Therapeutic Dose: Track Laying and Lessons from Amtrak

Sept. 1, 2009
With PAT and QbD, we built the high-speed trains before we laid down the rails—can we build the infrastructure?

Who among us hasn’t dreamed of returning to college or the start of their careers, knowing what they know now?

What wonderful things I could have accomplished in 1970 with my 2009 knowledge.
But wait a minute. In 1970, there weren’t any GMPs for me to refer to; there were no electronic balances, no mass spectrometers (unless you built one yourself), no Fourier Transform Infrared (FTIR) spectrometers, other than the slow grating types, no high pressure liquid chromatography (HPLC), no good gas chromatography (GC) devices, and, alas and alack, no Near Infrared (NIR) spectrometers.

Well, okay, there were NIRs, but they were primitive when compared with what we use today. Besides, Principal Components and Partial Least Squares weren’t commercially available until 1987.

It’s tempting to use the same hindsight with pharma and scientific manufacturing principles. Some people may look at the state of pharma today and ask why we couldn’t have started PAT and QbD initiatives 20 years ago.

No Wine Before Its Time

I think they’re asking the wrong question.  We are beginning to do PAT and QbD at precisely the right time in the industry’s history. Is it fair to chastise drug manufacturers for failing to do 21st century science in the middle of the 20th century? Especially since in-process pharmaceutical tests were carved in stone in the 1950s? Due to the conservative nature of the industry and its watchdog, “Ralph,” er, FDA, things stayed that way for a long time.  Now, pharma is ripe for change.[pullquote]

In 2002, the FDA decided to put science into pharmaceutical production and published Guidances espousing PAT, QbD, and Design Space, and things have gotten so much better.

Okay, maybe they haven’t gotten better. Why not? For nearly the same reason I couldn’t do 2009 analyses in 1967: lack of infrastructure.

With PAT and QbD, we are certainly not missing the technical infrastructure. Today, we have hardware and software that allow us to do wonderful things with production processes. We’re missing an equally important piece of infrastructure: people who appreciate and can harness the power of PAT and QbD. One look at the “Frequently Asked Questions” that industry has posed to ICH regarding QbD (accessible via shows the industry’s mindset.

For a parallel outside of pharma, consider that much maligned American institution: Amtrak.  Anyone who has ever taken a high speed train in Europe or Japan returns to the U.S. asking why we can’t have them here. The answer is clear: apart from a few miles of high speed track in the Northeastern U.S., we simply don’t have the infrastructure, although we can easily build the trains.

For pharma, the infrastructure for high-speed or PAT/QbD is the active engagement of support groups like QA and QC, IT, and finance management. Just because we are physically ready to generate millions of data points every day, doesn’t mean that anyone in our IT or quality departments will know what to do with them. Just because we want to do in-process release doesn’t mean our QA and Regulatory departments can handle the truth. And, just because we could potentially save billions of dollars with PAT or QbD doesn’t mean that management or the bean counters will hand out the funding required.
In short, we put the cart before the horse—or rather the train before the tracks—and didn’t get buy-in from the regulatory and other people in control of the industry before we went ahead and started building instruments taking measurements.

We will need to get everyone up to speed so that they can match the technical level that FDA and PAT/QbD specialists have achieved in the past couple years.

Maybe we can find some money in the government Stimulus Package for stakeholder outreach and management education? It needs to be on a fast track.

About the Author

Emil Ciurczak | Contributing Editor