BEST PRACTICES FOR CONTINUOUS IMPROVEMENT IN CMO PROCESS UNDERSTANDING
Stage 3 of FDA's new process validation guidance reiterates that the process of developing and monitoring CPPs and KPIs is
continuous and dynamic. The document states, "In addition to being a fundamental tenet of following the scientific method,
information transparency and accessibility are essential so that organizational units responsible and accountable for the
process can make informed, science-based decisions that ultimately support the release of a product to commerce"(3). With
sponsor–CMO relationships, this approach also requires ongoing analysis and refining, which can be performed by monitoring
To carry out a QbD approach, Stage 3 of the FDA's validation guidance must be applied across the entire manufacturing network,
including at CMO sites. Incorporating automation tools is important for achieving continuous and dynamic monitoring with CMOs
because it allows complex processes to be monitored with minimal human error and resources. The following summarizes best
practices and considerations for implementing a continuous improvement plan for process understanding based on the author's
Establish a baseline for how often the organizations should reevaluate parameters and processes. The sponsor in the case study described in this article reevaluated plans and parameters every six months. Process capability
and quality monitoring tools should be used to help develop and revise baseline metrics. This approach is consistent with
the notation that monitoring should be continuous and dynamic at CMO sites, as well as at owned manufacturing sites.
Use information from baselining exercises to increase process understanding at a CMO site. Develop, monitor and revise statistical plans and sampling plans for continuous quality improvement. Important elements include:
Who are the key people to review monitoring charts and involve in deviation investigation? What are the parameters? How often
should we evaluate these parameters?
Engage in proactive process monitoring that makes analytics easier to use, but with a system that ensures you are not simply
validating what you want to hear. Check assumptions and insist on doing it "the right way" on an on-going basis in spite of human nature. Data-driven monitoring
and decision making is an ongoing part of QbD. It is just not about simply clicking and seeing results that validate assumptions.
Conduct high-quality statistical analysis and use data-based decision making to interpret what has occurred in a process,
and then revise appropriately. A trained statistician, for example, can determine if a credible sample size was used and make a solid recommendation for
action. Outsource the analysis function if the skill set—or required time—is not available within your organization.
The sponsor and CMO must communicate, collaborate and be willing to adapt to new working approaches based on sound statistical
analysis. Data exchange can be a vehicle to increase communication between sponsors and CMOs, which is valuable for both parties.
Build flexibility into contracts. Sponsors and CMOs may need to renegotiate contracts to specify how, when and which data will be shared, or to accommodate
technological improvements that have become available since the contact was signed.
Kate DeRoche Lusczakoski, Ph.D., is a senior analytics specialist at Aegis Analytical Corp., 1380 Forest Park Circle, Suite 200, Lafayette, CO 80026,
. ditor's Note: Aegis Analytical produces monitoring by exception software.
1. E.S. Langer, BioPharm International 25 (2), 15–16 (2012).
2. V. McCurdy et al., Pharm. Eng. 30 (4), 12–32 (2010).
3. FDA, Guidance for Industry, Process Validation: General Principles and Practices (Rockville, MD, Jan. 2011).