Find a Sales office 
Have us call you
Email the sales teamMVA for Process Monitoring and Control in a PAT/QbD Environment
Statistical Process Control (SPC) has served various industries well as a quality control tool for many years. The ability to take samples offline, measure them and plot a point on a control chart is a simple and effective means for assessing a few independent process parameters.
But what happens when,
- A process needs to be monitored at a very high frequency?
- The process parameters being measured exceed the few that can be handled by simple charting procedures?
- The process variables are correlated?
Multivariate Statistical Process Monitoring (MSPM) may hold the key in this situation. This approach to process monitoring involves the use of multivariate models to simultaneously capture the information from as few as two process variables, up to thousands.
In a PAT/QbD environment, only CQA’s are being assessed based on CPP’s that have previously been found to define the Design Space. Within the multivariate Design Space, lies the Desired State of the product.
MSPM allows the simultaneous collection and analysis of many CPP’s from many different sources. The Unscrambler® (and The Unscrambler Online) can handle data imports from
- Vendor Proprietary Instrument Inputs
- Object Linking and Embedding for Process Control (OPC) inputs. More details can be found at The OPC Foundation
- Database inputs (from Oracle, SQL etc)
Modelling of such inputs may be achieved through techniques such as
- Batch modelling (commonly used for evolutionary process data that has several stages involved)
- Multiblock modelling (captures the most important information sources for a particular input, or group of inputs and applies an overall process model to the extracted data)
- Hierarchical modelling (uses a classification or first pass predictive model to analyse input data, then defines a more refined model to apply to the data for optimal prediction or classification performance)
- End-point modelling (a time independent alternative to batch modelling).
ICH Q10 approach of PAT
Unscrambler developed models can be readily implemented into CAMO’s 3 rd party control system partners platforms through the use of Unscrambler Predictor ( OLUP ) or Unscrambler Classifier ( OLUC ) products. The diagram represents a particular process setup for the Unscrambler Online Predictor/Classifier products.
The soon to be released Unscrambler Online can also be used to achieve the goals defined in ICH Q10, Pharmaceutical Quality System . The Unscrambler Online represents a flexible and adaptable solution for industry for the purposes of improved process understanding and therefore improved efficiency and reduction of manufacturing cost. The diagram below shows the Unscrambler Online principle for MSPM.
Unscrambler Online principle for MSPM
The Unscrambler® X and Online meet the requirements of electronic signatures and security 21 CFR Part 11 and full validation documentation is available allowing our clients to take advantage of the GAMP5© recommendations for using vendor prepared documentation.
 |
 |
 |
 |
 |
 |