Implementing a Control Strategy for Combination Product Design and Manufacturing

Tell us about the control strategy you've implemented with AstraZeneca and its external partners.

Over the last year, we worked to improve our control strategy for drug delivery devices. Our goal was to standardize the process to ensure the consistent performance and safety for device functions. We did this by formalizing how we identify critical aspects of the device design and manufacturing process and then implementing a risk-based approach to identify where controls are needed. This would ensure that those performance and safety functions achieved a suitable reliability target. As an example, for an autoinjector, we could identify that consistent injection time is a feature critical to the functional performance of the device. Next, we would drill down and identify over twenty parameters of the combination product that impact injection time (e.g. drug viscosity, autoinjector spring force, syringe needle diameter, etc). Then, we would analyze the impact of each of these attributes on injection time through say tolerance stacks, modeling or empirical testing and highlight the attributes with an outsized impact. Lastly after further refinement with a risk assessment we would determine where and how much control is necessary.

More generally, we standardized the process of identifying top level performance and safety functions and then drill down to understand the features that have a big impact on those functions. This gives us the framework with which to target the ideal quality level for each performance and safety-associated feature with ideally the lowest impact on the process.

What have you learned in implementing this strategy?

Across our industry, we’ve collectively seen a variety of strategies that companies implement to identify and handle performance & safety functions. We realized that establishing a single way of working was important. We could have implemented our framework as a standard operating procedure but that can be overly prescriptive. Instead, we handled this through guidance and training, so each function understood the “what, how and by whom” for implementing right-sized controls.

"There is no one clear industry aligned term for key performance and safety functions that are expected to have associated process controls, so each company approaches this in their own way."

How have these regulations, guidances and standards changed?

There’s been significantly more discussion in this area as part of guidances and standards in recent years. I suppose it started with IEC 60601 3rd edition. This was really the first instance of the phrase “essential performance”. It amounted to a design element that needs to be present to keep safety to an acceptable level. 

The principle around “Essential performance” lined up with a term the FDA used in a quality system regulation (QSR) from 1997 about “essential design outputs” (EDOs). Interestingly, there was only an expectation to identify EDOs – not to necessarily control them.

Recently, the FDA issued a guidance on “essential delivery design outputs” (EDDOs) which are defined as design elements that directly impact drug delivery. Unlike EDOs, safety functions are not within scope. There is yet another FDA term, “essential performance requirements” (EPRs). While EPRs have no guided definition, once the FDA identifies these for a type of device, they are typically expected to be controlled in manufacturing.

There is no one clear industry aligned term for key performance and safety functions that are expected to have associated process controls, so each company approaches this in their own way.

With all this going on, how can companies better identify design and process elements to control? 

Right-sizing your approach to assure performance & safety reliability is important so that you’re not implementing controls needlessly and consequently increasing the cost of your product. Implementing the right controls at the right places in manufacturing can be a tricky balance, but a thorough understanding of the design will facilitate this.

Organizational guidance to get everybody on the same page will bring consistency and build confidence. This way, the design teams know what downstream organizations will need identified and manufacturing teams will know what drives the control points. A focused list, rather than a laundry list, helps all teams involved understand what to focus on – something that becomes even more important when it comes to larger manufacturing scales.

I would also urge caution with defaulting to previously identified EPRs without a deeper understanding of the design. This leads to a lack of understanding for why those design elements are controlled the way they are and complicate future management of the product, for example with lifecycle design changes.

"We as an industry need to recognize that siloing information will encumber development and obfuscate the actual information necessary to right-size manufacturing controls."

How can companies facilitate collaboration between internal and external groups for large scale manufacturing? 

The design group with oversight over integration of the full system is best positioned to identify the key performance and safety functions and perform the in-depth analysis of all associated parameters. I would caution against the attitude that understanding design is simply not feasible if the design is not owned. At a minimum, the combination product system team needs to know the sensitive design elements behind the key safety & performance functions. This tends to include some level of understanding on how the device or combination product works to achieve those functions and, what controls already exist to ensure that they do so reliably. It can be a fine line to tread, but a level of trust and collaboration between design partners is necessary for a successful product. 

If it’s such a universal truth that collaboration and effective communication are the keys to getting things done effectively and efficiently, why do companies have such a hard time?

There is an inherent concern with compromising trade secrets or intellectual property which can make it challenging to navigate sharing an appropriate level of information for a productive design partnership. It’s important to set expectations up-front, and we as an industry need to recognize that siloing information will encumber development and obfuscate the actual information necessary to right-size manufacturing controls.

What about for internal partnerships, like between device and formulation teams?

Accelerated product development now seems like the status quo – this means that formulation and device development may need to work independently of each other during an early stage of the project. It is important however for the teams to communicate their assumptions and at least consider common risk areas, say for example a minimum level of free silicone in the primary container or identifying a specific range of viscosities or fill volumes.

"It’s important to recognize that what may be considered common sense does not necessarily translate to a common set of objectives across both design and manufacturing."

Are you seeing a trend in manufacturing companies integrating design functions and how does that impact collaborations?

Yes, in the combination product space we do see a trend with design and manufacturing integration, especially with organizations that offer their in-house designs for use with their clients’ molecules.

This tends to mean increased confidence that the device was designed with manufacturing in mind, particularly if the organization has a strong history in manufacturing that class of device. However, this will likely not account for the full set of combination product user requirements – a classic example is that your user population is only capable of a lower force than what was originally designed. This may also not account for controls from a systems perspective – for example the allowable variation in autoinjector spring force coupled with your higher-than-expected variation in your syringe extrusion force leads to a lower than tolerable reliability for your injection time. It’s important to ensure that the overall system is vetted. 

Anything else?

I would add that it’s important to recognize that what may be considered common sense does not necessarily translate to a common set of objectives across both design and manufacturing.

For example, on the design side, you tend to be more concerned about finding a design that works for patients and ensuring speed to market. So, taking a prefilled syringe as an example, you set a conservative extrusion force on a prefilled syringe. On the manufacturing side you’re more interested in minimizing costs and maximizing yields and might instead prefer a higher upper specification limit on that same force. Having an established framework will facilitate a collaborative discussion where teams can probe all sensitive parameters associated with injection force and perhaps with some compromises, they can be confident that they optimized their final choice of control limits. 

Speakers