Software Validation Breakout Session Summary

1
Software Validation Breakout Session Summary

• AAPS Workshop
• Scientific Approach to AIV
• March 4, 2003

2
Outline of Discussion

• Current approaches to analytical instrument
  software validation/qualification
• Validate software separately vs. qualify
  instrument as a system (holistic)
• Future improvements or optimizations
• Conclusions
• Misc.

3
General Observations

• Very open discussions
• Lots of opinions (over 15 pages)
• Good representation from most all segments, e.g.
  users, managers, QA, consultants, regulatory, but
  only limited vendor representation
• No clear consensus on major issues
• Some individuals clearly troubled by the negative
  tone directed at software validation

4
Current Practices

• As expected, the approaches utilized to test and
  release analytical instrument software covered
  the entire range of possibilities
• Separate and extensive testing (validation?) of
  the software
• Qualification of the software with the hardware
  and no separate software validation
• Contractor performed validation
• Some write their own scripts, some use vendor
  supplied scripts, some do both

5
Software Validation or Whole System Qualification

• Positive comments supporting a general validation
  practice for instrument software
• Validation, especially script writing is a good
  training exercise/experience
• A good validation can decrease the likelihood
  of producing bad data
• Separate the data acquisition process from the
  data reduction process

6
Software Validation or Whole System Qualification
II

• Positive Comments supporting a holistic system
  qualification approach
• May simplify, and therefore speedup, the process
  for qualification and release into production
• Covers the high risk elements of ensuring the
  instrument is appropriate for intended use
• May lower hurdle to upgrade software versions and
  thus increase innovation

7
Future Improvements or Optimizations

• More transparency of the vendor validation or
  qualification processes and results
• An elimination of, or a decreased emphasis on,
  software validation
• Software verified in PQ only
• Industry wide test scripts, or validation by 3rd
  party and available to industry

8
Future Improvements or Optimizations II

• Only run tests that address significant risks
• Regulate analytical instrument vendors as we do
  medical device vendors
• Establish user group/web page/or news group to
  exchange issues and/or views with respect to
  analytical instrument software
• Adoption of a holistic approach to instrument
  qualification

9
Summary

• Most want to make the process simpler
• Move to a holistic approach would be supported
  only if it was indeed simpler
• More vendor support for, or increased
  transparency of, the qualification process
• Clear difference of opinion on the need for a
  software validation process apart from instrument
  qualification

10
Misc.

• Set instrument categories for what needs more, or
  less, qualification or validation
• Are there clear examples of where the validation
  efforts have discovered high risk errors in the
  software
• What about systems from vendors that are not
  really designed for pharmaceutical work