Validation of Laboratory and Process Analysis Instrumentation

How to choose/make standards:

Lab-based standards are most often synthetic, with final forms tested for extraction of all the analyte Process samples are most often final product, scanned spectroscopically, then analyzed via a referee method. Process stds may be made synthetically, but several precautions need to be taken.

How to sample a process:
For a lab, merely assure a thorough cross-section. For a process, with many or all samples monitored, means need to be taken to match a scan with a physical sample

Proper calibration techniques:
Almost every company has SOPs on calibration of lab methods, so these will be quickly reviewed Process analyses will need statistical sampling combined, most likely, with Chemometric algorithms (all spelled out in USP 1119, FDA Guidances, EMA Guidances) Lab and process have terms in common: specificity, linearity, precision, accuracy, etc. Just the implementation of each differs.

Proper validation techniques:
Again, lab calibration techniques are well-spelled out with Guidances and have been honed over the decades. Using a standard (USP, ASTM, etc.) the sample is run side-by-side with known standards and accuracy is compared. For process analyses, validation is done on the entire sample (since real-time cannot allow chemical reactions and sample prep) The process sample will be judged by an equation that has been developed prior on other production lots and calibrated via referee methods.

When to re-calibrate/re-validate:
We will discuss how the equations, used to monitor the process, need to be updated to mirror changes in raw material supplies, process equipment changes, instrumentation repairs or updates.

Why should you Attend: Choosing a proper technique, alone, is time and money consuming. Improper calibration and validation of the equipment could mean failing good batches, while passing out-of-specification lots. This is not only expensive but could cause serious injuries to consumers.

Know both how and when to calibrate/re-calibrate/validate and, if necessary, when and how to re-calibrate.

Laboratory analyses have the luxury of being re-run, should there be a question of accuracy. Process analyses, being real-time, do not allow the luxury of retesting any one sample. The fate of the batch depends both on individual accuracy and the fact that many more samples are analyzed, allowing for any one out-liar to be mitigated by samples prior to and after the questionable unit, avoiding process changes.

Areas Covered in the Session:

• Types of analytical instrumentation utilized
  • Lab instruments: HPLC, UV, Infrared, others
  • In-process instruments: Near Infrared, Raman, TeraHertz
• Lab instrument calibration vs. in-process calibration
  • Lab: Standard solutions or powders resembling the analysis
  • Process: NIST or ASTM wavelength standards, NOT actual product samples
• Standards used for method calibration/generation
  • Lab: solutions or powders made for traceable standard chemicals (HPLC, UV/Vis, MIR)
  • Process: samples from either production run or synthetic units made out-of-range with known amounts of API used to generate an equation used in a process instrument
• Standards/samples used for validation
  • Lab: same as for original calibration
  • Process: materials from production line (or synthetic units) analyzed by compendial method
• When and how to re-calibrate and/or re-validate both lab and process units

• Lab Managers (QC and AR&D)
• Production Managers
• QA Personnel
• Anyone Involved with Process Analysis