In our first part of the series on clinical trial supplies in the September issue of Contract Pharma, we wrote that the preparation and management of clinical trial supplies is a complex service that integrates disparate functions and processes across the operational spectrum. Production of clinical trial supplies is at the leading edge of the outsourcing movement. From established mid-size pharmaceutical companies to virtual companies and university incubators, all look to specialty service providers for knowledge, experience, and controls required to execute these complex operations. The reasons for this outsourcing trend make sense.

Companies may only need these clinical supply services intermittently, or even only once. Full-time, in-house clinical supply capabilities are needed only for the deepest developmental pipelines, but every company needs to use these services eventually.

Production of clinical trial supplies is set-up intensively, not in terms of physical infrastructure, but in terms of intellectual, experiential, and control infrastructure. The competitive landscape and cash-flow pressures inherent in drug-development mean that the long lead times associated with developing new operations is a luxury most mid- to small-sized companies cannot afford.
Adding a new clinical trial supply group is like adding any strategic asset to a company’s portfolio. Development of any new service dilutes the focus and resources away from a company’s primary skill set.

The production of clinical trial supplies also requires the integration of established quality practices with additional quality requirements unique to the clinical trial supply environment. It’s important to remember that when you outsource clinical trial supplies, you’re also outsourcing the control of certain quality functions and outcomes specific to the clinical supply environment. ICH Q10 outlines expectations and working strategies for companies seeking to outsource various operational elements.

Quality in the Clinical Supplies World
Interpretation and application of GMP requirements is straightforward in established organizations with steady-state commercial operations; the risks in that environment are well understood, the solutions are formulaic, and interpretations of the requirements vary little from one company to the next. Not so in clinical trial supplies. The process output for a clinical supply run is not a discreet product, e.g. a million count tablet run. Instead, a clinical trial supply job is the custom solution to a unique clinical study protocol that might involve multiple active products of various strengths, comparators, and placebos across multiple study sites in multiple countries.

Quality in this context is no different from quality in any other production environment, although it is considerably more complex. Quality is the provision of goods and services that meet customer expectations. It is important to remember that the differentiating factor in clinical supplies is complexity, and it is a difference of degree, not dimension. Similarly, concerns about compliance with GMP vary by degree, not by dimension. The underlying principles of GMP are consistent:  GMP is the intersection of two processes—documented identification of risk and documented control of risk. Furthermore, it is important for quality professionals transitioning into clinical supply operations to remember that risk is phase-dependent, and control of those risks requires phase-appropriate measures. Like any pharmaceutical operation, quality requires one to identify the risks, and then apply controls commensurate with the risks. We have always said the FDA will look for two things. First, are you in control? And secondly, can you prove it? This is especially important in the manufacture of clinical trial supplies, because clinical trial supplies have additional requirements involved.

Quality Concerns Related to Clinical Supplies
Study design and the resulting clinical protocols drive all activity in the clinical supplies world. The common currency in bulk manufacturing is the lot number, and all activity is traceable through a root lot number. In this order scheme, a single item is being produced.

The study protocol represents the “lot number” for a clinical supply job. An order ticket for a bulk drug product might read “one million units of XYZ, 10mg tablets.” By contrast, a typical clinical study protocol might involve 300 patients at 10 study sites across three countries. The patients might receive a series of wallet kits containing a randomized combination of placebo, comparator, active one, and active two. All kits at all sites would be blinded, and traceable only by protocol number.

Fulfillment of a complex study protocol requires that packaging records create multiple items instead of the single item created by a bulk drug product record. Furthermore, the fulfillment of complex study protocols presumes that multiple drug products and labels can be present in a production suite at the same time. Appropriate control of these risks requires a thorough knowledge of the study design and intent, including study details such as: Randomization parameters; enrollment targets; visit/dosing schedule; number and location of sites, including number of countries; dosing strategy (variable or placebo); and treatment venue (home, or clinical).

Owing to the inherent complexity of clinical study protocols, there is a strong possibility that each piece will have its own unique identity versus the commercial environment where the manufacturing strategy is complete uniformity of each piece. This variation is highly (100%) predictable at the manufacturing site, so as to be completely randomized at the treatment site. This seems alien and non-GMP to the QA professional steeped in traditional bulk manufacturing, and it seems impossible to control in the conventional model of spatial/temporal segregation. Successful control of multi-product kitting requires strict spatial, but not necessarily temporal, control over manual operations, and strict engineering and validation controls for automation. Control of the uniqueness and randomness of a clinical supply job requires knowledge and understanding of the study protocol.

Label Control
Labels define and describe product identity and pedigree. This basic principle applies to both clinical supplies and commercial products, although the label attributes can be quite different. Just as the quality and risk management of kitting and assembly operations depends on understanding of the study protocol, quality and risk management of labels and labeling operations depends on a thorough understanding of label content. Unlike commercial labels, clinical trial labels may not provide direct pedigree to the component lots that make up multi-dose kits. Pedigree information for clinical supplies is controlled by the protocol number or study number, which should always be part of the labeling information. Pedigree will include the relationship and linkage between the protocol number, patient number, kit sequence number, and the constituent products that make up the kit.

Clinical trial labels will often contain data fields that vary with each label. The content of variable field data can take the form of patient ID numbers, QR codes, and other data formats that create unique attributes for each label in a study. Variable data in clinical labels may also include specific sequence information critical to the blinding and randomization attributes in the study protocol. In these cases both count and sequential continuity of labels is critical—being off by one label in a count doesn’t mean one wrong; it could mean all labels are wrong. It is not sufficient to simply verify that fixed data fields are present and correct.

Other quality issues specific to clinical trial supplies include just-in-time labeling at remote sites and control of relabeling and over-labeling of comparator products. The variety of possible labeling configurations is beyond the scope of this article, but the overall control strategy is the same: label content and counts must accurately and clearly support the intentions of the study protocol, and must be traceable in some fashion to the original pedigree of materials used. Control of the clinical labels requires knowledge and understanding of the specific label content and the meaning of both fixed and variable data fields in the context of the study design.

Operational Execution
Production and production controls for clinical trial supplies differs substantially from bulk drug product manufacturing in terms of setup times, run volumes, cycle time, and striking an effective balance between automated tasks and human tasks. Bulk pharmaceutical manufacturing and packaging represent large scale, repetitive operations. The goals of bulk pharmaceutical manufacturing are optimize output and to minimize variability. Production controls for bulk pharmaceutical products rely heavily on process development and process validation, coupled with statistical in-process testing and increasingly Process Analytical Testing (PAT). Set-up times are short compared to run times.

Production of clinical trial supplies is distinguished by small lots, rapid cycle times, and little if any repetition. Production runs are unique to each study. The time devoted to pre-production activities for clinical supplies will always be proportionally higher than actual production work. Pre-production work may occupy the vast percentage of your project time. This pre-production work includes study protocol review and approval, specification development, creation and approval of label proofs, record development, and distribution protocols. Actual production may be 10% or less of the total project time. Substantial project resources must be dedicated to these gateway activities. Short lead times are intrinsic to the clinical supply world. Forecasting in the conventional sense is difficult if not impossible, and quality infrastructure must be nimble. This is especially true in the areas of change control, inspection and release of inbound materials, review and release of finished lots, and shipment and distribution.

The conventional processes of line clearance and temporal segregation of multiple products take on new meaning for clinical trial supplies. Study designs may require multiple actives and placebos to be integrated into a single kit. In-process controls for these complex study designs require intensive focus on spatial and physical segregation of components to reduce the risk of errors.
Practical application of these controls includes the use of linear and partitioned work stations, masking templates, stop-and-go verification, and sequential clearances of each component. As with label controls, the scope and dimension of in-process controls will be dictated by the specific requirements for the study design. In-process controls are not formulaic, and effective controls must be specific to the risks for a given study protocol.

Additional Quality Considerations
Unblinding. Blinding is maintained through randomness across all treatment groups. Unblinding risks are present with any attribute consistent with only one treatment group. This could be seemingly innocuous; tilt of a label, color or size of shipper boxes, etc. Clinical supply teams must be trained in the overall concept of study blinding, and in the potential failure modes that could lead to unblinding.

Supply chain considerations. The investigational nature of clinical products corresponds to un-validated upstream manufacturing processes. Any disruption in the upstream supply of drug product can cause massive disruption and potentially interrupt the trial. Delayed drug product supply means delayed enrollments with corresponding cost implications. Additional supply chain controls must consider consistent and reliable supply of comparators products (if applicable), the network of domestic and off-shore distribution channels, certification of regional clinical depots, returns of unused clinical supplies, and in-trial and end-of-trial accountabilities.

Distribution chain considerations. Rapid deployment of clinical supplies requires shortened supply chain. Companies must carefully consider the advantages and risks of centralized distribution versus regional and local depots. Depots introduce new risks: multiple sites increase potential for waste, and a vast empire of depots becomes difficult to certify and manage. Similarly, competitive pressures may drive front-loading of distribution channels ahead of actual patient enrollment. Thorough understanding and control of in-transit risks is integral to quality distribution management, especially if the supplies are temperature sensitive.
Real-time feedback mechanisms for delivery confirmation, condition of received goods, and temperature conformity must be in place prior to engaging distribution channels. Similarly, awareness of import and customs regulations will minimize delays.
Comparators. Comparator products are defined as the commercially available standard of care. New drugs must be compared against current standard of care, not just placebo, and many study protocols involve comparators. These products can be difficult to acquire, and corporate intelligence often sets up barriers against their purchase. To complicate matters, country-specific and global trials may involve multiple brands and strengths which are not available in all countries. In certain cases, comparator products may not even be commercially available. In other cases, the comparator product may represent multiple drugs. The selection, procurement, pedigree of comparator products, and the shelf-life assignment to acquired comparators must also be considered.

Successful Clinical Supply Quality Teams
In part one of this series, we emphasized the importance of the project coordinator function to guide the process through basic questions to create a clear image of the overall project requirements. Similarly, the project quality team will evaluate those basic project requirements to understand the project-specific risks and create a risk management and control strategy to ensure the project is executed with a minimum of disruption. The project quality functions must be integrated with the project team from the earliest stages, and particularly with the project coordinator. Quality Assurance in clinical supplies is absolutely an engineering function, and successful clinical quality groups all share the same attributes:

• Solid grounding on the operational “book of knowledge” for clinical supplies. They understand the fundamental differences between lot numbering and protocol numbering. They understand the differences between spatial and temporal controls when multiple products are involved, and they are comfortable with the rapid cycle times common in clinical supply production.
• Solid grounding and operational knowledge of risk management principles and the ability to tease apart a process, identify the specific risks based on intimate process knowledge, and develop innovative, efficient, and effective controls appropriate to the specific study protocol.
• Solid, productive, and business-savvy relationships with the rest of the project team. Quality conversations nearly always start with “yes, let’s find a way”.
• Knowledge and commitment to the purpose and goals of the protocol, knowledge and commitment to the project team, commitment to the project coordinator, and total engagement with the sponsor.
• A commitment to live conversation and direct face-to-face interactions with team members and sponsors.
• A natural and default inclination towards written project plans and an innate drive to institutionalize what otherwise lives as “tribal knowledge”.
• A written quality agreement with sponsors: Good quality teams don’t want to make up rules and roles on the fly, and they realize up front that they won’t have time for that once the project is in motion, and certainly not when problems arise. Good quality teams anticipate the problems that will arise, and take proactive measures to preempt those problems and pre-define outcomes.
• Effective clinical quality teams are ever ready to provide GMP context and narrative into clinical supply operations that might be considered unconventional GMP; over-labeling, removal of original labels, side-by-side labels.

Like any pharmaceutical operation, quality and GMP compliance lie at the intersection of risk identification and risk control. Maintaining the uniqueness and randomness of a clinical supply job derives from knowledge and understanding of the study protocol. Control of clinical labeling derives from understanding the specific label content and the meaning of both fixed and variable data fields in the context of the study design. And finally, the effectiveness of in-process controls will be dictated not by formulaic solutions common to commercial manufacturing, but through controls specific to the study design. Traditional pharmaceutical quality teams need a new vocabulary and a new body of operational knowledge and risk awareness to be effective in the control of clinical trial supplies. Effective control of clinical supply risks derives from overall situational awareness, integration with the project team, and a deep operational background to understand how all the pieces fit together.