第1页:队列研究
队列研究Cohort Study
第2页:队列研究
队列研究
- Incidence studies
- Longitudinal studies
- Follow-up studies
- Prospective studies
- 发病率研究
- 纵向研究
- 随访研究
- 前瞻性研究
第3页:Outline
Outline
- Example
- Design and implement
- Measures of Risk
- Data sort and analysis
- Bias and Control
- Advantage,disadvantage
- 概述
- 实例
- 研究设计与实施
- 资料的整理与分析
- 偏倚及控制
- 优缺点队列研究Cohort study
第4页:队列研究简史
队列研究简史
- 定义与目的
- 原理与类型
- History of cohort study
- Definition and Purpose
- Principle and Types
第5页:历史
历史
- 起源OriginsGraunt (17th century) used cross-sectional mortality data toreconstruct life history using life-table methodsGraunt (17世纪)用横断面死亡资料,用寿命表方法构建生命史Farr (19thcentury) advanced the use of life-table methods a anindicator of population healthFarr(19世纪)将寿命表方法发展为人群健康的指标
- Insurance industry study 1870–1899
- 保险业研究1870–1899
第6页:结核(20世纪
Tuberculosis (20thcentury)结核(20世纪)WH Frost performed the first retrospective cohort studyin a cohort of 132 homes with tuberculosisWH Frost在有结核的132个家庭的队列开展第一个回顾性队列研究Used person-years to estimate attack rates用人年来估计罹患率历史History
第7页:WH Frost initiated
WH Frost initiated prospective cohort study oftuberculosis in Williamson county, TennesseeWH Frost在德州启动有关结核的前瞻性队列研究历史History
第8页:Framingham study
Framingham study
- Japanese atomic bomb survivors, 1946
- British physician study, 1950s
- Colorado Plateau uranium miners, 1950s
- Aniline-dye occupational cohort, 1954
- Asbestos exposure and lung cancer mortality 1965Retrospective cohort studiesProspective cohort studies
第9页:Nurses Health
Nurses Health
- British physician study
- Multi-center AIDS Cohort Study MACS, 1984–1999Current Studies
第10页:An epidemiologic
An epidemiologic design in which the incidence ofa disease (or condition) is compared among exposedand unexposed individuals是比较暴露与非暴力人群发病率的一种流行病学设计定义What’ the cohort study
第11页:Cohort Study
Cohort Study
- Begin with disease-freepatients
- Classify patients asexposed/unexposed
- Record outcomes in bothgroups
- Compare outcomes usingrelative risk
- 从没有疾病人群开始
- 将研究对象分为暴露与非暴露组
- 记录两组结局
- 用相对危险度比较结果
第12页:Cohort StudyKey
Cohort StudyKey Point:Presence or absence of riskfactor is determinedbeforeoutcome occurs.
第13页:What is队列
- What is a cohort?队列
- Cohort-Latin word for one ofthe 10 divisions of a Romanlegion
- A group of individuals
- sharing same experience
- followed up for a specifiedperiod of time
- Examples
- birth cohort
- occupational cohort chemicalplant workers
- 队列拉丁语原意是指古罗马军团中的一个分队
- 一组人群
- 有共同经历
- 随访一特定时间
- 例如
- 出生队列
- 职业队列
第14页:暴露指接触过某种物质
暴露指接触过某种物质
- 危险因素泛指能引起某特定不良结局,或使其发生的概率增加的因子,包括个人行为、生活方式、环境和遗传等多方面的因素。
- Exposurecontact some material of beingin some status
- Risk factoris a variable associated withan increased risk of disease orinfection. Including behavior,life style, environmental andgenetic factor and so on.
第15页:目的
目的
- Describe incidence ofoutcomes over time,naturalhistory of disease
- Determine causalrelationships between thoseoutcomes and exposures (risk orprognostic) factors.
- 描述随着时间变化结果的发生率,疾病自然史
- 确定结果与暴露(危险或预后)因素间的因果关系
第16页:目的
目的
- Determine a prognosis
- Evaluate new therapies anddiagnostics
- Evaluate screening procedures
- 确定预后
- 评价新的治疗与诊断方法
- 评价筛检过程
第17页:Framingham study
Framingham study
- Individuals 30–62 years old in community atrisk for disease
- Framingham, MA, 1948 to presentExamples of cohort study
第18页:Goal to elucidate
Goal to elucidate
- Questionnaire
- Physical examination
- Laboratory testingMulti-Centered AIDS Cohort Study
第19页:Source: partially
Source: partially adapted from WHO, 1993R1R0A1A0Incidence 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/ IR0= (A1/A0) / (R1/R0)Principle of a Cohort Study
第20页:Characteristics
Characteristics
- Observational
- Category by exposure
- Need control
- Prospective
- 观察性研究
- 根据暴露分组
- 需要对照
- 前瞻性
第21页:类型
类型
- Based on recruit time ofstudy subjectProspective study v.s.Retrospective Study
- Based on the type of cohortFixed cohort v.s.Dynamic cohort
- 依据研究对象召集的时间前瞻性研究VS回顾性
- 依据队列的类型固定队列VS动态队列
第22页:类型
类型
- Combined withcase-control studyNested case-control studycase-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
第23页:依据研究对
Prospective cohort studytimeExposureStudy startsDiseaseoccurrence依据研究对象召集的时间分类
第24页:Prospective cohort
Prospective cohort
- Identify cohort in thepresent
- Determine exposurestatus or possibleexplanatory/prognosticfactorsIn thepresentor inthe future
- 现在确定队列
- 确定暴露状态或可能暴露/预后因素在现在或将来依据研究对象召集的时间分类
第25页:Follow-up
Follow-up
- Ascertainment of outcomedone in future
- 随访确定结局可以是发病率或结局的风险可以是什么因素与结局有关
- 确认将来发生的结局依据研究对象召集的时间分类Prospective cohort study
第26页:Prospective Cohort
Prospective Cohort
- Question: Are non-steroidalanti-inflammatory drugs arisk factor for GI bleeds?
- 问题:非甾体固醇抗炎药是否为胃肠道出血的危险因素?
第27页:Example
Example
- Identify cohort :newdiagnoses of rheumatoidarthritis from October 17, 2002to October 17, 2003
- Determine exposurestatus :identify patientsprescribed NSAID’s and thosewho are not
- 确定队列:2002.10-2003.10诊断的类风湿关节炎患者
- 确定暴露状态:确定开与未开NSAID’s药物处方的病人
第28页:Determine outcome
Determine outcome
- 确定结局:随访所有的病人1年—确定在暴露及非暴露组出现多少胃肠道出血的病人Example
第29页:Prospective Cohort
Prospective Cohort
- time sequence strengthensinference about cause
- more accuratemeasurement of riskfactors (do not need toreconstruct past exposures)
- more completemeasurement of confounding优点
- 因果推断的时间顺序强
- 危险因素测量跟准确(不需要够姜过去的暴露)
- 更准确的测量混杂因素
第30页:Disadvantage
Disadvantage
- expensive and timeconsuming (inception andfollow-up)
- large numbers required tostudy rare outcomes
- difficult to study chronicdiseases with long latency缺点:
- 昂贵,费时间(启动及随访)
- 研究罕见结局是需大样本
- 难与研究慢性潜伏期长的疾病Prospective Cohort Studies
第31页:前瞻性队列研究应用条件
前瞻性队列研究应用条件
- 明确地检验假设
- 所研究疾病的发生率较高,一般不低于5‰
- 明确规定暴露因素和结局变量
- 可靠的测量手段
- 足够的观察人群和暴露情况
- 能完成随访的人群
- 足够的人、财、物力
第32页:Retrospective
Retrospective cohort studytimeExposureStudy startsDiseaseoccurrence
第33页:Historical
Historical
- Identify cohort in the past
- E.g., through records or administrative databases
- Determine exposure or prognostic factors in thepast
- Again, records or databases
- Identify outcome
- Outcome can be identified in past or present
- Outcome must be after previous two steps
第34页:Historical Cohort
Historical Cohort
- Question: Are non-steroidal anti-inflammatory drugs arisk factor for GI bleeds?
- 问题:非甾体固醇抗炎药是否为胃肠道出血的危险因素?
第35页:Historical Cohort
Historical Cohort
- Identify cohort–e.g., people diagnosed withrheumatoid arthritis between January andDecember 1992
- Determine exposure status. Which of these patientswere prescribed NSAID’s?
- Determine outcome. Did patient develop kidneydisease during the five years after inception intocohort? (follow-up lasted until December 1997)
第36页:Historical
Historical
- cohort easier to assemble(inception period in past)
- baseline measurementsalready available
- follow-up period alreadytaken place
- less costly and time-consuming优点:
- 队列易召集(在过去起始)
- 基线测量已经可以利用
- 已经随访
- 耗费少,省时间
第37页:Disadvantage
Disadvantage
- no control over the qualityof past measurements
- incomplete data sets
- control for confoundingmay be incomplete缺点
- 无法控制过去暴露的质量
- 资料不全
- 混杂因素控制不全Historical (Retrospective) Cohort Studies
第38页:Mixed Cohort
Mixed Cohort
- Question: Are non-steroidal anti-inflammatorydrugs a risk factor for GI bleeds?
- Identify cohort (patients who were diagnosedwith rheumatoid arthritis between October 2000and September 2001)
- Which of these patients are prescribed NSAID’s?
- How many of these patients will experience GIbleeds during next five years?
第39页:Fixed Cohort
Fixed CohortStartEnd
第40页:Cohort Study
Cohort Study
- Fixed Cohort
- A group of individuals recruited and enrolled ata uniform point in the natural history of adisease or by some defining event
- Cohort does not take on new members after it isassembled
- Examples
- Patients admitted to the ER with acute MI
- Survivors of Hiroshima bombings
- Children born to HIV-infected mothers
第41页:Types
Types
- Dynamic cohort
- A group of individuals recruited and enrolledthrough a mechanism that allows for in and outmigration of people
- Defined by characteristic other than disease, e.g.,geographic location, administrative unit
- Dynamic population
- Examples
- Framingham Study
- Kaiser Permanente
第42页:Dynamic Cohort
Dynamic CohortStartEnd
第43页:二硫化碳长期低剂量的暴露与冠心病的关系第
二硫化碳长期低剂量的暴露与冠心病的关系第二节实例研究
第44页:二硫化碳(CS
二硫化碳(CS
- 神经系统毒物,抑制酶的活性,影响脂蛋白代谢,造成心血管疾病
- 长期接触低浓度CS2可引起慢性中毒和动脉粥样硬化
- 短时间接触高浓度的CS2蒸气可急性中毒
第45页:研究因素长期低剂量的CS2暴露
研究因素长期低剂量的CS2暴露
- 定义在有CS2暴露但不至引起急性中毒的车间工作>5年20世纪60年代芬兰职业卫生研究所Hernberg和Tolonen教授做的前瞻性队列研究
第46页:二、确定研究结局
二、确定研究结局
第47页:三、确定研究现场和人群暴露组1942
三、确定研究现场和人群暴露组1942—1967年某粘纤厂25至64岁,343名男性工人有5年以上CS2暴露史对照组年龄±3岁出生地区相同工种的体力消耗相当在同一城市的造纸厂随机选择的343名男性工人
第48页:四,资料收集查阅档案记录用药情况
四、资料收集查阅档案记录用药情况、既往车间CS2的浓度等询问姓名、性别、年龄、工种及工作年限、吸烟、业余时间的体力活动情况实验室检查血糖、血脂、血清胆固醇水平、血压、心电图、心脏大小、体重及车间CS2浓度的动态变化
第49页:表4-1 暴露组和对照组的心肌梗死发生
表4-1 暴露组和对照组的心肌梗死发生:本页为数据表,下列为源文件提取的原文数据,供对照核查。
五、资料分析表4-1暴露组和对照组的心肌梗死发生率及RRCS2暴露组发生心肌梗死的RR为3.57,两组致死性心肌梗死发生率和总的心肌梗死发生率差异有显著性
本页含数据表,源文件由图片提取文本,数字与列边界粘连,无法安全自动切分为表格,需对照原课件人工还原。
第50页:表4-2 与不同临床类型冠心病的RR和
表4-2 与不同临床类型冠心病的RR和:本页为数据表,下列为源文件提取的原文数据,供对照核查。
CS2在不同临床类型冠心病的发生中作用程度不同ARRR临床类型5.253.57心肌梗塞3.214.69致死性心肌梗塞2.042.74非致死性心肌梗塞11.61.89心绞痛6.11.4心电图冠心样改变表4-2 CS2与不同临床类型冠心病的RR和AR比较
本页含数据表,源文件由图片提取文本,数字与列边界粘连,无法安全自动切分为表格,需对照原课件人工还原。
第51页:六、结论长期低剂量与冠心病发病和死亡存在
六、结论长期低剂量与冠心病发病和死亡存在因果关系CS2所致的冠心病,以致死性心肌梗死为主措施芬兰当局已于1972年把CS2的车间最高容许浓度从20ppm降至10ppm
第52页:Identify exposures
Identify exposures
- Idefine outcomes
- Identify study field andpopulation
- Sample size
- Data collectioan andFollow-up both groups
- Quality control
- 确定研究因素
- 确定研究结局
- 确定研究现场与研究人群
- 确定样本量
- 资料的收集与随访
- 质量控制第三节设计和实施Design and implement
第53页:确定研究因素
- Identify Exposure确定研究因素
- Main exposureBased on descriptive studyand case-control study
- Factor may effect outcomeconfounder, demographycharacteristic
- 主要暴露因素在描述性研究和病例对照研究的基础上确定
- 可能影响结局的因素混杂因素人口学特征等
第54页:Measuring Exposure
Measuring Exposure
- Content-Nature of the exposure; biologic mechanisms
- Quality
- Continuous-e.g., serum cholesterol
- Periodic-e.g., cigarettes, sexual contacts
- Singular-e.g., nuclear exposure
- Quantity
- Continuous and periodic exposures must bequantified
- Dose-response relationship
第55页:Measuring Exposure
Measuring Exposure
- Measurements
- Interview
- Medical examBlood tests or other specimens,Biomarkers, Other laboratory tests
- Records, medical records
- Sample storage
- Environment surveillance data
第56页:Measuring Exposure
Measuring Exposure
- Measuring exposure is one of the fundamentalactivities of a cohort study
- Exposure measurement must be comparable forall members of the cohort
- Carefully defined in advance of study
- Specific attention should be given to the accuracyand precision of proposed measurements
- Pilot studies often needed
第57页:Outcome Definition
Outcome Definition
- Primary outcome-the main event that will be relatedto the exposure
- Failure-time outcomes
- Death
- Disease occurrence
- Repeated measures
第58页:Outcome Definition
Outcome Definition
- Secondary outcomes-other events that are ofinterest and may corroborate the findings of themain outcome
- Using existed international standard
第59页:Examples of Outcomes
Examples of Outcomes
- Lung Cancer
- Heart Disease
- Motor vehicleinjury
- HIV infection
- Diabetes
- Diphtheria
第60页:Define Confounding
Define Confounding
- data collection
- data analysis
第61页:Study Population
Study Population
- Define Population at Risk using inclusion criteria
- Individuals with outcome of interest at time ofscreening and enrollment are not eligible for study
第62页:Selection
Selection
- The general population (i.e., the outcome of interesthas a high incidence rate)
- Special exposure groups (e.g., smokers, X-rayworkers
- Special occupational exposure group (uranium minersor asbestos workers. Cooperation, good record, regu.Exam, easy follow-up)
第63页:Selection
Selection
- Special resource groups (e.g., alumni, physicians,nurses, insured)
- Geographically or facility-defined groups (e.g., ThreeMile Island, hospitals with specialized maternity care)
第64页:Selection
Selection
- General population
- According to the exposure status group can bedivided into subgroups (exposure+and exposure-)
第65页:Specific comparison
Specific comparison groupExample: for effect of radiation if radiologistscohort group comparison group may beinternalizes.Textile workers for asbestos workers,Selection of Non exposed group
第66页:Study Populations
Study Populations
- Framingham study of cardiovascular disease
- Individuals 30–62 years old in communityat risk for disease
- Framingham, MA, 1948 to present
第67页:Framingham Study
Framingham Study Cohort AssemblyTotalNo.WomenNo. Men6,5073,4333,074Random Sample4,4692,4452,024Respondents740428312Volunteers4,3932,4181975Respondents free of CHD734427307Volunteers free of CHD5,1272,8452,282Total free of CHD
第68页:Study Populations
Study Populations
- Multi-Centered AIDS Cohort Study
- Goal to elucidate the natural history of HIV/AIDS
- 5000 gay men, volunteers
- 5 cities in US
- 1984–1999
- Extensive evaluations
- Questionnaire
- Physical examination
- Laboratory testing
- Repository
第69页:Sample Size
Sample Size
第70页:Follow-up Purpose
Follow-up Purpose
- Track subjects in both exposed andnon-exposed group
- Define outcome events
- Further collect data in exposures andconfounding
第71页:Follow-up
Follow-up
- Completeness and non-participation
- 90% rule of thumb
- All subjects must have an equal likelihood for detectingthe outcome
- Disease ascertainment must be comparable between theexposed and unexposed subjects
- Number of visits
- Reasons for additional evaluations
第72页:Follow-up mechanisms
Follow-up mechanisms
第73页:Follow-up Passive
Follow-up Passive
- Hospitals
- Disease Registries
- Clinics or physician offices
- Surveillance systems, e.g., National DeathIndex, CDC reportable conditions
第74页:Active surveillance
Active surveillance
- Systematic evaluations for outcome ofinterest
- Regular time intervals
- In all study subjects
- Regardless of active or passivesurveillance, the persons evaluating subjectsmust be blinded to exposure statusFollow-up
第75页:Follow upWhen
Follow upWhen
- Need to have time after the exposure for thedisease to develop.
- The induction period + incubation
第76页:Follow up periodHow
Follow up periodHow
- Individuals follow up to outcome(disease, death, syndrome)
- Cohorts follow up expected results
- Follow up interval depend on the incidence, latency.
第77页:THE ISSUE
THE ISSUE
- Some cohort (retrospective or prospective)studies extend over long periods of time.
- Difficult to track individuals and events.
第78页:If a large
If a large
- If loss to follow-up is differential between cohortgroups, and/or for reasons related to both theexposure and outcome, the validity of the studymay be in question.THE ISSUE OF FOLLOW-UP
第79页:调查员选择
调查员选择
- 调查员培训
- 制定调查员手册
- 监督
- Chose of investigator
- Training of investigator
- Formulation brochure forinvestigator
- Monitor质量控制Quality Insurance第三节设计与实施
第80页:Basic model
Basic model
- Person time calculate
- Rate calculate
- Effect estimation第四节资料的整理和分析Data sorting and analysis
- 资料的基本整理模式
- 人时的计算
- 率的计算
- 效应估计
第81页:表4-3 队列研究资料归纳整理表Dat
表4-3 队列研究资料归纳整理表Dat:本页为数据表,下列为源文件提取的原文数据,供对照核查。
一、资料的基本整理模式Basic model of data sorting合计total非病例non case病例casea+b=n1ba暴露组exposurec+d=n0dc非暴露组Non exposurea+b+c+d=tb+d=m0a+c=m1合计暴露组发病率=a/n1Incidenceofexposedgroup非暴露组发病率=c/n0Incidenceofnon-exposed表4-3队列研究资料归纳整理表Data sort table for cohort study
本页含数据表,源文件由图片提取文本,数字与列边界粘连,无法安全自动切分为表格,需对照原课件人工还原。
第82页:二、人时的计算
二、人时的计算
- 精确法
- 近似法
- 寿命表法
- Exact method
- Approximation method
- Life expectancy method
第83页:三、率的计算
三、率的计算
- 累积发病率
- 发病密度
- 标化死亡比
- 标化比例死亡比
- Cumulative incidence
- Incidence density
- Standardized mortality ratio,SMR
- Standardized proportionalmortality ratio, SPMR
第84页:变化范围0-1, ,0-1适用条件样本大
变化范围0-1range :0-1适用条件样本大suitable for : large sample人口稳定stable population整齐的资料even data报告时必须注明时间长短Note the time period when report累积发病率(cumulative incidence, CI )CI=观察期内发病(或死亡)Number of new case (or death)观察开始时的人口数Number of persons entering observation
第85页:发病密度变化范围0-∞
发病密度(incidence density)变化范围0-∞range: 0-∞适用条件观察时间长Suitable for : long observation人口不稳定Unstable population存在失访with lost follow up资料不很整齐data not evenID=观察期内发病(或死亡)人数Number of new (death) cases观察人时Observed person time
第86页:标化死亡比变化范围0-∞
标化死亡比(standardized mortality ratio, SMR)变化范围0-∞range:0-∞适用条件结局事件的发生率低suitable for:low incidence of outcome不宜直接计算率时not suitable calculate rate directlySMR=研究人群观察期内发病(或死亡)人数Number of new (death) cases in studied population标准人口预期发病(或死亡)人数Number of expected (death) cases in standard population
第87页:全人口某病的发病(死亡)率
全人口某病的发病(死亡)率×观察人口数New cases (death) of a disease in total population XNumber of observed persons预期发病(死亡)数的计算:Expected cases (death):
第88页:的意义被研究人群发生(死于)某病的危险性
SMR的意义被研究人群发生(死于)某病的危险性是标准人群的多少倍The times of risk of a disease (or death) in study populationover standard populationSMR=1研究人群某病发病(死亡)危险=标准人群If the SMR is quoted as a ratio and is equal to 1.0, then thismeans the number of observed deaths equals that of expectedcases.
第89页:研究人群某病发病(死亡)危险
SMR>1研究人群某病发病(死亡)危险>标准人群,是标准人群的SMR倍If higher than 1.0, then there is a higher number of deaths thanis expected, is the SMR times of standard populationSMR<1研究人群某病发病(死亡)危险<标准人群
第90页:标化比例死亡比
标化比例死亡比(standardized proportional mortality ratio, SPMR)变化范围0~∞适用条件不能得到历年人口资料仅有死亡人数、原因、日期和年龄SPMR=Actually death No.Expected death No.Range0~∞Suitable for:no yearly population data,only there is death No., cause,date and age of death
第91页:预期死亡数计算
预期死亡数计算(Expected death No):全人口中某病因死亡数全部死亡数×某单位实际全部死亡数Death due to a diseaseTotal No. of Death×Actually death No. in somedepartment
第92页:率的显著性检验
率的显著性检验
第93页:四、效应的估计
四、效应的估计
- Relative Risk (RR)
- Attributable Risk (AR)
- AR Percent (AR%)
- Population AR (PAR)
- PAR Percent (PAR %)
- Dose response relationship相对危险度归因危险度归因危险度百分比人群归因危险度人群归因危险度百分比剂量反应关系
第94页:相对危险度(, , ,)意义
相对危险度(Relative Risk RR)意义implicationE发病或死亡的危险是Ē的多少倍the times of the probability of the disease or death occurring in the exposed groupversus a non-exposed group.RR值暴露的效应暴露与结局关联强度RR effect of exposed the association between exposure and outcome暴露组率Rate inexposed非暴露组率Rate in nonexposed
第95页:表4-4 吸烟者与非吸烟者死于不同疾病
表4-4 吸烟者与非吸烟者死于不同疾病:本页为数据表,下列为源文件提取的原文数据,供对照核查。
意义implication吸烟者因肺癌死亡的危险是非吸烟者的10.7倍Smokers would be 10.7 times as likely as non-smokers to die oflung cancer吸烟者因心血管疾病死亡的危险是非吸烟者的1.7Smokers would be 1.7 times as likely as non-smokers to developcardiovascular diseaseCardiovascular disease170.321.7表4-4吸烟者与非吸烟者死于不同疾病的RRRR of death from different diseases in smoker and nonsmokerLung cancerDisease296.7550.12Smoker4.69Non-smoker10.7RR(1/10万人年)
本页含数据表,源文件由图片提取文本,数字与列边界粘连,无法安全自动切分为表格,需对照原课件人工还原。
第96页:表4-5 与关联强度RR and st
表4-5 与关联强度RR and st:本页为数据表,下列为源文件提取的原文数据,供对照核查。
表4-5 RR与关联强度RR and strengthen of association很强very strengthen10~<0.1强Strengthen3.0~9.90.1~0.3中Middle1.5~2.90.4~0.6弱Weak1.2~1.40.7~0.8无No association1.0~1.10.9~1.0关联强度RRstrengthen of association
本页含数据表,源文件由图片提取文本,数字与列边界粘连,无法安全自动切分为表格,需对照原课件人工还原。
第97页:的95%
的95%
- dcbaRRVar1111
- ln
- RRVarRRln96.1±lnln
第98页:归因危险度意义
- 归因危险度(Attributable Risk, AR )意义implication
- E与Ē人群比较,所增加的疾病发生数量The increased No. of diseases comparing an exposed population andan unexposed population
- AR值暴露因素消除后所减少的疾病数量Eliminating the exposure, the reduced No. of diseasesca或
- 1000
- ×
- RRIIIRRAR010nnIIARe
第99页:表4-6 与AR的区别The diff
表4-6 与AR的区别The diff:本页为数据表,下列为源文件提取的原文数据,供对照核查。
意义RR吸烟对肺癌的病因学意义较大AR戒烟对心血管疾病的预防作用较大即公共卫生意义较大表4-6 RR与AR的区别The difference between RR and ARCardiovascular diseaseLung cancerDisease1.710.7RR126.43170.32296.7545.434.6950.12ARNon-smokerSmoker(1/10万人年)ImplicationRR smoking has larger etiologysignificance of on lung cancerAR stop smoking has higher preventablesignificance on cardiovascular diseases, i.e.the public health significance
本页含数据表,源文件由图片提取文本,数字与列边界粘连,无法安全自动切分为表格,需对照原课件人工还原。
第100页:归因危险度百分比AR%(病因分值
归因危险度百分比AR%(病因分值
- %100%0×
- eIIIeAR%1001%×
- RRAR
第101页:人群归因危险度
人群归因危险度(population attributable risk, PAR)意义暴露人群与一般人群比较,所增加的疾病发生率的大小The increased disease in the exposed comparing with general populationPAR值暴露因素消除后所减少的疾病数量PAR the reduction in incidence after eliminate the exposurePAR= It-I0It:总人群率rate in total populationIo:非暴露组率rate in Non-exposed population
第102页:人群归因危险度百分比PAR%意义
人群归因危险度百分比PAR%意义implicationPAR占总人群全部发病(或死亡)的百分比The proportion of PAR in cases or (death) in total population或Pe:总人群的暴露比例proportion of exposed in total population
第103页:剂量反应关系
剂量反应关系
- List the incidence ofdifferent exposed levels
- Calculate the RR and ARof different exposed levelsusing the lowest asreference
- Carried out tendency testif necessary
第104页:表4-740 岁男子按初始血清胆固醇分组
表4-740 岁男子按初始血清胆固醇分组:本页为数据表,下列为源文件提取的原文数据,供对照核查。
结果血清胆固醇水平患冠心病的RR说明存在剂量效应关系表4-740-59岁男子按初始血清胆固醇分组冠心病6年发生情况血清胆固醇(mmol/L)人数病例数危险度平均年发病率RRAR<210454160.03520.00591.000.0000210-455290.06370.01061.810.0285>245424510.12030.02003.390.0851合计1333960.07200.0120
本页含数据表,源文件由图片提取文本,数字与列边界粘连,无法安全自动切分为表格,需对照原课件人工还原。
第105页:选择偏倚
选择偏倚
- 失访偏倚
- 信息偏倚
- 混杂偏倚第五节偏倚及其控制Bias and control
- Selection bias
- follow up bias
- information bias
- confounding bias
第106页:一、选择偏倚
一、选择偏倚selection biasSubjects are different withgeneral population ortargeted total populationin some important aspect,which result in theoutcome of research bias.研究人群在一些重要因素方面与一般人群或待研究的总体人群存在差异,而导致研究结果的偏倚。
第107页:产生原因
产生原因
- 选择对象的方法不当
- 最初选定参加研究的对象中有人拒绝参加
- 历史性队列研究中部分档案丢失或记录不全
- 志愿者队列
- 研究开始时未能发现早期病人等
- Unsuitable Selectionmethod for subject
- The selected subjectedrefuse to participate
- File lost or uncompletedrecord in historical cohortstudy
- Volunteer cohort
- Unidentified early patientin cohort
第108页:控制
控制
- 预防为主
- 抽样方法正确,
- 严格按规定标准选择对象
- Prevention first
- Correct sampling method
- Select subjected strictlybased criteria
第109页:二、失访偏倚
二、失访偏倚lost to follow-upThe bias due to migration,or died of non-outcomedisease or refuse tocontinue participant thestudy and exit the cohort研究对象因迁移、外出、死于非终点疾病或拒绝继续参加观察而退出队列所引起的偏倚。
第110页:控制
控制
- Design:select population easy to follow upincrease 10% sample size
- Implementadministration of interviewerformulate follow up plan andsurveillance measures middle termanalysis
- Sorting datamake up lacking items设计选择便于随访的人群在计算的研究样本的基础上扩大10%实施加强对随访员的管理制定随访计划和监测措施期中分析整理资料对于有缺项或漏项的对象进行补查
第111页:三、信息偏倚
三、信息偏倚Information bias在获取暴露、结局或其他信息时所出现的系统误差或偏差.Information bias is the systematic error or distortionoccurs in the data collection including exposure, outcomeor other information.
第112页:产生原因,疾病
产生原因cause疾病、暴露标准不明确检验仪器不精确、检验技术不熟询问技巧不佳记录错误,造假等The criteria of disease andexposure unclearEquipment not accuracy, testskill not adeptInterview skill, recordmistake and even cheating
第113页:控制,提高临床诊断技术
- 控制control提高临床诊断技术、明确各项标准选择精确稳定的测量方法事前调准仪器严格实验操作规程同等对待每个研究对象培训调查员,提高技巧,统一标准
- Increase diagnostic technique,clarified every criteria
- Selecting stable detect method
- Correct equipment
- Rigours standard manoeuvresprotocols
- Use the same standard forinterviewing study andcomparison groups
- Train interviewers according tostandard protocols
第114页:四、混杂偏倚
四、混杂偏倚confounding bias与所研究因素和结果均有联系的第三因素在暴露组与对照组的分布不均衡,混淆了研究因素和结果间的真实联系。
When a factor (confounder) associated with the exposure ofinterest is also associated with development of the disease or outcomeof interest independently of exposure. Therefore, a distorted estimateof the exposure effect results.
第115页:控制,设计阶段限制研究对象
控制Control设计阶段限制研究对象,匹配分析阶段分层分析、标准化或多因素分析DesignRestriction and matchingAnalysisStratified analysis ,standardize and multivariateanalysis
第116页:第六节优点与局限性
第六节优点与局限性Strengthen and limitation
第117页:一、优点、直接获得暴露组和非暴露组的发病
- 一、优点Strengthen直接获得暴露组和非暴露组的发病率或死亡率直接估计危险度符合时间顺序,验证病因的能力较强获得一种暴露与多种结局的关系收集的资料完整可靠,不存在回忆偏倚可研究疾病的自然史
- incidence or death rate ofE or Non-E group
- Directly estimate RR
- Temporal sequence
- One exposure, moreoutcome
- No recall bias
- natural history
第118页:二、缺点、不适于发病率很低的疾病的病因研
- 二、缺点Limitation不适于发病率很低的疾病的病因研究易发生失访偏倚耗时,耗人力、物力、财力设计要求严密,资料的收集和分析难度较大随访过程中,已知变量的变化或未知变量的引入增加分析难度
- Inefficient for rare diseases
- Bias of loss to follow-up
- Time, labor, material andmoney consuming
- Strict design, difficult in datacollection and analysis
- Unknown variable in followingup