A

If you include any output from WNL in a regulatory submission, then WNL likely should be validated.  Your company Quality Assurance policies/SOPs will determine if WNL should be validated.

A

Test data has been obtained from different sources, e.g. NIST, PK textbooks.  An explanation of the sources is explained in a document that comes with the Validation Suite (Computational Models Validation Report). The reference results have been generatedon a workstation at Pharsight and compared to independently-generated results, e.g., in SAS. The results that you generate in your environment will be compared to the reference results generated in Pharsight's environment.

A

The Validation Suite automatically compares the list of files that should be installed with the files that are actually installed on the machine to be qualified to identify any differences.

A

The Computational Models Validation Report included with the Validation Suite summarizes the results fo comparisons of output produced by WinNonlin to output produced in independent tests, e.g., using SAS.  Also you may perform your own additional independent verification of WinNonlin's results by comparint the output in the Validation Suite reference files with your own output.  Note: only SAS Windows versions 8.0 and 8.0.2 are supported. However, any SAS version should work given that it supports:
1) SAS transport file format (supported by SAS since version 5)
2) the ability to be run from the command prompt on a Windows PC

A

Yes, all of the output generated by WinNonlin on the machine to be qualified is placed in output files in a folder architecture that mirrors the reference files.  Therefore one could easily manually compare the output results against the reference results. Also, there is a log file that contains all of the results, so one could also use that to see the output results and to ensure that what is exported into the Validation Run Report has not been "edited" in any way.

A

For technical reasons, two tests cannot be automated and instructions are included for manual tests for Table Template 9 and the Multi-transform. Most WNL users do not utilize these features; therefore, they may not require validation for your intended use of WinNonlin.

Testing of the calculations is not a part of the test cases that are executed in the Validation Suite, as this is done at Pharsight as part of the internal validation of the computational engines within WinNonlin. The Validation Suite product includes a Computational Model Validation Report document which provides information regarding the testing of the calculations and statistics. For the testing that is summarized in this document, WinNonlin results were compared against results obtained by an independent reference (e.g., a SAS PROC).

A

A requirements document is not included in the Validation Suite. It is the user's responsibility to develop a requirements document based on the functionality that is intended to be used by the customer. With the descriptions of the test cases included in the Validation Suite, the user can develop traceability between intended functionality to be used and testing of that functionality by the Validation Suite. The user can then determine if any additional testing is required to satisfy the documented requirements.

A

The automated Validation Run Report summarizes the results of the test sets that are executed by the Validation Suite based on the selections of the user. The Validation Summary Report document template included with the Validation Suite is a document that "closes the loop" on the validation activities defined in the Validation Plan. If the user added additional testing (i.e., testing other than those executed by the Validation Suite) in the Test Plan, then the results of those additional tests would need to be included in the Validation Summary Report.