Pharmaceutical Production Batch Processes

1
Pharmaceutical Production -- Batch Processes

• Quality Control today made after batch
  completion
• Upsets are found too late
• No correction is possible
• Real Time Quality Control made during batch
  evolution
• Upsets found earlier
• Corrections possible
• Based on summarizing adequate data measured on a
  representative set of good batches
• control limits based on average batch trace
  3?

(measured properties)
2
Real Time Quality Control (Batch monitoring)

• Models the evolution of representative good
  batches
• Take all the variables and their correlation into
  account
• Use control charts for display, with limits
  computed from traces of good batches
• New batches monitored as they are evolving
• Problems detected and interpreted on the fly
• Culprit variables in problem batches clearly
  identified
• Quality of the whole batches predicted as it is
  evolving and at completion
• Based on multidimensional informative data
  measured during the batch evolution
  multivariate analysis

3
Batch Control Charts of a mixing process
(pharm)Two summary traces of good batch (green)
3? (red)
Good batches must evolve inside these limits
4
Monitoring the evolution of new Batches (On line)

• Bad batch (black) as it is evolving, seen outside
  the limits
• Deviations are interpreted with Contribution Plots

Bad batch
5
Control chart of the culprit variable
6
Benefits B-SPC same principles as MSPC

• Multivariate batch modeling
• brings the analysis of 3-way batch data
• to the same framework as continuous process data
• Allows interference as batch is evolving,
• and prediction of final quality
• Applies to
• evolving batches (individual observations),
• and whole batches as units
• Results shown as easily interpretable control
  charts
• Makes batch quality control more reliable, and
  faster
• Facilitates compliance inspection