 (29%) studies were of high quality, 16 (67%) were of moderate quality, and 1 (4%) was of low quality. Study populations included general adolescents as well as subgroups with specific health risks: insufficient physical activity (1/24, 4%), obesity or overweight (4/24, 17%), attention-deficit/hyperactivity disorder (1/24, 4%), cancer survivors (1/24, 4%), and at-risk youth (1/24, 4%). DHIs were categorized into 3 types: single-driver interventions (14/24, 58%), multimodal integrated interventions (7/24, 29%), and interaction-enhanced interventions (3/24, 13%). Most studies reported positive outcomes, including direct effectiveness (15/24, 63%), indirect effectiveness (8/24, 33%), and unclear effectiveness (1/24, 4%).Conclusions: This systematic review synthesizes evidence from diverse research designs and multiple types of DHIs, offering a more comprehensive perspective than previous reviews focused on single designs or tech

背景：青少年身体活动不足是一个主要的全球公共卫生问题。数字健康干预措施（DHIs）作为一种促进青少年身体活动的有希望的方法，越来越受到关注。然而，现有的系统综述主要集中在单干预形式或特定的研究设计上，而整合多个DHIs和不同研究设计的综述仍然很少。目的：本系统综述旨在综合来自不同DHIs和多个研究设计的证据，以评估其在促进青少年体育活动方面的有效性。方法：该综述方案在普洛斯彼罗（国际前瞻性系统评价登记册；CRD420251117923）注册。该系统综述检索了2014年1月1日至2025年6月30日之间发表的文献，检索范围包括Web of Science、PubMed、EBSCO、Scopus、Embase、Cochrane Library、ProQuest和谷歌Scholar。最终的搜寻工作于2025年8月3日完成。采用PICOS（人口、干预、比较、结果和研究设计）框架，本综述纳入了10-19岁的青少年，重点关注通过DHIs促进身体活动的循证研究。该综述仅限于同行评议的英语文献，排除了仅关注测量工具、不评估干预效果或不涉及青少年的研究。两位审稿人独立筛选研究并提取数据。研究质量评估使用乔安娜布里格斯研究所的工具。通过叙事综合和定性内容分析对研究结果进行综合。结果：共纳入24项研究，涉及约12183名青少年。研究设计包括10项随机对照试验、4项准实验研究、3项定量研究、3项横断面研究和4项混合方法研究。总体而言，7项（29%）研究为高质量，16项（67%）为中等质量，1项（4%）为低质量。研究人群包括一般青少年以及具有特定健康风险的亚组：身体活动不足（1/ 24,4%）、肥胖或超重（4/ 24,17%）、注意力缺陷/多动障碍（1/ 24,4%）、癌症幸存者（1/ 24,4%）和高危青少年（1/ 24,4%）。DHIs分为3种类型：单驱动干预（14/24,58%）、多模式综合干预（7/24,29%）和互动增强干预（3/24,13%）。大多数研究报告了积极的结果，包括直接有效（15/24,63%），间接有效（8/24,33%）和不明确的有效性（1/24,4%）。结论：本系统综述综合了来自不同研究设计和多种DHIs类型的证据，提供了比以往侧重于单一设计或技术格式的综述更全面的视角。结果表明，DHIs总体上提高了青少年的身体活动水平，尽管其效果因干预类型和研究设计而有很大差异。该综述填补了关键的研究空白，并强调了干预适应性和实施环境的关键作用。它还涉及实际问题，包括有特殊健康状况的青少年、数字健康不平等和技术依赖。尽管方法质量和随访不足存在局限性，但本综述为指导实际应用、政策制定和公平推广DHIs以加强青少年身体活动提供了重要证据。在全球青少年缺乏身体活动和健康差距扩大的背景下，它还概述了未来高质量研究的方向。试验注册：PROSPERO CRD420251117923；https://www.crd.york.ac.uk/PROSPERO/view/CRD420251117923。

{"title":"Digital Health Interventions to Promote Physical Activity Among Adolescents: Systematic Review.","authors":"Rui Shi Fan, Jia Jun Jiang, Qing Yuan Zhou, Xin Yue Zhang, Zhou Hang Wu, Liu Ji","doi":"10.2196/82395","DOIUrl":"10.2196/82395","url":null,"abstract":"Background: Insufficient physical activity among adolescents is a major global public health concern. Digital health interventions (DHIs) have gained increasing attention as a promising approach to promoting physical activity in adolescents. However, existing systematic reviews predominantly focus on single-intervention formats or specific study designs, while reviews that integrate multiple DHIs and diverse study designs remain scarce.Objective: This systematic review aims to synthesize evidence from diverse DHIs and multiple study designs to assess their effectiveness in promoting physical activity among adolescents.Methods: The review protocol was registered in PROSPERO (International Prospective Register of Systematic Reviews; CRD420251117923). This systematic review searched literature published between January 1, 2014, and June 30, 2025, across Web of Science, PubMed, EBSCO, Scopus, Embase, the Cochrane Library, ProQuest, and Google Scholar. The final search was completed on August 3, 2025. Using the PICOS (population, intervention, comparator, outcomes, and study design) framework, the review included adolescents aged 10-19 years and focused on evidence-based research promoting physical activity through DHIs. The review was limited to peer-reviewed English-language literature and excluded studies solely focused on measurement tools, those not evaluating intervention effectiveness, or those not involving adolescents. Two reviewers independently screened studies and extracted data. Research quality was assessed using the Joanna Briggs Institute tool. Findings were synthesized through narrative synthesis and qualitative content analysis.Results: A total of 24 studies were included, involving approximately 12,183 adolescents. Study designs comprised 10 randomized controlled trials, 4 quasi-experimental studies, 3 quantitative research studies, 3 cross-sectional studies, and 4 mixed methods studies. Overall, 7 (29%) studies were of high quality, 16 (67%) were of moderate quality, and 1 (4%) was of low quality. Study populations included general adolescents as well as subgroups with specific health risks: insufficient physical activity (1/24, 4%), obesity or overweight (4/24, 17%), attention-deficit/hyperactivity disorder (1/24, 4%), cancer survivors (1/24, 4%), and at-risk youth (1/24, 4%). DHIs were categorized into 3 types: single-driver interventions (14/24, 58%), multimodal integrated interventions (7/24, 29%), and interaction-enhanced interventions (3/24, 13%). Most studies reported positive outcomes, including direct effectiveness (15/24, 63%), indirect effectiveness (8/24, 33%), and unclear effectiveness (1/24, 4%).Conclusions: This systematic review synthesizes evidence from diverse research designs and multiple types of DHIs, offering a more comprehensive perspective than previous reviews focused on single designs or tech","PeriodicalId":16337,"journal":{"name":"Journal of Medical Internet Research","volume":"28 ","pages":"e82395"},"PeriodicalIF":6.0,"publicationDate":"2026-02-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147317396","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}

引用次数: 0

Physician, Restore Thyself? The Digital Gap in Physician Well-Being Support. 医生：康复？医生健康支持的数字差距。

IF 6 2区医学 Q1 HEALTH CARE SCIENCES & SERVICES

Journal of Medical Internet Research

Pub Date : 2026-02-27 DOI: 10.2196/93338

Jenny Castillo Cato

引用次数: 0

Artificial Intelligence in Health Professions Education: A Qualitative Study of Student Experiences. 卫生专业教育中的人工智能：学生体验的定性研究。

IF 6 2区医学 Q1 HEALTH CARE SCIENCES & SERVICES

Journal of Medical Internet Research

Pub Date : 2026-02-27 DOI: 10.2196/82432

Marina Guirguis, Salomon Fotsing, Joanne Fevry, Christine Landry, Diane Bouchard-Lamothe, Jennifer Lacroix, Alireza Jalali

Background: Artificial intelligence (AI) is increasingly integrated into education and healthcare, raising questions about how students use these technologies and how AI influences their learning. In health education, understanding these trends is particularly important because student learning directly impacts future clinical skills.

Objective: This study aimed to explore the use of AI tools by health sciences students at the University of Ottawa. More specifically, it sought to identify the most frequently used AI tools, describe students' usage habits, determine which tools support knowledge acquisition and skill development, and gather students' recommendations for effective strategies to raise awareness and train their peers on the responsible use of AI.

Methods: A qualitative approach was employed with students from ten health professions who reported using AI in their studies. Data were collected through semi-structured interviews and an open-ended qualitative online survey. Inductive thematic analysis within an interpretive paradigm was applied to capture patterns, perceptions, and emergent themes.

Results: 51 health professions students participated in the study. Most were women between the ages of 20 and 29. ChatGPT emerged as the most frequently used AI tool. Students perceived AI as a complementary tool that facilitated knowledge acquisition, skill development, writing and problem-solving. AI adoption was driven by curiosity, peer influence, and the desire to improve work efficiency. Students critically evaluated AI results, integrated the tools into their learning processes, and emphasized the importance of technical skills, critical thinking and digital literacy. Peer learning, hands-on demonstrations, and access to online resources were recommended for effective AI training.

Conclusions: This research demonstrates that health professions students actively use AI tools, particularly ChatGPT, to support learning, skill development and academic tasks. Although AI is valuable as an educational aid and its use varies by student and context, this highlights the need for structured guidance, critical evaluation skills and peer-supported training. These findings highlight the importance of thoughtfully integrating AI into educational programs to enhance learning outcomes, foster skill acquisition, ensure responsible and effective adoption.

Clinicaltrial:

背景：人工智能（AI）越来越多地融入教育和医疗保健领域，这引发了关于学生如何使用这些技术以及AI如何影响他们的学习的问题。在健康教育中，了解这些趋势尤为重要，因为学生的学习直接影响到未来的临床技能。目的：本研究旨在探讨渥太华大学健康科学专业学生使用人工智能工具的情况。更具体地说，它试图确定最常用的人工智能工具，描述学生的使用习惯，确定哪些工具支持知识获取和技能发展，并收集学生对有效策略的建议，以提高认识并培训他们的同龄人负责任地使用人工智能。方法：采用定性方法对来自10个卫生专业的学生进行研究，他们报告在学习中使用了人工智能。数据通过半结构化访谈和开放式定性在线调查收集。在解释范式内的归纳主题分析被应用于捕捉模式、感知和新兴主题。结果：51名卫生专业学生参与本研究。大多数是年龄在20到29岁之间的女性。ChatGPT成为最常用的人工智能工具。学生们认为人工智能是一种辅助工具，有助于知识获取、技能发展、写作和解决问题。人工智能的采用是由好奇心、同行影响力和提高工作效率的愿望驱动的。学生们批判性地评估人工智能的结果，将这些工具整合到他们的学习过程中，并强调技术技能、批判性思维和数字素养的重要性。同行学习、实践演示和访问在线资源被推荐用于有效的人工智能培训。结论：这项研究表明，卫生专业的学生积极使用人工智能工具，特别是ChatGPT，来支持学习、技能发展和学术任务。尽管人工智能作为一种教育辅助手段很有价值，其用途因学生和环境而异，但这凸显了对结构化指导、批判性评估技能和同伴支持培训的需求。这些发现强调了将人工智能纳入教育计划的重要性，以提高学习成果，促进技能习得，确保负责任和有效的采用。临床试验:

{"title":"Artificial Intelligence in Health Professions Education: A Qualitative Study of Student Experiences.","authors":"Marina Guirguis, Salomon Fotsing, Joanne Fevry, Christine Landry, Diane Bouchard-Lamothe, Jennifer Lacroix, Alireza Jalali","doi":"10.2196/82432","DOIUrl":"https://doi.org/10.2196/82432","url":null,"abstract":"Background: Artificial intelligence (AI) is increasingly integrated into education and healthcare, raising questions about how students use these technologies and how AI influences their learning. In health education, understanding these trends is particularly important because student learning directly impacts future clinical skills.Objective: This study aimed to explore the use of AI tools by health sciences students at the University of Ottawa. More specifically, it sought to identify the most frequently used AI tools, describe students' usage habits, determine which tools support knowledge acquisition and skill development, and gather students' recommendations for effective strategies to raise awareness and train their peers on the responsible use of AI.Methods: A qualitative approach was employed with students from ten health professions who reported using AI in their studies. Data were collected through semi-structured interviews and an open-ended qualitative online survey. Inductive thematic analysis within an interpretive paradigm was applied to capture patterns, perceptions, and emergent themes.Results: 51 health professions students participated in the study. Most were women between the ages of 20 and 29. ChatGPT emerged as the most frequently used AI tool. Students perceived AI as a complementary tool that facilitated knowledge acquisition, skill development, writing and problem-solving. AI adoption was driven by curiosity, peer influence, and the desire to improve work efficiency. Students critically evaluated AI results, integrated the tools into their learning processes, and emphasized the importance of technical skills, critical thinking and digital literacy. Peer learning, hands-on demonstrations, and access to online resources were recommended for effective AI training.Conclusions: This research demonstrates that health professions students actively use AI tools, particularly ChatGPT, to support learning, skill development and academic tasks. Although AI is valuable as an educational aid and its use varies by student and context, this highlights the need for structured guidance, critical evaluation skills and peer-supported training. These findings highlight the importance of thoughtfully integrating AI into educational programs to enhance learning outcomes, foster skill acquisition, ensure responsible and effective adoption.Clinicaltrial: ","PeriodicalId":16337,"journal":{"name":"Journal of Medical Internet Research","volume":" ","pages":""},"PeriodicalIF":6.0,"publicationDate":"2026-02-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147317398","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}

引用次数: 0

Team-Based Analysis of Large-Scale Qualitative Data: Tutorial Using a Nationwide SMS Text Messaging Poll of Youth. 基于团队的大规模定性数据分析：基于全国青少年短信调查的教程。

IF 6 2区医学 Q1 HEALTH CARE SCIENCES & SERVICES

Journal of Medical Internet Research

Pub Date : 2026-02-27 DOI: 10.2196/72526

Melissa DeJonckheere, Samantha A Chuisano, Marika Waselewski, Kendrin Sonneville, Tammy Chang

International registered report identifier (irrid): RR2-10.2196/resprot.8502.

国际注册报告标识符（irrid）： RR2-10.2196/resprot.8502。

引用次数: 0

Effectiveness of Telerehabilitation Interventions for Self-Management of Tinnitus: Update of a Systematic Review. 远程康复干预对耳鸣自我管理的有效性：一项系统综述的更新。

IF 6 2区医学 Q1 HEALTH CARE SCIENCES & SERVICES

Journal of Medical Internet Research

Pub Date : 2026-02-27 DOI: 10.2196/83529

Sara Demoen, Elise Van Kerchove, Annick Timmermans, Vincent Van Rompaey, Sarah Michiels, Annick Gilles

Background: Approximately 14% of the adult population has tinnitus, and current treatments are often costly and time-consuming. Telerehabilitation might reduce treatment costs without compromising effectiveness.Objective: Telerehabilitation is a quickly evolving research topic. Therefore, this systematic review update aims to give an overview of the research concerning the effectiveness of telerehabilitation interventions for self-management of tinnitus published between 2022 and 2025.Methods: This systematic review adheres to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses 2020) guidelines. PubMed, ScienceDirect, Scopus, Web of Science, and Cochrane Library were consulted for eligible studies concerning a study intervention of any possible form of self-management or telerehabilitation for adult patients with subjective tinnitus as a primary complaint. The risk of bias (RoB) and certainty of all included studies were assessed respectively by the Cochrane RoB2-tool and GRADE (Grading of Recommendations, Assessment, Development, and Evaluation) framework.Results: In total, 24 papers were included, of which 6 studied multiple telerehabilitation forms. Internet-based cognitive behavioral therapy with guidance by a psychologist or audiologist was examined in 5 studies (n=619), self-help manuals in 1 study (n=10), technological self-help devices in 3 studies (n=286), smartphone apps in 13 studies (n=23,788), and other internet-based interventions in 5 studies (n=442). These rehabilitation categories were proven to be effective in decreasing tinnitus severity and relieving tinnitus distress as measured by tinnitus questionnaires.Conclusions: The strength of this review is the gathering of recent studies on the very evolving topic of telerehabilitation for tinnitus. An important limitation of all included studies is that they raised some to great concerns of RoB. As a result, it is necessary to acknowledge that the overall certainty of the evidence ranged from low to moderate certainty. In addition, some crucial confounding parameters, such as the presence of hearing loss, hyperacusis, anxiety, depression, or sleeping problems, were not taken into consideration by all studies. This review gives an indication of the use of different telerehabilitation and self-management interventions for real-world clinical use, stating not only their possibilities but also their limitations. Overall, telerehabilitation was found to be effective in reducing tinnitus severity and distress. It forms a possible tool to improve the self-management capacities of the patient and the accessibility of tinnitus care as a replacement or an addition to in-person care. Nevertheless, barriers such as a lack of time, engagement, motivation, and openness of the patient, causing high dropout, should be taken into consideration. This review ac

背景：大约14%的成年人患有耳鸣，目前的治疗通常既昂贵又耗时。远程康复可以在不影响疗效的情况下降低治疗费用。目的：远程康复是一个快速发展的研究课题。因此，本系统综述旨在对2022年至2025年间发表的有关远程康复干预对耳鸣自我管理有效性的研究进行综述。方法：本系统评价遵循PRISMA（2020年系统评价和荟萃分析首选报告项目）指南。我们查阅了PubMed、ScienceDirect、Scopus、Web of Science和Cochrane Library，以获得有关主诉为主观性耳鸣的成年患者任何可能形式的自我管理或远程康复的研究干预的合格研究。所有纳入研究的偏倚风险（RoB）和确定性分别通过Cochrane rob2工具和GRADE（分级推荐、评估、发展和评价）框架进行评估。结果：共纳入论文24篇，其中6篇研究多种远程康复形式。在心理学家或听力学家的指导下，基于互联网的认知行为治疗在5项研究中（n=619），自助手册在1项研究中（n=10），技术自助设备在3项研究中（n=286），智能手机应用程序在13项研究中（n=23,788），其他基于互联网的干预措施在5项研究中（n=442）。这些康复类别被证明是有效的减少耳鸣严重程度和缓解耳鸣困扰，通过耳鸣问卷测量。结论：本综述的优势是收集了最近关于耳鸣远程康复这一不断发展的话题的研究。所有纳入的研究的一个重要局限性是，它们提出了一些对RoB的高度关注。因此，有必要承认，证据的总体确定性范围从低到中等确定性。此外，一些关键的混杂参数，如听力损失、听觉亢进、焦虑、抑郁或睡眠问题的存在，并没有被所有的研究考虑在内。这篇综述给出了不同的远程康复和自我管理干预在现实世界临床应用的指示，不仅说明了它们的可能性，也说明了它们的局限性。总之，远程康复被发现在减少耳鸣严重程度和痛苦方面是有效的。它形成了一种可能的工具，以提高患者的自我管理能力和耳鸣护理的可及性，作为替代或补充到亲自护理。然而，诸如缺乏时间、参与、动机和病人的开放性等障碍，导致高辍学率，应予以考虑。这篇综述强调了从基于互联网的认知行为疗法到对智能手机应用程序使用日益增长的兴趣的转变，进一步增加了治疗的可及性。试验注册：PROSPERO CRD 42021285450；https://www.crd.york.ac.uk/PROSPERO/view/CRD42021285450。

{"title":"Effectiveness of Telerehabilitation Interventions for Self-Management of Tinnitus: Update of a Systematic Review.","authors":"Sara Demoen, Elise Van Kerchove, Annick Timmermans, Vincent Van Rompaey, Sarah Michiels, Annick Gilles","doi":"10.2196/83529","DOIUrl":"10.2196/83529","url":null,"abstract":"Background: Approximately 14% of the adult population has tinnitus, and current treatments are often costly and time-consuming. Telerehabilitation might reduce treatment costs without compromising effectiveness.Objective: Telerehabilitation is a quickly evolving research topic. Therefore, this systematic review update aims to give an overview of the research concerning the effectiveness of telerehabilitation interventions for self-management of tinnitus published between 2022 and 2025.Methods: This systematic review adheres to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses 2020) guidelines. PubMed, ScienceDirect, Scopus, Web of Science, and Cochrane Library were consulted for eligible studies concerning a study intervention of any possible form of self-management or telerehabilitation for adult patients with subjective tinnitus as a primary complaint. The risk of bias (RoB) and certainty of all included studies were assessed respectively by the Cochrane RoB2-tool and GRADE (Grading of Recommendations, Assessment, Development, and Evaluation) framework.Results: In total, 24 papers were included, of which 6 studied multiple telerehabilitation forms. Internet-based cognitive behavioral therapy with guidance by a psychologist or audiologist was examined in 5 studies (n=619), self-help manuals in 1 study (n=10), technological self-help devices in 3 studies (n=286), smartphone apps in 13 studies (n=23,788), and other internet-based interventions in 5 studies (n=442). These rehabilitation categories were proven to be effective in decreasing tinnitus severity and relieving tinnitus distress as measured by tinnitus questionnaires.Conclusions: The strength of this review is the gathering of recent studies on the very evolving topic of telerehabilitation for tinnitus. An important limitation of all included studies is that they raised some to great concerns of RoB. As a result, it is necessary to acknowledge that the overall certainty of the evidence ranged from low to moderate certainty. In addition, some crucial confounding parameters, such as the presence of hearing loss, hyperacusis, anxiety, depression, or sleeping problems, were not taken into consideration by all studies. This review gives an indication of the use of different telerehabilitation and self-management interventions for real-world clinical use, stating not only their possibilities but also their limitations. Overall, telerehabilitation was found to be effective in reducing tinnitus severity and distress. It forms a possible tool to improve the self-management capacities of the patient and the accessibility of tinnitus care as a replacement or an addition to in-person care. Nevertheless, barriers such as a lack of time, engagement, motivation, and openness of the patient, causing high dropout, should be taken into consideration. This review ac","PeriodicalId":16337,"journal":{"name":"Journal of Medical Internet Research","volume":"28 ","pages":"e83529"},"PeriodicalIF":6.0,"publicationDate":"2026-02-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12988352/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147307034","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}

引用次数: 0

Impact of GPT-4-Generated Discharge Letters on Patients' Medical Comprehension: Prospective Crossover Study. gpt -4生成的出院信对患者医学理解的影响：前瞻性交叉研究

IF 6 2区医学 Q1 HEALTH CARE SCIENCES & SERVICES

Journal of Medical Internet Research

Pub Date : 2026-02-26 DOI: 10.2196/81243

Friederike Holderried, Alessandra Sonanini, Christian Stegemann-Philipps, Anne Herrmann-Werner, Philipp Spitzer, Martina Guthoff, Nils Heyne, Konstantin Sering, Martin Holderried, Felix Eisinger

Background: Patients often struggle to understand standard hospital discharge letters, increasing the risk of medication errors and misunderstandings. According to cognitive load theory (CLT), complex, information-dense texts can overload working memory and impair comprehension. Artificial intelligence tools that generate patient-centered versions could help reduce extraneous cognitive load and bridge this gap. However, evidence for their effectiveness remains limited.Objective: This study aimed to evaluate whether GPT-4 (OpenAI)-generated patient-centered letters improve standardized patients' retention and understanding of safety-relevant medical information compared with standard hospital discharge letters, and to explore potential effects on cognitive load as described by CLT.Methods: In this prospective, randomized, crossover study, 48 trained standardized patients received a conventional discharge letter for an assigned disease (out of 3) and its matching GPT-4-generated patient-centered letter. Participants read one version first, identified predefined safety-relevant "learning objectives," and then repeated the task with the alternate version. The primary outcome was the proportion of learning objectives fully, partially, or not reported. In a secondary analysis, results were stratified by content field (Medication, Organization, Prevention of Complications, Lifestyle/Disease Management) and Bloom taxonomy level ("Remember," "Understand").Results: The letter type significantly influenced comprehension (odds ratio [OR] 1.74, 95% CI 1.45-2.08; P<.001). Patient letters, compared with discharge letters, led to higher rates of fully (490/1073, 45.7% vs 413/1073, 38.5%) or partially (322/1073, 30% vs 287/1073, 26.7%) stated learning objectives and fewer omissions (261/1073, 24.3% vs 373/1073, 34.8%). Participants performed better on "Remember" than on "Understand" learning objectives, regardless of letter type (OR 3.33, 95% CI 1.96-5.88; P<.001). Compared with standard hospital discharge letters, patient letters consistently improved results at both cognitive levels ("Remember": 278/545, 51% vs 242/545, 44.4%; "Understand": 212/528, 40.2% vs 171/528, 32.4% fully stated). The effect of patient letters varied by content field (P<.001). The greatest improvements were observed for "Medication" (170/254, 66.9% vs 129/254, 50.8% fully stated) and "Organization" (78/158, 49.4% vs 62/158, 39.2% fully stated). Improvements in the content field "Prevention of Complications" were modest, and those for "Lifestyle/Disease Management" were even smaller across all conditions. A total of 24.3% (261/1073) of key information remained unrecognized.Conclusions: In this explanatory study, GPT-4-generated patient letters improved comprehension of safety-relevant discharge information among standardized patients, particularly regarding medication and or

背景：患者往往难以理解标准的出院信，增加了用药错误和误解的风险。认知负荷理论认为，复杂、信息密集的文本会使工作记忆过载，损害理解能力。产生以患者为中心的版本的人工智能工具可以帮助减少不必要的认知负荷，弥合这一差距。然而，证明其有效性的证据仍然有限。目的：本研究旨在评估与标准出院信相比，GPT-4 （OpenAI）生成的以患者为中心的信是否能提高标准化患者对安全相关医疗信息的记忆和理解，并探讨CLT对认知负荷的潜在影响。方法：在这项前瞻性、随机、交叉研究中，48名经过训练的标准化患者收到了一份针对指定疾病的常规出院信（共3份）和与之匹配的gpt -4生成的以患者为中心的出院信。参与者首先阅读一个版本，确定预先定义的与安全相关的“学习目标”，然后用备用版本重复任务。主要结果是学习目标完全报告、部分报告或未报告的比例。在二次分析中，根据内容域（用药、组织、并发症预防、生活方式/疾病管理）和Bloom分类水平（“记住”、“理解”）对结果进行分层。结果：信件类型显著影响理解（优势比[OR] 1.74, 95% CI 1.45-2.08）；结论：在本解释性研究中，gpt -4生成的患者信件提高了标准化患者对安全相关出院信息的理解，特别是在用药和组织方面。然而，它们在支持更高层次的理解（如风险预防或生活方式管理）方面效果较差。这些假设驱动的发现可以在CLT框架内解释，并可能激发对多模态、迭代支持的前瞻性评估。

{"title":"Impact of GPT-4-Generated Discharge Letters on Patients' Medical Comprehension: Prospective Crossover Study.","authors":"Friederike Holderried, Alessandra Sonanini, Christian Stegemann-Philipps, Anne Herrmann-Werner, Philipp Spitzer, Martina Guthoff, Nils Heyne, Konstantin Sering, Martin Holderried, Felix Eisinger","doi":"10.2196/81243","DOIUrl":"10.2196/81243","url":null,"abstract":"Background: Patients often struggle to understand standard hospital discharge letters, increasing the risk of medication errors and misunderstandings. According to cognitive load theory (CLT), complex, information-dense texts can overload working memory and impair comprehension. Artificial intelligence tools that generate patient-centered versions could help reduce extraneous cognitive load and bridge this gap. However, evidence for their effectiveness remains limited.Objective: This study aimed to evaluate whether GPT-4 (OpenAI)-generated patient-centered letters improve standardized patients' retention and understanding of safety-relevant medical information compared with standard hospital discharge letters, and to explore potential effects on cognitive load as described by CLT.Methods: In this prospective, randomized, crossover study, 48 trained standardized patients received a conventional discharge letter for an assigned disease (out of 3) and its matching GPT-4-generated patient-centered letter. Participants read one version first, identified predefined safety-relevant \"learning objectives,\" and then repeated the task with the alternate version. The primary outcome was the proportion of learning objectives fully, partially, or not reported. In a secondary analysis, results were stratified by content field (Medication, Organization, Prevention of Complications, Lifestyle/Disease Management) and Bloom taxonomy level (\"Remember,\" \"Understand\").Results: The letter type significantly influenced comprehension (odds ratio [OR] 1.74, 95% CI 1.45-2.08; P<.001). Patient letters, compared with discharge letters, led to higher rates of fully (490/1073, 45.7% vs 413/1073, 38.5%) or partially (322/1073, 30% vs 287/1073, 26.7%) stated learning objectives and fewer omissions (261/1073, 24.3% vs 373/1073, 34.8%). Participants performed better on \"Remember\" than on \"Understand\" learning objectives, regardless of letter type (OR 3.33, 95% CI 1.96-5.88; P<.001). Compared with standard hospital discharge letters, patient letters consistently improved results at both cognitive levels (\"Remember\": 278/545, 51% vs 242/545, 44.4%; \"Understand\": 212/528, 40.2% vs 171/528, 32.4% fully stated). The effect of patient letters varied by content field (P<.001). The greatest improvements were observed for \"Medication\" (170/254, 66.9% vs 129/254, 50.8% fully stated) and \"Organization\" (78/158, 49.4% vs 62/158, 39.2% fully stated). Improvements in the content field \"Prevention of Complications\" were modest, and those for \"Lifestyle/Disease Management\" were even smaller across all conditions. A total of 24.3% (261/1073) of key information remained unrecognized.Conclusions: In this explanatory study, GPT-4-generated patient letters improved comprehension of safety-relevant discharge information among standardized patients, particularly regarding medication and or","PeriodicalId":16337,"journal":{"name":"Journal of Medical Internet Research","volume":"28 ","pages":"e81243"},"PeriodicalIF":6.0,"publicationDate":"2026-02-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12982961/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147290006","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}

引用次数: 0