Russ Somma on QbD - "How Did You Launch Before, With a Catapult?"

The Competing Globally track also featured a presentation on Quality by Design, made by Russ Somma, President, SommaTech LLC, who worked for several years at Novartis and is also an editorial advisor on  Mr. Somma joked that the industry never had quality by accident, "We don't just lumber along and wait to fall apart, but the question is: Did you ask the right questions at the right time?" There's a need to better define process needs and required capacity early on.

"It's not a race to the end of development….we need a clear roadmap for development leading to tech transfer," he noted. Later on he described the need for PAT, but not as an add-on, without sufficient process understanding, or as an excuse to justify doing what you want (more on that later(.

He described the need to set drug product specs as well as the importance of:
a basic knowledge of:

  • excipient interactions and process understanding
  • raw material characteristics and variability
  • a target product profile-a “desired state” for product
  • stability of clinical forms, prototypes as well as drug substances.

CMC Project Teams are Critical
Mr. Somma said he was amazed that the CMC Project Team is a new concept for some drug companies ("What did you launch projects with before, a catapult?" he asked.)

He also noted the fact that product changes along the way can spell disaster, for example, if the product picks up moisture.

He suggested the following as some of the key questions to ask and attempt to answer as a starting point:

What are the properties that affect product performance?
What is formulation intended to do given drug substance properties?
What are the special requirements of the drug substance anddrug product?

He mentioned the importance of having a historical database for the product (one may start with broader ranges during early stages, he said, then tighten them.   These require a systematic reporting method referenced during clinical batches,pilot scale, scalup and validation.  This information then becomes  a part of knowledge store for product, and basis for SPC, facility design and maintenance.

It is important to determine the proven acceptable ranges, he noted:

Chart all process steps and controllable parameters
Provide a brief description of the process step and controlled parameter
Include Engineering units that are recorded
Anticipate result for exceeding proven acceptable range
Evaluate the risk of exceeding range

From this one can stablish the operating range to be used for process control, he said. It thus becomes like a staged-gate review process.

Next, one must develop process setpoint and establish equipment tolerances.  "You need narrow limits of values…know what potential failures could be and what you’d have to do about them, he said, noting that process analytical technologies provides a context for defining process-critical control parameters.

Don't Be Linear
"Don’t take linear approach and expect it to work in multidimensional world," he cautioned. "PAT is not a fancy term for getting away with what we want."  Critical sources of variability should be known, variability managed, prior quality attributes predicted, rationale for change

Many firms apply PAT as add on technology, when the underlying process understanding isn’t there, Somma said recalling a company that called him.  They were having trouble with their filling line.  Their IR sensors weren't picking up measurements.  It turned out that the tablet coating was releasing dust, coating the sensors and preventing them from functioning. Somma suggested that the process needed improvement and that it might be overheating. "“But we’re adding PAT to the coating line” was the response he got.  

Moving from Tacit to Explicit Knowledge
Mr. Somma sketched the different types of knowledge, noting that tacit knowledge needs to be replaced with explicit knowledge if scaleup is to proceed correctly and if improvements are to be realized across different functions within a company.  And that knowledge needs to be set forth in a database.

He cited tablet development as an example, since it is time consuming and iterative and scalup is less well understood than other processes.  He summarized new techniques in compression simulation and hybrid continuous process systems singling out MIT's  3DP Process used by the Pennsylvania-based company, Aprecia, and the Glatt Multicell GMC 30, semicontinuous process….

Quick Does Not Equal Cheap
If you want to be quick, Somma noted, you're not necessarily going to be cheap.  "YOu have to throw intelligence rather than money at the problem...and consider the product in its entire life cycle to formulate a proactive rather than reactive submission strategy," and to avoid fishing expeditions for data, and approval delays.

In short, he said, "Know what you do not know." U.S. pharma often fumbles over its own requirements and regulations.  "There are companies [that are already] doing things this way and if you don’t consider changing the way you do things, they’re going to eat your lunch…"  You'll find more about him on

Show Comments
Hide Comments

Join the discussion

We welcome your thoughtful comments.
All comments will display your user name.

Want to participate in the discussion?

Register for free

Log in for complete access.


  • <p> We are all riding a bandwagon of TLAs (three letter acromyms). However it is still interesting even with doing the right things, products in Pharma world are produced checking quality every step of the way. That itself tells us that there is gap between what is done in the lab and what is done on the manufactuirng floor.  </p> <p> Lack of QbD on the manufacturing floor is not a reflection of individual performance. It is due regulations dictating the industry. They have forced producers in a "quality by analysis" corner and we are stuck. Team work is needed to get out of this corner.    </p>


RSS feed for comments on this page | RSS feed for all comments