20
An Illustration of Rate Difference Estimation with SAS in Logistic Regression Yun Guo

An Illustration of Rate Difference Estimation with SAS in Logistic Regression

  • Upload
    kanoa

  • View
    55

  • Download
    2

Embed Size (px)

DESCRIPTION

An Illustration of Rate Difference Estimation with SAS in Logistic Regression. Yun Guo. Outline. Statistical methods for RD RD with logistic regression SAS realization Conclusion. Statistical methods. Wald/t approximation—no covariate adjust CMH (weighted estimate)-adjust strata - PowerPoint PPT Presentation

Citation preview

Page 1: An Illustration of Rate Difference Estimation with SAS in Logistic Regression

An Illustration of Rate Difference Estimation with SAS

in Logistic Regression

Yun Guo

Page 2: An Illustration of Rate Difference Estimation with SAS in Logistic Regression

Outline

• Statistical methods for RD

• RD with logistic regression

• SAS realization

• Conclusion

Page 3: An Illustration of Rate Difference Estimation with SAS in Logistic Regression

Statistical methods

• Wald/t approximation—no covariate adjust• CMH (weighted estimate)-adjust strata• Logistic Regression-adjust both

Page 4: An Illustration of Rate Difference Estimation with SAS in Logistic Regression

Statistical methods

• Logistic regression

,where, x is covariate, s is factor, t is treatment, sub i is indexing subject

(1)

exp( )

1 exp( )i i i

ii i i

x s tp

x s t

(2)

Page 5: An Illustration of Rate Difference Estimation with SAS in Logistic Regression

RD with logistic regression

• Odd Ratio through Logistic Regression

Simply subtract among different levels of factor

Say t(t=1) vs. t(t=0),

1 0log( ) log( ) log( ) exp( )1 1t t

p pOR

p p

Page 6: An Illustration of Rate Difference Estimation with SAS in Logistic Regression

RD with logistic regression

• Rate through Logistic Regression (say subject i=1)

1

0

1 0

exp( )

1 exp( )

exp( )

1 exp( )

i it

i i

i it

i i

t t

x sp

x s

x sp

x s

RD p p

(3)

Intercept, covariate as well as factor, not gone!!!!

Page 7: An Illustration of Rate Difference Estimation with SAS in Logistic Regression

RD with logistic regression

• How about the variance of estimate?

1

0

1 0

exp( )

1 exp( )

exp( )

1 exp( )

var( ) var( )

i it

i i

i it

i i

t t

x sp

x s

x sp

x s

RD p p

(4)

Very messy!!!

Page 8: An Illustration of Rate Difference Estimation with SAS in Logistic Regression

RD with logistic regression

Hereafter subscript 1/0 of est p index treatment.

Page 9: An Illustration of Rate Difference Estimation with SAS in Logistic Regression

Statistical methods

• Marginal approach (sum over all samples)

Page 10: An Illustration of Rate Difference Estimation with SAS in Logistic Regression

Statistical methods

Page 11: An Illustration of Rate Difference Estimation with SAS in Logistic Regression

Statistical methods

• To conclude, we have following 4 steps• 1, establish logistic regression • 2, get estimated rates, covariance of estimates,

estimated RD• 3, set up Delta method• 4, get CI

Page 12: An Illustration of Rate Difference Estimation with SAS in Logistic Regression

SAS realizationEXAMPLE DATA

Page 13: An Illustration of Rate Difference Estimation with SAS in Logistic Regression

SAS realizationPREPARE DATA

Page 14: An Illustration of Rate Difference Estimation with SAS in Logistic Regression

SAS realizationSTEP 1: Logistic regression

Page 15: An Illustration of Rate Difference Estimation with SAS in Logistic Regression

SAS realizationSTEP 1: Logistic regression

Page 16: An Illustration of Rate Difference Estimation with SAS in Logistic Regression

SAS realizationSTEP 1: Logistic regression

COV

EST

Page 17: An Illustration of Rate Difference Estimation with SAS in Logistic Regression

SAS realizationSTEP 2: RD estimate

Page 18: An Illustration of Rate Difference Estimation with SAS in Logistic Regression

SAS realization

Step 3: Delta method

Step 4: CI

Page 19: An Illustration of Rate Difference Estimation with SAS in Logistic Regression

Conclusion

1, When covariates adjust needed, a choice is to use L Regression

2, Other methods could be used too

Page 20: An Illustration of Rate Difference Estimation with SAS in Logistic Regression

Thank you!