China-PAR与WHO模型预测10年心血管病风险的准确性

目的 在我国社区人群中独立验证并比较中国动脉粥样硬化性心血管疾病风险预测研究（China-PAR）模型与2019年WHO开发的东亚人群心血管病风险评估工具预测10年心血管病风险的准确性。方法 2010年1月1日至2020年5月31日中国鄞州电子健康档案研究中基线无心血管病史的40~79岁人群。采用Kaplan-Meier法计算10年心血管病实际发生风险,并采用China-PAR模型、WHO复杂模型和简易模型分别计算预测风险。通过区分度C统计量、校准度χ²值和校准图评价模型的准确性。结果 共纳入225 811名研究对象,经过中位随访7.26年,共发生13 301例心血管病事件。China-PAR模型、WHO复杂模型和WHO简易模型的C统计量及其95%CI在男性中分别为0.741（0.735~0.747）、0.747（0.740~0.753）和0.739（0.733~0.746）,在女性中分别为0.782（0.776~0.788）、0.789（0.783~0.795）和0.782（0.776~0.787）。WHO复杂模型和简易模型预测男性10年心血管病风险整体分别低估了0.8%和4.4%,在女性中整体均低估了约15%;而China-PAR模型整体低估风险,在男性和女性中分别低估了19.5%和42.3%。结论 在我国社区人群总体风险评估中China-PAR和WHO模型预测10年心血管病风险均具有较好的区分度,但对于高风险组人群的预测准确性还有待改善,提示仍需进一步针对高风险组（如糖尿病或老年）人群建立单独的风险预测模型。     

China-PAR和PCE模型在新疆维吾尔族和哈萨克族人群中的应用研究

目的 对China-PAR(prediction for ASCVD risk in China)在新疆农村维吾尔族和哈萨克族人群进行外部验证,评估其预测该人群5年动脉粥样硬化性心脏病(atherosclerotic cardiovascular disease, ASCVD)发病风险的准确性,并将其与汇总队列方程(pooled cohort equations, PCE)进行比较。方法 本研究纳入3347名年龄在40～74岁之间,基线时无心血管疾病病史的研究对象。分别使用China-PAR和PCE计算个体5年ASCVD发病风险。模型再校准前后,分别通过模型的区分度、校准度和临床实用性来评价其预测准确性。结果 3347名研究对象中,女性1839人(54.94%)。在5年的随访中,共发生286例ASCVD事件。China-PAR和PCE的区分度均较好(C统计量：0.727～0.738)。男性中,China-PAR和PCE的C统计量分别为0.727(95%CI:0.689～0.766)、0.727(95%CI:0.684～0.770);女性中分别为0.738(95%CI:0.703～0.7...     

心血管疾病低风险群体颈动脉斑块危险因素及预测模型构建

目的 探讨心血管疾病低风险有颈动脉斑块人群的危险因素，构建列线图预测模型。方法 分别于2019年11月、2021年11月在广西壮族自治区(广西)柳州市两镇开展现况调查，现场招募481例成年人作为研究对象。使用传统危险因素负担分类评估心血管疾病发生风险；运用颈动脉超声评估研究对象的颈动脉斑块。采用Lasso回归模型进行特征筛选，利用logistic回归模型分析危险因素，最后建立列线图预测模型。采用受试者工作特征(receiver operator characteristic, ROC)曲线下面积(area under curve, AUC)评估预测模型的预测能力。利用Bootstrap自抽样法对模型进行内部验证，利用一致性指数(concordance index, C-index)评估其区分度，利用Hosmer-Lemeshow拟合优度检验评估其校准度。结果 研究对象共481人，有302例低风险群体，20.53%(62/302)的低风险群体存在颈动脉斑块。调整因素后的多因素logistic回归分析结果显示，年龄(OR=1.16, 95%CI:1.09～1.24)、高血压(OR=3.28...     

2型糖尿病发病风险预测模型常用评价方法

目的建立并最优化糖尿病2年发病风险预测模型,为预测模型的构建与优化提供分析框架。方法本研究数据来自2011—2012年北京市房山区心脑血管疾病慢性病队列调查,于2014年进行随访调查。问卷调查一般人口学特征、生活方式和疾病情况;并进行体格检查和实验室测定。排除基线调查为糖尿病的患者,最终纳入3 127名符合条件的研究对象。采用logistic回归建立糖尿病2年发病风险预测模型,调整原始模型参数构建新模型1、2、3,从整体表现、区分度、校准度3个方面评价模型预测性能,确定最优模型。绘制决策分析曲线评价最优模型净收益。采用R 3.6.1软件进行χ2检验。结果研究人群的糖尿病2年发病率为5.98%。建立的原始模型参数包括吸烟、降脂药使用史、口服葡萄糖耐量试验后2 h血糖、空腹血糖与体质指数共5个变量。新模型1在原始模型基础上纳入糖尿病家族史变量,与原始模型相比,预测能力改善最大为最优模型,净重新分类指数为0.044(95%CI:0.004～0.084),决定系数为0.325,Brier得分为0.046,受试者工作特征曲线的曲线下面积为0.880(95%CI:0.856～0.903),在最优切点处的灵敏度为85.0%,特异度为79.4%;Hosmer-Lemeshow检验无统计学意义(P0.05);内部验证曲线下面积为0.875(95%CI:0.870～0.879)。决策曲线分析结果显示,原始模型和最优模型的净收益均高于全人群策略,但二者净收益无较大差异。结论净重新分类指数可有效评估模型的改进程度,建议研究者在模型构建过程中选取恰当的指标或方法综合评估模型的性能,优化模型的预测能力。     

中国8个长寿地区65岁及以上老年人慢性肾脏病的6年发生风险预测

目的建立适用于中国≥65岁老年人慢性肾脏病(CKD)的发生风险预测工具。方法基于前瞻性队列研究,收集来自"中国老年健康影响因素跟踪调查"子队列-"老年健康生物标志物队列研究"2008/2009年至2014年和2012年至2017/2018年3742名研究对象的数据,对基线未患CKD的1055人追踪2次肾功能结局,使用Lasso方法筛选预测因素,采用Cox比例风险回归模型构建CKD发生风险预测模型并通过列线图工具实现模型可视化。采取bootstrap 1000次重复抽样的方法进行内部验证,并使用C指数和校准曲线评估模型的性能。结果研究对象年龄为(80.8±11.4)岁。经过4797人年随访,262例(24.8%)研究对象发生CKD。年龄、BMI、性别、文化程度、婚姻状况、退休金或保险、高血压、血尿酸、血尿素氮、TC和基线肾小球滤过率因素纳入6年CKD发生风险预测模型,模型校正C指数为0.766。校准曲线显示预测CKD发生概率和实际发生概率在高风险组中一致性较高,但在低风险组人群中一致性相对较差。结论本研究建立的CKD发生风险预测模型性能较好,列线图可作为可视化工具用于中国≥65岁老年人的6年CKD发生风险预测。     

基于决策树的结直肠癌风险预测模型研究

目的 构建决策树风险预测模型,为早期发现结直肠癌高危人群提供依据。方法 2020年10—12月采用序贯抽样方法抽取四川省5家医院结直肠癌患者为病例组,体检或就诊人群为对照组。采用自制问卷,收集调查对象的一般人口学特征、家族史、疾病史、用药史、生活方式及饮食习惯等内容。利用χ2检验、非参数检验和logistic回归分析筛选变量,基于决策树C5.0算法构建结直肠癌风险预测模型,并绘制受试者工作特征曲线、校准曲线和决策曲线,评估模型的区分度、校准度和临床适用性。2023年2—3月抽取结直肠癌患者和体检人群进行外部验证。结果 建模共纳入病例组325例,对照组322例。外部验证共纳入病例组43例,对照组58例。模型最终筛选出肠镜检查、工作中的体力强度、运动的持续时间、高血脂、是否吃辣及个人年收入为结直肠癌发病风险的预测指标。模型准确率为78.34%,AUC为0.845,最佳截断值及坐标为0.696(0.230,0.797),特异度为79.69%,灵敏度为76.95%。模型内部验证准确率为73.85%,AUC为0.789,最佳截断值及坐标为0.696(0.242,0.719),特异度为71.88%...     

基于巢式病例对照设计开发宫颈癌发病风险预测模型

目的 基于宁波市鄞州区域健康信息平台,使用巢式病例对照的研究设计构建宫颈癌发病风险预测模型。方法 2018年10月31日前建档的25~75岁无宫颈癌史的常住女性中,至少随访3年的过程中新发生的宫颈癌患者纳入病例组,以1∶10的比例匹配对照组,预测因子提取时间限制在结局发生前。变量选择采用Lasso-logistic回归,选择β不为0的变量拟合logistic回归模型并用Bootstrap法进行内部验证。模型的区分度用受试者工作曲线下面积（AUROC）评价,校准度用校准曲线图和Hosmer-Lemeshow检验来评价。结果 最终模型纳入的预测因子包括年龄、吸烟状况、宫颈炎史、HPV检测情况和液基薄层细胞检查情况。内部验证500次Bootstrap的AUROC为0.740（95%CI：0.739~0.740）,校准曲线与理想曲线几乎重合,Hosmer-Lemeshow检验P=0.991,模型区分度和校准度均较好。结论 本研究开发了一个简便且实用的宫颈癌发病风险预测模型,模型的可解释性强,内部验证区分度良好,校准度良好,可以用于一般人群,为个人对自身宫颈癌发病风险的评估提供依据。     

脑卒中手术患者医院感染风险预测列线图模型的构建

目的构建脑卒中手术患者医院感染风险预测列线图模型,为早期筛查医院感染高风险人群和制定预防策略提供一定的参考和依据。方法回顾性收集2016-2018年山东大学齐鲁医院脑卒中手术患者的临床相关资料,将患者按照7∶3的比例随机分为建模组(571例)和验证组(245例)。采用单因素和多因素Logistic回归探讨医院感染的独立危险因素,基于危险因素的回归系数构建脑卒中手术患者医院感染风险预测列线图模型。分别在建模组(内部验证)和验证组(外部验证)中采用受试者工作特征(ROC)曲线下面积(AUC)和校准曲线评估预测模型的区分度和校准度。结果共纳入816例脑卒中手术患者,医院感染213例,医院感染发生率为26.10%。Logistic回归分析显示,脑卒中类型、留置胃管、静脉血栓、手术风险分级(NNIS)、美国国立卫生研究院卒中量表(NIHSS)评分以及住院时间是脑卒中手术患者医院感染的独立危险因素(P<0.05)。依此构建的列线图模型在建模组和验证组中的ROC曲线下面积分别为0.849和0.858,具有良好的区分度;两组校准曲线显示列线图模型的预测值和实际观察值结果一致性良好(P=0.731、P=0.224)。结论本研究构建的个体化风险预测列线图模型有助于提高对脑卒中术后医院感染高危人群的筛查和早期诊断,尽早制定干预策略,以降低感染发生率。     

决策树模型在2型糖尿病患者脑梗死风险预测中的应用

目的 探讨决策树在预测2型糖尿病患者发生脑梗死风险中的应用,为进一步的预防和治疗提供参考依据.方法 采用C5.0决策树算法建立预测模型,并对其预测准确性进行评估,全部分析在SPSS Clementine12.0和SPSS13.0软件中完成.结果 所建立的预测模型对训练样本和测试样本的预测准确率分别为88.41%、85.00%,准确率较高.结论 采用C5.0决策树算法预测2型糖尿病患者发生脑梗死的风险是可行的,其预测结果可以为患者自我保健和医务人员制订治疗方案提供依据.     

德清县农村社区糖调节异常人群2型糖尿病发病风险预测

目的构建农村社区糖调节异常(IGR)人群2型糖尿病(T2DM)发病风险预测模型,为T2DM的预防提供理论依据和评估工具。方法收集2006-2014年浙江德清县农村社区28 251名非糖尿病成人健康数据,筛选出基线调查中5 656例IGR作为本研究队列人群,并按2:1随机分为建模队列(n=3 803)和验证队列(n=1 853)。采用SPSS 22.0软件进行统计分析,计数资料组间比较采用χ~2检验,基于建模队列采用Cox回归分析筛查T2DM相关危险因素,构建发病风险预测模型,根据纳入模型的危险因素系数β建立风险评分表,以受试者工作特征(ROC)曲线下面积(AUC)评价模型的预测能力。在验证队列中进一步验证,评价模型预测能力。结果最终纳入模型的变量有年龄、体质指数(BMI)、高血压史、蔬菜摄入量和食物偏好。建模队列的AUC=0.71(95%CI:0.66~0.77),验证队列的AUC=0.65(95%CI:0.58~0.73)。风险评分提示,总评分≥5分的对象发展为T2DM的风险显著增加。结论所构建的T2DM发病风险预测模型在农村社区IGR人群中有较好的预测能力,而且简单、使用方便,具备很好的应用价值。     

Exploring attribution of responsibility for visitor safety in a US national park

Attracting millions of visitors each year, US national parks represent one context in which unintentional injuries are recurrent and fatal. Given unique environmental and infrastructural risks, as well as varied recreational opportunities and social contexts, who is perceived as responsible for preventing visitor injuries, and how might this relate to risk perception? Further, how does attribution of responsibility relate to support for preventative risk management? Limited research has considered these questions yet increasing promotion of parks to diverse audiences suggests a need to build public support for risk management. Using quantitative and qualitative survey data, this study found that most visitors perceived themselves as responsible for their own safety, and perceptions of the uncontrollability of risks were positively related to these attributions; however, attribution failed to predict support for preventative risk management. Significant predictors included perceptions of risk, participation in high-risk activities, and travelling companions. Future directions for this research, including the role of voluntary risk-taking in recreational settings, are discussed.     

对医院财务风险的认识及防范

本文从市场风险和非市场风险两个方面对医院存在的财务风险进行分析。提出医院不能回避风险，但可以管理和控制风险。     

医院聘用护士流失风险管理策略

为降低因聘用护士流失给医院造成的损失。笔者通过对2001年至2003年医院聘用护士的流失原因分别从护士自身因素和医院因素二方面进行分析,提出采用流失风险识别、风险防范、风险控制等有效的风险管理策略,达到降低流失风险,确保护理工作连续有效。     

临床输血风险及风险管理

目前由于输血技术和管理等多方面的原因,临床输血存在较大风险,加强输血风险因素分析及风险宣传与管理,对保证输血安全具有极其重要的促进作用。     

From chance to choice: The use of a verbal analogy in the communication of risk

In risk communication to patients, Medical Doctors frequently use analogies attempting to enhance the otherwise poor comprehension of small probabilities by laypeople. Two experiments examined the effects of the verbal analogy (asking patients to figure out the probability of an outcome by imagining the chance to draw a winning ball from a jar) on individuals' probability perception by means of seven-point Likert scales. Study 1 showed that the employment of the verbal analogy reduced the perceived likelihood of the hypothetical event under judgment both in the positive (winning a 1 million Euro lottery) and in the negative domain (contracting malaria travelling to Africa). In Study 2, two possible versions of the verbal analogy (forced vs. weak) were tested on the perceived risk of contracting hepatitis A travelling to Africa. Results indicated that individuals' risk perception was decreased by the analogy only in the forced condition. The authors discuss possible explanations for this effect.     

Examining uncertainties in government risk communication: citizens' expectations

With the shift towards more participative approaches to risk communication and risk management, policymakers are expected to be more transparent with the public concerning existing uncertainties about health risks, such as lack of data or contradictory evidence. There is, however, a debate among researchers as to whether this transparency about uncertainties is indeed welcomed by the public and if it is effective in promoting trust in risk regulatory institutions. The qualitative study on which this article is based aimed to clarify citizens' perceptions of diverse sources of uncertainty in government risk communication and their expectations with respect to discussing uncertainties. Forty-seven adults from two major Canadian cities took part in focus groups or individual interviews in which they voiced their expectations about who should be involved in risk communication, why uncertainties should be communicated, what sources of uncertainties were preferred, as well as how and when these should be discussed. Their discussions showed that they expected the government to include citizens in evaluating and managing uncertainties pertaining to risks under individual control and directly informing consumer choices about health risks. In contrast, they questioned the relevance of systematically exposing uncertainties relating to risks perceived as outside individual control, and associated with very low probability scenarios of possible threats. Globally, they appraised the desirability of discussing uncertainties in risk communication in relation to the perceived utility of the information for decision-making. These findings indicate that risk communicators and managers need to consider ways in which discussions of uncertainties can empower citizens.     

Writing the risk of cancer: Cancer risk in public policy

In this paper we examine how cancer risk is written in cancer policy documents from the English speaking OECD nations. We offer an audit of the multiple ways in which cancer risk is conceptualised and presented in health policy and professional contexts with the long term aim of comparing this with lay conceptualisations. Our study sampled cancer policy documents produced by six nations, the World Health Organization and the International Union for Cancer Control since 2000 and analysed them iteratively through questions and codes. Whilst the documents contained a comprehensive range of concepts and locations for cancer risk, our analysis found two predominant representations: firstly, well established metaphors that depict cancer as uniformly dreadful and life threatening; and secondly, through a concentration on five behavioural risk factors (tobacco smoking, drinking alcohol, inadequate nutrition, sun exposure and physical inactivity) and one bodily state (overweight). We discuss the implications of this dual focus and of other tensions within ideas about cancer risk that we identified for risk communication.     

我国医疗风险管理概况

本文通过吸取国内外关于医疗风险管理研究最新成果的基础上,首先将与医疗风险管理和医疗风险等一组相关的概念进行了解释和比较,提出了医疗风险管理是一种综合性的管理。此后本文对我国的医疗风险的成因和分类展开讨论,旨在使医疗风险管理者在对风险进行管理时更有针对性和有效性。最后,根据我国医疗风险管理的现状,借鉴企业和金融业的风险管理方法和原则,提出了一些对我国医院进行医疗风险管理的建议,主要是从制度改革方面着手。     

The avian influenza pandemic: Discourses of risk,contagion and preparation in Australia

This paper examines the construction of avian influenza in Australian media and federal government policy, with a focus placed on discourses of contagion, preparedness and risk. The threat of an infectious disease outbreak, such as avian influenza, on social life is surrounded by a range of collective narratives which attempt to make it explicable. These narratives socially define the disease and provide explanations for its existence. The paper demonstrates that central to these narratives are depictions of the source of the outbreak and suggestions of appropriate responses to the threat. Methodologically, a narrative analysis of print media and government documents was conducted. This showed that conceptually both government and media discourses could be understood in terms of risk, contagion and blame. Furthermore, it was found that narratives linking the risk of avian influenza with globalised interconnectedness and contagion by the developing world underpin discourses of causation and frame the reactions to and preparation for a potential outbreak.