《观察性研究在临床研究中的应用.ppt》由会员分享，可在线阅读，更多相关《观察性研究在临床研究中的应用.ppt（130页珍藏版）》请在课桌文档上搜索。
1、1,观察性研究在临床研究中的应用,2,临床科研中常见的指标,疾病相关指标发病率指标患病率指标死亡相关指标,效应相关指标 相对危险度（RR）比之比（OR）需要治疗的人数（NNT）,3,Methods of study,Observational,Experimental,Theoretical,Descriptive study,Analytic study,Clinical trialField trialCommunity intervention,three types of basic methods,Case-control studyCohort study,Case report,
2、case series,Cross-sectional study,Ecological study,Routine data analysis,Screening,4,正确使用率的指标,发病率-患病率发病率-发生率预防领域的发病率-病因临床领域的发病率（发生率）-副作用、疗效、预后,5,正确使用率的指标,发病率构成的区别确定分母体现正确率如何应用构成,6,7,8,9,10,流行病学研究方法的分类（按设计类型）,描述性研究,病例报告、病例系列、常规资料分析、横断面研究、生态学研究、筛检,分析性研究,队列研究病例对照研究,实验性研究,临床试验现场试验社区干预试验,理论性研究,观察法,实验法,理论
3、法,流行病学模型研究流行病学方法研究,11,观察性研究,被动观察“事件”在“研究对象”中的分布/发生,干预性研究,主动对“研究对象”施以“干预”，观察它对“事件”发生的影响,12,观察性研究与随机对照试验,试验设计,患者人数,患者选择的限制,少,多,真实世界实际观察,随机、干预,少,严格,观察性研究,随机对照试验,13,随机对照试验面对的挑战,观察性研究的优势,观察性研究与随机对照试验互相补充,14,医学研究的许多问题靠观察性研究来探索。观察性研究更适用于发现治疗措施罕见的或较迟出现的副反应，更多地靠日常医学实践来实现。临床专业杂志发表的文章，百分之九十是观察性研究。设计好的观察性研究结果（s
4、trobe指南）与RCT结果一致,RCT vs.Observational study,15,门冬胰岛素30作为应用多年的预混胰岛素类似物的研究结果。更接近于临床的大规模观察性研究也能为临床治疗提供参考。研究证据丰富，不论RCT还是观察性研究都显示了对HbA1c、FPG和PPG的良好控制，而且低血糖发生率低。大规模观察性研究IMPROVE进一步显示，门冬胰岛素30可改善患者生活质量,16,17,18,从RCT到观察性研究,19,20,21,22,描述性研究,Descriptive Study,23,Descriptive studies often represent the first sc
5、ientific toe in the water in new areas of inquiry David A Grimes Lancet 2002;359:14549,24,Descriptive study describe the distribution of disease or heath status in relation to variables such as person,place and time,to prompt the clue of relationship between risk factors and disease or health,to for
6、mulate hypothesis.,描述性研究通过描述疾病或健康状况的三间分布情况，获得因素与疾病或健康状况间关系的线索，提出病因假设。,描述性研究的概念 Concept,25,现况研究病例报告病例系列分析个案研究历史(常规）资料分析随访研究生态学研究,Cross-sectional studyCase reportCase series analysisCase studyAnalysis of historical（routine）dataFollow-up studyEcological study,描述性研究的种类 Type,26,是对临床上某种罕见病的单个病例或少数病例的详细介绍;
7、包括:人口学特征、症状、体征、诊断、治疗、和随访等。,A detailed report of an unusual or novel occurrence of patients in clinical;including:demographic profile,symptoms,signs,diagnosis,treatment,and follow-up et al.,病例报告 Case Report,27,病例报告主要类型,特殊、罕见、新发疾病、常见疾病的特殊或罕见特点；疾病的病因或副作用的可能线索；没有预料的疾病与症状之间的联系；疾病诊治过程中出现的意外情况；独特的治疗方法或常见治疗
8、方法出现的新的治疗效果；解剖结构的位置或数量变化。,28,病例报告的应用,发现新的疾病或（新的）病因线索 如1985年，在Stroke杂志上报告了2例服用含有苯丙醇胺（PPA）的减肥药的患者发生出血性卒中。进一步研究证明减肥药或感冒药中含有的PPA是发生出血性卒中的独立危险因素。报告常见病的罕见表型、诊治过程中出现的意外现象。一位53岁的妇女因对布洛芬过敏而失去所有的睫毛和大部分头发，5年以后，医生用拉坦前列素滴眼液治疗其青光眼时，患者重获睫毛。,29,病例报告的应用,探讨病因和发病机理 如COL4A1是一种编码型胶原蛋白链的基因，其变异与常染色体显性遗传性脑穿通畸形和幼儿偏瘫有关，但有病例报
9、告报道了在一例散发的颅内出血复发患者中发现此基因的变异，提示了颅内出血散发病例的可能机理。,30,31,N Engl J Med 2005;352:2508-14,32,资料是回顾性的，准确性的限制，例数少，代表性差 没有分母，不能计算率 没有对照，没有比较,病例报告的局限性,33,病例报告不规范提供信息不全面StrobeConsort,34,35,36,一组相同疾病的临床资料进行整理、统计、分析、总结并得出结论.,病例系列分析 Case Series Analysis,A series of similar patients with an outcome of interest,in wh
10、ich to sort,statistics,analysis and summarize and derive a conclusion.,37,病例系列分析 Case Series Analysis,一般用来分析某种疾病的临床表现特征，评价预防、治疗措施的效果显示某些病变的自然进程的规律性用来确定开始流行或流行的存在方法为进一步研究提供线索,To analysis the clinical manifestation of the disease,to evaluation the effective of prevention and treatmentTo display the na
11、tural history of some diseaseAs a method to identify the beginning or presence of an epidemic To provide clue for further study,38,39,历史（常规）资料分析 Analysis of Historical or Routine Data,是指利用已有的疾病登记报告系统或者疾病监测系统，收集既往或当前的疾病或健康状态资料进行分析，描述疾病和健康状态的分布以及变动趋势。,Use the existed disease register or surveillance s
12、ystematic，collect and analyze the disease and health at present or in the past，describe the distribution and change of disease and health status,40,41,在某一特定时间，某一定范围的人群中，对某种疾病或健康状况及相关因素进行调查的一种方法。,A method to survey the disease or health status and related factors of in a certain population at a singl
13、e point in time or during a specific time interval.,现况研究 Cross-sectional Study,42,Schematic diagram of Cross-sectional Study,43,Time,44,开始时一般不设对照组特定时间或时期确定因果联系时受到限制对不会发生改变的暴露因素，可以可以作因果推断,二、特点 Characteristics,No control group Particular point of timeLimit to ensure the causal relationshipTo inference
14、 causal relationship for the exposures that wont be changed,45,用现在的暴露（特征）来替代或估计过去情况是有条件的 It need some conditions to use current exposures(features)to replace or estimate the exposure situation in past定期重复进行可获得发病率资料 Incidence data may been obtain by regularly repeat cross-sectional studies,二、特点 Chara
15、cteristics,46,现况调查的应用 Uses,用于描述疾病或与健康有关事件的分布 Describe the distribution of disease and health related events2.可用于研究影响疾病与健康的相关因素 Study the factors that effect the diseases and health3.考核、评价防治措施的效果 Evaluate the effect of medicine and prevention measurement,47,Department of Epidemiology Public Health Co
16、llege of Harbin Medical University,病例对照研究,Case control study,48,Cases,diagnosed“some disease”Control,without disease but comparable to casesCollecting exposure history by recall,lab test or review medical history.,第一节 基 本 原 理,病例对照研究基本原理,病例，确诊患某种特定疾病的病人 对照 不患该病但具有可比性的个体 通过询问、实验室检查或复查病史，搜集既往危险因素暴露史,49
17、,测量并比较两组各因素的暴露比例，经统计学检验该因素与疾病之间是否存在统计学关联,Measuring and comparing the exposure proportion in 2 groups and statistical testing if there is statistical association,病例对照研究基本原理,50,“病例”是以现在确诊的患有某特定疾病的病人，或是研究者所感兴趣的事件如健康、有效、痊愈、死亡、药物副作用等，也可以是某一疾病的症状或体征。CASE is Individuals with an event or condition of intere
18、st,such as disease,health,effect,cure,death,side effect，or sign or syndrome of a disease.,第一节 基 本 原 理,51,对照：是未患该病或不具有所感兴趣的事件但具有可比性的个体。CONTROLS：are individuals without the event or condition of interest.,第一节 基 本 原 理,52,“暴露”研究对象曾经接触过某些因素，如化学的、生物的或物理因素或具备某些特征，或处于某种状态，它可以是有害的，也可以是有益的。“Exposure”A state o
19、f contact or close proximity to a chemical,pathogen,radioisotope;exposure may be in a state such as hypertension.They maybe harmful or benefit.,第一节 基 本 原 理,53,时间(time),病例组case,对照组control,研究开始study begin,暴露exposure,暴露exposure,非暴露Non-exposure,非暴露 Non-exposure,调查方向direction of survey,图 5-1 病例对照研究示意图(De
20、mo graph of case and control study（Greenberg 2002）注：阴影区域代表暴露于所研究的危险因素的研究对象Note:grey areas is for the subject exposed to risk factor,病例对照研究,研究示意图（demonstration graph of study),54,病例对照研究资料整理表Data analysis for case control study,OR：(a/c)/(b/d)=ad/bc,55,病例对照研究的应用,广泛地探索疾病或事件的可疑危险因素 深入检验某个或某几个病因假说为进一步前瞻性研
21、究提供依据治疗方法或药物副作用的研究 疾病预后因素的研究 诊断试验、防治措施的效果的评价,第一节 概述,56,回顾性根据结果分类由果 因研究 观察法一般不能验证病因,病例对照研究特点 Characteristics,RetrospectiveCategory by outcomeFrom outcome to causeObservational methodCould not verify hypothesis in general,57,队列研究,Cohort study,58,是将人群按是否暴露某可疑因素及其暴露程度分为不同亚组，追踪其各自结局，比较不同亚组结局率的的差异，从而判定暴露与
22、结局间有无因果关联及关联大小的一种观察性研究方法。An epidemiologic design in which the incidence of a disease(or condition)is compared among exposed and unexposed individuals，to judge if there is causal relationship between exposure and outcome.,概念 Whats the cohort study,59,Cohort Study,Key Point:Presence or absence of risk
23、 factor is determined before outcome occurs.,60,Principle of a Cohort Study,Begin with disease-free personsClassify subjects as exposed/unexposedFollow up and record outcomes in both groupsCompare outcomes with relative riskAnalysis association between exposure and outcome,从没有疾病人群开始 将研究对象分为暴露与非暴露组随访
24、并记录两组结局相对危险度比较结果分析暴露与结局关系,61,Source:partially adapted from WHO,1993,R1,R0,A1,A0,Incidence rate among exposed IR1=A1(no.exposed cases)/R1(total person-time exposed)Incidence rate among unexposed IR0=A0(no.unexposed cases)/R0(total person-time unexposed)Incidence rate ratio(exposed vs.unexposed)=IR1/I
25、R0=(A1/A0)/(R1/R0),Principle of a Cohort Study,62,Features 特点,ObservationalSet up control From cause to outcome,appreciate temporality Confirm the casual relationship between exposure and outcome,观察性研究设立对照由因到果，符合时间顺序能确证暴露与结局间因果关系,63,类型 Types,Based on recruit time of study subjects Prospective v.s.Re
26、trospectiveBased on the type of cohort Fixed v.s.Dynamic,依据研究对象召集的时间 前瞻性 VS 回顾性依据队列的类型 固定 VS 动态,64,类型 Types,Combined with case-control study Nested case-control study Case-cohort study,与病例对照研究结合 巢式病例对照研究 病例队列研究,65,检验病因假设 评价治疗效果研究影响预后的因素研究疾病的自然史新药的上市后监测,应 用,明确调查目的和研究方法确定研究对象确定研究内容确定样本量,临床观察性研究的设计与实施D
27、esign and implementation of observational study,Clarify the objective and type of studySelection of study subjectsDetermine contents of studySample size and sampling method,67,临床观察性研究设计与实施Design and implementation of observational study,资料的收集资料的整理与分析常见偏倚及其控制研究的优点与局限性,Data collection Data analyzeBias
28、 and controlAdvantage and limitation,68,一、明确调查目的和类型 Clarify the objective and type of study,根据研究提出的问题，明确调查目的根据具体研究目的确定研究方法,Clarify the objective of study based on the issued problem Select the survey method reference to the aim.,69,哪种研究设计最好？,没有永远的最好，合适就好科学性证据级别局限性可行性成本伦理,70,研究方法的选择,Essentials of Epi
29、demiology in Public Health,2nd edition.,71,二、确定研究对象 Selection of study subjects,某个区域内的全体居民或其中一部分某些特殊群体,Entire or part of the residence in a regionCertain special groups of people,现况调查（普查）Census,72,二、确定研究对象 Selection of Study Subjects,现况研究（抽样调查）,明确抽样研究总体 保证样本代表性,Clarify total population of sampling s
30、tudy Ensure the representative of sample,Sample survey,73,病例对照研究：病例与对照的选择,病例与对照选择的基本原则 代表性 可比性（往往更重要）,选择方法 抽样 配比,74,来 源 Resource,总体人群中全部病例或总体随机样本人群中全部,代表性好工作开展比较困难耗费人力物力,比较合作 资料易得到且比较可靠与对照的可比性好代表性差,医院住院或门诊的病例,病例对照研究：病例与对照的选择,75,队列研究研究人群 Study Population 从目标人群中抽出的具有代表性的人 Sampling the representative s
31、ubjects from target population 未患所研究的“结局”Outcome free 分为暴露人群和非暴露人群 Categories as exposed and non-exposed population 确定人群纳入标准 Define Population at risk using inclusion criteria,第三节 设计与实施,76,暴露人群选择Selection of Exposed Population,Patients with expoure（biomarker）Special occupational exposure group(urani
32、um miners or asbestos workers.Cooperation,good record,regu.Exam,easy follow-up)Special exposure groups(smokers,X-ray workersThe general population(the outcome of interest has a high incidence rate),77,Special resource groups(e.g.,physicians,nurses,insured)Geographically or facility-defined groups(ho
33、spitals with specialized maternity care),暴露人群选择Selection of Exposed Population,78,非暴露人群选择 Selection of Non-exposed Group,内对照(internal controls)一群研究对象内部 Controls is selected with a one-sample(population-based)cohort of exposure,Example:Select the cohort(such as all residents of a given neighborhood)A
34、ll members of the cohort are then given first round questionnaires,and/or clinical examinations,and/or testing to determine exposureThe cohort is then divided into exposure categories based on those results,第三节 设计与实施,79,外对照(External controls)一群研究对象外部If everyone in a cohort is exposed(such as workers
35、 in an industry),a separate cohort as similar as possible to the exposed in terms of income,education,geography,and age should be sought,非暴露人群选择 Selection of Non-exposed Group,80,Study Populations Examples,Framingham study of cardiovascular diseaseIndividuals 30 62 years old in community at risk for
36、 diseaseFramingham,MA,1948 to present,81,Table 4-3 Framingham Study Cohort Assembly,82,三、确定研究内容 Contents,研究内容和指标依据调查目的而定The contents and indicators of study is based on the objectives.,83,主要内容包括：Main contents included研究的疾病（或某些特征）-诊断标准 Target disease（characteristics）diagnosis criteria 暴露(危险因素）-定义 Exp
37、osure（risk factor）-definition 影响暴露与疾病关系的因素（混杂因素），如人口学因素等 Confounders such as demography factors,84,研究暴露（结局）类型,连续变量身高、体重、血压、血糖分类变量名义变量种族、职业等级变量高血压分级、肿瘤分期二分变量性别、冠心病（有/无）,85,调查表 Questionnaire 定义 是记录研究（调查）内容的原始表格，是收集问题答案的工具（形式）。Definition An original table for recording the study contents,an instrument(
38、form)to collect answers to questions.,86,调查表的一般结构The Construction of Questionnaire,标题说明部分填表说明调查对象的一般情况调查的主要内容编码调查者,TitleDemonstrationGeneral situation of subjectsMain contentsCodeInvestigator,四、现况调查确定样本量 Determine the sample size,确定样本量预期现患率(P)允许误差（d）显著性水平（）,Determine the sample sizeExpected prevalen
39、ce(P)Allowable error（d）Significance level（）,88,计数资料（现患率）样本大小估计公式Formulas for categorical data（prevalence）,P 预期的现患率 Expected prevalence，q=1-p，d 容许误差 Allowable error，z 显著性检验的统计量 Parameter for Significance level，n 样本量 Sample size。,89,计量资料样本大小估计公式Formulas for Numerical data,n 样本量 Sample size，d 容许误差 Allo
40、wable error，S 总体标准差的估计值 Estimated of total standard deviation,90,病例对照研究样本含量的估计 Estimation of sample size,有关参数人群中被研究因素的暴露率 RR的近似值,OR值（查阅文献或预调查获得）值 把握度,91,方 法 method查表法(table method)公式法：(formula)病例组与对照组相等的公式（No.case=No.control）或近似公式(approximate equation),病例组的暴露率,对照组暴露率,92,病例组与对照组例数不等时的公式：（No.caseNo.co
41、ntrol）病例数（case）:对照数（control）=1:c对照组例数为（No.of control）：c*n,93,1:1匹配设计（paired match）总对子数（total pair）,m为暴露状况不一致的对子数 the total pairs for different exposure situation,94,计算样本量时需考虑的问题 Questions should pay attention to 抽样方法 Sampling methods 暴露组与非暴露组的比例 Odds of exposed to unexposed groups 失访率 Proportion of
42、lost follow up,队列研究确定样本量,第三节 设计与实施,95,样本大小估计 Estimation of Sample Size,查表 Table 公式计算（条件 暴露组和对照组样本含量相等）formula(sample size equal between exposed and control group,P1：暴露组预期发病率 expected incidence in exposure P0：对照组预期发病率 expected incidence in non-exposure,：两个发病率的平均值 average of 2 incidence,96,p0 Incidenc
43、e in exposed population or general population P1-P0 The difference of incidence between exposed and control population Significant level Power 1-,Factors Effected Sample Size,97,http:/biostat.mc.vanderbilt.edu/wiki/Main/PowerSampleSize,98,99,100,101,五、资料收集Data Collection,收集内容（研究内容）收集方法调查员培训,Contents
44、 Method Training of investigator,102,资料收集方法 Method of data collection,Lab detect or testInvestigate subjects（questionnaire）Use routine data,实验室测定或检查 对研究对象进行调查（调查表）利用常规资料,103,随访 Follow-up,随访目的、对象、内容和方法 Purpose，subjects,contents and method 观察终点 End point 观察终止时间 End of follow up 随访间隔 Interval between f
45、ollow up 随访者 Persons carried out follow up,第三节 设计与实施,104,Purpose of Following up Track subjects in both exposed and non-exposed group Define outcome events Further collect data in exposures and confounding Subjects of Following up All subjects in both exposed and non-exposed groups,105,Content of Fo
46、llow up same to base-line data,pay attention to outcomeFollow-up Methods interview,telephone self-administrate questionnaire physical exam medical recordsdata of surveillance,106,观察终点 End Point Study subjects occurred expected outcome such as disease,death,syndrome随访间隔 Interval Between Follow up Dep
47、ends on the speed of outcome观察终止时间 End of Follow up of Total Cohort Depend on the incidence,latency,labor,107,Follow-up,Completeness and non-participation90%rule of thumbAll subjects must have an equal likelihood for detecting the outcomeDisease ascertainment must be comparable between the exposed a
48、nd unexposed subjectsNumber of visitsReasons for additional evaluations,108,调查员培训 Training of investigator,掌握调查方法收集资料方法和标准一致性保证资料准确性,Grasp method of investigationAccordance of method and criteria of data collectInsurance the accuracy of data,109,资料整理 检查完整性和准确性 按标准归类、核实,六、资料的整理与分析,Data sort and analy
49、sis,data sort Check the integrity and accuracy of data Classify and check data by standard,110,资料分析 data analysis,计算各种率；定量资料可计算平均数计算标化率分析三间分布暴露与疾病关系,Calculate prevalence，average for quantitative dataCalculate standardized prevalenceAnalysis 3 dimension distributionAssociation between exposure and di
50、sease,观察性研究资料分析,暴露,非暴露,队列研究（评价暴露并随访疾病的发生）,比较 与,病例对照研究（评价疾病并探讨既往的暴露）,比较与,肺癌,非肺癌,112,Cross-Sectional Study,What came firstly?Obesity or Osteoarthritis？,Chicken or egg dilemma?,113,选择偏倚 失访偏倚 信息偏倚 混杂偏倚,偏倚及其控制 Bias and Control,Selection bias Follow up bias Information bias Confounding bias,114,一、选择偏倚 Sel
