页数:78页文件大小:936.0 KB来源: 流行病学ppt课件
本课件介绍描述性研究的基本方法,围绕现况研究的概念、普查与抽样调查的设计要点、样本量估计与抽样方法展开,并说明生态学研究、病例报告与个案调查的适用场景及其主要偏倚与控制。
描述性研究Descriptive Study
Methods of studyObservationalExperimentalTheoreticalDescriptive studyAnalytic studyClinical trialField trialCommunity interventionthree types of basic methodsCase-control studyCohort studyCross-sectional studyEcological study
Epidemiologyis the study of thedistributionanddeterminants of diseasefrequency inhuman population.Epidemiology: Principles and MethodsbyBrian MacMahon1996
The distribution
PersonPlaceTimeCasesDescriptive EpidemiologyWho? Where? When?
Descriptive studyis to subsetcollected data or some survey data,included the data from experiment,to groups by different places,different time or different populationcharacter and descriptive thefrequency of disease and health.Helps to build the relationshipbetween risk factors and disease orhealth, offer etiological clues.描述性研究通过描述疾病或健康状况的三间分布情况,找出某些因素与疾病或健康状况间的关系,提供病因线索。
目录
第一节现况研究概述
一、基本概念Conception又称横断面研究也称患病率研究时点或期间和特定范围内人群中的有关变量(因素)与疾病或健康状况的关系Cross-sectional study orprevalence studyDescribes the relationshipbetween disease or healthand variables (factors) in thepopulation at a particularpoint of time or period
TimeDZDZEE
TimeSchematic diagram of Cross-sectional StudyT1T2T
掌握目标群体中疾病或健康状况的分布提供疾病病因研究的线索二、目的与应用Mechanism Purposes and applicationsDescribes the distributingof disease or health eventswithin populationsTo provide etiologicalclues
This figure shows the infant death frequency(per 1000 live births) in different countries
确定高危人群
研究特点开始时一般不设对照组现况研究的特定时间在确定因果联系时受到限制对不会发生改变的暴露因素,可以提示因果联系三、特点与类型Characteristics and categoriesCharacteristicsNo control groupParticular point of timeLimit to ensure the causalrelationshipTo prompt causal relationshipfor the exposures that won’t bechanged
研究类型普查抽样调查三、特点与类型Characteristics and categoriesCategoryCensusSampling survey
The census of the lifetime alcohol drinking statusin America population
明确调查目的和类型确定研究对象确定样本量和抽样方法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第二节现况研究的设计与实施Design and implementation of cross-sectional studyTo make the purpose and category ofstudyTo select the objectsTo determine the sample size andsampling methodTo collect dataTo sort and analyze dataBias and controlAdvantage and limitation
一、明确调查目的和类型
二、确定研究对象To select the objects某个区域内的全体居民或其中一部分某一时点上的流动人员某些特殊群体Entire or part of thepopulation in a regionFloating population at aparticular point of time.Certain special groups ofpeople
三、确定样本量和抽样方法To determine the sample size and sampling method确定样本量预期现患率( P )允许误差(d)显著性水平(α)To determine the sample sizeExpected prevalence( P )Allowable error(d)Significance level(α)
计量资料样本大小估计公式Formulas for Numerical data计数资料样本大小估计公式Formulas for Categorical data
抽样方法Sampling Method非随机抽样Non-random sampling随机抽样Random sampling
非随机抽样Non-random sampling选择样本时,加入人主观因素,使总体中每个个体被抽取的机会是不均等的Non-random samples areselected by any kind ofprocedure that does not give allcases in the population equalchances to fall into the sample.
随机抽样Randomsampling总体中每一个对象都有同等机会被选入作为研究对象Every person in the target population has the same known(and non-zero) chance of being included in the survey
随机抽样Randomsampling单纯随机抽样系统抽样整群抽样分层抽样多级抽样Simple random samplingSystematic samplingClustering samplingStratified samplingMultistage sampling方法Methods
单纯随机抽样(简单随机抽样)最简单、最基本的抽样方法从总体N个对象中,利用抽签或其他随机方法抽取n个总体中每个对象被抽到的概率相等Simple random samplingSimplest and basic methodTo select numbernsubjectsbydrawinglotor by using atable of random numberfrom population with NunitsEqual chance for eachsubject inthe targetpopulation
图3-1总体与样本示意图Fig 3-1 Diagram of population and sample
系统抽样(机械抽样)1.将总体各个个体单位按某种标志排列、连续编号2.根据总体数N和确定的样本数n,计算抽样距离(N/n)Systemic sampling1.Rank every units by acharacterconstructandnumber them2.Define the size of thesampleand calculatesampling interval(k = N/n)
系统抽样(机械抽样)
图3-2系统抽样示意图Fig 3-2 Diagram of systemic sampling
Example: Systematic sampling
整群抽样Cluster samplingThe entire population ofinterest is divided intogroups, or clusters, and arandom sample of theseclusters is selected.单纯整群抽样(Simpleclustersampling)二阶段抽样(Twostagessampling)将总体分成若干群组,抽取其中部分群组作为观察单位组成样本
图3-3整群抽样示意图Fig 3-3 Diagram of cluster sampling
Section 4Section 5Section 3Section 2Section 1Example: Cluster sampling
分层抽样Stratified samplingFirst, stratify thepopulation into severalstrata by character of areas,ages, etc. then select objectsusing random sampling ineach strata.将总体单位按某种特征分为若干次级(层),然后从每一层内单纯随机抽样组成一个样本。
分层抽样Stratified sampling按比例分配(proportional allocation)最优分配(optimum allocation)
图3-4分层抽样示意图Fig 3-3 Diagram of cluster sampling
Example:stratified sampling
多级抽样将抽样过程分阶段进行,每个阶段使用的抽样方法往往不同,即将以上抽样方法结合使用,在大型流行病学调查中常用。
Multistage samplingMulti-stage samplingis a kind of complex sampledesign in which two or more levels of units areimbedded one in the other.
……一级抽样单位Primary units二级抽样单位Secondary units……多级抽样Multistage sampling
检查资料完整性和准确性按标准归类,核实五
Target Population or Sampling FrameCensus or Sampling SurveySampleDiseaseAndExposureNon-diseaseandExposureNon-diseaseandNon-exposureDiseaseandNon-exposure
44
Example:Job A (hazardous)100 Workers80 Healthy20 Respiratorysymptoms80 well10 wellJob B (hazardous)100 Workers95 Healthy5 Respiratorysymptoms95 well15 wellPoint YPoint XPrevalence rate of respiratorysymptoms in Job A20%(20/100)11%(10/90)Prevalence rate of respiratorysymptoms in Job B5%(5/100)14%(15/110)Ratio of prevalence rates, Job A/Job B:4.00.8
Cork and Kerry Diabetes & Heart Disease Study, 1998Prevalence of hypertension by age and sex
Example of Cross-Sectional StudyHypothesis:Obesity is a risk factor for knee osteoarthritisSample:100 retirees living at “University Village”
Medical exam + X-rays to diagnose osteoarthritisof the kneeOsteoarthritis+-40102030+-Obesity5050
Prevalence of osteoarthritis among obese subjects:40/50 = 0.8Prevalence of osteoarthritis among non-obesesubjects:20/50 = 0.4Prevalence ratio = 0.8/0.4 = 2.0
Obese subjects are two timesmore likely to have osteoarthritisof the knee than non-obesesubjects.
Cross-Sectional StudyWhat came firstly?Obesity or Osteoarthritis?
Chicken or egg dilemma ?
常见的偏倚偏倚的控制六、偏倚及其控制
确保随机化原则提高研究对象的依从性和受检率正确选择测量工具和检测方法培训调查员做好资料的复查复核工作辨析混杂因素偏倚的控制Control of biasPrinciple of randomizationTo improve the compliance ofsubjects and examination rateTo apply proper measurement toolsand test methodsTraining investigatorsReview data carefullyDistinguish and analyzeconfounding factors
优点Advantage常用抽样调查,结果有较强推广意义有来自同一群体的自然形成的同期对照组,结果具有可比性可同时观察多种因素三、研究的优点和局限性Advantage and limitationCommon sampling survey withstrong extension resultsWith a control group from thesame group over the sameperiod formed naturallyCan observe a variety offactors simultaneously
难以确定先因后果的时相关系不能获得发病率资料难以调查死亡病例、病程短、已痊愈的病例局限性LimitationDo not establish the truetemporal sequence of events.Does not yield incidence ortrue relative risk.They are not feasible for thecases of death, recovery andshort course.
第三节研究实例Example研究背景Background研究方法Study Method研究结果Results
“反应停”灾难Thalidomide Disaster
研究背景Background20世纪50年代,科学家推出一种新药,据说它能在妊娠期控制精神紧张,防止孕妇恶心,并且有安眠作用。
这药名叫“反应停”(沙利度胺、酞胺哌啶酮)。
Introduced in 1956 assedative (sleeping pill) and toreduce nausea and vomitingduring pregnancy,namedThalidomide
研究背景
分子式Molecularformula反应停Thalidomide
信访个人观察主要畸形%例数%例数50.710352.443手臂29.66028.123手臂与腿3.472.42手臂、腿、耳4.9103.73手臂与耳6.9148.57耳2.041.21腿2.553.73其他畸形20342合计反应停引起的主要畸形的构成比Ratio of deformities induced by Thalidomide
例数合并畸形例数合并畸形2无脑2幽门17心脏3十二指肠狭窄3小脑3十二指肠闭锁1眼2食道闭锁1鼻17肛门闭锁203个短肢畸形病例中合并其他畸形203 cases of phocomelia associated with other malformations
短肢畸形病例数反应停销售量(公斤)国家8207奥地利26258比利时3495769英国25140荷兰1160挪威237葡萄牙6113瑞士500030099西德10+7*25美国*反应停从国外购来(Thalidomide purchased from other countries)反应停销售量与短肢畸形数的关系Relationship between sales of Thalidomide and numbers of phocomelia
短肢畸形例数反应停销售量(占销售总量的比例)510152019581959196019611962年50100150西德反应停销售总量(虚线)与短肢畸形例数(实线)的时间分布Time distribution of Thalidomide sales (dotted line) and phocomelianumbers (solid line) in West Germany
“反应停儿童”的事件是一次惨痛的教训。
它提醒人们,任何新药在用于临床之前必须经过彻底检验。
Resulted in new drug testing rules!
概念研究目的研究类型优点与局限研究实例第
描述性研究的一种群体的水平上研究某种因素与疾病的关系A descriptive studyFocus on the relationshipbetween exposure factorsand disease based onpopulation一、概念Definition
以群体为观察和分析单位描述不同人群中某因素的暴露状况与疾病的频率,分析该暴露因素与疾病的关系一、概念DefinitionUnit of observation andanalysis is a populationrather than an individual.Descript the frequencyof exposure and disease toanalyze the relationshipbetween them in differentpopulation.
提供病因线索,产生病因假设评估人群干预措施的效果二、研究目的Aims of studyTo provide etiological cluesand generate hypotheses foranalytic studiesEvaluate the effect ofinterventions amongpopulation
生态比较研究生态趋势研究三、研究类型
应用较多的一种方法最简单的方法观察不同人群或地区某种疾病的分布,根据疾病分布的差异,提出病因假设生态比较研究Ecological comparison studyA method applied frequentlySimplestTo observe the differencebetween the distributions ofgroups or regions, togenerated the hypothesisaccording to the distribution.
生态趋势研究Ecological trend studyTo investigate the temporalrelationship between the exposureand disease by observing andcomparing the fluctuations ofaverage exposure levelandfrequency changes of disease in apopulation,finding thefluctuations trend.连续观察不同人群中某因素平均暴露水平的改变和(或)某种疾病发病率、死亡率变化的关系,了解变动趋势,比较暴露水平变化前后疾病频率的变化情况,判断某因素与某疾病的联系
经济,出结果快提供病因未明疾病的病因线索
适用于研究因素暴露变异范围小
出现生态学谬误难以控制混杂因素存在多重共线性问题难以确定因果联系局限性Ecological fallacyDifficult to controlconfoundingMulticollinearity problemsDifficult to determinecausal relationshipLimitations
25020015010050001600140012001000800600JapanDenmarkNew ZealandFed. Repub.Of GermanyFranceCanadaIsraelSwitzerlandUSAAustraliaYugoslaviaHong KongRomaniaFinlandPolandSpainHungaryNorwayUKItalySwedenIncidence Ratio per 100,000 WomenPer Capita Supply of Fat CaloriesPrentice RL, Kakar F, Hursting S, et al: Aspects ofthe rationale for the Women’s Health Trial. J NatlCancer Inst 80:802-814, 1988.)Correlation between dietary fat intake and breast cancer by country饮食脂肪摄入与乳腺癌发病
Is there
下载地址:{{downurl}}
{{{a}}}
{{#b}}{{{b}}}
{{/b}} {{/title}} {{#nolog}}登陆注册
{{/nolog}} {{#djsx}}点我刷新
{{/djsx}} {{#gocz}} {{/gocz}} {{#ffqr}} {{/ffqr}}