Bayesian consensus clustering for multivariate longitudinal data

In clinical and epidemiological studies, there is a growing interest in studying the heterogeneity among patients based on longitudinal characteristics to identify subtypes of the study population. Compared to clustering a single longitudinal marker, simultaneously clustering multiple longitudinal markers allow additional information to be incorporated into the clustering process, which reveals co-existing longitudinal patterns and generates deeper biological insight. In the current study, we propose a Bayesian consensus clustering (BCC) model for multivariate longitudinal data. Instead of arriving at a single overall clustering, the proposed model allows each marker to follow marker-specific local clustering and these local clusterings are aggregated to find a global (consensus) clustering. To estimate the posterior distribution of model parameters, a Gibbs sampling algorithm is proposed. We apply our proposed model to the primary biliary cirrhosis study to identify patient subtypes that may be associated with their prognosis. We also perform simulation studies to compare the clustering performance between the proposed model and existing models under several scenarios. The results demonstrate that the proposed BCC model serves as a useful tool for clustering multivariate longitudinal data.     

学校结核病影响因素及防控策略概述

结核病是一种严重的呼吸系统传染性疾病,在世界范围内广泛传播。某些特殊群体结核病发病风险较高,如学校学生及相关工作人员。学校学生聚集,人口密度大,有利于疾病的传播与扩散,导致聚集性疫情的发生。学校结核病疫情的出现主要与学生免疫力低、就诊延误现象、人口密度大、性别、年龄、知晓率等因素有关。为了尽早发现传染源,减少结核病在校园内的传播,使用合理的方法对密切接触者进行筛查极为重要。应明确各机构在学校结核病防控工作中的职责与重点,共同做好防控工作,减少学校结核病疫情的发生。     

基于数据挖掘的中医药治疗新型冠状病毒肺炎用药规律研究

新型冠状病毒肺炎(COVID-19)疫情发展迅速,危害严重,目前尚无特效药,而中医药在此疫情中发挥了积极的作用。该研究通过收集国家、各省、自治区和直辖市发布的诊疗方案,及中国知网、万方、维普三大数据库相关文献报道,全面系统整理现代中医治疗COVID-19的处方,并通过中医传承辅助平台V2.5,确定处方中用药的四气、五味、归经、使用频次、药物之间的关联规则及形成的新处方等。最终纳入有效处方96首,其中四气主要集中在温、寒、平,五味主要集中在苦、辛、甘,归经主要集中在肺、胃、脾经;使用高频次药物为甘草、杏仁、生石膏等;高频次药物组合为生石膏-杏仁,生石膏-甘草,甘草-杏仁等;核心组合有葶苈子-杏仁-生石膏,藿香-生姜-厚朴,麦冬-杏仁-黄芩等;形成新处方葶苈子、杏仁、生石膏、藿香、生姜、厚朴,麦冬、杏仁、黄芩、五味子、西洋参等。从药性到高频药物及新处方可以看出,我国现代中医治疗COVID-19的整体治则为扶正祛邪,兼顾气阴。     

湖南省新型冠状病毒肺炎聚集性疫情流行特征分析

目的 分析湖南省195起新型冠状病毒肺炎聚集性疫情的流行病学特征,为新型冠状病毒肺炎聚集性疫情防控提供科学依据。方法 收集2020年1—2月湖南省195起新型冠状病毒肺炎疫情资料,采用描述流行病学方法进行分析。结果 195起新型冠状病毒肺炎聚集性疫情,发病696 例,无死亡病例;报告事件数位居前三的是长沙市(51起)、岳阳市(31起)、邵阳市(22起)。疫情起数、发病数在1月31日均达到高峰。男女性别比1.02∶1;平均年龄44.7岁,其中15岁以下儿童42例(6.3%);无症状感染者55例(8.28%);病例以轻型和普通型为主,占79.5%。仅发生一代病例的事件为28起(14.4%),二代病例的事件154起(79.0%),三代病例的事件10起(5.1%),四代病例的事件3起(1.5%)。一代病例与二代病例代际间隔平均为6.2 d(95%CI:5.1~7.3),中位数5.0 d。家庭暴露聚集性疫情续发率范围0.7%~100%,中位11.8%;医疗机构续发率范围0.9%~20.0%,中位6.9%。127起(65.1%)为湖北输入型疫情。696例聚集性病例中,268例(38.5%)为首代病例,387例(55.6%)为二代病例,32例(4.6%)为三代病例,9例(1.3%)为四代病例。事件暴露方式多种多样,184起有同住暴露,185起有同车暴露,118起有聚会暴露,149起有聚餐暴露,165起有交谈暴露,8起有会议培训暴露。结论 湖南省新型冠状病毒肺炎聚集性疫情主要发生在家庭,应关注重点地区、重点人群、重点场所,落实各项防控措施,有效处置新型冠状病毒肺炎聚集性疫情,防止疫情进一步扩散。     

K-Means Clustering to Elucidate Vulnerable Subpopulations Among Medicare Patients Undergoing Total Joint Arthroplasty

BackgroundThe role of preoperative laboratory values for risk stratification following joint arthroplasty is currently ambiguous. In order to improve upon existing risk stratification within joint arthroplasty, this study sought to define novel phenotypes of total hip or total knee arthroplasty patients based entirely on preoperative laboratory measures. These phenotypes (“clusters”) were compared to elucidate statistically and clinically significant differences in outcomes.MethodsA total of 134,252 patients were gathered from the National Surgical Quality Improvement Program database between 2005 and 2015. “K-means” with 3 clusters was applied using 9 preoperative laboratory values: sodium, blood urea nitrogen (BUN), creatinine, albumin, bilirubin, white blood cell count, hematocrit, platelet count, and international normalized ratio of prothrombin values (INR). Outcome measures included 30-day readmissions, severe adverse events, and discharge to nonhome.ResultsCluster 2 was characterized by elevated preoperative BUN, creatinine, and INR and demonstrated almost twice the rate of adverse events (3.52% vs 2.20% and 2.22%), 30-day readmissions (6.39% vs 3.31% and 3.71%), and discharge to nonhome (47.97% vs 30.50% and 35.85%). Cluster 3 was characterized by a slightly higher risk of discharge to nonhome than cluster 1 and was overwhelmingly female (79.5% female, 35.8% discharge to nonhome). Cluster 1 represents the lowest-risk subgroup, experiencing the lowest rates of readmissions, adverse events, and discharge to nonhome.ConclusionPreoperative laboratory values, namely BUN, creatinine, and INR, are useful in identifying patients at risk of adverse outcomes. This analysis supports the existing surgical literature pushing for preoperative hydration as a targeted intervention to expedite recovery.     

一种度量细胞网络结构的方法

文题释义： 复杂网络：是指具有自组织、自相似、小世界、无标度中部分或全部特征的网络。小世界效应和无标度特性是复杂网络最显著的特点。小世界是指网络具有大的簇系数和较小的平均路径长度，无标度是指网络节点的连接度分布服从幂律分布。 细胞网络：在多细胞生物体中，单个细胞除了完成自身功能外，各个细胞间的通讯联络是必不可少的，如维持生物体的生长发育、各种生理活动的协调等。如果将每个细胞抽象为一个节点，每两个细胞间的通讯抽象为一条边，那么众多细胞及其之间的通讯将抽象成为一个巨大的细胞网络。 背景：在多细胞组成的生物体社会中，细胞之间的通讯必不可少。众多细胞之间相互通讯、相互联系，形成了一个复杂的细胞网络结构。对细胞网络结构相关属性值进行度量和评价，显得极其重要。 目的：基于复杂网络，提出了一种细胞网络结构的度量方法。 方法：结合文献研究及实际应用，搭建细胞网络结构的度量框架，分别从细胞节点的度、细胞网络的度分布、细胞网络的平均路径长度、细胞网络的簇系数等方面对细胞网络结构进行度量。以某小型实验为例，进行了实例验证。 结果与结论：细胞网络的度分布中，绝大部分细胞节点的度值比较小，只有较少量细胞节点的度值较高。细胞网络中细胞节点的度分布P(k)的幂律分布特点越是明显，则此细胞网络结构越是合理，细胞网络越正常；同时，许多细胞网络结构具有较小的平均路径长度。细胞网络的簇系数越大，表明该细胞网络具有较高的聚集特性。一般情况下，细胞网络结构越“紧密”，细胞网络结构的“集团化”特性越明显，细胞网络越正常。 ORCID: 0000-0002-2483-032X(张雨) 中国组织工程研究杂志出版内容重点：干细胞；骨髓干细胞；造血干细胞；脂肪干细胞；肿瘤干细胞；胚胎干细胞；脐带脐血干细胞；干细胞诱导；干细胞分化；组织工程     

济南市2014—2018年猩红热流行特征及时空聚集性分析

目的 研究济南市猩红热的流行特征,并对时空聚集性进行分析,为猩红热的防控工作提供依据。方法 从国家疾病监测信息报告管理系统中获取济南市2014—2018年猩红热疫情资料,采用描述性流行病学方法及时空重排扫描统计方法分析猩红热病例的流行病学和时空分布特征。结果 2014—2018年济南市累计报告猩红热7 297例,年均发病率20.46/10万。全年出现冬季和春季2个发病高峰。病例主要集中在1~9岁儿童,幼托儿童、学生及散居儿童发病数占总病例数的99.70%,男女性别比1.70∶1。市中区、历城区、槐荫区、天桥区、历下区是主要高发地区,占病例总数的83.50%。按年进行时空聚集性分析,每年均探测到3个时空聚集区,具有显著时空聚集性。结论 2014—2018年济南市猩红热处于上升趋势,学校和托幼机构是发病的高危场所,应根据时空分布及流行特征,制定相应的防控措施,控制猩红热的发生和扩散。     

Identifying synonymy between relational phrases using word embeddings

Many text mining applications in the biomedical domain benefit from automatic clustering of relational phrases into synonymous groups, since it alleviates the problem of spurious mismatches caused by the diversity of natural language expressions. Most of the previous work that has addressed this task of synonymy resolution uses similarity metrics between relational phrases based on textual strings or dependency paths, which, for the most part, ignore the context around the relations. To overcome this shortcoming, we employ a word embedding technique to encode relational phrases. We then apply the k-means algorithm on top of the distributional representations to cluster the phrases. Our experimental results show that this approach outperforms state-of-the-art statistical models including latent Dirichlet allocation and Markov logic networks.