y***y
发帖数: 563 | 1 请教版上大牛们一个问题。
如果我想设计一个实验，比较两组的survival rate,不给药的那一组（A)假设预计
survival rate是0.4，给药的那一组（B)希望达到survival rate 0.6,而且A和B是
paired sample,即每一个在B组吃药的病人，在A组都是有个相同性别，相近体重，相近
血压的不吃药的病人。那我需要如何计算sample size，以达到power=0.8呢？求高人指
点，非常感谢!
补充：
假设每组的病人不是一一对应的，那么假设我有一个希
望达到的hazard ratio (HR),那么如果有alpha=0.05， power=0.8, 和预期的log(HR),
通过公式可以计算出来需要的sample size.
我想知道的是，如果每组病人之间有个一一对应的关系，那么该如何修改计算sample
size的公式，我看到一个说法是原sample size乘以（1-rho),但是这样sample size比
之前传统方法的更小，我觉得好像不合理，因为相对于传统方法，这种paired design
降低了randomization,因此需要的sample size应该更大才对。 | a****g
发帖数: 8131 | 2 this is not a survival analysis
but a paired proportional comparison | P****D
发帖数: 11146 | 3 You didn't specify your hypothesis clearly so I am not sure these apply.
Just somewhere to start with.
------
Power Analysis of Non-Inferiority Tests of Correlated Proportions
Page/Date/Time 1 3/19/2013 12:43:35 PM
Numeric Results for a Non-Inferiority (One-Sided) Test of a Difference
Sample Equiv. Actual Treatment Standard Nuisance
Size Difference Difference Proportion Proportion
Parameter
Power (N) (De) (Da) (Pt) (Ps) (P01+P10) Alpha Beta
13 0.10000 0.20000 0.60000 0.40000 0.10000 0.05000
References
Lewis, J.A. 1999. 'Statistical principles for clinical trials (ICH E9) an
introductory note on an
international guideline.' Statistics in Medicine, 18, pages 1903-1942.
Lui, J., Hsueh, H., Hsieh, E., and Chen, J.J. 2002. 'Tests for equivalence
or non-inferiority for paired
binary data', Statistics in Medicine, Volume 21, pages 231-245.
Nam, Jun-mo. 1997. 'Establishing equivalence of two treatments and sample
size requirements in matched-pairs
design', Biometrics, Volume 53, pages 1422-1430.
Nam, Jun-mo and Blackwelder, W.C. 2002. 'Analysis of the ratio of marginal
probabilities in a matched-pair
setting', Statistics in Medicine, Volume 21, pages 689-699.
Report Definitions
Power is the probability of rejecting a false null hypothesis.
N is the number of subjects, the sample size.
De is the maximum difference between the two proportions that is still
called 'equivalent.'
Da is the actual difference between Pt and Ps. That is, Da = Pt-Ps.
Pt is the response proportion to the treatment (experimental or new) test.
Ps is the response proportion to the standard (reference or old) test.
The Nuisance Parameter is a value that is needed, but is not a direct part
of the hypothesis.
Alpha is the probability of rejecting a true null hypothesis.
Beta is the probability of accepting a false null hypothesis.
Summary Statements
A sample size of 13 subjects achieves NA% power at a 5% significance level
using a one-sided
equivalence test of correlated proportions when the standard proportion is 0
.40000, the maximum
allowable difference between these proportions that still results in
equivalence (the range of
equivalence) is 0.10000, and the actual difference of the proportions is 0.
20000.
【在 y***y 的大作中提到】
: 请教版上大牛们一个问题。
: 如果我想设计一个实验，比较两组的survival rate,不给药的那一组（A)假设预计
: survival rate是0.4，给药的那一组（B)希望达到survival rate 0.6,而且A和B是
: paired sample,即每一个在B组吃药的病人，在A组都是有个相同性别，相近体重，相近
: 血压的不吃药的病人。那我需要如何计算sample size，以达到power=0.8呢？求高人指
: 点，非常感谢!
: 补充：
: 假设每组的病人不是一一对应的，那么假设我有一个希
: 望达到的hazard ratio (HR),那么如果有alpha=0.05， power=0.8, 和预期的log(HR),
: 通过公式可以计算出来需要的sample size.
| y*****w
发帖数: 1350 | 4 This is not a paired sample. By "paired", you have the SAME patients that
had pre- and post-treatment measures. | w*******9
发帖数: 1433 | 5 It's paired if he can argue that the factors he matched by could almost
explain off all the variability of the response. Otherwise, it's not paired
and there is a great potential of mistakenly exaggerating the treatment
effect.
【在 y*****w 的大作中提到】
: This is not a paired sample. By "paired", you have the SAME patients that
: had pre- and post-treatment measures.
| Y******Y
发帖数: 8753 | 6 这些是matched sample. 的确需要用paired data的analysis
【在 y*****w 的大作中提到】
: This is not a paired sample. By "paired", you have the SAME patients that
: had pre- and post-treatment measures.
| y***y
发帖数: 563 | 7 谢谢楼上各位的回复。我知道,假设每组的病人不是一一对应的，那么假设我有一个希
望达到的hazard ratio (HR),那么如果有alpha=0.05， power=0.8, 和预期的log(HR),
通过公式可以计算出来需要的sample size.
我想知道的是，如果每组病人之间有个一一对应的关系，那么该如何修改计算sample
size的公式，我看到一个说法是原sample size乘以（1-rho),但是这样sample size比
之前传统方法的更小，我觉得好像不合理，因为相对于传统方法，这种paired design
降低了randomization,因此需要的sample size应该更大才对。
请大家指正。 | k*******a
发帖数: 772 | 8 我觉得这个可以看作个stratified cox PH model
可以把每个pair 看作个statum
这样的话，我觉得sample size应该和以前一样，因为每个pair提供的information和普
通的两个人提供的informaiton一样 | w*******9
发帖数: 1433 | 9 paired data variability 小所以需要的size小，这是为什么要pair. 比如x=rnorm(
100), y=x+0.01, 如果是paired, pvalue=0. 如果是unpaired, pvalue会很大。
你说的paired survival sample size 很有意思，你看见的sample size
modification 能给个链接吗
？我想学习学习。
),
design
【在 y***y 的大作中提到】
: 谢谢楼上各位的回复。我知道,假设每组的病人不是一一对应的，那么假设我有一个希
: 望达到的hazard ratio (HR),那么如果有alpha=0.05， power=0.8, 和预期的log(HR),
: 通过公式可以计算出来需要的sample size.
: 我想知道的是，如果每组病人之间有个一一对应的关系，那么该如何修改计算sample
: size的公式，我看到一个说法是原sample size乘以（1-rho),但是这样sample size比
: 之前传统方法的更小，我觉得好像不合理，因为相对于传统方法，这种paired design
: 降低了randomization,因此需要的sample size应该更大才对。
: 请大家指正。
| w*******9
发帖数: 1433 | 10 放狗搜了下，发现这个paired survival data基本上都是假定parametric copula，比
如bivariate exponential之类的。网上是有些方法算sample size的。 | y*****w
发帖数: 1350 | 11 By "matched sample", you are right, paired data analysis should be applied.
However, this is still not a strictly matched sample by any means. LZ would
have to prove that the samples were matched through some mechanism (age and
gender matching, as a simple example), not arbitrarily.
LZ can check out the following article by Peter Austin. It's about
propensity score matching, with some contents on survival analysis (K-M
estimates and Cox models) in the context of matched pairs. Check out Page
1131.
The Journal of Thoracic and Cardiovascular Surgery, Volume 134, Number 5
【在 Y******Y 的大作中提到】
: 这些是matched sample. 的确需要用paired data的analysis
|
|
