{"title":"Methodological Challenges for the Responsible Use of AI in Systematic Reviews: Risk of Bias Assessment","authors":"Farzad Maleki, Manoochehr Karami","doi":"10.1111/jebm.12665","DOIUrl":"10.1111/jebm.12665","url":null,"abstract":"","PeriodicalId":16090,"journal":{"name":"Journal of Evidence‐Based Medicine","volume":"17 4","pages":"712-713"},"PeriodicalIF":3.6,"publicationDate":"2024-12-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142846692","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}
Lenka Tenžera, Tatjana Milenković, Srdjan Ljubisavljević
{"title":"Holistic Approach and Systematized Researcher Guidelines for the Homeopathic Treatment of Lac Caninum Migraines","authors":"Lenka Tenžera, Tatjana Milenković, Srdjan Ljubisavljević","doi":"10.1111/jebm.12668","DOIUrl":"10.1111/jebm.12668","url":null,"abstract":"","PeriodicalId":16090,"journal":{"name":"Journal of Evidence‐Based Medicine","volume":"17 4","pages":"708-711"},"PeriodicalIF":3.6,"publicationDate":"2024-12-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142846683","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}
This study aims to investigate the occurrence of adverse events associated with topiramate by analyzing data from the FDA Adverse Event Reporting System. The goal is to provide a basis for the safe clinical use of topiramate.
Methods
Adverse event data from the FDA Adverse Event Reporting System, from its inception through the first quarter of 2024, were extracted. Signal detection was conducted using three methods: the reporting odds ratio, the medicines and healthcare products regulatory agency method, and the Bayesian confidence propagation neural network. Adverse events were statistically analyzed according to the preferred term and system organ class classifications from the Medical Dictionary for Regulatory Activities version 27.0. Positive signals were then compared against the drug label and the Important Medical Event list.
Results
A total of 12,168 adverse event reports involving topiramate as the primary suspect were analyzed, resulting in the extraction of 244 positive signals across 15 system organ classes. Among these, 21 signals were identified as serious adverse reactions not included in the drug label, encompassing 5 system organ classes. Notable signals included hypospadias, spina bifida, abortion spontaneous, renal tubular dysfunction, uveitis, retinal detachment, and choroidal effusion. Additionally, signals such as osmotic demyelination syndrome and Arnold-Chiari malformation were identified as requiring further monitoring.
Conclusion
This study identified several unexpected and serious adverse reaction signals that align with previously reported cases. These findings underscore the need for ongoing study, focused attention, and vigilant monitoring during the clinical use of topiramate.
{"title":"Postmarketing Safety Surveillance of Topiramate: A Signal Detection and Analysis Study Based on the FDA Adverse Event Reporting System Database","authors":"Kai Lin, Mengjiao He, Zuoqi Ding","doi":"10.1111/jebm.12667","DOIUrl":"10.1111/jebm.12667","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Objective</h3>\u0000 \u0000 <p>This study aims to investigate the occurrence of adverse events associated with topiramate by analyzing data from the FDA Adverse Event Reporting System. The goal is to provide a basis for the safe clinical use of topiramate.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>Adverse event data from the FDA Adverse Event Reporting System, from its inception through the first quarter of 2024, were extracted. Signal detection was conducted using three methods: the reporting odds ratio, the medicines and healthcare products regulatory agency method, and the Bayesian confidence propagation neural network. Adverse events were statistically analyzed according to the preferred term and system organ class classifications from the Medical Dictionary for Regulatory Activities version 27.0. Positive signals were then compared against the drug label and the Important Medical Event list.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>A total of 12,168 adverse event reports involving topiramate as the primary suspect were analyzed, resulting in the extraction of 244 positive signals across 15 system organ classes. Among these, 21 signals were identified as serious adverse reactions not included in the drug label, encompassing 5 system organ classes. Notable signals included hypospadias, spina bifida, abortion spontaneous, renal tubular dysfunction, uveitis, retinal detachment, and choroidal effusion. Additionally, signals such as osmotic demyelination syndrome and Arnold-Chiari malformation were identified as requiring further monitoring.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusion</h3>\u0000 \u0000 <p>This study identified several unexpected and serious adverse reaction signals that align with previously reported cases. These findings underscore the need for ongoing study, focused attention, and vigilant monitoring during the clinical use of topiramate.</p>\u0000 </section>\u0000 </div>","PeriodicalId":16090,"journal":{"name":"Journal of Evidence‐Based Medicine","volume":"17 4","pages":"795-807"},"PeriodicalIF":3.6,"publicationDate":"2024-12-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142846701","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}
<p>Guidelines are increasingly regarded as the “reference standards” in clinical decision making. Randomized controlled trials (RCTs) top the evidence hierarchy used in guidelines synthesis. If the RCTs show positive results, their inclusion and exclusion criteria become the treatment's indications (Class I–II). Contrariwise, the characteristics for the subgroups who experience more harm than benefit become the contra-indications (Class III). RCTs involving human participants could not be designed to show a treatment is ineffective or harmful by the Declaration of Helsinki [<span>1</span>]. All the contra-indications known for a treatment are from either totally unexpected outcomes from RCTs or rare adverse events that only emerge under long term surveillance.</p><p>RCTs were developed to provide reliable objective assessment of the effectiveness of medical treatments [<span>2</span>]. Because most medical treatments are only modestly effective, very large sample sizes and prolonged follow-up are needed for adequate statistical powers to detect treatment effects, making RCTs prohibitively expensive to run. To maximize the chance of “positive” results, RCTs set very stringent inclusion and exclusion criteria so that only patients most likely to benefit and least likely to suffer harm from the investigated treatments are enrolled. RCTs are mostly funded by commercial companies even if they are conducted by academic institutes and healthcare facilities. The stated reason of many RCTs is to compare the safety and efficacy of treatment options for a disease to help patients. The unstated reason that compels and motivates companies to fund costly RCTs is the legal requirement for regulatory approval before they could market medical products. Benefiting patients and protecting them from harm come as the incidental consequences of companies’ fiduciary duties to their shareholders. The tool (RCTs) used by companies to serve their own interests is “co-opted” by learned societies for synthesizing guidelines intended to help doctors make clinical decisions. The “misuse” of any tool for unintended purposes would likely result in sub-optimal outcomes.</p><p>While RCTs might excel in “internal validity,” they disappoint in “external validity”—their results might not be generalizable to patient populations outside [<span>3, 4</span>]. About 80% of RCTs excluded ≥50% the patients screened for enrollment [<span>5</span>]. For the excluded (unstudied) patients, the RCTs are “agnostic” on their benefit: risk balance for the treatment (Figure 1a). No evidence for benefit (Figure 1a) is not equivalent to evidence of no benefit (Figure 1b). Many patient groups are under-represented in RCTs for factors (e.g. age, co-morbidities, etc.) or protected characteristics (gender, ethnicity, etc.) which may not influence their response to a treatment. A default assumption of no benefit for all patient groups omitted from RCTs would unfairly deprive numerous potential beneficiari
{"title":"Misuse of Guidelines Could Disadvantage and Harm Patients","authors":"Ernest W Lau, Hendrik Bonnemeier, Benito Baldauf","doi":"10.1111/jebm.12666","DOIUrl":"10.1111/jebm.12666","url":null,"abstract":"<p>Guidelines are increasingly regarded as the “reference standards” in clinical decision making. Randomized controlled trials (RCTs) top the evidence hierarchy used in guidelines synthesis. If the RCTs show positive results, their inclusion and exclusion criteria become the treatment's indications (Class I–II). Contrariwise, the characteristics for the subgroups who experience more harm than benefit become the contra-indications (Class III). RCTs involving human participants could not be designed to show a treatment is ineffective or harmful by the Declaration of Helsinki [<span>1</span>]. All the contra-indications known for a treatment are from either totally unexpected outcomes from RCTs or rare adverse events that only emerge under long term surveillance.</p><p>RCTs were developed to provide reliable objective assessment of the effectiveness of medical treatments [<span>2</span>]. Because most medical treatments are only modestly effective, very large sample sizes and prolonged follow-up are needed for adequate statistical powers to detect treatment effects, making RCTs prohibitively expensive to run. To maximize the chance of “positive” results, RCTs set very stringent inclusion and exclusion criteria so that only patients most likely to benefit and least likely to suffer harm from the investigated treatments are enrolled. RCTs are mostly funded by commercial companies even if they are conducted by academic institutes and healthcare facilities. The stated reason of many RCTs is to compare the safety and efficacy of treatment options for a disease to help patients. The unstated reason that compels and motivates companies to fund costly RCTs is the legal requirement for regulatory approval before they could market medical products. Benefiting patients and protecting them from harm come as the incidental consequences of companies’ fiduciary duties to their shareholders. The tool (RCTs) used by companies to serve their own interests is “co-opted” by learned societies for synthesizing guidelines intended to help doctors make clinical decisions. The “misuse” of any tool for unintended purposes would likely result in sub-optimal outcomes.</p><p>While RCTs might excel in “internal validity,” they disappoint in “external validity”—their results might not be generalizable to patient populations outside [<span>3, 4</span>]. About 80% of RCTs excluded ≥50% the patients screened for enrollment [<span>5</span>]. For the excluded (unstudied) patients, the RCTs are “agnostic” on their benefit: risk balance for the treatment (Figure 1a). No evidence for benefit (Figure 1a) is not equivalent to evidence of no benefit (Figure 1b). Many patient groups are under-represented in RCTs for factors (e.g. age, co-morbidities, etc.) or protected characteristics (gender, ethnicity, etc.) which may not influence their response to a treatment. A default assumption of no benefit for all patient groups omitted from RCTs would unfairly deprive numerous potential beneficiari","PeriodicalId":16090,"journal":{"name":"Journal of Evidence‐Based Medicine","volume":"17 4","pages":"705-707"},"PeriodicalIF":3.6,"publicationDate":"2024-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11684497/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142837167","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}
Ilari Kuitunen, Ville T. Ponkilainen, Rasmus Liukkonen, Lauri Nyrhi, Oskari Pakarinen, Matias Vaajala, Mikko M. Uimonen
<p>Systematic reviews and meta-analyses are a key part of evidence synthesis and are considered to provide the best possible information on intervention effectiveness [<span>1</span>]. A key part of the evidence synthesis is the critical appraisal of the included studies [<span>2</span>]. The risk of bias is typically assessed by using Cochrane's risk of bias original tool or the revised risk of bias 2.0 tool, both of which are outcome specific tools for randomized controlled trials (RCTs) [<span>3, 4</span>]. Risk of bias assessments are time-consuming in evidence synthesis projects [<span>5</span>]. Additionally, they have been shown to be susceptible to biases, even in top-tier medical journals and Cochrane reviews [<span>6-8</span>]. The interrater agreement has also shown to be varying between reviewers [<span>9</span>]. Therefore, there is a clear need for improvements in both the quality and efficiency of these evaluations.</p><p>The rise of large language models, such as OpenAI's ChatGPT, has led to an increase in the use of these in research. While challenges such as authorship disputes and data fabrication have arisen, these tools show great promise when used appropriately [<span>10</span>]. Two previous studies have evaluated the performance of ChatGPT in risk of bias assessments [<span>11, 12</span>]. One focused on ROBINS-I tool and found rather low agreement in it [<span>11</span>]. Another small study focused on risk of bias (RoB) 2.0 tool, and concluded that currently ChatGPT should not be used, but further studies would be needed [<span>12</span>]. The aim of our current study was to evaluate the performance of the most recent version of OpenAI's large language model ChatGPT-4o in the risk of bias assessment.</p><p>We conducted a systematic assessment of the performance of ChatGPT-4o in Cochranes RoB 2.0 tool analyses. First, we searched PubMed on July 31, 2024 for the most recent 50 meta-analyses published in top-level medical journals (<i>Lancet</i>, <i>JAMA</i> or <i>BMJ</i>). The results were uploaded to Covidence software for a screening process. Then, two authors (IK and OP) screened the reviews and included meta-analyses of interventions, which included only RCTs, and had used Cochrane RoB 2.0 tool as their risk of bias assessment tool. A total of six reviews were included (Figure S1). Then a third author (MV) extracted a total of 100 risk of bias assessments from these included reviews. A fourth author (LN) uploaded these 100 studies in pdf format to ChatGPT-4o with a standardized short prompt which was written to the text field. The prompt was: “Perform a risk of bias analysis according to the Cochrane group RoB2 guidelines for the following article and perform the assessment for the main outcome of the trial. Report results only as high, some concerns, low, no information for domains 1–5 and an overall assessment.” The complete list of the included RCTs and extracted risk of bias assessments and ChatGPT-4o assessments i
{"title":"Evaluating the Performance of ChatGPT-4o in Risk of Bias Assessments","authors":"Ilari Kuitunen, Ville T. Ponkilainen, Rasmus Liukkonen, Lauri Nyrhi, Oskari Pakarinen, Matias Vaajala, Mikko M. Uimonen","doi":"10.1111/jebm.12662","DOIUrl":"10.1111/jebm.12662","url":null,"abstract":"<p>Systematic reviews and meta-analyses are a key part of evidence synthesis and are considered to provide the best possible information on intervention effectiveness [<span>1</span>]. A key part of the evidence synthesis is the critical appraisal of the included studies [<span>2</span>]. The risk of bias is typically assessed by using Cochrane's risk of bias original tool or the revised risk of bias 2.0 tool, both of which are outcome specific tools for randomized controlled trials (RCTs) [<span>3, 4</span>]. Risk of bias assessments are time-consuming in evidence synthesis projects [<span>5</span>]. Additionally, they have been shown to be susceptible to biases, even in top-tier medical journals and Cochrane reviews [<span>6-8</span>]. The interrater agreement has also shown to be varying between reviewers [<span>9</span>]. Therefore, there is a clear need for improvements in both the quality and efficiency of these evaluations.</p><p>The rise of large language models, such as OpenAI's ChatGPT, has led to an increase in the use of these in research. While challenges such as authorship disputes and data fabrication have arisen, these tools show great promise when used appropriately [<span>10</span>]. Two previous studies have evaluated the performance of ChatGPT in risk of bias assessments [<span>11, 12</span>]. One focused on ROBINS-I tool and found rather low agreement in it [<span>11</span>]. Another small study focused on risk of bias (RoB) 2.0 tool, and concluded that currently ChatGPT should not be used, but further studies would be needed [<span>12</span>]. The aim of our current study was to evaluate the performance of the most recent version of OpenAI's large language model ChatGPT-4o in the risk of bias assessment.</p><p>We conducted a systematic assessment of the performance of ChatGPT-4o in Cochranes RoB 2.0 tool analyses. First, we searched PubMed on July 31, 2024 for the most recent 50 meta-analyses published in top-level medical journals (<i>Lancet</i>, <i>JAMA</i> or <i>BMJ</i>). The results were uploaded to Covidence software for a screening process. Then, two authors (IK and OP) screened the reviews and included meta-analyses of interventions, which included only RCTs, and had used Cochrane RoB 2.0 tool as their risk of bias assessment tool. A total of six reviews were included (Figure S1). Then a third author (MV) extracted a total of 100 risk of bias assessments from these included reviews. A fourth author (LN) uploaded these 100 studies in pdf format to ChatGPT-4o with a standardized short prompt which was written to the text field. The prompt was: “Perform a risk of bias analysis according to the Cochrane group RoB2 guidelines for the following article and perform the assessment for the main outcome of the trial. Report results only as high, some concerns, low, no information for domains 1–5 and an overall assessment.” The complete list of the included RCTs and extracted risk of bias assessments and ChatGPT-4o assessments i","PeriodicalId":16090,"journal":{"name":"Journal of Evidence‐Based Medicine","volume":"17 4","pages":"700-702"},"PeriodicalIF":3.6,"publicationDate":"2024-12-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11684499/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142828877","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}
Pengfei Ding, Dingding Zhang, Haiping Ling, Tao Tao, Yongyue Gao, Yunfeng Wang, Huasheng Zhang, Lingyun Wu, Chunhua Hang, Wei Li
Objective
The objective of this study was to determine whether insulin resistance (IR) could be used as a predictor of poor prognosis at 3 months after subarachnoid hemorrhage (SAH).
Methods
The study included patients aged 18 years or older with a confirmed diagnosis of SAH due to ruptured aneurysm from January 2021 to March 2024. Patients with confirmed diabetes mellitus and taking glucose-lowering drugs, or taking lipid-lowering drugs, or SAH not due to ruptured aneurysm, or comorbid systemic diseases were excluded. Patients were classified into good prognosis (modified Rankin scale [MRS] 0–2) and poor prognosis (MRS 3–6). Receiver operating characteristic curve (ROC), least absolute shrinkage and selection operator (LASSO) analysis, and multivariate logistic regression analysis were used to determine the potential of triglyceride-glucose (TyG) index and the triglyceride to high-density lipoprotein cholesterol (TG/HDL) ratio as predictors of poor prognosis. Finally, a prognostic prediction model based on IR was constructed.
Results
A total of 358 patients were included in this study. Poor prognosis patients had higher age, BMI, hypertension percentage, glucose, triglycerides, TyG index and TG/HDL ratio, and lower HDL. ROC, LASSO, and multivariate logistic regression analysis revealed that age, glucose, TyG index, and TG/HDL ratio had significant potential to predict the prognosis of SAH patients. The prognostic prediction model constructed by combining age, glucose, TyG index, and TG/HDL ratio had high discriminatory power (area under the curve [AUC] = 0.80), satisfactory calibration curves, and good clinical utility.
Conclusion
IR is strongly associated with the prognosis of SAH patients, and the combination of age, glucose, TyG index, and TG/HDL ratio can provide a new direction for future treatment.
{"title":"Insulin Resistance Predicts Prognosis in Patients With Subarachnoid Hemorrhage","authors":"Pengfei Ding, Dingding Zhang, Haiping Ling, Tao Tao, Yongyue Gao, Yunfeng Wang, Huasheng Zhang, Lingyun Wu, Chunhua Hang, Wei Li","doi":"10.1111/jebm.12660","DOIUrl":"10.1111/jebm.12660","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Objective</h3>\u0000 \u0000 <p>The objective of this study was to determine whether insulin resistance (IR) could be used as a predictor of poor prognosis at 3 months after subarachnoid hemorrhage (SAH).</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>The study included patients aged 18 years or older with a confirmed diagnosis of SAH due to ruptured aneurysm from January 2021 to March 2024. Patients with confirmed diabetes mellitus and taking glucose-lowering drugs, or taking lipid-lowering drugs, or SAH not due to ruptured aneurysm, or comorbid systemic diseases were excluded. Patients were classified into good prognosis (modified Rankin scale [MRS] 0–2) and poor prognosis (MRS 3–6). Receiver operating characteristic curve (ROC), least absolute shrinkage and selection operator (LASSO) analysis, and multivariate logistic regression analysis were used to determine the potential of triglyceride-glucose (TyG) index and the triglyceride to high-density lipoprotein cholesterol (TG/HDL) ratio as predictors of poor prognosis. Finally, a prognostic prediction model based on IR was constructed.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>A total of 358 patients were included in this study. Poor prognosis patients had higher age, BMI, hypertension percentage, glucose, triglycerides, TyG index and TG/HDL ratio, and lower HDL. ROC, LASSO, and multivariate logistic regression analysis revealed that age, glucose, TyG index, and TG/HDL ratio had significant potential to predict the prognosis of SAH patients. The prognostic prediction model constructed by combining age, glucose, TyG index, and TG/HDL ratio had high discriminatory power (area under the curve [AUC] = 0.80), satisfactory calibration curves, and good clinical utility.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusion</h3>\u0000 \u0000 <p>IR is strongly associated with the prognosis of SAH patients, and the combination of age, glucose, TyG index, and TG/HDL ratio can provide a new direction for future treatment.</p>\u0000 </section>\u0000 </div>","PeriodicalId":16090,"journal":{"name":"Journal of Evidence‐Based Medicine","volume":"17 4","pages":"771-781"},"PeriodicalIF":3.6,"publicationDate":"2024-12-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142828878","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}
Giacomo Visioli, Maria Pia Pirraglia, Alessandro Lambiase, Magda Gharbiya
<p>Misrepresentation of scientific findings can lead to an overestimation of a medical issue, a phenomenon exacerbated when the scientific community is eager for information on a novel pathogen. The COVID-19 pandemic has led to an unprecedented growth in research output, including numerous studies on potential ocular manifestations [<span>1</span>]. The identification of conjunctivitis as an early symptom of COVID-19 naturally prompted questions about whether SARS-CoV-2 could affect other ocular structures [<span>2</span>]. Initial reports suggesting retinal involvement generated significant interest and debate within the medical community. While such inquiries were legitimate, it is important to approach them with scientific rigor to avoid drawing unwarranted conclusions.</p><p>Four years ago, we conducted a study on retinal findings in 46 patients with severe COVID-19 pneumonia. Our conclusion was unequivocal: we found no retinal alterations attributable to SARS-CoV-2 infection [<span>3</span>]. Instead, the observed changes were likely due to systemic conditions such as hypertension or diabetes. Surprisingly, an analysis of the past 2 years' citations revealed that 41.7% misrepresented our article, citing it to claim we supported COVID-19-related retinal findings, despite us stating the opposite. These distortions occurred across journals regardless of their prestige, as detailed in Table 1.</p><p>This case study underscores a critical issue in scientific interpretation: the assumption that simultaneous occurrence indicates causation [<span>4</span>]. Observing retinal abnormalities in patients with COVID-19 does not necessarily mean that SARS-CoV-2 is the etiological or predisposing factor. Especially during a pandemic, when a significant portion of the global population is infected, coincidental occurrences are statistically more likely.</p><p>One method to establish a causal relationship is to demonstrate an increased incidence of a condition that correlates specifically with the infection [<span>5, 6</span>]. However, after 4 years of extensive research, no definitive evidence has emerged to support an increased incidence of retinal pathology directly linked to SARS-CoV-2 [<span>7</span>]. Findings such as retinal thrombosis or cotton wool spots are more plausibly explained by systemic conditions or comorbidities common in severely ill patients rather than a direct pathogenic role of the virus [<span>8</span>].</p><p>Furthermore, it remains unclear whether the ocular findings reported in COVID-19 patients represent a problem of significant medical relevance. In many cases, these retinal changes are minor, asymptomatic, and do not necessitate specific treatment. Overstating such findings can misdirect scientific focus and may lead to unnecessary alarm among patients. Moreover, there were some highly cited early reports during the pandemic that claimed to identify retinal abnormalities in COVID-19 patients, but subsequent scrutiny revealed t
{"title":"Misleading Citations and Publication Bias in COVID-19 in Ophthalmology","authors":"Giacomo Visioli, Maria Pia Pirraglia, Alessandro Lambiase, Magda Gharbiya","doi":"10.1111/jebm.12664","DOIUrl":"10.1111/jebm.12664","url":null,"abstract":"<p>Misrepresentation of scientific findings can lead to an overestimation of a medical issue, a phenomenon exacerbated when the scientific community is eager for information on a novel pathogen. The COVID-19 pandemic has led to an unprecedented growth in research output, including numerous studies on potential ocular manifestations [<span>1</span>]. The identification of conjunctivitis as an early symptom of COVID-19 naturally prompted questions about whether SARS-CoV-2 could affect other ocular structures [<span>2</span>]. Initial reports suggesting retinal involvement generated significant interest and debate within the medical community. While such inquiries were legitimate, it is important to approach them with scientific rigor to avoid drawing unwarranted conclusions.</p><p>Four years ago, we conducted a study on retinal findings in 46 patients with severe COVID-19 pneumonia. Our conclusion was unequivocal: we found no retinal alterations attributable to SARS-CoV-2 infection [<span>3</span>]. Instead, the observed changes were likely due to systemic conditions such as hypertension or diabetes. Surprisingly, an analysis of the past 2 years' citations revealed that 41.7% misrepresented our article, citing it to claim we supported COVID-19-related retinal findings, despite us stating the opposite. These distortions occurred across journals regardless of their prestige, as detailed in Table 1.</p><p>This case study underscores a critical issue in scientific interpretation: the assumption that simultaneous occurrence indicates causation [<span>4</span>]. Observing retinal abnormalities in patients with COVID-19 does not necessarily mean that SARS-CoV-2 is the etiological or predisposing factor. Especially during a pandemic, when a significant portion of the global population is infected, coincidental occurrences are statistically more likely.</p><p>One method to establish a causal relationship is to demonstrate an increased incidence of a condition that correlates specifically with the infection [<span>5, 6</span>]. However, after 4 years of extensive research, no definitive evidence has emerged to support an increased incidence of retinal pathology directly linked to SARS-CoV-2 [<span>7</span>]. Findings such as retinal thrombosis or cotton wool spots are more plausibly explained by systemic conditions or comorbidities common in severely ill patients rather than a direct pathogenic role of the virus [<span>8</span>].</p><p>Furthermore, it remains unclear whether the ocular findings reported in COVID-19 patients represent a problem of significant medical relevance. In many cases, these retinal changes are minor, asymptomatic, and do not necessitate specific treatment. Overstating such findings can misdirect scientific focus and may lead to unnecessary alarm among patients. Moreover, there were some highly cited early reports during the pandemic that claimed to identify retinal abnormalities in COVID-19 patients, but subsequent scrutiny revealed t","PeriodicalId":16090,"journal":{"name":"Journal of Evidence‐Based Medicine","volume":"17 4","pages":"703-704"},"PeriodicalIF":3.6,"publicationDate":"2024-12-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11684496/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142828803","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}
Yaxuan Ren, Xufei Luo, Ye Wang, Haodong Li, Hairong Zhang, Zeming Li, Honghao Lai, Xuanlin Li, Long Ge, Janne ESTILL, Lu Zhang, Shu Yang, Yaolong Chen, Chengping Wen, Zhaoxiang Bian, ADVANCED Working Group
Background
The application of large language models (LLMs) in medicine has received increasing attention, showing significant potential in teaching, research, and clinical practice, especially in knowledge extraction, management, and understanding. However, the use of LLMs in Traditional Chinese Medicine (TCM) has not been thoroughly studied. This study aims to provide a comprehensive overview of the status and challenges of LLM applications in TCM.
Methods
A systematic search of five electronic databases and Google Scholar was conducted between November 2022 and April 2024, using the Arksey and O'Malley five-stage framework to identify relevant studies. Data from eligible studies were comprehensively extracted and organized to describe LLM applications in TCM and assess their performance accuracy.
Results
A total of 29 studies were identified: 24 peer-reviewed articles, 1 review, and 4 preprints. Two core application areas were found: the extraction, management, and understanding of TCM knowledge, and assisted diagnosis and treatment. LLMs developed specifically for TCM achieved 70% accuracy in the TCM Practitioner Exam, while general-purpose Chinese LLMs achieved 60% accuracy. Common international LLMs did not pass the exam. Models like EpidemicCHAT and MedChatZH, trained on customized TCM corpora, outperformed general LLMs in TCM consultation.
Conclusion
Despite their potential, LLMs in TCM face challenges such as data quality and security issues, the specificity and complexity of TCM data, and the nonquantitative nature of TCM diagnosis and treatment. Future efforts should focus on interdisciplinary talent cultivation, enhanced data standardization and protection, and exploring LLM potential in multimodal interaction and intelligent diagnosis and treatment.
背景：大语言模型（large language models, LLMs）在医学领域的应用越来越受到关注，在教学、研究和临床实践中，特别是在知识提取、管理和理解方面显示出巨大的潜力。然而，llm在中医中的应用尚未得到深入的研究。本研究旨在全面概述法学硕士在中医领域应用的现状和面临的挑战。方法：采用Arksey和O'Malley五阶段框架，对2022年11月至2024年4月期间的5个电子数据库和谷歌Scholar进行系统检索，识别相关研究。从符合条件的研究中全面提取和组织数据，以描述LLM在中医中的应用并评估其性能准确性。结果：共确定了29项研究：24篇同行评议文章，1篇综述和4篇预印本。发现了两个核心应用领域：中医知识的提取、管理和理解，以及辅助诊断和治疗。专门为中医开发的法学硕士在中医执业医师考试中准确率达到70%，而中国通用法学硕士的准确率为60%。一般的国际法学硕士不通过考试。EpidemicCHAT和MedChatZH等模型在定制中医语料库上进行了培训，在中医咨询方面的表现优于普通法学硕士。结论：尽管具有潜力，但中医法学硕士仍面临数据质量和安全问题、中医数据的特异性和复杂性、中医诊疗的非定量化等挑战。未来应注重跨学科人才培养，加强数据标准化和保护，挖掘法学硕士在多模态交互和智能诊疗方面的潜力。
{"title":"Large Language Models in Traditional Chinese Medicine: A Scoping Review","authors":"Yaxuan Ren, Xufei Luo, Ye Wang, Haodong Li, Hairong Zhang, Zeming Li, Honghao Lai, Xuanlin Li, Long Ge, Janne ESTILL, Lu Zhang, Shu Yang, Yaolong Chen, Chengping Wen, Zhaoxiang Bian, ADVANCED Working Group","doi":"10.1111/jebm.12658","DOIUrl":"10.1111/jebm.12658","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>The application of large language models (LLMs) in medicine has received increasing attention, showing significant potential in teaching, research, and clinical practice, especially in knowledge extraction, management, and understanding. However, the use of LLMs in Traditional Chinese Medicine (TCM) has not been thoroughly studied. This study aims to provide a comprehensive overview of the status and challenges of LLM applications in TCM.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>A systematic search of five electronic databases and Google Scholar was conducted between November 2022 and April 2024, using the Arksey and O'Malley five-stage framework to identify relevant studies. Data from eligible studies were comprehensively extracted and organized to describe LLM applications in TCM and assess their performance accuracy.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>A total of 29 studies were identified: 24 peer-reviewed articles, 1 review, and 4 preprints. Two core application areas were found: the extraction, management, and understanding of TCM knowledge, and assisted diagnosis and treatment. LLMs developed specifically for TCM achieved 70% accuracy in the TCM Practitioner Exam, while general-purpose Chinese LLMs achieved 60% accuracy. Common international LLMs did not pass the exam. Models like EpidemicCHAT and MedChatZH, trained on customized TCM corpora, outperformed general LLMs in TCM consultation.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusion</h3>\u0000 \u0000 <p>Despite their potential, LLMs in TCM face challenges such as data quality and security issues, the specificity and complexity of TCM data, and the nonquantitative nature of TCM diagnosis and treatment. Future efforts should focus on interdisciplinary talent cultivation, enhanced data standardization and protection, and exploring LLM potential in multimodal interaction and intelligent diagnosis and treatment.</p>\u0000 </section>\u0000 </div>","PeriodicalId":16090,"journal":{"name":"Journal of Evidence‐Based Medicine","volume":"18 1","pages":""},"PeriodicalIF":3.6,"publicationDate":"2024-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142801123","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}
This study aimed to evaluate the acceptability and effects of internet-based cognitive behavioral therapy (iCBT) or CBT-oriented interventions compared with control groups on depressive symptoms, remission of depression, and quality of life (QOL) in adolescents.
Methods
We searched English and Chinese databases for randomized controlled trials up to October 10, 2024 that investigated the effects of iCBT compared with controls in adolescents exhibiting elevated depressive symptoms or diagnosed with depression. Standardized mean differences (SMDs), relative risks (RRs), and 95% confidence intervals were applied to evaluate the pooled effects of outcomes.
Results
A total of 19 RCTs involving 3574 cases were included in this study. We found small effects on depressive symptoms severity at different time points (posttest: SMD = –0.49 [–0.66, –0.33]; 3-month follow-up [FU3]: SMD = –0.21 [–0.30, –0.11]; FU6: SMD = –0.18 [–0.35, –0.02]; FU12: SMD = –0.38 [–0.56, –0.20]). We also found a significant effect in depression remission rate at the posttest (RR = 1.74 [1.36, 2.21]) and a significant effect in QOL at the posttest (SMD = 0.30 [0.07, 0.54]). However, the result regarding acceptability was nonsignificant (RR = 1.22 [0.76, 1.97]). No significant publication bias was found in these results.
Conclusion
iCBT or internet-based CBT-oriented interventions can effectively reduce depressive symptom severity and improve depression remission rate and QOL in depressed adolescents. These results are preliminary and require further validation through future systematic reviews.
{"title":"The Acceptability and Effects of Internet-Based Cognitive Behavioral Therapy on Depressive Symptoms and Remission in 13- to 17-Year-Old Adolescents: A Systematic Review and Meta-Analysis","authors":"Pei Liu, Mingjing Situ, Xiaoxia Duan, Huan Sun, Meiwen Wang, Yujie Tao, Tingting Luo, Sixun Li, Zhuo Wang, Yi Huang","doi":"10.1111/jebm.12657","DOIUrl":"10.1111/jebm.12657","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Objective</h3>\u0000 \u0000 <p>This study aimed to evaluate the acceptability and effects of internet-based cognitive behavioral therapy (iCBT) or CBT-oriented interventions compared with control groups on depressive symptoms, remission of depression, and quality of life (QOL) in adolescents.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>We searched English and Chinese databases for randomized controlled trials up to October 10, 2024 that investigated the effects of iCBT compared with controls in adolescents exhibiting elevated depressive symptoms or diagnosed with depression. Standardized mean differences (SMDs), relative risks (RRs), and 95% confidence intervals were applied to evaluate the pooled effects of outcomes.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>A total of 19 RCTs involving 3574 cases were included in this study. We found small effects on depressive symptoms severity at different time points (posttest: SMD = –0.49 [–0.66, –0.33]; 3-month follow-up [FU3]: SMD = –0.21 [–0.30, –0.11]; FU6: SMD = –0.18 [–0.35, –0.02]; FU12: SMD = –0.38 [–0.56, –0.20]). We also found a significant effect in depression remission rate at the posttest (RR = 1.74 [1.36, 2.21]) and a significant effect in QOL at the posttest (SMD = 0.30 [0.07, 0.54]). However, the result regarding acceptability was nonsignificant (RR = 1.22 [0.76, 1.97]). No significant publication bias was found in these results.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusion</h3>\u0000 \u0000 <p>iCBT or internet-based CBT-oriented interventions can effectively reduce depressive symptom severity and improve depression remission rate and QOL in depressed adolescents. These results are preliminary and require further validation through future systematic reviews.</p>\u0000 </section>\u0000 </div>","PeriodicalId":16090,"journal":{"name":"Journal of Evidence‐Based Medicine","volume":"17 4","pages":"740-757"},"PeriodicalIF":3.6,"publicationDate":"2024-11-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142710477","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}
Xiaohui Pan, Hongbo He, Yuqian Bao, Yan Bi, Luyuan Chen, Xiaoping Chen, Hui Fang, Wenhuan Feng, Ling Gao, Lixin Guo, Yifang Guo, Yaling Han, Qi Hua, Nanfang Li, Quanmin Li, Yan Li, Yong Li, Xialian Li, Jing Liu, Huijuan Ma, Jianjun Mu, Kailei Nong, Huiqian Shang, Yunfeng Shen, Zhongwei Shi, Fang Sun, Ningling Sun, Jun Tao, Jiguang Wang, Xinling Wang, Jing Wu, Xinhua Xiao, Liangdi Xie, Jing Xu, Jing Xu, Hongying Ye, Dongni Yu, Hong Yuan, Huijie Zhang, Jian Zhang, Lili Zhang, Yuqing Zhang, Jiaqiang Zhou, Xinli Zhou, Dalong Zhu, Tiehong Zhu, Sheyu Li, Zhiming Zhu, Diabetes and Obesity Research Group of Chinese Diabetes Society (CDS) and Chinese Hypertension League (CHL)
Both hypertension and type 2 diabetes are attributable to premature death, cardiovascular and kidney diseases with largely overlapping population. Followed the GRADE approach, this expert consensus aimed to reduce the cardiovascular and kidney death and disability due to hypertension and minimize the treatment burden in adults with type 2 diabetes. Through online survey and discussion, a multidisciplinary team comprehensively prioritized seven key guideline questions. Informed by the evidence synthesis and online discussion, the team developed 12 recommendations under the GRADE Evidence-to-decision (EtD) framework. The recommendations covered the screening of hypertension in adults diagnosed with type 2 diabetes but not hypertension and the monitoring, lifestyle interventions, and medications in those diagnosed with type 2 diabetes and hypertension.
{"title":"Chinese expert consensus on the management of hypertension in adults with type 2 diabetes","authors":"Xiaohui Pan, Hongbo He, Yuqian Bao, Yan Bi, Luyuan Chen, Xiaoping Chen, Hui Fang, Wenhuan Feng, Ling Gao, Lixin Guo, Yifang Guo, Yaling Han, Qi Hua, Nanfang Li, Quanmin Li, Yan Li, Yong Li, Xialian Li, Jing Liu, Huijuan Ma, Jianjun Mu, Kailei Nong, Huiqian Shang, Yunfeng Shen, Zhongwei Shi, Fang Sun, Ningling Sun, Jun Tao, Jiguang Wang, Xinling Wang, Jing Wu, Xinhua Xiao, Liangdi Xie, Jing Xu, Jing Xu, Hongying Ye, Dongni Yu, Hong Yuan, Huijie Zhang, Jian Zhang, Lili Zhang, Yuqing Zhang, Jiaqiang Zhou, Xinli Zhou, Dalong Zhu, Tiehong Zhu, Sheyu Li, Zhiming Zhu, Diabetes and Obesity Research Group of Chinese Diabetes Society (CDS) and Chinese Hypertension League (CHL)","doi":"10.1111/jebm.12655","DOIUrl":"10.1111/jebm.12655","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <p>Both hypertension and type 2 diabetes are attributable to premature death, cardiovascular and kidney diseases with largely overlapping population. Followed the GRADE approach, this expert consensus aimed to reduce the cardiovascular and kidney death and disability due to hypertension and minimize the treatment burden in adults with type 2 diabetes. Through online survey and discussion, a multidisciplinary team comprehensively prioritized seven key guideline questions. Informed by the evidence synthesis and online discussion, the team developed 12 recommendations under the GRADE Evidence-to-decision (EtD) framework. The recommendations covered the screening of hypertension in adults diagnosed with type 2 diabetes but not hypertension and the monitoring, lifestyle interventions, and medications in those diagnosed with type 2 diabetes and hypertension.</p>\u0000 </section>\u0000 </div>","PeriodicalId":16090,"journal":{"name":"Journal of Evidence‐Based Medicine","volume":"17 4","pages":"851-864"},"PeriodicalIF":3.6,"publicationDate":"2024-11-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142621992","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号