[EpiData-list] Append data from laboratory

epidata-list at lists.umanitoba.ca epidata-list at lists.umanitoba.ca
Mon Jul 28 12:53:13 CDT 2008


Dear Ziad,

My sugestion is first make two different rec files... one wiht the data you
are entering and a second one from one or more laboratory results. Usually
is easier to keep different files from different sources. Lets say... one
from clinical evaluations, one from CD4+ T cells and Viral Loads, one from
the hamatology laboratories, one froam immunology laboratories and etc,
because thy come from different sources.

The most important thing here is to keep in all files with a unique
identfier, or a combination of fields that uniquely connect each lab result
and its owner (the patient).

Usually this connection is done later... right before the data cleaning,
before the analysis. But if you see a necessity to bind tha data for a beter
controls of what is going on with your patinets (and your data) then...

Make a qes file with all the fields whith the data is the xls spreadsheet.
If the spreadsheet has wierd column names... first backup the file, later
change the column names to the name which you want in the rec file. Make
shure that the two recfiles have the very same variable names (the one with
CD4 and VL and the one wiht the CD4 and VL and all clinical data).

Example:

xls "unique number" name CD4+ "Viral laod" "Collection date"
         1                    Ziad   560      <80            last sunday
........

to

 xls    v1  v2      v5    v6      v7
         1  Ziad   560  <80    last sunday
........

CD4VL.qes

*****************
My lab results
*****************
v1 Unique aidentifier ######
v2 Name <A     >
v5 CD4 cell count ####
v6 Viral load #########
v7 Collection date <dd/mm/yyyy>

Also important that these variables must have the vary same name in the
destination recfile with the clinical and lab results data...

 AIDS.qes

*****************
My patients data
*****************
v1 Unique aidentifier ######
v2 Name <A     >
 v3 Date of HIV diagnosis <dd/mm/yyyy>
v4 Date of TB diagnosis <dd/mm/yyyy>
v5 CD4 cell count ####
v6 Viral load #########
v7 Collection date <dd/mm/yyyy>

.....

Now that you got this, you must save your spread as a TXT separeted by TAB
or by commas (I like TAB better :))

Now, from in EpiData import this spread sheet using the CD4LV.qes as
structure. Chek you data, if it is OK then go on...

Now you r going to combine you dataset side by side using the unique
identfier. This option in Epidata is very ituitive. IMPORTANT IS TO BACKUP
ALL THE FILES BEFORE YOU START. Sometimes the variable format are tricky
from xls and you might have to make it several times before you got it
right!

A tip is allways use Stata to import data into EpiData, because in this
situation Epidata keeps all the file structure and all the labels. From
stata you dont have to make a qes file before. But this doen not seem to be
you case because your data is allready on xls.

Another easier way to do it is to use a software that transport and tranform
almost every datase format such as StatTransfer. In this situation from .xls
to .rec in a plink of you eyes. And then conbine in EpiData entry or your
analysis software!

May the force be with you!
-- 
Abraço forte e que a força esteja com você,
Pedro Emmanuel


More information about the EpiData-list mailing list