Hi DRNOC,
While reviewing code related issues, we were closely looking into the
“_Potential_Code_Errors.pdf” and observed that “Bad Record Percent” column is
not representing the actual percentage of records and I wanted to share my
concerns with you because I believe there are potential bugs or
mis-representations of information.
Here are my assumptions and comments:
1. “Total Records” represents the number of total records with a particular
code type (such as 09, 10, CH)
2. “Total Codes” represents the number of Total Unique Codes with Potential
Code Errors
3. “Bad Record Percent” = “Total Codes” / “Total Records” X 100
* Computing the percentage this way does not represent the actual
percentages of records that are associated with the bad codes. I believe this
percentage is not meaningful.
* I would expect the percentage should be computed as:
i. “Bad
Record Percent” = “Total Records associated with Bad Codes” / “Total Records” X
100
* I think this also has implications in the EDC report as well
Any thoughts?
Regards,
Abu Saleh Mohammad Mosa, PhD, MS, FAMIA
Assistant Research Professor of Health Informatics
Core Faculty of Institute for Data Science and Informatics (PhD Program)
Adjunct Assistant Professor of Electrical Engineering and Computer Science
Center for Biomedical Informatics, MA213D, DC018.00
University of Missouri School of Medicine
One Hospital Drive, Columbia, MO 65212
Phone: 573-882-8349
Email: [email protected]<mailto:[email protected]>
Web: medicine.missouri.edu | hmi.missouri.edu
_______________________________________________
Gpc-dev mailing list
[email protected]
http://listserv.kumc.edu/mailman/listinfo/gpc-dev
