The Pharmaceutical Inspection Co-operation Scheme (PIC/S) has finalized a risk-based inspection planning tool for inspectorates to use in applying science- and risk-based principles to planning GMP inspections. The tool is contained within a PIC/S document titled, "A Recommended Model for Risk-Based Inspection Planning in the GMP Environment," which was published in Dec. 2011 and became effective on Jan. 1, 2012 (1). This article describes the history behind the new methodology.
PIC/S' work to develop a quality risk management (QRM) tool designed to facilitate risk-based GMP inspection planning began in 2007 through an Expert Circle. The International Conference on Harmonization's (ICH) Q9 QRM guideline, which had been finalized in 2005, provided a firm regulatory basis for developing such a tool as well as other risk-based initiatives.
The Expert Circle first met in Paris in July 2007 to create a QRM work program for inspectorates, which involved developing a QRM training program and related guidance for inspectors that would enable them to inspect QRM-related activities at manufacturers in a harmonized manner. Another task was the development of QRM models for inspectorates that would address the planning and conduct of inspections, as well as their follow-up, and the management of quality defects.
In Jan. 2008, a working group (part of the Expert Circle) comprising inspectors from various drug regulatory agencies assembled. Over the next three years, the group developed an approach to risk-based inspection planning that was relatively novel, simple to use, science-based, and highly flexible in design. It became apparent early on, however, that the remit of this working group was very broad in scope. Hence, it was decided during the third meeting of the Expert Circle, in Malta in Sept. 2008, that the working group would focus only on the GMP and GDP risk-based inspection planning aspects of the mandate. This shortened goal proved a wise decision because it allowed efforts and resources to be dedicated to, what was proving to be, quite a difficult task. The aim was to develop a tool that would allow the frequency and scope of GMP and GDP inspections to be determined using a formalized risk-based approach, a concept directly based on ICH Q9. Several different types of risk scoring models were developed and assessed, but each was found to be problematic, for one reason or another.
For example, a key development at the Maltese meeting was the identification of a set of nine risk-indicating factors that could form the basis for risk assessment of GMP and GDP sites. These factors related to:
Coming up with an effective and practical scoring system for these various factors, however, proved to be a challenge. The group had to always bear in mind that any risk-scoring approach developed for the tool had to be one that could be easily used by a large and diverse group of inspectorates, ranging from those in European, African, and Asian countries, to the inspectorates of certain parts of North and South America, as well as Australia and New Zealand.
By the fourth meeting, in Paris in April 2009, the group saw some breakthroughs. Key concepts underpinning the tool were crystallized into a risk-scoring model and a tool design that were, while not yet optimized, capable of meeting the broad requirements set out for the tool. These requirements included that the tool should represent a simple and science-based QRM methodology that may be used by inspectorates when planning the frequency and scope of inspections. The tool also had to be based on the principles of ICH Q9 and the concept of ranking sites on the basis of an estimated risk that they may pose to patients, consumers, animals, and users of the drug products. The model also had to take into account the risk to product quality, such as the risk of producing non-compliant batches as a result of high levels of process complexity.
The risk-scoring model
The risk-scoring model that arose out of the Paris 2009 meeting centered around the concept that sites would be risk-rated on the basis of three main attributes: complexity, criticality, and compliance. These attributes encompassed almost all of the nine risk factors previously identified as being important:
The complexity and criticality scores for a site would then be combined using a matrix to obtain what is termed an intrinsic risk rating for the site. This rating refers to the inherent risk that is associated with a site, its processes and products, regardless of its compliance status. It reflects the complexity of the site, its processes and products, as well as the criticality of the products or services provided by the site from a supply perspective. Complexity and criticality often remain fairly constant, regardless of the compliance status of the site. Therefore, one usually cannot assess this risk on the basis of inspection deficiencies or compliance history.
Once the intrinsic and compliance risks associated with a site have been estimated, they are combined using another matrix to generate a relative risk rating for the site. It is this risk rating that is considered when deciding the frequency of the next routine inspection at the site.
With regard to the scope of the next routine inspection at the site, the tool requires the inspector who last inspected the site to consider certain items before making his/her recommendation for the scope of the next inspection. These items include, for example:
Another significant change in the tool that arose from the 2009 Expert Circle meeting was the decision to focus the tool on GMP inspections only, because it proved too difficult to design one tool that could effectively be applied to both GMP and GDP inspections at the same time.
Following the 2009 meeting, the tool was refined to reflect the lessons learned to date, and a pilot testing phase on the tool began. Between Oct. 2009 and March 2010, several inspectorates evaluated the tool as part of their national inspection planning programs. The tool was also presented for critical analysis to another PIC/S Expert Circle, relating to APIs, which met in Dublin in May 2010. Useful lessons were gained from each of these meetings, and the penultimate version of the tool was presented for review at the final meeting of the Expert Circle in Warsaw, in Sept. 2010, where the tool was adopted.
A PIC/S-wide consultation on the tool ran from Dec. 2010 to March 2011. Afterward, additional refinements were made to further improve the tool. The final version was reviewed and formally adopted by the PIC/S Committee in December of that year.
The finalized QRM tool is contained within the PIC/S document 037-2 (1). While there are, of course, certain limitations associated with the use of this tool, it is hoped that the methodology provided will enable inspectorates to apply more science- and risk-based principles to the planning of GMP inspections. There are several important design features of the tool that are not discussed in this brief article, but the aforementioned PIC/S document provides a comprehensive overview of the tool, its design, and how it works.
There is also evidence that the tool can easily be customized for application in other areas. For example, it is likely that pharmaceutical companies will be able to customize the tool for application in their own auditing programs, such as in relation to active substance suppliers.
The author would like to acknowledge the contribution made by the PIC/S inspectors who worked on the development of the inspection planning tool between 2007 and 2012.
Kevin O'Donnell, PhD, is Market Compliance Manager of the Irish Medicines Board (IMB).
1. PIC/S website, "A Recommended Model for Risk-Based Inspection Planning in the GMP Environment," http://www.picscheme.org/, accessed Oct. 16, 2012.