The current economic crisis will undoubtedly spur pharmaceutical firms to curb spending. One strategy that may help achieve this goal is to incorporate process analytical technology (PAT) into manufacturing processes. PAT may reduce costs by helping companies control process variability, improve yields, reduce waste, and produce high-quality therapies consistently. Companies that have not yet embraced PAT may find its potential to reduce expenses a compelling argument in its favor during this time of financial difficulty.

Analytical tools such as near-infrared (NIR) spectroscopy enable PAT to enhance process understanding and control, but they require analytical expertise. “A high-level analytical scientist should be involved in designing, developing, and implementing a PAT system,” say Duncan Low, scientific executive director, and Cenk Undey, principal engineer, both at Amgen (Thousand Oaks, CA). And, laboratory professionals will need training to understand process instruments, process-sampling systems, and data-acquisition systems used in the manufacturing process, says Rod Woods, director of PAT at MannKind (Valencia, CA).

In addition, PAT “frequently uses statistical tools that chemists and engineers generally don’t use,” Woods cautions, including modeling techniques such as principal components analysis and multiple linear regression. Teaching people the tools is easy, he says, but normally requires additional learning from experience with the statistical method and with the application. The model for the application must be refined through experience and accumulated data.

For each analytical instrument installed on a production line, a company must have a scientist who understands the data it collects. Online testing “creates an overwhelming amount of data, and you can very quickly get lost in it,” adds Jason Kamm, principal at Tunnell Consulting. “You need skilled people who can read the spectra and understand what’s important and what’s not.”

The challenge of using PAT may not be as daunting as it seems, given that most pharmaceutical companies have the necessary knowledge to implement it. Many firms already use NIR and other analytical devices in quality-assurance and research applications. The industry has a pool of experienced chemists and engineers who can interpret analytical data, says Woods. “PAT implementation will go ahead despite the fact that it requires a skilled scientist to implement,” says Low.

Using PAT to make decisions and adjust a manufacturing process doesn’t require as much analytical expertise as does implementing PAT initially. “Once the PAT application is fully developed in a commercial environment, process operators frequently make their go–no-go decisions on simplified graphic or alphanumeric output,” Woods explains. “The raw data from the application require only periodic review from the development group or people with suitable skills.” If standard maintenance procedures are in place, analytical experts need only be available for troubleshooting, say Low and Undey.  

On the other hand, a company needs extensive knowledge of the manufacturing process and PAT tools before it can use PAT applications to automatically control a process. Although this technique can provide fine control of difficult steps, it is a challenge that few manufacturers choose to confront, says Woods. Yet, applications such as chemical synthesis with unstable materials, keeping a continuous process in a state of control, and working with valuable actives might justify the cost and effort required for automatic control through PAT.

In the past, Big Pharma’s budgets have generally been large enough to compensate for waste and inefficiency in manufacturing processes. This may no longer be a viable strategy, given the faltering economy.

“I think PAT tools will become more necessary,” says Kamm. “If manufacturers are going to keep their costs down, enhance process understanding, and deal with regulatory constrictions in the new world, they’ll have to [adopt PAT tools].”