最新药物流行病学实验流行病学PPT文档.ppt
学习目的,实验流行病学的定义实验流行病学的特点实验流行病学的分类实验流行病学的设计实验流行病学的强势与弱势,推荐参考文献吴涛,詹思延,李立明.流行病学实验研究发展历史.中华流行病学杂志 2004;25(7):633-636.,历史,发展,1747年 苏格兰医师James Lind首次对比研究了用桔子、柠檬及其他干预治疗坏血病的疗效,On May 20,1747 he is ready to begin the first clinical nutrition experiment.,group 1 drank one quart of cider a day group 2 gargled with sulfuric acid group 3 had two spoonfuls of vinegar,3 times a day group 4 drank 1/2-pint seawater a day group 5 drank barley water group 6 ate two oranges and 1 lemon a day,How did he conduct his experiment?,What happened?Only sailors who ate the oranges and lemon felt better.In fact,in only six days they felt great and were able to start working again.The other sailors in the experiment felt worse.,1816年 法国Hamilton医师首次报道了爱丁堡的一项大型对照实验,评价放血疗法的效果,这是迄今为止有关采用交替法产生对照的最早记载之一。It had been so arranged,that this number was admitted,alternately,in such a manner that each of us had one third of the whole.The sick were indiscriminately received,and were attended as nearly as possible with the same care and accommodated with the same comforts.One third of the whole were soldiers of the 61st Regiment,the remainder of my own(the 42nd)Regiment.Neither Mr Anderson nor I ever once employed the lancet.He lost two,I four cases;whilst out of the other third treated with bloodletting by the third surgeon thirty five patients died.,1898年 丹麦医师Fibiger通过半随机对照实验验证血清治疗白喉的效果。8 out of 239patients in the serum treated group and 30out of 245in the control group died.No formal statistical analysis was performed but no objection can be raised against the statistical significance of the numbers,which were deemed correct by an inspector of the sick benefit association(in 1900Pearson invented the 2 test,10 which would have shown P=0.0003).The rate of serum sickness was high at 60%.,Tables-2-by-2 unstratified11:44:19,2011-4-22|+-|Total-+-+-+|8 231|239-|30 215|245-+-+-Total|38 446|484Tests of significanceFisher exact test(one tailed):0.000187 Fisher exact test(two tailed):0.000289 Uncorrected chi-square:13.24 p-value:0.000275 Yates corrected Chi-square:12.04 p-value:0.000522 Measures of exposure effect 95%CIRisk ratio:0.27 0.13,0.58Odds ratio:0.25 0.11,0.55Risk difference:-0.09-0.14,-0.04Proportional attributable risk:-2.66-6.82,-0.71Population proportional attr.risk:-0.56-0.76,-0.26,1948年 英国医学研究委员会领导开展了世界第一个临床随机对照实验(Randomized controled Trial,RCT),由英国统计学家Hill评估了链霉素治疗肺结核的疗效。,First“modern”RCT-1948,The hero was Austin Bradford Hill,Francis TM,Napier JA,Voight RB et al Evaluation of the 1954 field trial of poliomyelitis vaccine Final Report Ann Arbor University Michigan 1957,1954 field trial of poliomyelitis vaccine,Results of vaccine trials,The randomized,controlled experiment,The Observed Control study,Source:Thomas Francis,J r.,“An evaluation of the 1954 Poliomyelitis vaccine trials-summary report,”American Journal of Public Health vol 45(1955)pp.1-63.,实验流行病学(experimental epidemiology)以人群为研究对象的实验研究:又称:流行病学实验(epidemiological experiment)干预研究(intervention study),实验流行病学(experimental epidemiology),将来自同一总体的研究人群随机分为实验组和对照组,研究者对实验组人群施加某种干预措施后,随访并比较两组人群的发病(死亡)情况或健康状况有无差别及差别大小,从而判断干预措施效果的一种前瞻性、实验性研究方法。,实验流行病学研究原理示意图,结局,结局+,实验组(干预措施),对照组(对照措施),样本,目标人群,结局+,结局,主要特点对照前瞻干预随机,为什么要设立对照?因果性实验中的对照法则设立对照,为什么要设立对照?影响实验流行病学研究效应的主要因素不能预知的结局霍桑效应(Hawthorne effect)安慰剂效应(placebo effect)潜在的未知因素的影响,设立对照的方式标准方法对照安慰剂对照自身对照交叉对照,随访结局指标前因后果,暴露研究者施加,盲法(blinding)减少研究对象和研究者主观因素的影响(信息偏倚),Just kidding:双盲实验,患者家属:医生,流行病学家采用的双盲试验是这个意思吗?,医生,患者,护士,单盲(single blind),双盲(double blind),三盲(triple blind),开放试验(open trial),不可能实现盲法外科手术锻炼饮食教育,即使是药物也很难完全做到盲法药物的气味、味觉、感觉副作用检验结果对药物进行化验,来自总体的随机抽样人群随机分配到实验组和对照组,控制混杂的方法研究设计阶段随机化限制匹配资料分析阶段分层分析多变量分析,可比的对照有效的对照,自身前后比较,比较,已知结果,研究对象,试验组(干预组),对照组,发病时序,工作时序,资料的收集,数据分析,研究设计,结果解释,实验流行病学研究的实施,研究设计,明确研究目的选择研究对象确定实验现场设立对照确定样本大小随机化分组应用盲法,研究设计,明确研究目的选择研究对象确定实验现场设立对照确定样本大小随机化分组应用盲法,对干预措施有效的人群预期发病率较高的人群干预对其无害的人群能将实验坚持到底的人群依从性好的人群,排除(exclusions)减少结果偏倚(保证内部真实性)影响结果外推(影响外部真实性),研究设计,明确研究目的选择研究对象确定实验现场设立对照确定样本大小随机化分组应用盲法,相对稳定、足够的数量较高、稳定的发病率疫苗:近期未发生该病流行较好的医疗卫生条件领导重视、群众配合,研究设计,明确研究目的选择研究对象确定实验现场设立对照确定样本大小随机化分组应用盲法,有效对照安慰剂对照交叉对照,研究设计,明确研究目的选择研究对象确定实验现场设立对照确定样本大小随机化分组应用盲法,干预前结局指标,样本量干预措施实施前后的变化,样本量,样本量(1-),样本量单侧检验或双侧检验研究对象分组数量,实例,计数资料假设对照组的发病率为40%,通过干预措施发病率下降到20%才有推广使用价值,规定=0.01(双侧),=0.05,问两组要观察多少人?,干预措施实施前后的变化(1-)单侧检验或双侧检验研究对象分组数量,EpiCalc 2000,EpiCalc 2000,实例,计量资料假设对照组血清胆固醇水平为215mg/dl,合理膳食可以使干预组较对照组降低15mg/dl,已知从其他资料获得胆固醇标准差为25mg/dl,规定=0.05(双侧),=0.05,计算各组样本数?,EpiCalc 2000,EpiCalc 2000,研究设计,明确研究目的选择研究对象确定实验现场设立对照确定样本大小随机化分组应用盲法,简单随机分组 分层随机分组整群随机分组,简单随机分组,EpiCalc 2000,分层随机分组,总体,层,层,层,可按年龄、性别、种族、教育水平等分层,在各层内再进行前述的简单随机分组,整群随机分组,以家庭、学校、医院、村庄或居民区等为单位随机分组要保证组间可比性(群间变异越小越好),明确研究目的选择研究对象确定实验现场设立对照确定样本大小随机化分组应用盲法,研究设计,单盲 双盲三盲,数据分析,资料整理表结局指标频率指标计量指标效应指标,数据分析,频率指标治疗措施有效率治愈率病死率生存率复发率预防措施抗体阳性率发病率感染率,资料整理表结局指标频率指标计量指标效应指标,计量指标治疗措施血压胆固醇预防措施抗体几何平均滴度,数据分析,资料整理表结局指标频率指标计量指标效应指标,组间差异的统计学检验,数据分析,保护率效果指数,资料整理表结局指标频率指标计量指标效应指标,保护率(protective rate,PR)保护功效(protective efficacy,PE)归因保护/预防比例(attributable protective proportion),指标涵义对照组如果接受干预措施,则会减少的发病或死亡的比例试验组中归因于干预措施而减少的发病或死亡的比例,效果指数(index of effectiveness,IE),偏倚问题,偏倚(bias)使实验流行病学研究的结果系统的偏离真相,即低估或高估某干预措施的效果。,偏倚问题,选择偏倚信息偏倚混杂偏倚,研究初期严格、正确地执行随机化分配及盲法(隐藏分组情况),可以最大限度地减少实验流行病学研究中的选择偏倚反之,实验流行病学研究同样不能幸免选择偏倚,偏倚问题,选择偏倚信息偏倚混杂偏倚,随机分配后实验流行病学研究中涉及的选择偏倚主要来自于随机分配后的“退出”(withdrawal)不合格(ineligibility)不依从(noncompliance)失访(loss to follow-up),降低效率易引入选择偏倚,偏倚问题,选择偏倚信息偏倚混杂偏倚,随机分配后“退出”引起偏倚的原因退出者发生结局事件的概率不同于继续随访者(信息截尾)信息截尾的程度在不同干预组间不同,意向性分析(intention to treat analysis),RCTs Basic AnalysisDichotomous(Disease Yes/No)Outcome,N,Randomize,Rt=a/n1,Rc=c/n0,RR=Rt/Rc and ARR=Rc-Rt,Measures of Effect used in RCTs,EER=30/100=30%,CER=40/100=40%,EER=Experimental(or treatment)event rateCER=Control(or baseline)event rateRR=EER/CER=30/40=75%RRR=1 RR=25%ARR=CER-EER=40 30=10%NNT=1/ARR=1/0.10=10,Statistical Analyses-ITT,Intention-to-Treat Analysis Gold StandardCompares outcomes based on original randomization scheme regardless of eligibility,non-compliance,cross-overs,and lost-to-follow-upPer Protocol(PP)Analysis Compares outcomes based on actual treatment received among those who were compliant(analysis drops non-compliant)Asks whether the treatment works among only those that complyAs Treated(AT)AnalysisCompares outcomes based on actual treatment received regardless of original assignment.Equivalent to analyzing the data as a cohort study!Asks whether the treatment works among those that took it.Both PP and AT approaches ignore original randomization and are therefore subject to BIAS!,ITT vs.Per-Protocol vs.As Treated,Effect of ITT vs.PP vs.AT analyses on an RCT of coronary artery bypass surgery versus medical treatment in 767 men with stable angina.(Lancet 1979;i:889-93).,偏倚问题,选择偏倚信息偏倚混杂偏倚,确定偏倚(ascertainment bias)施加干预的人被干预的人判断结局的人分析数据的人,选择偏倚信息偏倚混杂偏倚,偏倚问题,Randomization beautifully solved the problem of confounding 随机分组以最美丽最简单的方式解决了混杂的问题Placebo and blinding reduce selection and information bias 安慰剂和盲法协助控制选择偏倚和信息偏倚The scientifically most rigorous method of establishing causal relationship and forms the gold standard for evaluation of clinical effectiveness.随机对照临床试验是在人群中建立因果关系最可靠的方式,因而成了评估医学干预效果的金标准,研究中的伦理学问题,研究对象人研究必须具有科学依据公平选择研究对象获得社区的知情同意(informed consent)对照组的选择和“善后”处理较长试验期限导致“延误”问题,优缺点,优 点(strength),随机分组,能够较好地控制偏倚和混杂 为前瞻性研究,因果论证强度高 有助于了解疾病的自然史 获得一种干预与多种结局的关系,难以保证有好的依从性 难获得一个随机的无偏样本 容易失访 费用常较观察性研究高 容易涉及伦理道德问题,缺 点(weakness),一些经典的实验流行病学研究实例,下课!,