基于AR模型和支持向量机的急性低血压预测

ICU中,急性低血压的发生严重威胁着患者的生命安全,临床上对其预测性判断主要依靠医生经验。为实现急性低血压预测,利用PhsioNet的MIMIC II数据库ICU监护中的患者临床记录,对发生与未发生急性低血压两者间的平均动脉压信号进行AR模型的功率谱估计,运用医学信息学理论,选取功率谱幅度的中位数、平均值、最大值、标准偏差和极差用于支持向量机分类预测器的学习和训练,建立分类预测模型。预测模型对测试集进行分类预测,得到预测正确率为87.5%,表明相对于直接提取患者平均动脉压信号的统计特征参数作为预测特征,本方法具有更好的预测效果,有利于实现急性低血压提前预测。     

基于双谱分析—BP神经网络的ICU急性低血压发生预测模型研究

目的:ICU中,急性低血压的发生严重威胁着患者的生命安全,能够及时准确地对其预测具有重要临床意义。为了提高医生对患者发病提前判断的准确性,本文研究了一种基于双谱分析和BP神经网络的急性低血压发生的预测模型。方法:应用双谱分析提取动脉血压数据特征,构建特征向量,利用BP神经网络训练出分类预测模型,实现对急性低血压发生的提前预测。结果:经过不断的优化调整,本文最终构造了一个三层的BP神经网络预测模型,具有良好的自动预测能力。结论:实验表明,本方法能达到比较好的分类预测效果,可为ICU中急性低血压发生的提前预测和干预提供辅助参考。     

BP神经网络在急性低血压发生预测模型中的应用

背景：ICU监护中术后急性低血压并发症的发生严重威胁着患者的生命安全,目前临床上主要依靠医生的经验进行预见性判断。 目的：为实现急性低血压发生的自动检测和提前预报,运用医学信息学理论,探讨一种预测急性低血压发生的模型。 方法：对发生与未发生急性低血压两者间平均动脉压信号进行小波多尺度分解,并选取各层小波系数的统计特征值中位数和最大值作为信号特征参数,提出了基于BP神经网络方法对提取的信号特征参数进行分类预测,并在MATLAB环境下进行仿真实验。 结果与结论：实验结果表明,利用BP神经网络方法对急性低血压发生的预测是可行的。     

决策树算法应用于MIMIC-Ⅲ数据库的ICU患者急性肾损伤预测研究

目的 急性肾损伤(acute kidney injury,AKI)是重症监护病房(intensive care unit,ICU)最常见的并发症和致死因素之一.准确预测具AKI风险的患者,明确与AKI发生相关的关键因素,可为临床决策与风险患者干预提供有效指导.方法 采用公开的重症监护室数据库MIMIC-III,提取30020例患者记录(包括AKI患者17222名,Non-AKI患者12798名),收集其住ICU期间基本信息、生理生化指标、药物使用、合并症等临床信息.将患者按4:1比例随机划分训练集和独立测试集,应用逻辑回归、随机森林与LightGBM 3种机器学习方法,分别建立24 h、48 h与72 h 3个时间点的AKI预测模型,采用十折交叉验证法,对各种模型进行训练与测试,预测患者是否发生AKI,并获取重要特征.此外,利用24 h预测模型,在一周时间窗口内对ICU患者进行每隔24 h预测.结果 3种学习模型中,LightGBM性能最优,其24 h、48 h和72 h模型预测AKI的受试者工作特征曲线(receiver operator characteristic curve,ROC曲线)下面积(area under curve,AUC)值分别为0.90、0.88、0.87,F1值分别为0.91、0.88、0.86,在每隔24 h预测时,提前1 d、2 d和3 d预测AKI的成功率分别为89％、83％、80％.已住院时长、体质量、白蛋白、收缩压、碳酸氢盐、葡萄糖、白细胞计数、体温、舒张压、血尿素氮等是预测ICU患者AKI的重要特征,仅使用24个重要特征,模型仍能取得良好的预测性能.结论 基于ICU患者的基本信息、生理生化指标、药物使用及合并症等临床信息,应用机器学习模型,可对其是否发生AKI进行多时间点的有效预测,并明确其关键风险因素.     

灰色系统理论在医院统计预测中的应用

在医院统计预测中应用灰色系统理论，建立一阶单变量的总出院人数预测模型GM（1，1）和关联度分析，可用于医院统计指标的预测。本文通过对某医院总出院人数建模的实例应用，说明灰色预测模型较传统的预测方法更具有科学性和实用性。     

基于XGBoost的危重症患者住院时间分类预测模型和风险因素研究

为预测危重症患者在重症监护病房的住院时间(length of stay in intensive care unit, ICU LOS),并探索实验室指标对ICU LOS的影响,本研究基于危重症患者的25个临床指标构建XGBoost模型,对患者是否发生超过3 d的ICU LOS进行预测,并基于SHAP模型对最佳性能模型进行解释性评估。结果显示,XGBoost模型准确率为87.9%。相比于其他预测模型,XGBoost模型在准确率、敏感度和区分度上均有明显优势。同时,SHAP模型增加了集成模型的可解释性和可靠性。研究表明,XGBoost模型可有效识别ICU LOS较长的患者,辅助医生优化临床治疗方案,改善患者预后状况。     

灰色系统理论在医院统计预测中的应用

在医院统计预测中应用灰色系统理论，建立一阶单变量的总出院人数预测模型GM（1，1）和关联度分析，可用于医院统计指标的预测。本文通过对某医院总出院人数建模的实例应用，说明灰色预测模型较传统的预测方法更具有科学性和实用性。     

ICU中发生急性低血压的预测方法研究进展

在ICU监护中,急性低血压的发生可能引起严重的后果,甚至威胁患者的生命安全,如何检测和提前预报急性低血压发生已成为医学界必须重视的临床问题。近年来,医疗监护技术和信号分析处理方法的迅速发展,促进了对急性低血压发生的预测方法的相关研究,研究内容主要侧重于两个方面:一是通过研究与可能发生急性低血压相关的生命体征参数,发现动脉压、心率及血氧饱和度等可以作为有效的指标参数;二是研究这些参数的变化趋势,通过一定的时间窗和阈值判断,可达到提前1小时预测急性低血压的发生。将数字信号处理技术和临床监护参数相结合,是实现智能化监护技术发展的方向,在继续丰富监护数据库的基础上,研究者们正致力于应用医学信息学方法认识急性低血压发生的规律,寻求提前预测急性低血压发生的方法,进而设计智能化预测软件。以上研究有利于实现急性低血压提前预测,提前干预,大大降低抢救风险,具有重大的临床研究应用价值。     

综合性医院重症监护病房病原菌分离情况分析

目的 探讨重症监护病房(ICU)医院内感染的临床特点及病原菌种类、分布情况,为临床合理使用抗菌药物、预防和控制医院感染提供参考和依据.方法 采用前瞻性监测与回顾性调查相结合的方法,对ICU患者的临床资料进行统计分析.结果 ICU病人标本中分离出病原菌593株,得出菌种分布与感染情况.结论 重症监护病房医院内感染发生率高,以呼吸道感染为主,主要病原菌以革兰阴性非发酵菌为主,加强ICU患者感染的控制,可减少ICU医院内感染的发生.     

ICU治疗结局的影响因素及远程ICU监护在延迟收住患者中的应用初探

目的:分析重症监护病房患者治疗结局的影响因素,初步探讨远程重症监护对延迟收住患者死亡率的影响。方法:以2017年5月至2019年9月ICU收治的481例患者作为研究对象,其中2017年5月至2018年6月期间收治的248例患者均为常规收住和监护模式,回顾性分析ICU治疗期间存活和死亡患者的临床资料,采用二元Logistic回归分析法分析ICU患者死亡的危险因素;根据死亡危险因素制定远程监护措施,以2018年7月至2019年9月收治的233例患者作为研究对象,以急性生理与慢性健康评分Ⅱ(Acute physiology and chronic health evaluation,APACHEⅡ)≥16分作为ICU入住标准,实施ICU远程监护措施,分析两种模式下对ICU延迟入住患者治疗时间及死亡率的影响。结果:常规ICU收住模式下的248例患者中,存活172例,死亡76例,死亡率为30.65%;经二元Logistic回归分析显示,入院时血WBC、中性粒细胞、PCT水平以及ICU住院时间、急诊住院时间均是影响ICU不良治疗结局的独立危险因素(P<0.05)。与常规ICU收住模式相比,经远程ICU监护下的患者ICU住院时间缩短,ICU患者死亡率降低(P<0.05)。结论:在ICU资源配置不足的情况下,ICU延迟收住所致患者急诊住院时间过长是影响患者不良治疗结局的独立危险因素。通过建立并实施远程ICU监护措施对于强化急诊治疗,改善患者治疗结局有积极的作用。     

Serum extracellular vesicle-derived ASS1 is a promising predictor for the occurrence of HEV-ALF

Development of biomarkers for predicting the occurrence of hepatitis E virus related-acute liver failure (HEV-ALF) is conducive to prevention and early intervention. Serum samples from 250 HEV-ALF patients, 250 patients with acute hepatitis E (AHE) and 250 health controls (HCs) were collected. We assessed the predictive ability of extracellular vesicle (EV)-derived argininosuccinate synthase 1 (ASS1) levels for HEV-ALF occurrence. Serum EVs were successfully isolated. EV-derived ASS1 levels in the HEV-ALF patients were significantly higher than those in the AHE patients and HCs. In HEV-ALF patients, EV-derived ASS1 levels were positively correlated with the number of failed organs and disease progression. The logistical regression showed that EV-derived ASS1 level is an independent risk factor for HEV-ALF, and orthogonal partial least squares discriminant analysis (OPLS-DA) also suggested that EV-derived ASS1 level has high predictive capability. Besides, the area under the curve (AUC) of EV-derived ASS1 level to predict HEV-ALF occurrence was 0.728 (0.684–0.772) with the sensitivity and specificity being 72.80% and 64.80%, which had a high decision-making ability. Furthermore, there existed no significant difference between the age ≥60 and age <60 groups in EV-derived ASS1 levels. Serum EV-derived ASS1 level is a promising predictor for the occurrence of HEV-ALF.     

慢性乙型肝炎或乙肝肝硬化重叠急性戊型肝炎临床特点分析

目的研究慢性乙型肝炎或乙肝肝硬化基础上重叠急性戊型肝炎（简称乙戊重叠）的临床特点及戊肝病毒（HEV）对乙肝病毒（HBV）的影响。方法将2002年12月至2006年12月解放军第三。二医院收治的所有625例戊肝患者分为急性戊肝组（AHE组，437例）和乙戊重叠组（S组，188例），并进一步将乙戊重叠组分为慢性乙肝重叠戊肝组（CHB＋AHE，130例）和乙肝肝硬化重叠戊肝组（LCB＋AHE，58例），回顾性分析其特点并对32例乙戊重叠患者急性期和恢复期的HBVDNA水平进行随访。结果与AHE组相比，s组的总胆红素水平、重型肝炎发生率、痊愈患者平均住院日和病死率均明显升高，而谷丙转氨酶、白蛋白和凝血酶原活动度均明显降低；而且LCB＋AHE组的TBil水平，并发症如腹水、腹膜炎、肝性脑病和糖代谢紊乱等的发生率要明显高于CHB＋AHE组。对32例s组患者随访发现，20例（62．5％）出现HBVDNA水平下降，平均下降值为2，1log10；6例（18．8％）HBVDNA始终阴性，4例（12，5％）HBVDNA无变化，2例（6，2％）HBVDNA轻度升高。结论乙戊重叠感染会导致患者病情加重，但HEV对于HBV的复制有一定的抑制作用。     

2009 H1N1 influenza and experience in three critical care units

Aim: We describe futures of ICU admission, demographic characteristics, treatment and outcome for critically ill patients with laboratory-confirmed and suspected infection with the H1N1 virus admitted to the three different critical care departments in Turkey.Methods: Retrospective study of critically ill patients with 2009 influenza A(H1N1) at ICU. Demographic data, symptoms, comorbid conditions, and clinical outcomes were collected using a case report form.Results: Critical illness occurred in 61 patients admitted to an ICU with confirmed (n=45) or probable and suspected 2009 influenza A(H1N1). Patients were young (mean, 41.5 years), were female (54%). Fifty-six patients, required mechanical ventilation (14 invasive, 27 noninvasive, 15 both) during the course of ICU. On admission, mean APACHE II score was 18.7±6.3 and median PaO₂/FIO₂ was 127.9±70.4. 31 patients (50.8%) was die. There were no significant differences in baseline PaO₂/FIO₂ and ventilation strategies between survivors and nonsurvivors. Patients who survived were more likely to have NIMV use at the time of admission to the ICU.Conclusion: Critical illness from 2009 influenza A(H1N1) in ICU predominantly affects young patients with little major comorbidity and had a high case-fatality rate. NIMV could be used in 2009 influenza A (H1N1) infection-related hypoxemic respiratory failure.     

Prediction of acute hypotensive episodes by means of neural network multi-models

This work proposes the application of neural network multi-models to the prediction of adverse acute hypotensive episodes (AHE) occurring in intensive care units (ICU). A generic methodology consisting of two phases is considered. In the first phase, a correlation analysis between the current blood pressure time signal and a collection of historical blood pressure templates is carried out. From this procedure the most similar signals are determined and the respective prediction neural models, previously trained, selected. Then, in a second phase, the multi-model structure is employed to predict the future evolution of current blood pressure signal, enabling to detect the occurrence of an AHE.The effectiveness of the methodology was validated in the context of the 10th PhysioNet/Computers in Cardiology Challenge—Predicting Acute Hypotensive Episodes, applied to a specific set of blood pressure signals, available in MIMIC-II database. A correct prediction of 10 out of 10 AHE for event 1 and of 37 out of 40 AHE for event 2 was achieved, corresponding to the best results of all entries in the two events of the challenge. The generalization capabilities of the strategy was confirmed by applying it to an extended dataset of blood pressure signals, also collected from the MIMIC-II database. A total of 2344 examples, selected from 311 blood pressure signals were tested, enabling to obtain a global sensitivity of 82.8% and a global specificity of 78.4%.     

Aversive learning in patients with unilateral lesions of the amygdala and hippocampus

The present study applied a visual half field paradigm with emotional facial expressions in patients with selective unilateral amygdalo-hippocampectomy (AHE) to elucidate the contributions of the left and right medial temporal lobe and amygdala to emotional learning. Electrodermal indicators of aversive learning were studied in 14 left AHE and 12 right AHE patients, as well as 13 controls matched in sex and age. In a differential conditioning paradigm with negative (CS+) and positive (CS-) facial expressions, CS+ were associated with an aversive vocalization (US, 95 dB, 3 s). During extinction, stimuli were presented laterally and preattentively using backward masking. Appropriate CS durations yielding preattentive presentation were individually determined prior to conditioning. In contrast to controls, both left and right AHE patients failed to show an autonomic conditioning effect following left visual field presentations of masked negative CS+ during extinction. AHE patients also showed no clear differential acquisition. Moreover, right AHE patients poorly recognised that negative valence was an affiliating dimension of the CS-US compound.     

Impact of Intensivist and Nursing Staff on Critically Ill Patient Mortality: A Retrospective Analysis of the Korean NHIS Cohort Data, 2011–2015

PurposeCritical care medicine continues to evolve. However, critical care cases require increasing amount of medical resources. Intensive care unit (ICU) mortality significantly impacts the overall efficiency of healthcare resources within a system of limited medical resources. This study investigated the factors related to ICU mortality using long-term nationwide cohort data in South Korea.Materials and MethodsThis retrospective cohort study used data of 14905721 patients who submitted reimbursement claims to the Korean Health Insurance Service between January 1, 2011 and December 31, 2015. A total of 1498102 patients who were admitted to all ICU types, except neonatal and long-term acute care hospitals, were enrolled.ResultsOf the total 1498102 participants, 861397 (57.5%) were male and 636705 (42.5%) were female. The mean age at admission was 63.4±18.2 years; most of the subjects were aged over 60 years. During the 5-year period, in-hospital mortality rate was 12.9%. In Cox analysis, both in-hospital and 28-day mortality rates were significantly higher in male patients and those of lower socioeconomic status. As age increased and the number of nursing staff decreased, the mortality risk increased significantly by two or three times. The mortality risk was lower in patients admitted to an ICU of a tertiary university hospital and an ICU where intensivists worked.ConclusionThe number of nursing staff and the presence of an intensivist in ICU were associated with the ICU mortality rate. Also, increasing the number of nursing staff and the presence of intensivist might reduce the mortality rate among ICU patients.