Instructions for using OpenEpi Software:

In order to familiarize yourself with the OpenEpi software, I’d like you to input the following 2x2 tables into the software and print out the results. These are the 5 tables used throughout lesson 12. I think that working with the printouts in front of you will help you to better understand the material. Please circle each of the Statistical Results that I present below on your OpenEpi printout. This should help you to recognize where to find the information you need from the software printout depending on the study design and the measure of effect being estimated

***Please note that the orientation of the tables below are the same as in ActivEpi (i.e., exposure on the top, disease status on the left). When you enter the data into OpenEpi, you need to be very careful since the orientation is different (i.e., exposure on the left, disease status on the top).

Instructions for inputting data into OpenEpi:

Go to www.openepi.com

For OR, RR:

Click on twobytwo table (in the counts folder)

Click on Enter new data (Be sure your Pop-up blocker is turned OFF!)

Input data (Be careful about the 2x2 layout!!!!!)

Click on Calculate

Click on Show all results (unless case-control study, you can limit to that)

Print out results and circle each of the statistical results presented below.

For IDR:

Click on compare two rates (in the Person-time folder)

Click on Enter new data (Be sure your Pop-up blocker is turned OFF!)

Input data (The layout for this data will be the same as ActivEpi.)

Click on Calculate

Print out results and circle each of the statistical results presented below.

Note: You can change the confidence level by clicking on settings. Go to the far right and change to 90%, 99% etc. You are supposed to be able to change the table layout to be consistent with ActivEpi here too, however, I was unable to do this.

Page 12.4 Expo #1, 2

Page 12.5 Expo #1,2

Large Sample Cohort Study

Measure of Effect = RR

Results:

Mantel-Haenszel chi-square = 6.996

p-value (1-tailed) = 0.004086

p-value (2-tailed) = 0.008171

Risk Ratio = 2.083

95% CI = (1.185, 3.661)

(Note: If you did not get these results in OpenEpi, check to make sure that you input the data properly. These tables show Exposure status on the top and Disease status on the left. OpenEpi accepts data the opposite way.)

Page 12.4 Expo #4

Page 12.5 Expo #1

Small Sample Clinical Trial

Measure of Effect = RR

Results:

Fisher-Exact p-value (1-tailed) = 0.08625

Fisher-Exact p-value (2-tailed) = 0.1725

Risk Ratio = 3.2

95% CI = (0.8918, 11.48)

Page 12.6 Expo #1, 3

Large Sample Case-control Study

Measure of Effect = OR

Results:

Mantel-Haenszel chi-square = 4.669

p-value (1-tailed) = 0.01536

p-value (2-tailed) = 0.03072

Odds Ratio = 3.157

95% CI = (1.099, 9.073)

Page 12.6 Expo #2

Small Sample Case-control Study

Measure of Effect = OR

Results:

Fisher-Exact p-value (1-tailed) = 0.2199

Fisher-Exact p-value (2-tailed) = 0.4399

Odds Ratio = 3.5

95% CI = (0.473, 25.9)

Page 12.7 Expo #1, 2

Large Sample Cohort Study

Measure of Effect = IDR

Results:

z-score = 3.994

p-value (1-tailed) = 0.00003248 (equivalent to chi-square p-value)

p-value (2-tailed) = 0.00006496 (equivalent to chi-square p-value)

Incidence Density Ratio (Rate Ratio) = 3.464

95% CI = (1.809, 6.634)

You will notice that OpenEpi provides you with more tests than are presented above. I have shown these since these are the test results recommended by David Kleinbaum. There are only slight differences in all of these tests, so most of the time they will all lead you to the same conclusions.

Unless you specify otherwise, you will get information on both the Risk Ratio and the Odds Ratio. The software has no idea what type of study you are inputting. You must determine yourself which measure of effect to use. Look carefully for the correct values.

When you go through lesson 12 in ActivEpi, don’t be too concerned with converting 1-tailed chi-square tests to 2-tailed tests. You can get this from most computer programs. I just want you to realize that a relationship exists between the chi-square and the normal distributions. Also, don’t worry about calculating all these values for the chi-square tests and the confidence intervals by hand. Again, I want you to see the formulas, and understand where the numbers come from, but I am perfectly happy with you learning how to input the data into software such as OpenEpi and learning how to interpret what results are presented. Of course, you need to understand the difference between a one-tailed test and a two-tailed test, when we use an exact test, which large sample test is recommended, and how to interpret a p-value and confidence interval.