Afnan: Scientific Sampling

Oct. 4, 2011
Process knowledge demands a very different approach to sampling.

For years now, we’ve heard that the relationship between FDA and industry is adversarial. If that’s true, the hostility hasn’t served regulators, the industry, or the consumer. 

Recently, Murray Lumpkin, FDA’s deputy commissioner, was quoted as saying that the relationship between industry and FDA cannot be one of partnership, a statement that would seem to support the status quo.

I disagree. Instead, he seems to be advocating a relationship based on mutual respect, focusing on objective oversight and emphasizing the audit function.

Objectivity is critical, and, if the questions I keep hearing from industry are any indication, it is still far off. “How will I know if my process won’t work as intended, before product is distributed?” is a popular one asked by operators, QC professionals, managers, and inspectors. 

Or, “How do you know that a process is capable considering the wide range of product specifications and limited sample size?” Pharmaceutical manufacturers tend to answer that one the same way: “Look at hundreds of batches over a period of time, and you’ll see capability when none fail release testing.”

FDA’s process analytical technology (PAT) guidance attempted to remove subjectivity from the relationship between regulator and regulated. It emphasized the need for process understanding, and the fact that a process “can only be considered well understood when product quality attributes can be accurately and reliably predicted.”

Looking for examples from other industries, FDA saw consensus standards as the way to achieve that objectivity and nudge manufacturers toward true process understanding.

A small group at FDA became active in the standards development organization, ASTM. Technical committee E-55 was established specifically to focus on standards relevant to the pharmaceutical industry. The committee also established liaisons with other technical committees of interest and relevance such as E-11 (quality and statistics).

Just a few weeks ago, a highlight of ASTM’s E-55 committee’s meeting in Europe was a presentation on E-2709 by FDA’s Alex Viehmann. The standard, which focuses on pharmaceutical sampling, was driven by FDA’s 2011 Process Validation guidance and, even though I was not involved in its writing, I couldn’t help but see it as the fruit of several years of hard work.

The way to process understanding has always been simple: replace subjectivity with objectivity founded on science and best, not common, practice.

In pharma today, common practice relating to process performance tends to be a mix of tradition and compendia and, with time, becomes de facto best practice. E-55 reminds us that common isn’t necessarily best.

E-2709—Standard Practice for Demonstrating Capability to Comply with a Lot Acceptance Procedure—was developed within ASTM E-11 and finalized by the pharmaceutical statistician community. (Development was led by James Bergum of BMS, Thomas Murphy of T.D. Murphy Statistical Consulting, and FDA’s Viehmann. The standard is based on Dr. Bergum’s research and work.) It, in turn, has driven another standard, now in the works, on dose uniformity.

E-2709’s scope is to provide “a general methodology for evaluating single- or multiple-stage lot acceptance procedures which involve a quality characteristic measured on a numerical scale.” The standard is designed to allow manufacturers to show compliance with in-process, validation, or lot-release specifications.

It defines, for example, within n% confidence whether the future sample taken from the same batch has greater than n% chance of passing the standard test.

It thus eliminates subjectivity, making science, rather than tradition, a basis for action.

I would urge you to study ASTM E-2709, or, better yet, to join ASTM to explore this and other standards. Keep your eyes out for the next standard on dose uniformity, and whatever you do, don’t mistake accepted tradition for best practice.

About the Author

Ali Afnan | PhD