A Collection of Epidemiologic Routines
Version 1.0, 8/June/1990
Kevin M. Sullivan
Division of Nutrition
Centers for Disease Control
1600 Clifton Rd. NE, MS A08
Atlanta, GA 30333
Over the years I have developed a number of programs that
perform specific epidemiologic tasks and have combined
several of these together in Epi_Pak. The routines included
MOWAN - calculates exact confidence interval for the odds
ratio from a single 2x2 table.
PAR - depicts the relationship between the Population
Attributable Risk and prevalence of exposure for a given
PROPCI - uses several confidence interval procedures for
RATECI - calculates approximate confidence interval for a
SAMPSIZE - performs samples size and power calculations.
SCREEN - calculates several screening indices from a 2x2
I have been meaning to incorporate these into the program
dEpid but have not had the time. More detailed descriptions
of these programs are provided below.
See the MOWAN.DOC file for more details on this routine.
Please note that the FORTRAN code is not included with
This program requests the entry of a risk ratio value and
then provides a graphic depicting the relationship between
the Population Attributable Risk (PAR - the proportion of
cases of a disease that are "attributable" to an
exposure/risk factor) and the prevalence of exposure. For
example, studies of the association between aspirin and Reye
syndrome have estimated that Reye syndrome is 20 times more
likely to develop in children who use aspirin. In 1980
approximately 50% of children received aspirin for the
treatment of colds and fever (prevalence of exposure = 50%).
Using PAR, you can see that for a risk ratio of 20, if the
prevalence of exposure is 50%, then about 90% of Reye
syndrome cases that occur in the population would be
attributed to the use of aspirin. If the prevalence of
aspirin use was reduced to 10%, then about 65% of the cases
of Reye syndrome that occur would be attributable to the use
One option provided is whether you would like graphics.
If you answer (Y)es, a graph is provided using the higher
ASCII characters. If you answer (N)o, a character-based
graph is presented to the screen.
Please see the PROPCI.DOC file for information on this
This program provides an approximate confidence interval
for a rate (i.e., person-time information). The confidence
interval formula is:
a | a
- + Z * | ---
N \| N*N
a=number of cases
Z=Z value, e.g., for 95% confidence interval Z = 1.96
For example, say there were 15 cases during 19,017
person-years of follow-up. The point estimate would be 7.9
cases per 1000 person years with upper and lower 95%
confidence limits of 3.9 and 11.9 per 1000 person-years,
Please see the SAMPSIZE.DOC file for information on this
Please see the SCREEN.DOC file for information on this
Should you have any suggestions or comments, please
NON-WARRANTY. EPI_PAK is provided "as is" and without any
warranty expressed or implied. The user assumes all risks of
the use of EPI_PAK. EPI_PAK may not run on your particular
hardware/software configuration. We bear no responsibility
for any mishap or economic loss resulting therefrom the use
of this software.
COPYRIGHT CONDITIONS. You may make and distribute copies of
EPI_PAK provided that there is no material gain involved.
USE AT YOUR OWN RISK. All risk of loss of any kind due to
the use of EPI_PAK is with you, the user. You are
responsible for all mishaps, even if the program proves to
be defective. This program makes certain assumptions about
the data. These assumptions affect the validity of
conclusions made based on the output from this program.
Please acknowledge EPI_PAK in any manuscript that uses