Test Data For PAMCOMP
(applies to version 1.41 or higher)
1. Introduction
These test data should let you test the correct installation of the PAMCOMP software and serves as an example how to compute person-years and standardized mortality ratios. In case of problems please contact taegerd@uni-muenster.de.or visit the PAMCOMP homepage at http://medweb.uni-muenster.de/institute/epi/pamcomp/pamcomp.html.
2. Creating a new PAMCOMP file
After installation of the software start PAMCOMP. Now you must first create a new PAMCOMP file which will hold all necessary tables and information to compute person-years and SMRs. Goto FILE|NEW and create a new PAMCOMP file, f.i. pamtest.pam.
3. Importing the cohort data
The next task is to import the cohort data. You will find attached with this manual test cohort data as an ASCII file named cohort.txt. You can open it with notepad and will see following
There are 6 variables separated by a semicolon
Variable | Comment |
id | identification variable; ranges from 1 to 50 |
dob | date of birth; format: ddmmyyyy |
doh | date of hire; format: ddmmyyyy |
eps | entry point of study; format: ddmmyyyy |
tps | termination point of study; format: ddmmyyyy |
icd | indicates of person is dead (=1) or not (=0) |
To import these data go to TABLES|MAIN TABLE|ADD|ASCII
A dialog window will appear where you can select the cohort.txt file. Afterwards following ASCII import window will appear.
Select COLUMNS HEADERS IN FIRST ROW and the data type LONG for the variables ID and ICD . For the other variables choose the DATE data type and as date format DDMMYYYY (to select variables click on the column headers). For the variable DOH it should look like
Now click on IMPORT and the cohort data will be included into the pamtest.pam file.
4. Calculating person years
To calculate the person years go to CALCULATE|PERSON YEARS. You must fill in the same as in the figure below.
Now click on COMPUTE and the person years will be calculated with following result
There will be 8 age classes [10-14], [15-19], [20-24], [25-29], [30-34], [35-39], [40-44], [45-49] and 6 calendar classes [1975-1979],[1980-1984],[1985-1989],[1990-1994],[1995-1999],[2000-2004].
4. Importing reference data
The next task is to import the reference data. Because your person-years matrix is of dimension 8 x 6 (age classes x calendar classes) your matrix of rates must be of the same type. If for instance in your reference population there were 456 people in the age class [10-14] in calendar class [1975-1979] and also 5 deaths, your rate for this matrix cell will be 5/456=0.010965.
You will find attached with this manual two test reference data sets as ASCII files. The first is named rates_eng.txt for PC operating systems with the dot as decimal operator and the second is named rates_deu.txt for PC operating systems with the comma as decimal operator. Choose that what will represent the decimal separator of your operating system,
You can open f.i. rates_eng.txt with notepad and will see following
There are 8 rows representing the age classes and 6 columns representing the calendar classes. The value for the cell in the upper left is 5/456=0,010965.
To import these data go to TABLES|RATE TABLE|ADD|ASCII and import each column as DOUBLE format data. Afterwards you will see the rate file as
5. Calculating Standardized Mortality Ratios
After person-years computation and importing of reference data you can compute SMRs. To do go to COMPUTE|SMR. Following dialog will appear
Choose as
Rate file: rates_deu (or rates_eng)
Person Years table: age*calendar
ICD variable: icd
ICD value: 1
Method: Byar
Significance level: 0.05
so that is looks like
Press COMPUTE and the SMR as well as the death distribution will be calculated.
The death distribution in this case looks like
and the SMR
If these are the correct figures the PAMCOMP software is installed properly. If you get different numbers for the SMR and the confidence bounds, check if your Null Value for the SMR is 100. If not you can change it via the menu TOOLS|OPTIONS.
(c) 2001-2002 Dirk Taeger
v1.01 of this file
All samples are hypothetical data