Evaluating QbD's Impact on the Outsourced Relationship

August 1, 2011
Pharmaceutical Technology
Volume 2011 Supplement, Issue 4

An Industry Roundtable Moderated by Patricia Van Arnum and Rich Whitworh. Contract service and technology providers share their perspectives on the influence of quality by design in the expectations between sponsor companies and outsourcing providers.

PharmTech: From your view as a contract development and management organization (CDMO) or contract manufacturing organization (CMO), has quality by design (QbD) influenced the expectations between suppliers and pharmaceutical companies?


McQuaid (Almac): Historical relationships were commonly based on providing task-orientated services with a large degree of knowledge and process understanding still residing within the client organization. Small and mid-cap pharma companies have been quicker to move away from the functional outsourcing model to leverage the growing expertise in leading CDMOs. As the larger pharma companies seek to externalize more development activities, there are signs that they are also interested in leveraging the capabilities of leading CDMOs. The expectations today are that relationships will be far more collaborative in nature with clients demanding not only technical execution of tasks, but also that the CDMO takes fuller responsibility for technical direction and decision-making throughout the overall development process. The CDMO accumulates a lot of product-specific knowledge through the execution of QbD activities and becomes an important partner for the pharmaceutical client. Effective management and communication of knowledge between client and CDMO will have a heavy influence on the success of the relationship.

Jordan (Catalent): From the perspective of a solutions provider, it has been recognized by customers that adherence to the QbD process contributes to a robust approach throughout all phases of the development process. Successful QbD implementation requires a greater degree of collaboration and sharing of knowledge and experience to ensure identification of these critical areas and the implementation of a suitable feedback loop and control strategy to maintain product quality.

Adoption of QbD provides a common language for the development and commercialization of products. Applying the same terminology (i.e., critical quality attributes [CQAs], critical process parameters, design space, and control strategy) when discussing strategies helps to prevent any misunderstandings. Having shared goals and understanding will lead to quicker and more effective decisions, which in turn will speed up the process of developing new products. This could ultimately lead to stronger partnerships built on trust and mutual understanding.

Herring (DSM): In general, the QbD initiative has not had a significant impact on current relationships nor the expectations from our customer base due to the large volume of technology-transfer activity that we conduct. However, with customers that are in the early stages of manufacturing development, we have seen a stronger emphasis in deploying QbD and process analytical technology (PAT) methods. The ability to design and execute this work has become an expectation from these customers. This is evident in the structure of projects with respect to the volume and scope of development activity, as well as the amount of interaction required to execute the project. Building relationships off of technical capability, execution, and continuous improvement is always a good approach for a CMO.

Figure 1: From left to right, John McQuaid, technical development manager with Almac; Andy Jordan, product development manager at Catalent Pharma Solutions; Bruce Herring, manager of operations excellence at DSM Pharmaceuticals; Dwight Lewis, vice-president of quality operations and regulatory affairs at IRIX Pharmaceuticals; and Tom Beil, vice-president of quality and regulatory affairs at SAFC.

Lewis ( IRIX): The QbD initiative has influenced the relationship that IRIX, as a CMO, has with its clients and it has changed the expectations of some clients relative to the service provided for active pharmaceutical ingredient (API) projects in clinical development. There is increased emphasis on expanded data analysis during laboratory-scale route scouting and during initial scale-up runs used to provide API for use in toxicology studies and Phase I clinical support. This expanded data analysis often accompanies project deliverables such as experimental plans using statistical design of experiments (DoE ) and multivariate designs much earlier in the development timeline than previously. Active scientific communication and exchange relative to these efforts require technical involvement from process engineering and manufacturing technical resources much earlier to provide adequate risk assessment to cover the key process areas that are critical for control. This does not always mean that a 'design space' is created to support a QbD submission, but the final process-control strategy based on a criticality assessment for the API is still the goal.

Beil (SAFC): It is not a universally accepted practice with all of our customers, but we work with a technology-transfer process that moves the transfer toward QbD.

Process understanding

PharmTech: Greater process understanding is at the core of QbD. In what ways can a CDMO/CMO facilitate that understanding?

McQuaid (Almac):Due to the nature of the CDMO business, employees will typically have worked on many more compounds than their counterparts who have stayed solely within one large pharmaceutical company. The knowledge and experience of working with different materials and dosage forms can be invaluable to our pharmaceutical clients during design of experiments (i.e., selection of critical investigation factors and definition of CQAs). There are also differing opinions between regulatory agencies and pharmaceutical companies as to what the QbD initiative translates to in practical terms. CDMOs can often offer a balanced opinion on how to implement QbD by having had the experience of working with a wide range of parties in this particular area.

CDMOs within larger integrated contract service providers are well placed to provide pharmaceutical organizations with whatever support they need to complete their required process understanding, whether that be understanding API manufacturing processes, characterization of input materials, formulation design and scale-up experiments, design and execution of process-optimization experiments, or even applying QbD principles to analytical method design and validation. The quality system of the CDMO must be capable of supporting QbD activities with up-to-date templates for conducting risk assessments and capturing the rationale behind design of experiments (DoE) activities. Project management of QbD activities is also an important function to offer as there are a lot of stakeholders from many different disciplines involved. Facilitating knowledge transfer and ensuring that decisions are made within the individual project constraints are critical.

Jordan (Catalent): A shared knowledge-management system effective across the product life cycle enhances product and process understanding. A central store of key product information supplemented by all parties involved in development will allow for more efficient and effective decisions. The manufacturer must provide detailed assessments of key product attributes during process

development and scale-up. This information can then be accessed when the product is in commercial manufacture and an excursion occurs, which is outside the limits within the batch manufacturing records, but may still be within the design space for the product. If this information is easily accessible, then a decision on how to proceed can be taken more efficiently and expeditiously. During the development process, this knowledge can be used when working on similar products, which will simplify and ensure a more targeted approach and could ultimately reduce project timelines. A knowledge-management system will encourage a culture of cross-functional collaboration and bidirectional flow of information as teams engaged in development and commercial manufacture update and share their knowledge.

Herring (DSM): CMOs execute a wide range of processes across numerous drug formulations while using excipients, raw materials, and APIs with various characteristics. The cumulative experience and learnings from these diverse activities can be tapped into by CMOs and used to build the design space. CMOs often perform formulation development, scale-up, and process validation for small or virtual customers with limited expertise and understanding of regulatory expectations around QbD/PAT. CMO scientists, engineers, and other technical experts transfer this expertise to customers during these activities. CMOs also have a wide range of experience with factors such as material transfer, equipment variables, manufacturing environment, and material portfolio. All these factors are inputs to the design space, and the CMO is in the best position to categorize and quantify these inputs to help the build the appropriate design model with the customer.

Lewis (IRIX): A CMO can facilitate a client's understanding through sharing case studies from previous projects and synthetic development experiences. An example would be sharing the development of a control strategy with an API process for limiting genotoxic impurities through developing controls for raw material or intermediate quality attributes rather than using controls in the API specification. Many clients want to know what a CMO's experience with developing physical property controls consists of, particularly for solid dosage drug-product development support. A CMO needs to maintain an easy-to-reference knowledge base of experience in controlling particle size, bulk density, crystallinity, polymorphism, hydration, and solvation of APIs.

Beil (SAFC):One of the problems the pharma and biopharma industries have is the perception that 'no one can do my product as well as I can.' With so many products developed using a similar platform, the customer should be working on its next product's 'intellectual property' while allowing the CMO to standardize and optimize the process. All too often when I worked in pharma and now as a CMO, I have witnessed the transferring company forcing the CMO to change a process to fit the company's unique platform, only to come back to a more optimized process first offered up by the CMO.

The influence of QbD on specific factors

PharmTech: How does QbD influence other factors in an outsourced project, such as the type and frequency of communication, the level and type of testing, and costs?

McQuaid (Almac): Successfully outsourcing QbD activities requires quality communication. By its very nature, large amounts of data may be generated, so expectations around data sharing and presentation need to be clearly defined from the outset. Expectations around the format and detail of development reports should be shared and defined up-front.

During early-phase development, QbD is more about good application of science and well-documented investigations rather than systematically trying to understand and control all the CQAs of a given process. This early-phase development learning then forms the basis of later design decisions for investigations aimed at gaining full process understanding. Therefore, the application of QbD in early development should not have all that much bearing on the level and type of testing and costs.

We would typically execute more complex experimental designs later in development to gain a full understanding of a process and fully define its design and control space. This takes more time, and costs more than traditional drug-product development approaches. However, if this investment is made at the right time, it represents value for money and can offer savings in the long term by preventing issues in the commercial supply chain.

Jordan (Catalent): The greater collaboration and knowledge sharing that is required leads to more efficient and targeted communications (i.e., more frequent earlier in the life cycle, less frequent later). With better understanding and control over key product attributes, cost of manufacture could decrease over the product life cycle compared with more traditional development processes.

Herring (DSM): QbD certainly drives a higher level of communication and interaction between the pharmaceutical company and the CMO. The amount of data exchange and transparency required to execute QbD functions strengthens the project management relationship. Protocol development, execution, and testing is a joint exercise between a customer and CMO. Also, the increased manufacturing activity in the design space required to challenge the process boundaries increases the quantity of batches produced in development. This model appropriately shifts costs to the front end of the product-development life cycle, where they can be mitigated and optimized. Once the process is filed, significant changes are much more costly and time-consuming.

Lewis (IRIX): Communication is a key factor in an API development project applying QbD principles. Weekly teleconferences are typical with written technical updates required at least one day prior to each teleconference. The CMO's project team and the client's technical staff function as a team in these efforts. During DoE work, data are often communicated immediately to the client for review and assessment versus further experimentation plans. The amount and type of experimental processing and analytical testing required often expands beyond the original scope of the initial project. This requires that both the CMO and the client keep track of the efforts, results, and expenditures versus the scope of the business agreement. A flexible project scope design is typically warranted where variances within acceptable percentages are allowed without formal approval.

Beil (SAFC): The companies that best understand the QbD concept also tend to have a technology-transfer process with various stage-gate reviews and communication networks established. A less controlled arrangement up front may start quicker, but they never finish on time or on budget.

Performance metrics

PharmTech: Performance metrics, such as quality, reliability of supply, and timeliness are all important factors in evaluating CDMOs and CMOs. What additional metrics may be factored into a project that is being developed or manufactured in a QbD paradigm or how may QbD alter the evaluation process?

McQuaid (Almac): A CDMO needs to deliver on these typical contractor metrics; they are the basis of any good service provider–client relationship. Also, a measure of the effectiveness of communication and the flexibility and problem-solving capabilities of the contract organization should be factored in. This can often be difficult to quantify as it is not entirely tangible and is often based on an individual's expectation, but it is critical to the success of projects of this nature and also the effectiveness of the collaborative relationship.

Jordan (Catalent): One metric could be 'right first time' new product introduction. An enhanced product and process understanding backed up with development data that has been focused on those areas critical to product quality should result in reduced variability and ultimately lead to a smooth technology transfer as the process is scaled up to the commercial batch size.

Herring (DSM): QbD is all about understanding the manufacturing process with the premise that this understanding will lead to a more capable and predictable process. The automotive industry has taught us volumes about understanding and controlling process capability. The application of statistical process control and capability indices (i.e., Cp, Cpk) are entirely appropriate to pharma processes as well, especially during in-process measurements. Monitoring variation and then subsequently mitigating or controlling that variation to design specifications will ensure that the output of that process is predictable. A CMO's knowledge and capability to execute QbD/PAT initiatives should be an important part of the evaluation process for any customer looking to develop a product that requires new drug application (NDA) submission.

Lewis (IRIX): Clients seeking to use QbD principles look for a CMO that is able to develop statistical designs and conduct statistical analyses of large bodies of data in a timely fashion. Many clients want a CMO that is amenable to a "man in the lab" during execution of designed experiments and a "man in the plant" during actual development and demonstration manufacturing runs.

Beil (SAFC): It is easy to judge the CDMO or CMO, what is more difficult is to judge yourself. Quite often the product is still in development during the transfer process. I have seen the conditions of in-process specifications set but not the analytical methodology. This can lead to misleading results and inaccurate controls.

QbD in today's marketplace

PharmTech: As a CDMO/CMO, have you observed more interest by pharmaceutical companies in QbD as part of a project?

McQuaid (Almac):There has been an increase in requests to consider QbD during development projects. For new clients, once we understand the target product profile and the manufacturing processes that may be used to manufacture the client's product, we can usually give examples of what an appropriate QbD approach may look like in terms of scope of work. Where we have a strategic relationship with a client, this process is simpler because there is already an understanding of what is expected.

Jordan (Catalent): QbD is becoming more prevalent in discussions as industry adoption accelerates.

Herring (DSM): We have observed a strong QbD interest from customers, particularly the technical groups, which are in early-stage development. With projects where we are getting into the value chain much later, there is not a significant interest in pursuing a QbD approach although there may be small pockets of QbD or PAT activity.

Lewis (IRIX): IRIX has experienced increased interest is using QbD principles to assess synthetic route selection, determination of critical process parameters, and critical quality attributes. Prospective risk assessments are now being requested earlier in the clinical development timeline rather just before registration and process validation of the API.

Beil (SAFC): We work with a wide range of companies on many scales of developed products, and we always prefer to partner with the companies that have some level of QbD in their program.

QbD at work

PharmTech: Do you have an example from your own work in which QbD came into play as part of an outsourcing project?

McQuaid (Almac): We have examples where clients have asked us to develop their 'to-be-marketed' formulations. It has been clear from the start of some projects that a QbD approach was desired, so formulation development, analytical development, and process scale-up were all conducted using risk assessment and DoE processes.

Herring (DSM): We are in the final phase (NDA submission) of a QbD project for a new customer in which we provided full formulation, finish dosage, and scale-up activities for a new molecule. The QbD activity was very successful for our customer. We collaborated to identify critical quality characteristics as well as critical process parameters, conducted multivariate analysis, and established in-process specifications to ensure robustness during manufacturing as well as all requirements for the NDA filing.

Lewis (IRIX): IRIX provided development and manufacturing services for a project using QbD principles to develop and assess risk for manufacturing of a Phase III development API. In preparation for a registration campaign, a two batch demonstration campaign was planned. Prior to the demonstration campaign, IRIX conducted a proven acceptable range (PAR) study for optimization of a hydrogenation step using a new catalyst identified through screening by the catalyst manufacturer. The PAR study compared the process performance of the new catalyst in reaction conditions designed to model the manufacturing process and equipment while using the previous catalyst as a control. IRIX also worked with the client to conduct studies using failure mode and effects analysis (FMEA) of each process step and to incorporate recommendations from those studies in the batch records for the demonstration campaign. A further set of FMEA studies will be conducted prior to the registration campaign using the data collected during the demonstration campaign as well as continued research and development laboratory studies into key areas of concern relative to process control and economics.