《病例对照》PPT课件

病例对照研究,Case Control Study,Epidemiology is the study of the distribution and determinants of disease frequency in human population. Epidemiology: Principles and Methods by Brian MacMahon 1996,Descriptive Epidemiology,Concept Map,Case-Control Cohort study,Experimental Epidemiology,c.s. Ecology scre,Descriptive Epidemiologic Process,WHO was affected? WHERE were they affected? WHEN were they affected?,HOW and WHY?,Descriptive Studies: Uses,Hypothesis generating Suggesting associations Evaluating intervention effect Studying health situation of population,Analytical Epidemiologic Process,HOW the way? WHY reason?,Causal inference,第一节 基本原理 Basic Principle,基本原理 研究特点 研究史 应用,Basic Principle Characteristics History Application,Case control study is also called Case-comparison study, Case-referent study and Retrospective study which study the relationship between exposure and disease. Comparing with cohort study, case control study is saving on time, works, cheap and quick in obtaining the results. This method is suitable to explore risk factor of disease, effect of treatment and side effect of medicine and so on.,第一节 基 本 原 理,Cases, diagnosed “some disease” Control, without disease but comparable to cases Collecting exposure history by recall, lab test or review medical history.,第一节 基 本 原 理,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,基本原理,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”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 interest, such as disease, health, effect, cure, death, side effect ，or sign or syndrome of a disease.,第一节 基 本 原 理,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,“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 of 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,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(Demo 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,第一节 基 本 原 理,研究示意图（demonstration graph of study),表1 病例对照研究资料整理表 Table 1 data analysis for case control study,OR：(a/c)/(b/d)=ad/bc,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,第一节 基 本 原 理,病例对照研究特点 Characteristics,Retrospective Category by outcome From outcome to cause Observational method Could not verify hypothesis in general,1843年Guy向统计学会报告-最早病例对照研究 分析职业暴露与肺结核 发生的关系 1844年Louis的著作 最早出现病例对照研究 的概念 1926年Lane Claypon 报告 生殖因素与乳腺癌关系研究,研究简史 History,第一节 基 本 原 理,1843 Guy reported Statistics Association- the earliest case control study Analyzed the association: occupational exposure &TB 1844 works of Louis case control study appeared first time. 1926 Lane Claypon reported association: reproduction factor & breast cancer,1947年Schreck和Lenowitz 包皮环切和性卫生与阴茎癌的关系 1947年Hartwell 输血与肝炎关系的研究 1950年Doll和Hill 吸烟与肺癌的研究,第一节 基 本 原 理,1947 Schreck and Lenowitz peritomy, sex health and carcinoma of penis 1947 Hartwell Transfusion and hepatitis 1950 Doll和Hill smoking and lung caner,20世纪60年代以来 Since 60s of 20 Century,第一节 基 本 原 理,孕妇服用反应停（thalidomide）与婴儿短肢畸形 母亲吸烟与先天性畸形 早产儿吸入高浓度氧与晶体后纤维组织增生症,Thalidomide RLF,20世纪60年代以来 Since 60s of 20 Century,小剂量电离辐射与白血病 母亲早孕期服用雌激素与少女阴道腺癌等之间的关系 经期使用月经棉与中毒性休克综合征,Low dose ionization radiation and Leukemia Estrogen using pregnancy and adenocarcinoma of v Tampon & toxic shock syndrome,应用 Application,1、广泛地探索疾病或事件的可疑危险因素 2、深入检验某个或某几个病因假说 3、为进一步前瞻性研究提供依据,Exploring the possible risk factor widely Deeply test a or several hypothesis Provide basis for cohort study further,应用 Application,4、治疗方法或药物有害作用的研究 5、疾病预后因素的研究 6、诊断试验、防治措施的效果的评价,4. Study on the effect of treatment, side effect of drug 5. Study the factor of prognosis 6. The evaluation of diagnostic testing and prevention measure,第二节 研究类型 Classification,按研究目的分类 按病例、对照的来 源分类 按病例与对照是否 匹配分类 衍生的研究类型,Based on purposes Based on source of case and control Based on matching or not New derived types,一、按研究目的分类Based on purpose,1、探索性病例对照研究 2、验正性的病例对照研究,Exploring case control study Testing case control study,二、按病例、对照的来源分类 Classification on source of subjects,医院基础上的病例 对照研究 病例和对照均来自医院 或诊所 社区基础上的病例 对照研究 病例和对照均来自于事 先界定的目标人群社区,Hospital-based case- control study cases and controls are selected from a pre-determined group of hospitals or clinics Population-based case-control study cases and controls are selected directly from an a-priori defined target population,Incidence type case-control All cases are incident (new) cases arising from an identifiable target population in specified period of time. Controls are all or a random sample from the same target population. Prevalence type case-control All cases are prevalent cases (new and existing) from an identifiable target population in a specified period of time. Controls are all or a random sample from the same target population.,第二节 研究类型,1、病例与对照不匹配 从设计所规定病例和对照人群中，分别抽取一定量的研究对象 对照组人数病例组人数 对照应代表产生病例人群,三、按病例与对照匹配与否分类： Classification based on matching,C C are unmatched sampling c and c respectively from target population No. of control group No. of case group Control can represent population in which cases come from,病例与对照不匹配 Case and control not matching,病例组 Case group,对照组 Control g,目标人群 Target population,第二节 研究类型,第二节 研究类型,2、病例与对照匹配 Matched case control study,匹 配/配 比 要求：对照在某些因素或特征上与病例保持一致 目的：对两组进行比较时排除匹配因素的干扰 分 类 频数匹配 个体匹配,Matching Requirement：controls are similar to cases in certain characteristics Purpose：to eliminate the interference of matched factors Typing frequency matching, Individual matching,第二节 研究类型,频数匹配(成组匹配） 匹配因素所占的比例在对照组与病例组一致,frequency matching (Group matching ) selection of controls in proportion to the distribution of certain characteristics of the cases.,个体匹配 以病例和对照个体为单位进行匹配,给每一个病例选择一个或几个对照，使对照在某些因素或特征方面与病例相同或基本相同.,individual matching selection of one or more controls for each case, who have the same or similar specified characteristics as that case.,病例与对照匹配 Matched 1:1个体匹配 (individual matching),某些特征 Some characteristics,配对 (pair matching),第二节 研究类型,病例与对照匹配 Matched 1:R个体匹配 (1:R individual matching),某些特征 Some characteristics,1:2, 1:3 1:R,第二节 研究类型,匹配 matching,定性指标 qualitative indicator,第二节 研究类型,匹配 matching,定量指标 quantitative indicator (1),53 years old (5059age group）,5059 years,第二节 研究类型,匹配 matching,定量指标 quantitative indicator (2),第二节 研究类型,第二节 研究类型,匹配法注意事项 慎重选择匹配因素 可疑病因不作为匹配因素 比例一般为1:1，最多不超过1:4 避免 “匹配过度”,Notes for matching carefully select matching factor suspect cause shouldnt as matching factor generally the odds for is 1:1, it better no more than 1:4 avoiding overmatching,提高研究效率 控制混杂因素,increasing efficiency control confunding,匹配的目的 Purpose of matching,“匹配过度（overmatching）” 把不必要的项目列入匹配,力图使病例与对照尽量一致, 就可能陡然丢失信息. 增加工作难度, 结果反而降低了研究效率 。 refers to the unnecessary or inappropriate use of matching in a cohort or casecontrol study， which increases .,三、衍生的研究类型 derived study design,病例对照与队列研究作为病因研究的主要方法，有其各自的优势与不足，而且这些优势与不足相互补充，因此，提出了一些新的研究类型，结合使用这两种方法，扬长避短。,第二节 研究类型,case control study and cohort study are the main method for etiology study. Both the 2 methods have advantage and limitation. Therefore new methods are developed to adopt good points and avoid short-comings,Nested case control study Case cohort study Case case study Case crossover study Case time control study,巢式病例对照研究 病例队列设计 单纯病例研究 病历交叉研究 病例时间对照设计,基本原理 按队列研究方式进行 选择一队列，收集基线资料，采集所研究的生物学标志的组织或体液标本储存备用,Basic principle,第二节 研究类型,巢式病例对照研究（nested case control study）,In the way of cohort study Select a cohort， collect baseline data， collect biological tissue or body liquid specimen，store them for test late,随 访 随访到出现能满足病例对照研究样本量的病例数为止。,Follow up Follow up until enough cases occurring for case control study,匹配 按病例进入队列的时间、疾病出现时间与性别、年龄等匹配条件，从同一队列选择1个或数个非病例作对照，抽取病例与对照的基线资料并检测收集的标本,Matching Based on the time of enter cohort, occurring disease, age a gender of cases, one or more controls are selected and withdraw the data and specimen of case and control of base line.,资料处理 按匹配病例对照研究方法处理资料,Data analysis Data analysis methods are same as case control study,1992年Ross用巢式病例对照研究: 上海地区肝癌与尿中黄曲霉素生物学标志关系 18244名中年男性队列中发现22例肝癌 每例配5或10个对照 检测研究开始时的尿样： 发现黄曲霉素B1及其代谢产物和DNA加成物的OR值经 调整混杂因素后，为3.8(1.212.2)。 黄曲霉素作为肝癌致病因素最直接的证据,第二节 研究类型,病例-队列研究（case-cohort study）,基本原理 研究开始时，在队列中随机选取一组样本作为对照组 观察结束时，队列中出现被研究疾病所有病例作病例组 与随机对照组进行比较,Basic principle Controls: a random sample from the cohort at risk at the start of study. Cases: all incident cases in a given risk period Comparing the above 2 groups,第二节 研究类型,这种研究模式，可同时研究几种疾病，不同疾病有不同病例组，但对照组都是同一组随机样本,This study method can research several disease in the same time, there are different cases for each disease, however the control group is the same, the random sample.,病例-队列与巢式病例对照研究的区别Difference between case cohort and nested case control study,对照是随机选取，不与病例进行匹配 随机对照组中成员如发生被研究疾病，既为对照，又同时为病例 1个随机对照组可以同时和几个病例组比较分析,Controls are random and not matched with case If the disease occur in the control group, it will be both the case and control in the same time One control group can be used for several case group to compare and analysis,第二节 研究类型,病例-队列与巢式病例对照研究的优点Advantage of case cohort and nest case control study,因果关系清楚 资料可靠 论证强度高 省时省力省钱 适合于分子流行病学研究,Clearly Causal relationship Data is reliable The high strengthen of evidence Saving on time, works and fund Suitable to molecular Epi,第二节 研究类型,背 景 1994年Piegorseh、Begs等提出 遗传与环境的关系 交互作用 病例对照研究和队列研究的效率费用,Background 1994 Piegorseh,Begs et al proposed genetic environment interaction the efficiency and expense of case and control study,病例病例研究（case case study）,第二节 研究类型,应用前提条件 在正常人群中基因型与环境暴露各自独立发生 所研究疾病为罕见病(此时可用OR来估计RR值),Prerequisite The distribution of genotype and environmental exposure are independent in general population the disease is rare (ORRR),基本原理,第二节 研究类型,第二节 研究类型,研究示意图Demonstration graph,单纯病例研究,环境暴露exposure,基因型genotype,病人 patient,+,-,+,-,+,-,第二节 研究类型,OR=ad/bc 当ORge1时,有正相乘模型交互作用 positive interaction 当ORge=1时, 无相乘模型交互作用 no interaction 当ORge1时, 有负相乘模型交互作用 negative interaction,E+,E-,G-,G+,b,a,d,c,表2 环境暴露与基因型 Table 2 environmental exposure and genotype,解决问题 1991年美国Maclure提出病例交叉设计 日常生活中一些突发事件之后,常会伴随某些结果的发生。究竟是这些突发事件导致了结果的发生,还是仅仅由于机会所致?,病例交叉研究(case-crossover design),第二节 研究类型,Question The case crossover design was introduced in 1991 by Maclure M To study the transient effect of brief exposure on the occurrence of a rare acute-onset disease. Is the onset due to accidence or chance?,暴露与某急性事件有关 比较同一研究对象暴露情况 在急性事件发生前一段时间与未发生事件的某段时间如果暴露与少见的事件(或疾病)有关,那么刚好在事件发生前一段时间内的暴露频率应该高于更早时间内的暴露频率。,Exposure is related to some acute onset disease Comparing the exposure situation in the same subject These events should occur much more frequently during a period immediately prior to the onset of disease than at a similar period which is more distant from the disease onset,基本原理 basic principle,第二节 研究类型,研究对象 Study subject 病例和对照,两部分的信息均来自于同一个体 “病例部分” 危险期,疾病或事件发生前的一段时间 “对照部分” 对照期,危险期外特定的一段时间 研究是对个体危险期和对照期内的暴露信息进行比较,Study subjects The information of case and control comes from on individual. The “case part” is during the hazard period, that is prior the onset of disease or event. The “control part” is information outside the hazard period. It is actually the information comparison between hazard period and general.,第二节 研究类型,例如 据报道某种药物可以引发猝死，如果该报道正确，则应该可以观察到服用此药物后一段时间内猝死增多，或者说在猝死前几天或几周内应有服药增多的报道。,第二节 研究类型,第三节 研究设计与实施 The design and implement,研究步骤 病例与对照的选择 样本含量的估计 资料的来源与收集方法,Steps The Selection of case and control The estimation of sample size Data resource and collection method,提出假设 根据以往疾病分布的记录或现况调查得到的结果，结合文献资料，提出病因假设,Formulating Hypothesis Based on the past records or results of cross-sectional study， combined the literature to formulate hypothesis,一、研究步骤 Steps,明确适宜的研究类型 （Decide study type）,选择病例与对照比较的方法：匹配或成组比较法，或两者同时使用 如用匹配法，需确定病例与对照的比例及匹配条件,Comparing method of case and control：pair matching， grouping matching or both If matching，define the condition of matching,如果研究目的是广泛地探索疾病的危险因素，可以采用不匹配或频数匹配的方法 罕见病，或符合规定的病例数很少时，则用体匹配方法。,For exploring study，use nonmatchig or frequency match For rare disease，use individual matching,病例与对照的选择Selection of case and control,对照的选择是整个研究的关键之一 研究得出的结论是否可靠，首先要看对照的选择是否合理。,Selection of control is the key of a study If the conclusion is reliable of study depends on the reasonability of the selection of control,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,Basic principle of control selection： representativeness comparability （more important） Selection method sampling matching,病例的选择 Selection of Case,选择病例的要求 诊断可靠 使用金标准 选择确诊的新病例 回忆偏倚小 代表性好 容易合作 被调查因素改变少,Requirement for case selection reliable diagnosis golden standard Selection new cases few recall bias representiveness cooperation few change of exposure,来 源 Resource,总体人群中全部病例或总体随机样本人群中全部,Total case or total sampling cases in study population,代表性好 工作开展比较困难 耗费人力物力,病例,case,Better representiveness Difficulty to conduct Time, working consuming,医院住院或门诊的病例,Cases of inpatient or outpatient,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,Cooperation Data easy collection and reliable Comparabality Worse representiveness,对照的选择 Control Selection,原 则 候选对象必须来自于产生病例的总体 意味着对照一旦发生所研究的疾病，就能成为病例组的研究对象。,Principle Candidate should come from the same population with cases This means that control become case in case he/she suffered the disease of cases.,对照与病例比较的方法 成组比较法 匹配法,Comparing method of case and control group comparison matching,频数匹配 个体匹配,Frequency matching Individual matching,来 源 研究的总体人群 或抽样人群 医院中患有其它 疾病的病人 亲属、邻居、同事、 同学等,Source Total study population or sampling patient with other disease in hospital Relatives, neighbor, colleague, classmate,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,Parameters Exposure rate in population Approximate RR or OR ( Literature, or pilot study) value power,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 Exposure rate in case,对照组暴露率Exposure rate in control,例：为研究某市肺癌与吸烟的关系，欲进行一次病例对照研究。 某市普通人群中吸烟率 P0=30% OR=5.0 =0.05 把握度0.90 问需要多少病例与对照?,、值查表，Z为1.64， Z为1.28 代入公式得： 病例组与对照组，各需28人。 如代入近似公式计算，所得结果与原公式很接近(n=29.5),第三节 研究设计与实施,病例组与对照组例数不等时的公式：（No.caseNo.control） 病例数（case）:对照数（control）=1:c 对照组例数为 （No. of control）： c*n,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,研究因素的选定与 测量 根据研究的目的或具体的目标，确定调查变量的数目和每一个变量的具体项目。,The selection and measure of exposure Based on purpose and detail aim to decide the No. of variable and items for each variable,每项变量都要有明确的 定义。 研究中应尽可能地采用定量或半定量的量度。,Each variable should have its clear definition Use quantitative or semi- quantitative variable as much as possible,资料来源与收集,资料的收集在病例对照研究中十分重要 方式方法不恰当，收集的资料就不可靠 统计处理无法纠正的系统误差,Data analysis is very important in case and control study. Unsuitable method of data collection will result in unreliable data There is no statistics method that can correct the bias,资料来源 医院病案记录 疾病登记报告等摘录 致病因素数据需检测病人的标本或环境获得 询问调查获得,Data resources medical records disease register measure of risk factor or specimen of patient interview,第四节 资料的整理与分析 Data sort and analysis,资料整理 资料分析 描述性统计 统计推断,Data Sort Data analysis Descriptive statistics Statistics inference,描述性统计 描述一般特征 均衡性检验,Descriptive statistics Description of general features Test of homogeneous,均衡性检验 检验两组在研究因素以外其它主要特征有否可比性 两组间非研究因素均衡可比，才能认为两组暴露率差异与发病有关,Tests for homogeneity Test the comparability of other main features in 2 groups Only other main features are homogenous, the difference of odds of exposure is related to disease,均衡性检验，应把两组的这些特征逐一加以比较，必要时作显著性检验,Homogenous test compare each feature and carry out significant test if necessary.,统计性推断 成组比较法资料的分析 1:1配对资料的分析 混杂因素作用的估计与分层分析,Statistics inference data analysis for group comparison data analysis for 1:1 match the estimate of confunding and stratified analysis,Data analysis for unmatched study 2test test if the factor is statistically associated with disease 2X2 table antel-Haenszel method,成组比较法资料的分析 2检验 是检验研究因素与疾病之间有否统计学联系？ 可用传统的四格表公式 也可用Mantel-Haenszel (M-H)方法,四格表法 2X2 table method,M-H method,两种方法计算结果一致，表明吸烟与食管癌有联系，但联系强度如何，要计算OR。 The results of 2test and M-H Method are same. Smoking is associated with esophagus cancer. However strength of association is need to calculate OR.,OR计算及意义 病例组中暴露人数与非暴露人数的比值除以对照组中暴露人数与非暴露人数的比值 The odds ratios is that the odds of exposure in the cases to the odds of exposure in the controls.,与RR一样，优势比反映暴露者患某种疾病的危险性较无暴露者高的程度。 前提条件 所研究疾病的发病率（死亡率）很低 病例对照研究中所选择的研究对象代表性好，则OR值就很接近甚至等于RR值,Similar to RR, OR reflect the risk of disease for exposed to unexposed. Prerequisite the incidence (mortality） is very the representativeness of subject is good,结果表明吸烟口服避孕药患心肌梗死的危险性是不口服避孕药2.20倍，95%的可信范围是在1.263.84之间,2=6.63,1:1 配对资料的分析 匹配资料是由病例与对照结合成对子，分析结果时不应把对子拆开分析 先将资料列成表5-4的格式 注意表内的数字a、b、c、d是病例与对照配成对的对子数,2=11.28, OR=4.33 表明男性吸烟者患食管癌的危险性是不吸烟者的4.3倍 OR95%可信限,Table 4 Smoking history in 93 pairs male esophagus cancer and control,混杂因素作用的估计与分层分析 配比法 分层分析,The estimation of effect of confounder and stratified analysis Matching Stratified analysis,按饮酒与否分层 计算OR值,年龄小于40岁与大于40岁口服避孕药的OR分别为2.80，和2.78，均高于不分层OR(2.20) 表明年龄是混杂因素，年龄似可掩盖口服避孕药的作用,进行齐性检验 齐性检验 合并与OR值 OR值无明显差别，则年龄是混杂因素 合并OR值 OR值有显著差异，则年龄是明显混杂? 合并OR值 本例齐性检验无显著差异，计算合并OR与2,计算合并OR与2值 公式（M-H法）： 合并 合并,第五节 资料的整理与分析,=2.79,经年龄分层调整后，OC2与OR（即合并2与合并OR）虽较未调整的2 (6.60)与OR(2.20)为高，且统计学上的显著性。 年龄与心肌梗死之间有显著的关联 年龄是口服避孕药与心肌梗死之间的混杂因素，似能掩盖口服避孕药的作用,选择偏倚 Selection bias 信息偏倚 Information bias 混杂偏倚 Confounding bias,第五节 常见偏倚及其控制 Bias and control,一、选择偏倚 Selection Bias Subjects have different probability of being selected according to exposures or outcomes of interest, creating a biased measure of association (i.e. odds ratio, relative risk). 研究对象的选择由于暴露（或结局）不同，被选择的机会不同，使得测量暴露与结局的关系时产生偏倚。,伯克森偏倚 Berksons Bias 指以医院为基础的病例对照研究中，目标疾病的入院率因待研究的暴露因素的存在与否或暴露水平的不同而存在差异，由此而引起的误差。 If hospital based cases/controls study， the of rate of admission of target disease differ with exposed situation， OR will be biased.,现患病例新发病例偏倚 Neyman bias 检出证候偏倚 Detection signal bias 时间效应偏倚 Time effect bias,选择偏倚的控制 Control of selection bias 尽量合理地选择病例与对照 Selection case and control appreciate 尽可能地从多家(类)医院选择病例与对照 Select controls from more hospitals,二 、信息偏倚 information bias 回忆偏倚 recall bias 病例对照研究的主要弱点，很难完全避免 The main weakness of case control study and it is difficult to avoid totally.,三、 混杂偏倚 Confounding bias 病例对照研究遇到很多混杂因素 不易识别 not easy to recognize 不易确定 not easy to confirm 控制 Control 混杂因素作为匹配因素 matching 限制 restriction 分层分析 stratify analysis 多因素分析 multivariate analysis,第六节 研究实例 Example,研究背景 Background 研究方法 Study Method 研究结果 Results,研究背景,美国波士顿Vincent纪念医院妇产科医生Herbst发现 19661969年收治7例阴道腺癌患者 年龄1522岁 过去年龄均大于25岁 7例在1522岁之间 7例全是腺癌 阴道癌占女性生殖系统癌的2% 阴道腺癌仅占阴道癌的5%10%，非常罕见,第三节 研究实例,Arthur L. Herbst, M.D., Assistant，Clinical Professor of Obstetrics and Gynecology, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts, is interviewed by Sidney L. Arje, M.D., Vice President for Professional Education, American Cancer Society, Inc., New York, N .Y.,http:/caonline.amcancersoc.org/cgi/reprint/22/5/292.pdf,乙烯雌酚,阴道腺癌,第二节 研究类型,Herbst对阴道腺癌危险因素进行探索 7例患者加上另一个医院例患者作为病例组 每个病人配4个对照，共32个对照 调查员用调查表对病例、对照与她们的母亲进行了调查，经统计学处理后的主要结果见表5-2,研究方法 Method,第三节 研究实例,Herbst studied the risk factor of vagina adeicarcinoma 7 patient plused another diagnosed in other hospital as cases Each case 4 controls were matched Investigator invest cases and their mother stilbestrol,研究结果 Results,研究结果 Results,在比较的因素中，只有三个因素有显著差别 母亲怀孕期间使用过己烯雌酚激素治疗 (P0.00001) 母亲以前流产史(P0.01) 此次怀孕阴道出血史(P0.05),Only 3 risk factor showed significant difference mother used of stilbestrol during pregnancy (P0.00001) history of abortion previously (P0.01) vagina bleed in this pregnancy (