Logistics by design: a framework for advanced therapy developers to create optimal Logistics Platforms

In by


Simon Ellison*, Ryan McCoy*, Marc Bell, Kelly Frend & Stephen Ward

Advanced therapeutic medicinal products (ATMPs) are delivering a new wave of treatment options for unmet healthcare needs. Their impact however, will be severely truncated if supply chain infrastructure is unable to robustly and cost-effectively connect patient to product. Guaranteeing logistical success is becoming an ever-increasing focal point as the field rapidly delivers. Establishing a seamless development approach, including logistics, will be critical in delivering a successful commercial logistics strategy. Paramount to facilitating ATMP commercial realisation is a support structure for logistics planning. In this article we present ‘Logistics by Design’ (LbD) – a framework for logistics-based decision making, based in-part, on Quality by Design principles. This is accompanied by case studies that illustrate the value in applying LbD principles early in the development lifecycle, thereby de-risking the probability of logistics strategy failure.
* Joint First Authors

Submitted for review: Oct 8 2018 Published: Dec 6 2018

DOI: 10.18609/cgti.2018.102
Citation: Cell Gene Therapy Insights 2018; 4(10), 1019-1040.
Open access

Download PDF

Advanced therapeutic medicinal products (ATMPs), of which cell and gene therapies are an integral part, offer a new paradigm for the treatment of unmet healthcare needs. Clinical successes are translating into commercial realization at ever increasing rates and the number of new therapies entering trials are growing year on year [13]. However, compelling clinical results and attainment of marketing authorization (MA) aren’t necessarily synonymous with commercial success or longevity. Since amendments in European Commission (EC) regulation were enacted for ATMPs (2007/1394), 4 of the 12 therapies awarded MA in the European Union (EU) have since been withdrawn [4]. Understanding why therapies haven’t endured commercial success [57] and applying these learnings to de-risk future development programs, will safeguard the realization of ATMPs in the years ahead and enable a sustainable future; when there aren’t just a handful of approved products, but hundreds, servicing millions of patients worldwide.

Possible reasons for commercial failure include high manufacturing cost of goods sold (CoGs), lack of adoption by the clinical community [8], commitment to reimbursement by healthcare providers [9,10] or the complexity of logistics and supply chain infrastructure [11]. Significant progress has occurred in recent years towards establishing closed processing, using automated systems to support robust manufacture and reduce manufacturing CoGs for ATMPs [12,13]. Furthermore, initiatives such as the UK network of Advanced Therapies Treatment Centres, are seeking to address barriers with respect to clinical delivery of these novel therapies.

However, the glue that binds all these elements together, the supply chain, remains pretty immature for the ATMP sector. In part, this is because ATMPs cover a huge diversity of final product and starting material types, meaning a one-size-fits-all commercial logistics strategy is challenging to implement. The authors recognize that ATMP developers can undoubtedly learn from or build upon the logistic models and principles employed by other industries. For example, synergies certainly exist between viral based vaccine production and gene therapies, or tissue engineered ATMPs and organ transplantation. However, the complexity in ATMP manufacturing strategies (centralized vs decentralized, autologous vs allogeneic), product characteristics (e.g., fresh versus frozen, living vs ‘dead’) and clinical utilization (e.g., out-patient vs in-patient, surgical vs non-surgical) means, in the short-term at least, tailoring of the commercial logistics strategy to the specific needs of each therapy is required. In time, as the field matures, it is expected that this would condense down to several ‘platform’ like solutions, perhaps reflective of different therapy or sub-therapy types (e.g., gene, single cell, organoid or tissue engineered), or different clinical applications reflecting different models of therapy access or availability (e.g., cartilage repair vs stroke). The other challenge is that the majority of early-stage development in the ATMP space is driven by academic and small to medium size enterprises (SMEs), which often don’t have the resources and infrastructure in place, like large pharma, to support supply chain activities and thus there is a real need to provide a framework for early stage developers to support their decision making.

Ideally, supply chain management should be a part of every company’s overarching management structure. Supply chain management is the oversight relating to the flow of physical material (starting materials, reagents, consumables, clinical samples etc.), information/data (patient information, batch manufacturing records, quality excursions, clinical data), logistical (transportation and storage of goods) and financial movements as they progress from suppliers, through manufacturing and onto the consumer (e.g., healthcare provider, patient). Figure 1 provides examples of the key stakeholders and movement of materials/data between them. The authors recognize that there are interdependencies that bridge between the stakeholders within the physical, data and business development groups. For example, data associated with delivery of a shipment to the manufacturing centre, in combination with physical receipt and assessment of the material by manufacturing staff, may feed into a performance assessment database held by the business team. However, for simplicity of visualization, examples of such group interdependencies are not shown.

Figure 1. Supply chain management is part of a company’s overarching management structure. It can be considered as the oversight relating to all activities associated with physical materials (yellow arrow), information/data (orange arrow) and financial movements (green arrow) as they progress from suppliers, through a manufacturing center and onto the consumer. Examples of key stakeholder/activities are shown by symbols within the arrows. Logistics is a sub-component of the supply chain management activities and is specific to the process of planning, executing, and directing procedures for the well-organized transportation and storage of goods (yellow arrow), including services and related information/data (orange arrow), from the point of origin to the point of consumption. Exemplars interdependent activities between key stakeholders are represented by the arrows between symbols and reflect ‘transport’ activities for physical materials (yellow, solid) and ‘data transfer’ activities (orange dotted) for data and information management.
CRO: Contract research organization; MC: Manufacturing Centre. Healthcare providers may represent hospitals, clinics, treatment centers or infusion centers and may be different the run through the same of different institutions.

Logistics is often used interchangeably with supply chain, but technically is a sub-component of these activities. It is specific to the process of planning, executing and directing procedures for the organization of transportation and storage of goods, from the point of origin to the point of consumption, including the services and any related information/data (Figure 1). A strong relationship between therapy developer and logistics expert, whether in-house or contracted in, is highly valuable. Whatever mode is selected, then an integrated project approach is a necessity, to ensure technical capability, understanding, and operational synergies are combined by both the development and logistics teams; to ensure a unified vision for a seamless commercial logistics strategy is established early in the development lifecycle. This way potential bottlenecks can be reduced or ideally designed out and optimal solutions yielded; thereby allowing all stakeholder needs to be met by the time of commercial realization. This in turn, will ensure a smooth and efficient connection of the therapy to the patient.

This concept stands for all the critical logistical movements within a therapy’s value chain (e.g., starting material donation, monitoring samples, vector, product, etc.), as failure in any of these movements, be it a temperature deviation or delayed delivery, removes or reduces the therapy’s value: either to the patient, due to efficacy impacts; to the therapy developer, due to lack of clinical data or income; and to the healthcare system, due to the continued ill health of the patient and therefore burden on the healthcare system.

A potential reason why logistics and the supply chain may not get the full attention it deserves, pertain to the fact it has not traditionally been embedded in ATMP sector development strategies, resulting in a disconnection from the development team. Risk management of logistics in pharmaceutical and medical products is not a new concept [14]. However, ATMPs pose additional challenges to the movement of material on a global scale. The strategies being considered to de-risk these activities are likely to be more expensive to implement either through the technology selection required, complex clinical infrastructure to be navigated or increased shipment volumes per batch/dose (e.g., multiple processing site scenarios for autologous therapies) [11,15]. Thus, ignoring logistics until late in the development cycle means opportunities will be lost to create a lean and efficient supply chain that adds value to a therapy developer’s commercial offering.

Forward-looking companies are seeking competitive advantage by establishing platforms and integrated operations. This will accelerate and support future generations of therapy developers in connecting patients to products. However, it is still essential that all therapy developers see the value of logistics platform development and pro-actively support cross-communication between technical experts from the outset.

Key factors defining logistics complexity, practical implementation and economic burden are derived from decisions made across multiple parts of the organization, as well as expectations set by regulatory bodies pre- and post-MA. Thus, only through the engagement of the manufacturing, regulatory, clinical, heath economic and business stakeholders in collaboration with internal supply chain management and external logistics providers, will a unified and aligned vision of the required logistics platform be identified and more importantly enabled. Achieving this early in the development lifecycle allows subsequent testing during clinical assessment and de-risks development activities, safeguarding a fit-for-purpose capability upon commercial realization. The significance of failing to do so, can only be realized once commercial reality hits. The ramifications of which will subsequently be dependent on the company’s ability to adapt post-MA, which can be highly challenging. As illustrated by the well-documented rise, fall and rise again of Dendreon’s Provenge [5].

To support cell and gene therapy developers, this article presents, to the authors’ knowledge, an inaugural framework termed ‘Logistics by Design’ (LbD) for logistics strategy development within the context of a company’s supply chain management undertaking. Employing Quality by Design (QbD) principles, LbD seeks to provide structure to, and de-risk, logistics planning and implementation activities and, importantly, combine it seamlessly to the therapeutic development strategy and resulting chemistry, manufacturing and controls (CMC) dataset. To illustrate the potential of LbD, the authors have built case studies for its application in the context of an autologous ex vivo gene therapy product type (e.g., CAR-T/TCR therapies). The authors have selected to focus on this therapy type, as the logistical complexity of the inbound apheresis starting material, for a clinical market with a global footprint, based on a UK focused centralized manufacturing model, offers the greatest potential to demonstrate the value of this framework.

The Logistics by Design Framework

As with clinical and manufacturing development pathways, we believe the key to logistics success is designing in ‘quality’ from the outset. By doing this, challenges in delivering the logistics strategy can be identified early and there will be sufficient time to consult with key stakeholders (e.g., manufacturing, clinical teams and providers) and tailor the development program to address any high risk or cost drivers. Therefore, we propose a structured logistics development pathway, which runs parallel to clinical and manufacturing development pathways, but is joined up (Figure 2) and contains six key stages:

  • Mapping of and risk identification for the commercial logistics vision – application of LbD principles
  • Building collaborations (technology selection and testing)
  • Infrastructure planning
  • Field validation
  • Scaling for commercial operations
  • Commercial deployment

Figure 2. Product development lifecycles are typically captured in the context of clinical (green) and manufacturing (yellow) development pathways. We propose the formal incorporation of a Logistics pathway (red) that runs parallel to these activities, whereby commercial logistics strategy planning starts prior to Phase 1 clinical trials and the subsequent early phase trials are utilized as test beds for logistic delivery and infrastructure planning. Later stage Phase 3 trials are then formally utilized to validate the commercial logistics platform developed, prior to scaling for commercial operation and deployment.


Stage 1: logistics mapping and risk identification for the commercial logistics vision: application of LbD principles

The LbD framework is built on QbD principles, a concept first outlined by quality expert Joseph M Juran [16] and similar to QbD focuses on risk mitigation, as opposed to cost reduction. QbD is a systematic approach to development that begins with predefined objectives, emphasizes product and process understanding and process control, based on sound science and quality risk management. QbD concepts were formally incorporated into pharmaceutical development in 2005 through the publication of the International Council for Harmonisation (ICH) Q8, Q9 and Q10 documents and adopted by the FDA through the publication of Guidance for Industry Process Validation: General Principles and Practices (Rev. 1, 2011).

In the context of pharmaceutical development (which includes cell and gene therapy), pre-defined product characteristics are captured in Target Product Profiles (TPP) that define the labelled use with respect to safety and efficacy. They usually detail a minimal acceptable limit of performance as well as an ideal performance level. A Quality Target Product Profile (QTPP) is then established that provides a “prospective summary of the quality characteristics of the drug product that ideally will be achieved to ensure the desired quality, considering safety and efficacy”. Thus, the QTPP describes the design criteria for the product and should therefore form the basis for development of the critical quality attributes (CQAs), which in turn allows identification of the critical process parameters (CPPs) and ultimately a control strategy to be devized.

Identification of the CPPs that impact the CQAs may then be elucidated through several approaches including risk assessments, prior knowledge and established science to name a few. Once the CPPs are identified, it is important to establish the operational range or ‘design space’ for each. This can be achieved using several supporting tools, such as Design of Experiments (DoE), which allow the impact of changing these variables to be characterized in a multivariate context. This in turn allows identification of a suitable control strategy, whereby an operational range for these CPPs are defined, which gives assurance that the specified CQAs of the product will be met.

Translating the QbD philosophies described above, we have created a Logistics by Design (LbD) framework to support the cell and gene therapy community. The proposed framework aims to provide guidance on a structure to logistics strategy planning and the value of early implementation.

Figure 3 illustrates the transposition of QbD into LbD principles. As with QbD, from the outset, it is important to set pre-defined objectives that capture the commercial vision of the therapy developer. As with the TPP, we have termed this a Target Logistics Profile (TLP), this document aims to define the overarching objectives of a commercial logistics strategy with respect to supporting business goals, supplying market needs, maintaining regulatory compliance and facilitating clinical adoption. Whereby a minimal and ideal level of performance will be defined for each objective. A Focused Target Logistics Profile (FTLP), like a QTPP, is then established to provide a prospective summary of the traits for a commercial logistics strategy that will need to be achieved, including all components of the value chain, to ensure successful delivery of the therapy to the patient whilst maintaining chain of custody and identity. Thus, the FTLP describes the design criteria for the logistics strategy. This will then form the basis for identification of the critical logistics attributes (CLAs). CLAs can be defined as any physical, temporal, informatic and operational property or characteristic that should be maintained within a required range or tracked and traced, to ensure the desired logistics strategy is fulfilled. CLAs may or may not be directly linked to product CQAs. Where they are linked, failure to deliver the CLA would result in the product CQA falling outside the desired range and thus loss of product. For example, excursion outside of controlled temperature range, which in turn leads to loss of cell viability of the product. Other CLAs may not be directly linked to product CQAs but may be equally important. For example, CLAs associated with maintaining chain of identity or chain of custody. Failure to manage these appropriately may not impact product quality (CQAs) but could impact the ability for the patient to receive material, or worst still, result in the incorrect product being administered. Application of risk assessment tools, prior knowledge, established science or cost of logistics modelling at this point can then be utilized to interrogate the proposed logistics strategy to identify the critical logistics parameters (CLPs) or cost drivers. CLPs are those parameters whose variability or failure would impact a CLA and therefore should be monitored or controlled to ensure the desired logistics strategy is fulfilled, and thus in turn, a control strategy can be devized for these CLPs. Identification of the design space or operating ranges for these CLPs may then be elucidated through practical assessment using supporting tools, such as Design of Experiments (DoE).

Figure 3. Transposition of Quality by Design (QbD) principles into the Logistics by Design (LbD) concept with accompanying definitions for key parts of the framework.


Without having a defined TLP or FTLP, we have utilized a root-cause failure (Ishikawa) diagram for identifying CLPs associated with a generic, all encompassing, ATMP logistics strategy (Figure 4). This figure illustrates the vast number of potential CLPs for a commercial logistics strategy. For ease of visualization and characterization, CLPs constituting potential root-cause failures were divided into 5 principle areas relating to Manufacturing, Clinical, Market, Shipping and Data Management activities. It is important to note that CLPs are often interweaved into complex networks of dependencies that are not always directly obvious and therefore thinking about logistics on a ‘global’ scale from point of origin to point of receipt will be critical to success. Case Study A provides an example of how a single decision made by the manufacturing development team, to cryopreserve the final product, requires substantial in-depth analysis, preparation and verification of subsequent downstream logistical and clinical undertakings to maximize the chance of successful delivery to the patient.

Figure 4. Exemplar root cause failure (Ishikawa) diagram for a generic logistics strategy. Root-cause failures were divided into five principle areas relating to Manufacturing, Clinical, Market, Shipping and Data Management Activities. Primary causes of failure (black Text) were then broken down in secondary causes (blue). Examples of the selection options/therapy requirements for each of the root-cause failures is shown in red text.


Stage 2: building collaborations (technology selection & testing)

Having defined and identified all the logistic activities required to execute the complete value chain, this stage of the framework is focused on building (and managing) collaborations essential to seeing the logistics strategy realized. Based on current logistic provider and market capacity, if a therapy is successful (thousands of doses a year), it is likely to take up a substantial proportion of global capability. Thus, choosing the correct level of collaborative working for each activity will be essential in balancing quality of service and cost considerations. Logistic providers should be viewed as technical experts, with ‘joint planning’ and ‘collaborative’ relationship levels sought for complex and high-risk elements of the logistics chain [17]. Their expertise and experience can then be harnessed to add maximum value. Furthermore, by having this level of co-engagement and investment in the therapy, providers can appreciate the landscape ahead and ‘co-evolution’ of companies can occur synergistically, amplifying the probability of success.

Each provider has hubs that are strategically placed geographically, preferred lane maps for given routes (e.g., preferred airline carriers, based on reliability and service quality metrics – which may impact on flight availability), variable shipping strategies (e.g., some cryo-shippers do not fit on narrow body planes and therefore limit flight options) or individual relationships with key technology providers in the sector, the culmination of which all impact their ‘freedom to operate’. Thus, performing due diligence early in the development lifecycle and subsequently identifying the right provider for your needs, will be critical in ensuring the service desired is firstly, achievable and secondly the one delivered.

Part of these collaborative working relationships will include the selection and testing of specific technologies, whether it be the physical product shippers or complex integrated data management systems. If selected early in the development lifecycle, these can then be assessed in the context of early stage clinical trial environments to ensure that at least for low operational volumes, they meet required needs. Based on the vision the solution can be built, tested and optimized to the needs of the key stakeholders iteratively, to ensure commercial realization, a fit-for-purpose solution is in place.

Stage 3: infrastructure planning

Early insight into the challenges and technological feasibility of the proposed commercial logistics strategy, as provided in stages 1 and 2 of the framework above, will provide developers with an opportunity to revisit any of the original decisions based on data driven assessments of performance. If performance was sub-optimal then alternative solutions can be employed and tested throughout this phase.

Once the agreed pathway forward is characterized, then planning can commence for establishing the required infrastructure to support field validation (stage 4) as part of pivotal Phase 3 studies and ultimately the full-scale commercial deployment (stage 6).

Activities included at this stage of development may include for example the identification and sourcing of appropriately located warehousing capability within a specific geographic or temporal footprint. For example, having an allogeneic off-the-shelf cryopreserved product for an acute clinical condition such as stroke, where it may be important to administer the therapy within a small timeframe window (<12h), means it may be more desirable to have several smaller cryo-storage hubs, appropriately distributed globally, than one big master centre. Or alternatively, other examples may include the installation and set-up of electronic data management services (server set-ups, software installations, technology sourcing, procurement and implementation). In both cases, identifying where these should be based, the order they come on-line, and the lead times associated with their enactment, all need to be accounted for. By starting the process at this point in the development lifecycle, it ensures the opportunity to capture the full market potential of the product is realized from the outset of commercial activities: having demand outweigh the ability to supply, is potentially the worst scenario to be in logistically.

Stage 4: field validation

This stage of the framework is focused on large-scale validation of the proposed logistics strategy during the Phase 3 pivotal trials with a view to gathering ‘in-the-field’ data.

Stage 5: scaling for commercial operations

This stage of the framework focuses on the scale-out of the logistics strategy to cover the full commercial footprint. Exemplars of activities that may be actioned at this point in the development plan include:

  • Bringing on board additional warehousing
  • Lane mapping to ensure delivery windows are achievable based on clinical availability, manufacturing schedules, etc.
  • Implementation of training in key procedures and processes to new market sectors
  • Translation of documents into languages for new market sectors that were not covered as part of the pivotal clinical trials
  • Embedding novel technologies (e.g., controlled thawing devices) at new clinical sites
  • Bringing on-line additional manufacturing facilities to meet expected market demand

  • Stage 6: commercial deployment

    This stage of the process involves executing the developed commercial logistics strategy, undertaking continual performance monitoring with a view to identifying points of failure (where and why) and where appropriate, implementing further mitigation strategies to correct these in real time. Additionally, once substantial volumes of data are generated, to put in place iterative review procedures with the intention to identify areas warranting further streamlining of processes and procedures.

    Case Study A

    CLAs and CLPs: mapping the complex networks of dependencies on a ‘global’ scale, from point-of-origin to point-of-receipt – a critical step in developing successful logistics strategies

    Cryopreservation of ATMPs during manufacturing has numerous advantages in terms of implementing the downstream logistics undertaken. It extends product shelf-life, removes the substantial pressures associated with the time critical nature of ‘fresh’ materials and allows for flexibility in patient readiness for clinical administration procedures. Cryopreservation is likely to remain the go to option until a stable cell/therapy at ambient temperatures or alternative preservation strategies are developed.

    Managing and handling cryopreserved products has significant implications from logistical and clinical perspectives with respect to successfully connecting product to patient. Thus, mapping the impact on the CLAs and CLPs within the logistics network is paramount to ensuring the logistics strategy will be successful in meeting the needs of the TLP. Exemplars of key stakeholders, CLAs and CLPs impacted, as initially described in the root-cause failure Ishikawa diagram (Figure 4) are tabulated in Table 1 shown in Figure 5 and described below respectively, as a consequence of the ‘manufacturing’ stakeholder deciding to introduce a cryopreserved final product.

    Table 1. Exemplar CLAs and CLPs relating to different key stakeholders that may be impacted by the decision of the manufacturing stakeholder to introduce a cryopreserved product.
    ManufacturingFinal product storage conditionsTemp (-196oC)
    ShippingPackaging selectionPhysical parameters (size, weight)
    Operational parameters (dry shipper validated hold times)
    Business (market)/shippingClinical site locations (destinations)/airline providersAvailable routes (aeroplane type)/aeroplane hold size
    Clinical siteDriver accessCorridor sizes / stairs etc.
    Pharmacy storageFreezer infrastructure available/footprint
    Dry shipper handlingAvailability of pallet truck
    Data managementTracking devicesBluetooth, GPS or BLE
    Variation in clinical site data formatStandardize or filter multiple formats
    Systems integrationWhen to adopt cell orchestration platforms
    These are based on the Ishikawa tier 1 headings shown in Figure 4. There are other key CLAs and resulting CLPs that need to be assessed in order to create a viable logistics platform.

    Shipment of cryogenically stored frozen products occurs in dry-shippers. Dry-shipper technology selection will determine a range of subsequent CLPs, including the weight and size of the dry shipper. These will not only influence the cost of the shipment (a highly critical factor when proposing to do thousands or tens of thousands of shipments a year), but also how it will be handled on the ground. The total weight of full dry-shippers and any protective secondary containers can easily exceed 20–25 kg and are often transported on pallets. This minimizes heavy lifting operations for personnel within the logistics chain, but also ensures the vessel is maintained level during the shipment process. The consequence however is multi-factorial. Firstly, it may limit route availability for airline transport (the holds of some aircraft, particularly regional jets common in the USA and other markets, are not sufficiently large to accommodate some dry-shipper designs). This, in turn, may lengthen transport time to the clinical site as follow-on transport options are required. If this results in shipment times exceeding the validated shipping window of the dry-shipper, clinical site reach may be constrained, impacting market access.

    Figure 5. Decision-making by one part of the organization has the ability to significantly influence other parts of the system, and potentially the ability to treat patients effectively.


    Secondly, movement of pallets require the use of pallet trucks and pathways that support their use (e.g., no stairs, sufficient width for manoeuvrability etc.). These may seem like small details, but every clinical site will be different. When dealing with high-cost therapies that have potentially curative properties for the patient, it may be imperative that clinical teams have ready access to a pallet truck to move a dry-shipper from pharmacy to patient bedside. These are the kind of risks that can be easily addressed given the appropriate amount of foresight. Thirdly, it is well recognized that hospital pharmacies, particularly those outside of major research hospitals, are not well setup to support cryopreserved products with their own liquid nitrogen storage capability. Consequently, pharmacy storage is expected to consist of leaving the product in the dry-shipper until point of use. Pharmacies are typically small footprint working areas and not amenable to the storage of large numbers of dry-shippers on pallets, which constrains them in terms of the number of products that can be held concurrently whilst awaiting patient administration. Overarching these physical material movement activities, will be equally important data management processes. Due consideration will need to be given to factors such as the types of tracking devices used, based on the need for real time versus offline data capture and how these integrate into systems and processes. For example, if pharmacies have the potential to transfer between dry-shippers (should they be close to the end of their validated hold-times and the patient is not ready for administration), then it will be essential that the temperature tracking devices are real time in nature, have alarm set-points programmed well in advance of ‘failure temperatures’ and that the alarm messages feed into automated systems that inform the correct stakeholders to take action. Thus, to enable such activities, it may be important that the current clinical site IT infrastructure can support novel software solutions.

    The exemplars above only demonstrate a portion of the thought processes, for a couple of the pathways represented in Figure 5. However, the key message here is that all logistics decisions are likely to have both positive and negative impacts at different points throughout the supply chain. Therefore, understanding what the CLPs are and how they impact the CLAs, by mapping the logistics vision early in development, will enable the implementation of control strategies to mitigate risks or manage expectations with respect to clinical and market reach.

    Case Study B

    Mapping inbound shipments of fresh apheresis starting material: understanding material collection, shipment & manufacturing considerations in terms of CLAs & CLPs


    Successful commercial logistics strategies will be built on the timely coordination of clinical, supply chain and manufacturing operations. Mapping material flows and applying risk assessment tools to the projected landscape of a commercial logistics strategy, early in the development lifecycle, provides significant opportunity to identify and subsequently ‘design-out’ any barriers to later stage translation and realization. Or alternatively at least, inform and manage expectations of key stakeholders.

    To illustrate this, we have developed the following case study. To illustrate the link to the root-cause failure diagram in Figure 4, we have highlighted relevant CLAs in italics throughout the case study. Employing a centralized manufacturing model (CLA – Manufacturing Strategy) that utilizes the Cell and Gene Therapy Catapult manufacturing centre in Stevenage, we mapped inbound shipments of fresh apheresis material from several national and international starting locations, as is required for any such centre (CLA – Market Size/Clinical Locations); 1 national (London), 6 EMEA (Amsterdam, Brussels, Copenhagen, Paris, Prague, Tel Aviv and Warsaw), 3 North American (Boston, Houston and Los Angeles) and 1 APAC (Tokyo). For the national (UK) routes, road-based transport was employed, whilst for the EMEA routes for Paris and Brussels, a combination of the Eurostar (arrival at London St Pancras) and road transport was utilized (CLA – Shipping – route). For all other EMEA, APAC and North American routes, a combination of road and air transport (arrival at London Heathrow, LHR) was employed (CLA – Shipping – route). Logistic pathways for these different scenarios were deconstructed into key ‘activities’ to allow a more granular analysis to be undertaken (Figure 5). We examined the impact of two different scheduled courier collection times (12pm and 4pm local time – CLA – Clinical Site – Consignment Collection Time) on the total transit/shipment times and the local time of receipt at the manufacturing centre (CLA – Manufacturing Facility – Operating Hours). An overview of data captured is illustrated in Figure 6.

    Figure 6. Key ‘activities’ associated with the logistic pathways for the inbound shipment of apheresis material originating from national (UK) or EU, APAC and Northern American clinical sites to the Manufacturing Centre (MC) at Stevenage. For the originating locations studied in Case study B, Paris and Brussels used rail transport via Eurostar (UK entry port, St Pancras), whilst all other locations used air transport (UK entry point, London Heathrow). In this study, shipments from all EU and North America originating destinations had ‘recovery’ times (time taken for customs clearance and for the courier to retrieve the package from the airline/Eurostar – including subsequent transport to the courier hub in London if required) that were based on historical data. The estimated road transit times (to the Cell and Gene Therapy Catapult Stevenage manufacturing center from the courier hub/UK entry point) were tailored to the time of day at which a journey commenced. The other variable components of the logistics chain contributing to the total transit time, were then the flight duration (the length of which was a function of the physical distance between the originating destination and LHR), the flight pre-processing time (e.g., the ‘check-in’ or airline/Eurostar handover time required prior to the scheduled departure time) and the transfer time between the clinical collection site and airline/Eurostar handover point.

    Based on the starting locations studied, the total transit time ranged from 90 min (1.5 hours) to 2155 min (35 hours 55 min) for the 12 pm collection time and 120 min (2 hours) to 1915 min (31 hours 55 min) for the 4 pm collection time respectively. Interestingly, all European starting point material could be delivered within 24 hours, for either collection times; and even from several US sites, for the 12 pm collection time slot.

    Total transit times were within a 2-hour window for each destination (7 out of 12) irrespective of the collection time (Table 1). Where differences (>2 hours) were observed (Brussels, Prague, Copenhagen, Tokyo and Tel Aviv) this was attributable to constraints associated with available flight/rail times for these destinations, as a function of the collection time from clinic, for example only a single flight was available from Prague and Copenhagen each day at 19.55 and 20.35, respectively. In these instances, there may be an advantage to scheduling clinical apheresis sessions later in the day (if feasible) to reduce the total shipment times.

    Flight/train schedules are by far the most constrained element of any logistics pathway. Airlines are key stakeholders in ensuring successful delivery of products, yet therapy developers and logistics companies, even though working to influence recovery of dry shippers, etc. have no influence over their flight schedules and decision-making pathways. Therefore, if their use is essential, developers and logistic companies must ensure other elements of the logistics pathway, pre- and post- flight, are aligned accordingly and mitigation strategies for potential risks of failure in-place.

    For perishable starting materials, such as the apheresis shipment modelled herein, the goal is to minimize the overall shipment time. Therefore, the time windows for each part of the pathway are typically squeezed to remove any slack and maximize the quality of the starting material upon receipt at the manufacturing centre. For example, Figure 6 illustrates the cumulative shipment time for each of the routes studied, which provides an approximation as to the bare minimum that the shelf-life of the inbound material needs to be validated for. The validated shelf-life however, should also allow for potential scenarios where delays occur during shipment. As flight/train times are highly constrained and therefore lie on the critical path of any logistics pathway for fresh material, it is thus valuable to understand the risk and impact associated with delays (e.g., weather, or flight delay) causing the missing of pre-planned departures. Once known, potential mitigation strategies can then be devized or employed. For example, very simplistically, as shown in Table 2, we identified the next available scheduled flight/train for each destination relative to that originally proposed. As illustrated, missing the scheduled departure time and being able to connect with the subsequent flight/train, would typically result in an extension to the shipping time between 2 and 5 hours, but in some extreme cases, such as for Prague, it could result in a minimum delay of 17 hours 50 min based on an original collection time at the clinic of 12 pm.

    Table 2. Difference in the total cumulative shipping time between 12pm and 4pm local collection times as a function of originating location.
    Collection locationTransit time between 12pm and 4pm collection times (min)
    Los Angeles-35
    Tel Aviv240

    For the later collection time investigated (4 pm), missing the planned flight resulted on average of a delay between 9 and 12 hours, with a worst case scenario of 20 hours 30 min. Knowing this early in the product development lifecycle would allow therapy developers to subsequently ensure starting material shelf-life is validated for an appropriate timeframe, or if this is not feasible, would then support implementation of suitable control strategies to be employed from the clinical (e.g., ask patient to arrive day before and stay overnight in clinic) or logistical perspectives (build in extra time for drivers to reach the clinical site) to de-risk the chance that the flight departure time is missed. A real-world example of such a scenario is the well-quoted example of Dendreon (Provenge) where the product was originally launched with an 8-hour shelf life. This meant for the USA patient population to be served, three manufacturing sites were required. The resulting cost and complexity drove up the market price and contributed to the products initial challenges. Subsequent development work enabling the shelf life to be extended to 18 hours, meant Dendreon was able to remove one manufacturing center and, as part of other efforts, relaunch Provenge as a viable commercial therapy that is now treating ~4,000 patients per year [18].

    It is important to understand that there are multiple stakeholders in each part of the process and when mitigating the impact of delays all their roles should be considered. For example, if customs clearance is required, as per the APAC and North American locations studied herein, then LHR is a good entry point from a Customs perspective as they operate 24/7. However, if the reverse shipping routes were employed for perishable starting materials/products then it is important to understand that for example Tokyo and Tel Aviv Customs operations only run between 08:30 and 17:00 Monday to Friday and 08:00 and 22:00 Monday to Friday and Sunday, respectively. Thus, if a scheduled departure time is missed (as described previously), but rescheduled for a later flight, it doesn’t automatically mean a seamless continuation of the journey once at the destination. If Customs operations are closed, further delays may be incurred, which in turn could ultimately mean loss of material/product if suitable validated shelf-life is not in place. Other exemplar factors to consider may include differences in summer and winter flight schedules, impact of bad weather or geo-political events. Therefore, engaging with providers that have overarching knowledge of the logistical landscape, early in the product development lifecycle, is essential in supporting manufacturing and business decision-making processes to ensure processes and procedures developed are commercially realisable.

    In this study, post-flight recovery and subsequent road transit to the manufacturing centre was assumed to occur in a continuous fashion. Based on the projected time for each of these activities and assuming on-time flight/train arrival at LHR, the anticipated arrival time at the manufacturing centre was calculated (Figure 7). If the facility is not operating a 24/7 model, thereby accommodating acceptance of starting material and immediate commencement of manufacturing at any time of the day, then with this level of understanding, therapy developers can start to consider the impact of various process development and business decision-making outcomes on the commercial viability of different operating strategies for the facility. For example, what does the validated apheresis shelf-life need to be for a global footprint, if there is a requirement for hold times overnight at the manufacturing centre, or does facility operation allow the processing of 1, 2 3 or more apheresis collections if they all arrive within a 1 or 2 hour window. Choosing to operate a 24/7 facility is not without challenges as it has a cost impact. Thus, mapping the commercial logistics strategy will be highly valuable in supporting data driven decision-making and enabling key stakeholders to have oversight of the advantages and disadvantages to strategies employed reducing the likelihood of failure upon commercial realization.

    Table 3. Route mapping: understanding the impact of missing constrained scheduled departure times on total shipment times.
    DestinationOriginal collection time (local)Scheduled
    departure time
    (flight or rail)
    First available ‘back-up’ optionNew departure timeMinimum additional shipment time
    (h and min)
    Brussels12:0014:56Later flight – same day16:562h 00 min
    Amsterdam12:0015:55Later flight – same day20:304h 35 min
    Paris12:0017:13Later flight – same day19:132h 00 min
    Prague12:0019:55Later flight – next day13:4517h 50 min
    Warsaw12:0015:30Later flight – same day20:004h 30 min
    Copenhagen12:0020:35Later flight – next day07:5011h 15 min
    Houston12:0016:25Later flight – same day20:203h 55 min
    Boston12:0019:15Later flight – same day22:503h 35 min
    Los Angeles12:0016:55Later flight – same day21:354h 40 min
    Tokyo12:0001:55Later flight – same day11:209h 25 min
    Tel Aviv12:0007:35Later flight – same day16:359h 00 min
    Brussels16:0017:56Later train – same day18:561h
    Amsterdam16:0020:30Later flight – next day07:2010h 50 min
    Paris16:0019:13Later flight – same day21:001h 47min
    Prague16:0019:55Later flight – next day13:4517h 50 min
    Warsaw16:0020:00Later flight – next day07:3011h 30 min
    Copenhagen16:0020:35Later flight – next day07:5011h 15 min
    Houston16:0020:20Later flight – next day16:5020h 30 min
    Boston16:0022:50Later flight – next day07:408h 50 min
    Los Angeles16:0021:35Later flight – next day15:3518h 00 min
    Tokyo16:0001:55Later flight – same day11:209h 25 min
    Tel Aviv16:0007:35Later flight – same day16:359h 00 min
    Based on the originally mapped routes, assuming that departure conditions for the original scheduled flights weren’t met, the next available flight/train was determined to determine the minimum impact on total shipment time.


    Figure 7. Cumulative shipment time maps, broken down into the key activities as described in Figure 6 for inbound shipments of fresh apheresis material from one national (London), six EMEA (Amsterdam, Brussels, Copenhagen, Paris, Prague, Tel Aviv and Warsaw), three North American (Boston, Houston and Los Angeles) and one APAC (Tokyo) destinations based on local collection times of 12 pm (A) or 4 pm (B) from clinical centers. For the national (UK) routes, road-based transport was employed, whilst for the Paris and Brussels EMEA routes, a combination of the Eurostar (arrival at London St Pancras) and road transport was utilized. Whilst for all other EMEA, APAC and North American routes, a combination of road and air transport (arrival at London Heathrow, LHR) was employed. Tables to the right of the graphs show the expected arrival time at the Manufacturing Centre (MC) and the cumulative shipment time in hours and min.


    LbD gives therapy developers, for the first time, the tools to identify risks, utilize providers’ expertise, coordinate the solution, and create logistics platforms that connect their therapies to their patients. There may not be easy solutions to the challenges identified herein but having the knowledge and making data-driven based decisions early in the development lifecycle, will maximize the chance of developing commercially successful logistics strategies. The authors recognize that establishing a de-risked commercial logistics strategy through application of the LbD framework, will just be the beginning. Following on from this, it will be critical that therapy developers minimize the cost contribution of logistics to maximize their commercial viability. The authors envision a scenario where the outputs of the LbD framework could be utilized as a baseline for developing ‘cost of logistics’ modelling tools, which can in turn be wrapped around or integrated into the framework outputs. Thereby supporting therapy developers in understanding the financial implications of changes to their logistical or supply chain strategies.

    financial & competing interests disclosure

    The authors have no relevant financial involvement with an organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock options or ownership, expert testimony, grants or patents received or pending, or royalties. No writing assistance was utilized in the production of this manuscript.


    1. Ginn SL, Amaya AK, Alexander IE, Edelstein M, Abedi MRL. Gene therapy clinical trials worldwide to 2017: An update. J. Gene Med. 2018; 20(5): e3015. CrossRef

    2. Catapult, C.a.G.T., UK Clinical Trials Database. 2017. CrossRef

    3. Medicine, A.f.R., ARM Annual Data Report 2017. 2017. CrossRef

    4. Cuende N, Rasko JEJ, Koh MBC, Dominici M, Ikonomou L. Cell, tissue and gene products with marketing authorization in 2018 worldwide. Cytotherapy 2018; 20(11): 1401–13. CrossRef

    5. Jarosławski S, Toumi M. Sipuleucel-T (Provenge®)—Autopsy of an Innovative Paradigm Change in Cancer Treatment: Why a Single-Product Biotech Company Failed to Capitalize on its Breakthrough Invention. Vol. 29, 2015. CrossRef

    6. Abou-El-Enein M, Elsanhoury A, Reinke P. Overcoming Challenges Facing Advanced Therapies in the EU Market. Cell Stem Cell 2016; 19(3): 293–7. CrossRef

    7. uniQure, uniQure announces it will not seek marketing authorization renewal for Glybera in Europe, 2017.

    8. Gardner J, Webster A, Barry J. Anticipating the clinical adoption of regenerative medicine: building institutional readiness in the UK. Regen. Med. 2018; 13(1): 29–39. CrossRef

    9. Driscoll D, Farnia S, Kefalas P, Maziarz RT. Concise Review: The High Cost of High Tech Medicine: Planning Ahead for Market Access. Stem Cells Trans. Med. 2017; 6(8): 1723–9. CrossRef

    10. Jørgensen J, Kefalas P. Reimbursement of licensed cell and gene therapies across the major European healthcare markets. J. Mark. Access Health Policy 2015; 3: 10.3402/jmahp.v3.29321. CrossRef

    11. Rutherford C, Barry J, Campbell JDM, Turner M. The Importance of Understanding & Designing Cellular Therapy Supply Chains. Cell Gene Ther. Ins. 2017; 3(10): 873–89. CrossRef

    12. McCall, Phacilitate SIG Automation – Rapporteurs’ Report. 2017. CrossRef

    13. Marshall D, Ward S, Baradez M-O. Requirement for smart in-process control systems to deliver cell therapy processes fit for the 21st century. Cell Gene Ther. Ins. 2016; 2(6): 683–9. CrossRef

    14. George K, Nagali V, Rassey L. Expect the unexpected: Reduce corporate exposure and create value through supply chain risk management. 2011.

    15. Lowdell MW, Bain O. Supply Chain Management in the Delivery of ATMPs for Trials & as Licensed Medicines. Cell Gene Ther. Ins. 2017; 3(10): 843–51. CrossRef

    16. Juran JM. Juran on quality by design : the new steps for planning quality into goods and services. 1992; New York: Free Press, NY, USA; Maxwell Macmillan Canada; Maxwell Macmillan International; Toronto, Canada.

    17. Audy JF, Lehoux N, D’Amours S, Rönnqvist M. A framework for an efficient implementation of logistics collaborations. Int. Trans. Op. Res. 2012; 19(5): 633–57. CrossRef

    18. Hayes E. Under Sanpower, Can Provenge Finally Have A Happy Ending? 2017: https://pharmaintelligence.informa.com/resources/product-content/provenge-finally-have-a-happy-ending


    Simon Ellison1*, Ryan McCoy2*, Marc Bell3, Kelly Frend4 & Stephen Ward5
    1Cell and Gene Therapy Service Director, World Courier
    2Lead Technical Scientist, Cell and Gene Therapy Catapult, UK
    3Manufacturing Development Specialist, Cell and Gene Therapy Catapult, UK
    4World Courier, Specialty Services Manager
    5Chief Manufacturing Officer, Cell and Gene Therapy Catapult, UK
    *Joint first authors.