研究设计概览
The aim of research should be to
establish the truth and research design
aims to minimise or exclude the threats
to the internal validity of the study.
There are only a handful of ways to do
a study properly but a thousand ways
to do it wrong,-- Sackett (1986)
偏倚 Bias
与真值的系统偏差
Sackett DL,Bias in analytic research,J
Chron Dis 1979; 32,51
常见偏倚
选择性偏倚所选择的研究对象(样本)系统性地有别于所要推断的总体人群。例如志愿者偏倚、电话调查
介入偏倚过多关照试验组导致疗效高估、病人依从性差导致疗效低估
随访偏倚失访病人的特征有别于保持随访病人的特征失访可能与治疗有关
知情偏倚医生如果知道病人接受何种处理将会影响他对病人病情的判断
测量偏倚对所关心结局的判断带有主观性,则判断往往会引人而异混杂 Confounding
分组/处理因素与所关心的结果同时受到另外一个因子(混杂因子)的影响
例如:大医院的病死率高于小医院
选择大小医院和病死率高低均与疾病严重程度有关
疾病严重程度即混杂因子
随机化医学研究中的随机化平衡了比较的两组间显在的和潜在的混杂因素的影响,因此是控制混杂的重要手段
在设计和分析时,可以采用分层或配对的办法对已知混杂因子进行控制
但在分析时采用分层手法有一定的局限性,
这时需依靠基于模型的统计方法予以解决。
偶然性 Chance
由于存在偶然性,单纯随机化分组也有可能造成比较的两组间的不均衡
样本数的不均衡
某些混杂因子的不均衡关于研究设计
统计学对医学研究的最重要贡献在于研究设计而不是数据分析
从统计学角度出发的研究设计主要涉及研究的结构和数据收集过程的方方面面,包括度量的选择及其频数
做适于研究目的的设计大多数研究只是想回答一些简单的问题,但这并非意味它们的设计也是简单的
没有合适的研究设计就不可能有意义的研究结果再多花一点时间在你的设计上吧。对所要研究问题的认识越深入、考虑的方面越多、越细,你的设计越接近于完美。
研究设计的分类
观察性研究( observational)还是实验性研究( experimental)
前瞻性研究( prospective)还是回顾性研究
( retrospective)
纵向研究( longitudinal)还是横断面研究
( cross-sectional)
研究设计的类型
Experimental
Prospective
Longitudinal
Observational
Prospective Retrospective
Longitudinal Cross-sectional Longitudinal Cross-sectional
处理 /干预 随机化设立对照 设盲实验性研究之四大要素实验性研究的结构
研究因子 Factors
研究因子的水平 Levels
均衡设计
组内比较与组间比较
析因设计 Factorial design
Type of
Study
Alternative
Name
Unit of
Study
Randomised Controlled
Trials
Clinical Trials
Community
Intervention
Trials
Patients
Communities/
Groups/
Clusters
Cluster
Randomised
Trials
population
group 1
group 2
Outcome
Outcome
new treatment
control treatment
随机对照试验 Randomized controlled trials
Advantages
Ability to prove causal relationships is
strongest of all study types
Randomisation minimises the risk of
confounding variables
The experimental design permits greater
flexibility,efficiency and powerful statistical
manipulation
Disadvantages
Often based on specific volunteer
samples and findings may not be
generalisable to the wider population
Not feasible/ethical for major
Public Health questions (e.g,
Smoking and Lung Cancer)
随机化 Randomization
防止选择性偏倚、控制混杂
统计分析的基础
常用随机化的方法
单纯随机化 Simple randomization
1,抛硬币
2,利用随机数字表
3,利用随机数字发生器
区组随机化 Block randomization
分层随机化 Stratified randomization
分层变量的选择。年龄、性别?、中心
其他随机化方法整群随机化 Cluster randomization
观察性研究 Observational study
虽然实验性研究是推断研究因子效果的金标准,但是可行性和伦理性上的制约使我们采用观察性研究
病例 -对照研究 Case-control study、定群研究 Cohort
study、横断面研究过去 现在 将来
Cohort
Cases
And
Controls
Cross-Sectional
观察性研究
With observational studies the investigator
does not control the treatment or exposure
but attempts to make valid comparisons
between people with or without disease or
between those naturally exposed or
unexposed to a factor of interest.
定群研究 Cohort studies
Cohort studies entail observing two or more
groups differing in exposure to a potential cause
of disease overtime to compare the incidence of
disease in each group.
It is prospective with the exposure measured
before outcome,enabling the direction of events to
be established that is necessary if we want to say
anything about causation.
Only one risk factor can be assessed for each study
but multiple outcomes can be measured.
Typically these studies require large
numbers,especially if the incidence of the
disease/outcomes is low and a long time
scale,Hence they are likely to be expensive.
研究对象选择是否具有代表性
失访失访的性质(是否与结果变量有关?)
其他问题习惯改变、工作改变偏重于高危人群调查病例 -对照研究 Case-control studies
With case-control study the comparison is made
between people with and without the disease in
order to identify differences in previous exposures
in an attempt to identify the cause of the event.
Retrospective
简单、成本低、适用于发病率极低的病种
Multiple exposures can be assessed,and they are
primary method of studying new or unusual
outcomes.
缺点对照难找(医院对照、健康对照)
病例的确定回忆性偏倚 Recall Bias
匹配、匹配过度、一对多匹配横断面研究
研究对象的选择
应答率健康者偏倚
因果推断因病致贫还是因贫致病?
观察性研究的局限性
The principal problem of observational
studies is that although externally valid
their internal validity may be undermined
by previously unrecognized confounding
factors that may not be evenly distributed
between intervention groups.
适当的统计方法只能控制显在的混杂因子,
但却对潜在的混杂因子无能为力。