目的慢性肝病患者常表现出独特的血流动力学异常与代谢紊乱,术后液体管理面临诸多挑战,尤其是术后入住重症监护室(intensive care unit,ICU)的重症患者,亟待深入探究术后液体治疗方案与预后之间的关系。方法本研究基于MIMIC-IV数据库中2...目的慢性肝病患者常表现出独特的血流动力学异常与代谢紊乱,术后液体管理面临诸多挑战,尤其是术后入住重症监护室(intensive care unit,ICU)的重症患者,亟待深入探究术后液体治疗方案与预后之间的关系。方法本研究基于MIMIC-IV数据库中2414名慢性肝病、接受手术治疗,并术后转入ICU的患者,对纳排后最终得到的2143名患者数据进行回顾性队列研究。采用多变量调整Logistic回归模型,分析术后转入ICU首日液体治疗方案与术后7天死亡风险的关联,并通过限制性立方样条(restricted cubic spline,RCS)分析剂量-反应关系。结果多因素分析指出限制性补液为独立保护因素,相较于非限制性补液组,限制性补液显著降低了术后7天死亡率(6.4%vs 12.4%,OR=0.44,95%CI:0.22~0.88,P=0.021)。减少了机械通气的使用(42.9%vs 72.3%,OR=0.29,95%CI:0.24~0.35,P<0.001)和ICU停留时长(1.86 d vs 3.47 d,OR=0.81,95%CI:0.78~0.84,P<0.001)。RCS曲线显示,术后首日液体入量与术后7天死亡风险呈现J型曲线关系,拐点为1850 mL,超过该阈值后,术后7天死亡风险随之增加。亚组分析结果表明,限制性补液的保护作用在不同年龄、合并症群体中均呈现出一致性。结论慢性肝病患者术后首日采取限制性补液方案可有效降低短期死亡风险,且液体入量与7天死亡风险呈非线性剂量效应关联,液体入量超过1850 mL时,死亡风险显著升高。展开更多
Background:Given the strikingly high diagnostic error rate in hospitals,and the recent development of Large Language Models(LLMs),we set out to measure the diagnostic sensitivity of two popular LLMs:GPT-4 and PaLM2.Sm...Background:Given the strikingly high diagnostic error rate in hospitals,and the recent development of Large Language Models(LLMs),we set out to measure the diagnostic sensitivity of two popular LLMs:GPT-4 and PaLM2.Small-scale studies to evaluate the diagnostic ability of LLMs have shown promising results,with GPT-4 demonstrating high accuracy in diagnosing test cases.However,larger evaluations on real electronic patient data are needed to provide more reliable estimates.Methods:To fill this gap in the literature,we used a deidentified Electronic Health Record(EHR)data set of about 300,000 patients admitted to the Beth Israel Deaconess Medical Center in Boston.This data set contained blood,imaging,microbiology and vital sign information as well as the patients'medical diagnostic codes.Based on the available EHR data,doctors curated a set of diagnoses for each patient,which we will refer to as ground truth diagnoses.We then designed carefully-written prompts to get patient diagnostic predictions from the LLMs and compared this to the ground truth diagnoses in a random sample of 1000 patients.Results:Based on the proportion of correctly predicted ground truth diagnoses,we estimated the diagnostic hit rate of GPT-4 to be 93.9%.PaLM2 achieved 84.7%on the same data set.On these 1000 randomly selected EHRs,GPT-4 correctly identified 1116 unique diagnoses.Conclusion:The results suggest that artificial intelligence(AI)has the potential when working alongside clinicians to reduce cognitive errors which lead to hundreds of thousands of misdiagnoses every year.However,human oversight of AI remains essential:LLMs cannot replace clinicians,especially when it comes to human understanding and empathy.Furthermore,a significant number of challenges in incorporating AI into health care exist,including ethical,liability and regulatory barriers.展开更多
Background Acute kidney injury(AKl)is a common and serious complication following coronary artery bypass grafting(CABG),with reported incidence rates ranging from 4%to 28%.Red cell distribution width(RDW),calculated a...Background Acute kidney injury(AKl)is a common and serious complication following coronary artery bypass grafting(CABG),with reported incidence rates ranging from 4%to 28%.Red cell distribution width(RDW),calculated as the standard deviation of erythrocyte size divided by mean corpuscular volume,reflects the heterogeneity in red blood cell volume.Easily obtained from routine blood tests,RDW is increasingly recognized as a prognostic marker for various adverse clinical outcomes.Previous studies have also indicated a correlation between elevated RDW levels and declining renal function,as measured by estimated glomerular filtration rate.However,the specific relationship between RDW and the risk of AKI in patients undergoing CABG remains inadequately characterized.This study aimed to investigate the association between RDW and AKI incidence in this surgical population.Methods I n this secondary retrospective cohort study,data of patients who underwent CABG were extracted from the Medical Information Mart for Intensive Care(MIMIC)-IV database(2008-2019).The primary outcome was the occurrence of AKI as a postoperative complication following CABG,which was identified after intensive care unit(ICU)admission.RDW was measured within 24 hours on admission to ICU.AKI was defined according to Kidney Disease Improving Global Outcomes(KDIGO)criteria.Multivariable logistic regression models were used to adjust for demographics,comorbidities,and critical laboratory markers.Patients were stratified into RDW tertiles(T1:≤13.1%;T2:13.1%-13.8%;T3:≥13.8%).Multivariate logistic regression and subgroup analyses were applied to explore the association of the RDW with the risk of AKI.Furthermore,we also examined the association of RDW with AKI by employing restricted cubic splines(RCS).Results A total of 3,388 patients who underwent CABG were included in this study,of whom 2563(75.65%)had AKI.According to the multivariate regression models,the RDW demonstrated a significant positive correlation with the risk of AKI after comprehensive covariate adjustment.When each unit of RDW was increased,the risk of AKI in patients with CABG increased 10%(OR:1.10,95%CI:1.03-1.18,P=0.0063).In addition,When the RDW was analyzed by tertiles,patients in the high tertile(OR:1.33,95%CI:1.08-1.64,P=0.0072)presented a progressively higher risk of AKI than those in the low tertile,with a significant dose-response trend(P for trend<0.01).The subgroup analysis demonstrated that the positive association between elevated RDW and increased risk of AKI following CABG was robust and consistent across most major patient subgroups defined by demographics and comorbidities.The strength of this association was notably amplified in patients with pre-existing renal disease and in older patients(>65 years)and White race(P<0.05),suggesting a particularly prominent relationship between RDW and AKI risk in these specific populations.The lack of statistical significance in peripheral vascular disease subgroups required further investigation(P>0.05).Conclusions In ICU patients who received coronary revascularization by CABG,an elevated RDW demonstrated an positive correlation with the risk of AKI.[S Chin J Cardiol 2025;26(4):228238]展开更多
Background Acute kidney injury(AKI)is a common and serious complication following coronary artery bypass grafting(CABG),with an incidence rate ranging from 4%to 28%.The estimated plasma volume status(ePVS)-a novel mar...Background Acute kidney injury(AKI)is a common and serious complication following coronary artery bypass grafting(CABG),with an incidence rate ranging from 4%to 28%.The estimated plasma volume status(ePVS)-a novel marker calculated from routine hematocrit and hemoglobin levels-reflects both volume overload and hemodilution,which are potential contributors to renal impairment.Nevertheless,the relationship between ePVS and AKI in patients undergoing CABG remains poorly understood.Therefore,this study aimed to investigate the association of ePVS with the risk of AKI in adult patients who underwent CABG.Methods This retrospective cohort study utilized data from the Medical Information Mart for Intensive Care(MIMIC)-IV database,covering the period from 2008 to 2019.The primary outcome was the occurrence of AKI following admission to the intensive care unit(ICU).Hematocrit and hemoglobin levels were measured within 24 hours after ICU admission.The ePVS was calculated using the Strauss-derived Duarte formula:ePVS=[100-hematocrit(%)]/hemoglobin(g/dL).AKI was defined in accordance with the Kidney Disease Improving Global Outcomes(KDIGO)criteria.Multivariable logistic regression models were used to adjust for demographics,comorbidities,and critical laboratory markers.Patients were stratified into three groups based on the ePVS tertiles(low:≤6.30;middle:6.30-8.08;high:>8.08).Multivariate logistic regression and subgroup analyses were applied to explore the association of the ePVS with the risk of AKI.Furthermore,we also examined the association of ePVS with AKI by employing generalized additive models.Results A total of 3,388 patients were included in the final analysis,of whom 2,573(75.94%)developed AKI.Following multivariable adjustment,each unit increase in ePVS was associated with a 9%increase in the odds of AKI(OR:1.09,95%CI:1.05-1.14;P<0.001).When analyzed by ePVS tertiles,the highest tertile demonstrated a significantly elevated AKI risk compared with the lowest tertile(OR:1.48,95%CI:1.18-1.86,P=0.0007),with a significant dose-response relationship observed across tertiles(P for trend<0.001).Subgroup analyses further indicated that the association between ePVS and AKI was more pronounced among patients with pre-existing renal or peripheral vascular disease and was statistically significant only in White patients.Conclusions ePVS was independently associated with an increased risk of AKI in adults undergoing CABG.These findings supported the potential utility of ePVS as a simple,economical clinical tool for early identification of patients at high risk for AKI following cardiac surgery.展开更多
目的本研究旨在评估血清肌酐/白蛋白(creatinine to albumin ratio,CAR)与坏死性胰腺炎患者死亡率之间的关系。方法本研究使用重症监护医疗信息市场(MIMIC-IV,3.1)数据库的数据进行了回顾性研究。研究包括分析各种死亡率变量并获得人院...目的本研究旨在评估血清肌酐/白蛋白(creatinine to albumin ratio,CAR)与坏死性胰腺炎患者死亡率之间的关系。方法本研究使用重症监护医疗信息市场(MIMIC-IV,3.1)数据库的数据进行了回顾性研究。研究包括分析各种死亡率变量并获得人院时的CAR值。使用ROC截断值确定CAR的最佳阈值。采用受试者工作特征分析(Receiver Operating Characteristic analysis)研究CAR对坏死性胰腺炎患者死亡率的预测能力、敏感性、特异性和曲线下面积(Area under curve,AUC),根据ROC截断值将将CAR≥0.44的坏死性胰腺炎患者分为高CAR组,CAR<0.44的坏死性胰腺炎患者分为低CAR组。使用Kaplan-Meier生存曲线评估高和低CAR组患者与死亡率之间的关系。结果共纳人156名符合标准的坏死性胰腺炎患者,采用ROC曲线评估CAR对坏死性胰腺炎患者死亡率的预测能力,截断值为0.44。根据截断值分为高CAR组和低CAR组,两组之间年龄、有无糖尿病、平均动脉压、血氯、血钾、血钙、血糖、尿素氮、肌酐、总胆红素、谷丙转氨酶、谷草转氨酶、白蛋白、白细胞计数、血小板计数、氧分压和CAR之间P<0.05,有统计学差异。两组之间住院天数、30天存活时间、60天存活时间和90天存活时间之间P<0.05,有统计学差异。R0C结果显示曲线下面积0.782,95%CI(0.678,0.886);敏感性,0.655,95%CI(0.482,0.828);特异性,0.874,95%CI(0.816,0.932);阳性预测值,0.543,95%CI(0.378,0.708);阴性预测值,0.917,95%CI(0.868,0.966)。CAR与急性生理与慢性健康评分(Acute Physiology and Chronic Health Evaluation Ⅱ,A-PACHEI)评分进行ROC比较,CARAUC值方面,0.782,95%CI(0.678,0.886),APACHEIIAUC值方面,0.642,95%CI(0.526,0.703)。对低CAR和高CAR两组坏死性胰腺炎患者30天、60天和90天的使用Kaplan-Meier法绘制生存曲线,log rank检验P<0.05,高CAR组坏死性胰腺炎的患者病死率明显高于低CAR组坏死性胰腺炎的患者。结论CAR可作为坏死性胰腺炎患者死亡率的独立预测因素,为临床评估病情严重程度、指导治疗和预后判断提供重要参考。展开更多
在ICU电子病历脓毒症早期预测任务中,传统机器学习方法因难以捕捉稀疏表格数据中的动态特征交互而导致性能受限。为此,本研究提出改进的LF-Transformer深度学习模型,通过构建基于乘法算术块的特征交互增强方法,结合交互候选生成器(ICG)...在ICU电子病历脓毒症早期预测任务中,传统机器学习方法因难以捕捉稀疏表格数据中的动态特征交互而导致性能受限。为此,本研究提出改进的LF-Transformer深度学习模型,通过构建基于乘法算术块的特征交互增强方法,结合交互候选生成器(ICG)的动态top-k特征选择机制,有效提升了模型对稀疏医疗数据的表征能力。采用MIMIC-IV数据集验证模型性能,实验结果表明改进模型在脓毒症预测任务中AUROC达到0.841,特异度0.763,敏感性0.759,显著优于传统方法。该研究成果为开发ICU脓毒症智能预警系统提供了有效的算法支持,具有一定的临床实践价值。In the task of early sepsis prediction using ICU electronic medical records, traditional machine learning methods suffer from performance limitations due to their inability to capture dynamic feature interactions in sparse tabular data. To address this, we propose an improved LF-Transformer deep learning model. By constructing a feature interaction enhancement method based on multiplicative arithmetic blocks and integrating a dynamic top-k feature selection mechanism via an Interaction Candidate Generator (ICG), the model significantly improves its representational capability for sparse medical data. Validated on the MIMIC-IV dataset, experimental results demonstrate that the enhanced model achieves an AUROC of 0.841, specificity of 0.763, and sensitivity of 0.759 in sepsis prediction, outperforming traditional methods significantly. This research provides effective algorithmic support for developing intelligent sepsis early warning systems in ICUs, offering practical clinical value.展开更多
文摘目的慢性肝病患者常表现出独特的血流动力学异常与代谢紊乱,术后液体管理面临诸多挑战,尤其是术后入住重症监护室(intensive care unit,ICU)的重症患者,亟待深入探究术后液体治疗方案与预后之间的关系。方法本研究基于MIMIC-IV数据库中2414名慢性肝病、接受手术治疗,并术后转入ICU的患者,对纳排后最终得到的2143名患者数据进行回顾性队列研究。采用多变量调整Logistic回归模型,分析术后转入ICU首日液体治疗方案与术后7天死亡风险的关联,并通过限制性立方样条(restricted cubic spline,RCS)分析剂量-反应关系。结果多因素分析指出限制性补液为独立保护因素,相较于非限制性补液组,限制性补液显著降低了术后7天死亡率(6.4%vs 12.4%,OR=0.44,95%CI:0.22~0.88,P=0.021)。减少了机械通气的使用(42.9%vs 72.3%,OR=0.29,95%CI:0.24~0.35,P<0.001)和ICU停留时长(1.86 d vs 3.47 d,OR=0.81,95%CI:0.78~0.84,P<0.001)。RCS曲线显示,术后首日液体入量与术后7天死亡风险呈现J型曲线关系,拐点为1850 mL,超过该阈值后,术后7天死亡风险随之增加。亚组分析结果表明,限制性补液的保护作用在不同年龄、合并症群体中均呈现出一致性。结论慢性肝病患者术后首日采取限制性补液方案可有效降低短期死亡风险,且液体入量与7天死亡风险呈非线性剂量效应关联,液体入量超过1850 mL时,死亡风险显著升高。
文摘Background:Given the strikingly high diagnostic error rate in hospitals,and the recent development of Large Language Models(LLMs),we set out to measure the diagnostic sensitivity of two popular LLMs:GPT-4 and PaLM2.Small-scale studies to evaluate the diagnostic ability of LLMs have shown promising results,with GPT-4 demonstrating high accuracy in diagnosing test cases.However,larger evaluations on real electronic patient data are needed to provide more reliable estimates.Methods:To fill this gap in the literature,we used a deidentified Electronic Health Record(EHR)data set of about 300,000 patients admitted to the Beth Israel Deaconess Medical Center in Boston.This data set contained blood,imaging,microbiology and vital sign information as well as the patients'medical diagnostic codes.Based on the available EHR data,doctors curated a set of diagnoses for each patient,which we will refer to as ground truth diagnoses.We then designed carefully-written prompts to get patient diagnostic predictions from the LLMs and compared this to the ground truth diagnoses in a random sample of 1000 patients.Results:Based on the proportion of correctly predicted ground truth diagnoses,we estimated the diagnostic hit rate of GPT-4 to be 93.9%.PaLM2 achieved 84.7%on the same data set.On these 1000 randomly selected EHRs,GPT-4 correctly identified 1116 unique diagnoses.Conclusion:The results suggest that artificial intelligence(AI)has the potential when working alongside clinicians to reduce cognitive errors which lead to hundreds of thousands of misdiagnoses every year.However,human oversight of AI remains essential:LLMs cannot replace clinicians,especially when it comes to human understanding and empathy.Furthermore,a significant number of challenges in incorporating AI into health care exist,including ethical,liability and regulatory barriers.
文摘Background Acute kidney injury(AKl)is a common and serious complication following coronary artery bypass grafting(CABG),with reported incidence rates ranging from 4%to 28%.Red cell distribution width(RDW),calculated as the standard deviation of erythrocyte size divided by mean corpuscular volume,reflects the heterogeneity in red blood cell volume.Easily obtained from routine blood tests,RDW is increasingly recognized as a prognostic marker for various adverse clinical outcomes.Previous studies have also indicated a correlation between elevated RDW levels and declining renal function,as measured by estimated glomerular filtration rate.However,the specific relationship between RDW and the risk of AKI in patients undergoing CABG remains inadequately characterized.This study aimed to investigate the association between RDW and AKI incidence in this surgical population.Methods I n this secondary retrospective cohort study,data of patients who underwent CABG were extracted from the Medical Information Mart for Intensive Care(MIMIC)-IV database(2008-2019).The primary outcome was the occurrence of AKI as a postoperative complication following CABG,which was identified after intensive care unit(ICU)admission.RDW was measured within 24 hours on admission to ICU.AKI was defined according to Kidney Disease Improving Global Outcomes(KDIGO)criteria.Multivariable logistic regression models were used to adjust for demographics,comorbidities,and critical laboratory markers.Patients were stratified into RDW tertiles(T1:≤13.1%;T2:13.1%-13.8%;T3:≥13.8%).Multivariate logistic regression and subgroup analyses were applied to explore the association of the RDW with the risk of AKI.Furthermore,we also examined the association of RDW with AKI by employing restricted cubic splines(RCS).Results A total of 3,388 patients who underwent CABG were included in this study,of whom 2563(75.65%)had AKI.According to the multivariate regression models,the RDW demonstrated a significant positive correlation with the risk of AKI after comprehensive covariate adjustment.When each unit of RDW was increased,the risk of AKI in patients with CABG increased 10%(OR:1.10,95%CI:1.03-1.18,P=0.0063).In addition,When the RDW was analyzed by tertiles,patients in the high tertile(OR:1.33,95%CI:1.08-1.64,P=0.0072)presented a progressively higher risk of AKI than those in the low tertile,with a significant dose-response trend(P for trend<0.01).The subgroup analysis demonstrated that the positive association between elevated RDW and increased risk of AKI following CABG was robust and consistent across most major patient subgroups defined by demographics and comorbidities.The strength of this association was notably amplified in patients with pre-existing renal disease and in older patients(>65 years)and White race(P<0.05),suggesting a particularly prominent relationship between RDW and AKI risk in these specific populations.The lack of statistical significance in peripheral vascular disease subgroups required further investigation(P>0.05).Conclusions In ICU patients who received coronary revascularization by CABG,an elevated RDW demonstrated an positive correlation with the risk of AKI.[S Chin J Cardiol 2025;26(4):228238]
基金supported by the Guangdong Provincial Medical Science and Technology Research Fund Program(No.2024112010149612)。
文摘Background Acute kidney injury(AKI)is a common and serious complication following coronary artery bypass grafting(CABG),with an incidence rate ranging from 4%to 28%.The estimated plasma volume status(ePVS)-a novel marker calculated from routine hematocrit and hemoglobin levels-reflects both volume overload and hemodilution,which are potential contributors to renal impairment.Nevertheless,the relationship between ePVS and AKI in patients undergoing CABG remains poorly understood.Therefore,this study aimed to investigate the association of ePVS with the risk of AKI in adult patients who underwent CABG.Methods This retrospective cohort study utilized data from the Medical Information Mart for Intensive Care(MIMIC)-IV database,covering the period from 2008 to 2019.The primary outcome was the occurrence of AKI following admission to the intensive care unit(ICU).Hematocrit and hemoglobin levels were measured within 24 hours after ICU admission.The ePVS was calculated using the Strauss-derived Duarte formula:ePVS=[100-hematocrit(%)]/hemoglobin(g/dL).AKI was defined in accordance with the Kidney Disease Improving Global Outcomes(KDIGO)criteria.Multivariable logistic regression models were used to adjust for demographics,comorbidities,and critical laboratory markers.Patients were stratified into three groups based on the ePVS tertiles(low:≤6.30;middle:6.30-8.08;high:>8.08).Multivariate logistic regression and subgroup analyses were applied to explore the association of the ePVS with the risk of AKI.Furthermore,we also examined the association of ePVS with AKI by employing generalized additive models.Results A total of 3,388 patients were included in the final analysis,of whom 2,573(75.94%)developed AKI.Following multivariable adjustment,each unit increase in ePVS was associated with a 9%increase in the odds of AKI(OR:1.09,95%CI:1.05-1.14;P<0.001).When analyzed by ePVS tertiles,the highest tertile demonstrated a significantly elevated AKI risk compared with the lowest tertile(OR:1.48,95%CI:1.18-1.86,P=0.0007),with a significant dose-response relationship observed across tertiles(P for trend<0.001).Subgroup analyses further indicated that the association between ePVS and AKI was more pronounced among patients with pre-existing renal or peripheral vascular disease and was statistically significant only in White patients.Conclusions ePVS was independently associated with an increased risk of AKI in adults undergoing CABG.These findings supported the potential utility of ePVS as a simple,economical clinical tool for early identification of patients at high risk for AKI following cardiac surgery.
文摘目的本研究旨在评估血清肌酐/白蛋白(creatinine to albumin ratio,CAR)与坏死性胰腺炎患者死亡率之间的关系。方法本研究使用重症监护医疗信息市场(MIMIC-IV,3.1)数据库的数据进行了回顾性研究。研究包括分析各种死亡率变量并获得人院时的CAR值。使用ROC截断值确定CAR的最佳阈值。采用受试者工作特征分析(Receiver Operating Characteristic analysis)研究CAR对坏死性胰腺炎患者死亡率的预测能力、敏感性、特异性和曲线下面积(Area under curve,AUC),根据ROC截断值将将CAR≥0.44的坏死性胰腺炎患者分为高CAR组,CAR<0.44的坏死性胰腺炎患者分为低CAR组。使用Kaplan-Meier生存曲线评估高和低CAR组患者与死亡率之间的关系。结果共纳人156名符合标准的坏死性胰腺炎患者,采用ROC曲线评估CAR对坏死性胰腺炎患者死亡率的预测能力,截断值为0.44。根据截断值分为高CAR组和低CAR组,两组之间年龄、有无糖尿病、平均动脉压、血氯、血钾、血钙、血糖、尿素氮、肌酐、总胆红素、谷丙转氨酶、谷草转氨酶、白蛋白、白细胞计数、血小板计数、氧分压和CAR之间P<0.05,有统计学差异。两组之间住院天数、30天存活时间、60天存活时间和90天存活时间之间P<0.05,有统计学差异。R0C结果显示曲线下面积0.782,95%CI(0.678,0.886);敏感性,0.655,95%CI(0.482,0.828);特异性,0.874,95%CI(0.816,0.932);阳性预测值,0.543,95%CI(0.378,0.708);阴性预测值,0.917,95%CI(0.868,0.966)。CAR与急性生理与慢性健康评分(Acute Physiology and Chronic Health Evaluation Ⅱ,A-PACHEI)评分进行ROC比较,CARAUC值方面,0.782,95%CI(0.678,0.886),APACHEIIAUC值方面,0.642,95%CI(0.526,0.703)。对低CAR和高CAR两组坏死性胰腺炎患者30天、60天和90天的使用Kaplan-Meier法绘制生存曲线,log rank检验P<0.05,高CAR组坏死性胰腺炎的患者病死率明显高于低CAR组坏死性胰腺炎的患者。结论CAR可作为坏死性胰腺炎患者死亡率的独立预测因素,为临床评估病情严重程度、指导治疗和预后判断提供重要参考。
文摘在ICU电子病历脓毒症早期预测任务中,传统机器学习方法因难以捕捉稀疏表格数据中的动态特征交互而导致性能受限。为此,本研究提出改进的LF-Transformer深度学习模型,通过构建基于乘法算术块的特征交互增强方法,结合交互候选生成器(ICG)的动态top-k特征选择机制,有效提升了模型对稀疏医疗数据的表征能力。采用MIMIC-IV数据集验证模型性能,实验结果表明改进模型在脓毒症预测任务中AUROC达到0.841,特异度0.763,敏感性0.759,显著优于传统方法。该研究成果为开发ICU脓毒症智能预警系统提供了有效的算法支持,具有一定的临床实践价值。In the task of early sepsis prediction using ICU electronic medical records, traditional machine learning methods suffer from performance limitations due to their inability to capture dynamic feature interactions in sparse tabular data. To address this, we propose an improved LF-Transformer deep learning model. By constructing a feature interaction enhancement method based on multiplicative arithmetic blocks and integrating a dynamic top-k feature selection mechanism via an Interaction Candidate Generator (ICG), the model significantly improves its representational capability for sparse medical data. Validated on the MIMIC-IV dataset, experimental results demonstrate that the enhanced model achieves an AUROC of 0.841, specificity of 0.763, and sensitivity of 0.759 in sepsis prediction, outperforming traditional methods significantly. This research provides effective algorithmic support for developing intelligent sepsis early warning systems in ICUs, offering practical clinical value.
