The regulatory authorities worldwide largely driven by the Federal Drugs Administration in the U.S and its European counterpart, are making increasing demands in the interest of patient safety.
Pharmaceutical companies and organizations in the health sector are to apply more complete and complex techniques mainly for process validation but also for the monitoring and the evaluation of the performance of production processes (traditionally called SPC – Statistical Process Control).
There will also be greater demands of the use of statistical procedures for method or measurement validation, notably to check for measurement imprecision and bias. To meet these demands, it is good practice to follow these three 'critical checks'.
Bonus Video - Learn more on 'Critical Checks' in our dedicated webinar series, starting with: Critical Checks for Pharmaceuticals and Healthcare [Part #1]: Validating Your Data Integrity and Process Performance
1st Critical Check:
Data integrity, is your Data reliable?
It is absolutely critical that the data used for all statistical analyses be irreproachable and have complete integrity. This is why it is recommended to conduct statistical techniques for establishing if there is excessive imprecision or bias in the measurement system.
Gage R&R and Gage Linearity Studies can conducted to check for the size of the measurement error but only designed experiments can provide potential solutions for improving it.
These latter studies only demonstrate short term data integrity. It is also necessary, but often forgotten, to conduct ongoing stability studies, sometimes on a daily basis, to check that the measurement process stays on track over the longer term.
2nd Critical Check:
Awareness around the interaction between Measurement error and Process stability
There are fortunately statistical software packages available today that facilitate this task for practitioners. The implementation of the Six Sigma methodology has also made the use of statistical methods to check your process performance and validate the health of your measurement system in general far more accessible.
After validating the data quality preferably on an ongoing basis, the process stability has to be checked using control charts to determine if there are any special causes or shifts in the process.
Following my experience in the health sector, many practitioners do not seem to be aware of the interaction between the techniques employed for data integrity and process stability. All is not what it seems. Sometimes a shift in a control chart if not due to a process change but due to a shift in the measurement system.
That is why ongoing verification of the measurement system can be so critical.
3rd Critical Check:
How to monitor your process performance
In process performance monitoring, many practitioners are faced with a minefield of different permutations, statistics and situations and a plethora of different control charts to track their processes and variable statistical techniques for process performance and capability.
The first issue is to unlock your understanding of the different Capability Statistics (process performance) by grasping the meaning of the two different types of standard deviation.
For continuous process monitoring, the second main issue revolves around whether the data is normally distributed or not or whether only one or more than one sample is collected from a batch or at a given time in the production process. Very frequently no underlying distribution fits the data set, sometimes it is complicated to determine the reasons why.
Watch How To Validate Your Data Integrity and Process Performance
This presentation, a first in a series, helps you to find your way through whatever data is thrown at you. Watch now: Critical Checks for Pharmaceuticals and Healthcare [Part #1]: Validating Your Data Integrity and Process Performance