Download Homework 6

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Data assimilation wikipedia , lookup

Transcript
Homework #6 – PH6415
Do the analyses for these problems using SAS (except problem 1 – do this by hand).
Copy your program and SAS output into Microsoft Word and answer any remaining
questions by hand.
Also remember, a SAS data set consists of rows and columns. Each row represents an
individual unit (such as a person, plant, country, etc.) on which data was collected. Each
column represents a variable.
See examples from SAS tutorials.
*****************************************************
HOMEWORK ASSIGNMENT: SUPPLEMENTAL AND PAGANO’S BOOK
*******************************************************
Problem 1:
1. Recall this data set from last week’s homework. You will now analyze it using Cox
proportional hazard regression. The SAS dataset called breast.txt contains survival data for
45 patients diagnosed with breast cancer and entered into a study. The following variables
are on the dataset:
Variable
Surv
Status
Meta
Trt
Description
Follow-up time in months after enrollment in study (surgery)
Indicator variable (Status =1 if patient died, Status = 0 if patient censored)
Meta=1 if tumor had a positive marker for metastasis, otherwise Meta=0;
Trt=1 if new therapy, Trt=0 if standard of care
You can read the data set into SAS with the following code:
data breast;
infile 'C:\breast.txt' dsd dlm=' ' firstobs=2;
input surv status meta trt;
run;
( a)
Using Cox-regression (PROC PHREG) estimate the hazard ratio for those receiving
new treatment compared to standard treatment adjusted for the marker of metastasis.
Give a 95% CI for the HR estimate.
(b)
Using Cox-regression (PROC PHREG) estimate the hazard ratio for those with a
positive marker of metastasis compared to those without a positive marker for
metastasis adjusted for treatment effect. Give a 95% CI for the HR estimate.
( c)
Use an interaction in the Cox-Regression model to test if the treatment effect is
different for those who have a positive marker for metastasis compared to those who
do not. What do you conclude?
Problem 2:
Chapter 21, page 512, number 9


These data are attached to the assignment in
bladder.txt. You will need to enter the data into
your SAS program using the infile statement.
Data coded as follows (try to recode on your own, use
problem 2 above as reference):
o Placebo: group=1
o Drug: group=2
o Time until first tumor in months: time
o Recurrence: censor=1
o No Recurrence or censored: censor=0
o Number: Number of tumors removed (1 or 2 or
more)
A. Using Cox-regression (PROC PHREG) estimate the hazard ratio for those receiving the
new drug compared to placebo adjusted for the number of tumors. Give a 95% CI for the
HR estimate.
B. Using Cox-regression (PROC PHREG) estimate the hazard ratio for those who had one
tumor removed compared to two or more (placebo group only). Give a 95% CI for the
HR estimate.
C. Use an interaction in the Cox-Regression model to test if the treatment effect is different
for those who have a one tumor compared to those who have one or more tumors. What
do you conclude?