Future Supply Chains Enabled by Continuous Processing

In this section, we consider "What might the benefits of more flexible, responsive continuous manufacturing-based end-to-end (E2E) supply and innovation chains bring to the healthcare/pharmaceutical manufacturing sector?"

In our analysis of current E2E supply chains, we observe the following potential opportunities from moving to continuous process manufacturing:

1. Greater product and volume flexibility enabling multiple supply chain models, more tailored to specific market needs
2. Significant inventory reduction opportunities through a more responsive E2E SC (Supply Chain)
3. Improved quality
4. Rapid scale-up post clinical trials, perhaps redefining the nature of clinical trials and subsequent commercial production
5. Reduction in the management burden and overhead structural costs they generate in the in the current supply chain paradigm where multiple human to human hand-offs are required to deliver product to patients

These five themes are further developed below, outlining where the opportunity sits.

Greater Product and Volume Flexibility Enabling Multiple Supply Chain Models, More Tailored to Specific Market Needs

It is now generally accepted that the pharmaceutical industry is progressively moving away from the large volume "blockbuster" drug production model and requires future production and supply chain models that can deliver significantly greater product variety and volume flexibility. Indeed, this may involve product delivery models developed or tailored to serve relatively niche markets where patient populations are significantly smaller than today’s norms. This, together with advances in stratified and personalized medicines, will require levels of product customization that make the batch centric production models of today incapable of economically supplying these product varieties (SKUs) at the smaller volumes required, and at the speed increasingly demanded by end-users (patients and payers) without the costly "buffer" of huge inventory.

The potential opportunities for "continuous processing" centric manufacturing supply chains include:

• New capabilities for firms to meet as yet unmet patient needs, by developing capabilities to supply niche markets that are currently uneconomical to serve because of the small product volumes linked to specific patient populations.
• The volume flexibility afforded by continuous processing, unconstrained by batch size, has major implications for materials requirements and inventory, shortening dramatically the shelf-life of products that are often determined by minimum batch volumes.
• Better adaption to market supply and country versions: variety of small markets (in terms of units produced, not value) can drive a substantial accumulated volume. Impact of individualized medicine can require a variety of strengths that need to be produced, each as multiple country versions. In the batch world of today, this is difficult to handle and negatively impacts on the Cost of Goods Sold, inventory and shelf life requirements, as well as operational costs linked to increased complexity.
• Late customization and deferment models is another alternative supply scenario, where final product definition is achieved in-market within a distributed and geographically dispersed "final stage" manufacturing activity; in this scenario, both upstream and downstream stages may preferentially support continuous process technologies that support "material flow" supply dynamics as opposed to the current "batch" campaign model.

Within medium-to-large volume-scale manufacture, two key questions emerge; the volume-scale where the transition from batch to continuous becomes attractive, and whether the desired volume flexibility may be achieved by various combinations of batch and continuous processing. As continuous manufacturing overcomes the discretization of batch sizes, it opens up a range of volume options not otherwise possible. Furthermore, the development of post-dosing manufacturing capabilities might afford further levels of late-customization when applied to a common base product. If the latter is made continuously, the volume options for "minimum order quantities," a key criterion in supply chain design, become potentially unconstrained.

One potential application area is in supporting patient "dose flexibility". Whereas dose flexibility is already a reality in injectable products, oral liquids, and semisolids, where the patient is provided this flexibility, this may also be deliverable in discrete dose formats. For example, liquid forms produced in continuous mode, can potentially provide varying formulations, process controlled to conform to specification. Alternative technologies would be required for solid-dose forms, such as additive production process models, ink-jet styled dose control strategies, or novel "multi-dose" pack-formats and pack-devices to deliver this dose flexibility functionality.

Late product customization, integrated with individualized pack labeling, will be fundamental in supporting potential developments in individualized and personalized medicines. The availability of faster and more flexible supply chains, as enabled by continuous processing, may also enable products and dosage forms that inherently have shorter shelf lives.

Significant Inventory Reduction with a More Responsive E2E SC

The opportunities for a significant reduction in inventory through continuous processing results from chemical processing models offering reduced process steps and process equipment that provides significant volume flexibility. This potentially leads to substantial reductions in inventory which inturn enable:

• Moving to more of a "demand-driven" replenishment model rather than the current long-term forecasting approaches with wider opportunities for manufacturing and supply chain integration.
• The ability to operate on substantially shorter lead times for product replenishment by significantly reducing intermediate and finished goods stock levels.
• Consideration of chemical routes involving short-lived unstable intermediates within continuous manufacturing, normally avoided in batch-chemistry, opening up alternative synthesis routes.

For these benefits to be effectively realized, the industry will need to confirm the expected improvement in supply chain robustness and resilience ensuring complete confidence in the supply chain delivering medicines to patients.

Improved Quality

Continuous processing can lead to substantially less rework, assuming rapid start-up to steady state and that recycling is routinely possible within agreed operating and regulatory frameworks. As consistency is one of the hallmarks of continuous processing, a well-designed and effectively run continuous process can deliver a highly consistent product, leading to lower variance and more reliable performance.

Continuous processing has, among other aspects, one fundamental differentiator from batch processes, which can have a significant impact on the supply scenario. This aspect is process control and the capability of enforcing process conditions at a micro-level, which has a fundamental impact on development processes, on quality and on supply.

Batch processes, for example, operate under the paradigm that the totality of material is transformed in a reactor of some sort, which holds the totality of material all at once. This makes the reactor size dependent on the desired batch size. Reactor size, however, drives the enforceability of process conditions of the entirety of material on a micro-level. An illustration of this is in an exothermic chemical reaction where heat management is critical to controlling the reaction. The heat generation is endogenous to the material and the heat control can only be obtained by cooling the walls of the reactor. Temperature is dependent on the distance of the point of interest to the wall. The larger that distance, the smaller the impact of the wall temperature, in our case, the cooling effect on the reactive conditions, such that local overheating is a real possibility, as even with the best temperature control in the reactor wall, the freedom of the reaction to exhibit local overheating (that is not even noticed) is high.

Similar examples exist for other processes, which may also have complex numerical scenarios involved, aka nonlinearities in material laws and heat flow or property propagation in general terms. Examples include batch crystallizers, wet granulation in shear mixers (high or low), including the fluid bed drying of tableting processes, as the special distribution of properties within a single tablet is not always uniform. In summary, the larger the reactor is (or shall we better say the process equipment to describe in an abstract way everything from a chemical reactor to a powder handling system), the less control we have over the real conditions that transform the material. Consequences of that statement are among others that the quality of the transformation is only loosely controlled, and a robust or in other words forgiving product or formulation is needed to limit the impact of this. Again in our exothermic reaction example with poor control at the micro-level of reactive conditions like temperature and mixing, local overheating may occur and even degrade the product through a deteriorating follower reaction or decomposition. In the batch world, these by-products will be diluted into the entire batch and will raise impurities. Similar effects can be named for almost any other unit operation: in wet granulation local over granulation would come to mind, in tablet formation capping as a consequence of inhomogeneity of process conditions and so on, the list is long. So, in other words, a different scale of unit operation has a fairly high chance of quality attributes being scale dependent. One can even drill this down to a tablet size being the determining factor for the last "material transformation" in the pharmaceutical delivery chain, the dissolution in the patient’s stomach. The consistency potential of continuous processing is thus significant both in terms of quality but also as set out below the opportunities for more rapid scale-up.

Rapid Scale-Up Post Clinical Trials

In terms of the challenges in commercial scale-up of continuous processes, once the manufacturing process has been established within the clinical trial regime, these are widely recognized as being significantly smaller and less expensive than for traditional batch processes.

The cost of bringing products to registration can be significantly reduced when using continuous manufacturing for the design of experiments (DoE) during development to support a quality-by-design (QbD) filing. By way of exemplification, GSK have demonstrated significant reduction in scale-up time and cost during development by switching from batch to continuous granulation.1 When a product transfers to commercial manufacture, it is anticipated that this will achieve reductions in operating costs, work in progress, and footprint. This switch from batch to continuous granulation also provided evidence of significantly reduced variance in the size distribution of the granules. There are also potentially reduced material requirements for scale up in continuous processing trails. In a batch-based operation, the scale is defined by the size of the process equipment that is used. It defines the amount of material that is exposed to a homogeneous application of processing conditions. Any variation of processing conditions requires the production of this quantity of material at the given set-point. This multiplies the material consumption per set-point with the batch size and hence leads to huge material consumption to prove the validity of different set-points of the processes. In continuous processing, the variation of set-points can take place "on the fly" and hence allows a much faster set-point screening. It practically replaces the minimum amount of material per processing condition from the batch defined amount to an amount that is given by the transient time it takes to change from one set-point to another. These can be significantly smaller in a continuous processing setting and hence the amount of materials required for an array of set-points can be substantially reduced. This can be seen as an easier and less costly scale-up model.

A specific example of reduced material requirements for scale-up is in the quantity of API(Active Pharmaceutical Ingredient) required to fully develop and scale up to commercial scale. In a batch process, quantities are typically not available at the phase II stage of development. Performing multivariable factorial DoE (a key component of QbD development) using large-scale batch processes is time consuming because each data point in a multi-step process that could take several days or weeks to generate. In contrast, a comprehensive DoE with multiple data points could be performed in less than a day with continuous manufacturing.

Reduction in Management Overhead Costs

As discussed earlier, significant managerial costs, often "hidden" in company manufacturing or supply chain overheads, are driven by the managerial resources required to operate the supply of products through the manufacturing supply chain. A more continuous flow operating model would require systemized linkages across production and supply operations, minimizing human to human hand-offs, and driving the need for better demand signal detection through the E2E supply chain. Currently, these overheads "allocated" to products may equate to around 50% of the product cost, and although constitute a multitude of cost factors, are primarily linked to the combination of expensive and under-utilized assets and associated depreciation charges, but also the management costs involved in the oversight of these complex interactions.

In quantifying the opportunity, Table 1 provides an assessment of the potential scale of the opportunities across the theme areas that can be targeted over the medium term.

Table 1. Scaling the Potential Opportunities Across the End-to End Supply Chain

The potential benefits longer term can be even greater than those set out above but recognize the transformation journey is unlikely to be realized quickly because of the factors set out in later sections in this paper. Indeed, current industry performance does not represent an optimized batch model - far from it, and some observers will question whether an industry that is unable to run what many would regard as simpler batch processes effectively, can deliver efficient continuous manufacturing processes. However, a key difference is that continuous processes impose disciplines that are optional in a batch model.