ASSIGNMENT #4

In this assignment, will use a sample of the National Trauma Data Bank,

the dataset is in the course directory with the name NTDB.

Up to this point you have just copied and run the statements. In this

assignment you will be writing some simple code of your own.

Part A: Exploring your data (1 point)

(Review assignment 1 if you don't remember how to write the SAS code to do .)

1. Print the contents of the dataset.

Part B: Data Management (2 points)

We are now going to create dummy variable to indicate whether or not people have died. People who have died will have a disposition as ('Died') but it would be more helpful if it was coded as a 1 and everyone else a 0.

We'd also like to determine if patients are hypotensive (sbp < 90) by creating a dummy variable; hypotensive patients coded as a 1 and everyone else a 0.

LIBNAME COH611 "/courses/dd4f9595ba27fe300" ACCESS=READONLY;

The following statements show how to fix the data for these two variables.

Data example;

Set coh611.ntdb;

If disposition in ('Died') then dead = 1;

Else dead = 0;

If sbp1 = . then hypotensive= .;

Else if sbp1 <90 then hypotensive = 1;

Else hypotensive = 0;

Run;

Add statements to do the following:

Create variable named youth that is set to 1 if age (AGE) is less than 9 and 0 otherwise . (HINT: Look in the output of your contents procedure to determine which this is.

Part C: Data Analysis (3 points)

Test for a relationship between hypotension and death using a Chi-square test

What is the odds ratio? What is the p-value?

Test for a relationship between sbp1 and pulse1 using a correlation

What is the correlation coefficient? What is the slope model?

Test for the relationship between youth (the variable you created) and death using a Chi-square test

What is the odds ratio? What is the p-value?

Part D: Conclusion (4 points)

Discuss the results from the three tests conducted in Part C. Did you find a relationship or not? What statistics support the relationship (or lack thereof)?