Bioequivalence and Bioavailability Forum

Hi Ken,

❝ How do we prove that R has audit trail ?

You can’t – there is none (like in SAS, Matlab, S-Plus, Phoenix/WinNonlin, ). Have a look at this docu­ment.* The only software having an audit-trail out of the box are all LIMS and some CDS.

❝ For example, if we use Bear or PowerTOST in R, how do we provide to inspector audit trail ?

Tell the inspector that there is none. Have a set of SOPs which you follow. Provide a transparent work­flow documenting all steps performed.

---

• The R Foundation for Statistical Computing
  R: Regulatory Compliance and Validation Issues
  A Guidance Document for the Use of R in Regulated Clinical Trial Environments
  Vienna, Austria (August 8, 2013)
  online

Hi Ken, Helmut, all,

I think quite a few CROs I have talked to have doubts about R and audit trails. Understandably. Per se, there is just no audit trail built into R, but I am sure we could make one.

I think I could write a windows application that takes the following as input:
-Name + company of user
-R script
-path to R
-time, date
etc.

As output the following will be generated for a session:
-name of user
-R script
-R output
-name of person granting rights to user
-path to R
-unique session number
-time, date
etc.

All output -which will be the audit trail and result simultaneously- will be associated with checksums.

The windows application will then be able to authenicate any such input posthoc. For example, an inspector or auditor will immediately be able to test if the session output has been manipulated.
If the user tampers with the output, then application will capture it and output an error.

Any interest in this?