Supply Chain

The Art of Supply and Demand Planning

Supply Chain

11:39, January 18 2018


Barrett Glasser, Biogen, explores how supply and demand planning can be optimized to improve clinical supply chain processes

As clinical studies and demand signals become more complex, and supply networks and operations more diverse, the need for supply chain organizations to ‘do more with less’ continues to be an emphasis. Those groups which set themselves up for success in this dynamic landscape are rewarded with competitive advantage and better outcomes for patients. The following article outlines approaches to the necessary demand and supply planning, modeling, and execution required to achieve and sustain high-level clinical supply chain performance.

Aligning with Customers

At its core, clinical supply chain management involves the art and science of matching supply to demand. However, given the dynamic and ambiguous nature of a company’s clinical development landscape at any particular time, planning investigational medicinal product (IMP) manufacturing can seem like building a car as you drive it! Long lead time activities, be they procurement of rare raw materials for API and DS production, or gathering of the necessary stability data to support global clinical labeling, can come into direct conflict with challenging goals and the need to serve patients as quickly as possible. Little, if any, room is left for errors of misunderstanding or misinterpretation. This makes it exceedingly important to have defined and formalized touch-points between clinical operations and clinical supply chain organizations. Here are a few tips for consideration:

A). Numerous companies conduct clinically-focused meetings analogous to commercial Sales and Operating Plan (S&OP) meetings. These are often called Clinical Demand and Operating Meetings (CD&OP), where teams document best program-level assumptions and established needs looking out 18-24 months into the future. Take the time, at least quarterly, to not only discuss anticipated material needs, but also to collaborate around decision points and to endorse current plans.

B). At a more tactical, usually study-specific level, documented agreements for IMP can help get operations off on the right foot. Alignment around kit configurations and labeling intentions are obvious, but it’s also important to define other elements, such as supply levels intended for site initiations, interactive response system (IxRS) threshold expectations, the amount of IMP to be dispensed to patients, etc. While some of this is established in the protocol, certain supply plans may need to be established before the protocol is finalized. Work off a protocol concept, if possible, and strive to get this agreement in place at least sixmonths before study start.

C). Supply chain organizations only have one supply plan, widely communicated across their network, at a time. In order to have confidence in this plan, it’s important to know what scenario is the correct one to plan for – multiple (equally likely) scenarios is not an option!  Always align on one enrollment and utilization forecast as the best guess, for planning purposes.

Aligning on Business Principles

Supply chain risk is a fact of life, but it’s also an emerging discipline. In order to mitigate the risk of supply disruptions, based on elements, such as unexpectedly high demand, onsite temperature excursions, and damaged shipments, most companies will plan to build an inventory buffer of some sort.  The tradeoff, of course, is cost. Some advice:

A). Agree, across your organization, on how to set inventory buffers. Do you use a percent overage that is dictated by certain parameters? Do you plan for a certain forward coverage, in terms of weeks or months on hand? Do you do both, as some organizations do? Or are you able to take a more scientific approach, based onthe standard deviation of your ongoing demand? Regardless of your decisions, going through the exercise of standardization will ensure you have (and document) sound principles underlying your approach.

B). Sometimes external events impacting supply continuity will be unforeseen and beyond anyone’s control (think floods, volcanoes and the like). Having a ready-made playbook for when natural disasters or other external events occur can help minimize reaction time and risk to patients (search for an excellent online article on ‘Mitigating Supply Chain Risk’ by Chad Presher).

Also, in addition to determining business principles related to the management of risk, clinical supply chain organizations need to decide what standards to hold themselves to. Understanding expected performance is fundamental to understanding when things are veering off track.Once these expectations are defined, documented, and widely understood across the organization, active measurement of key performance indicators becomes not only possible, but a necessary step toward maturing as a clinical supply chain organization. After all, as W. Edwards Deming once said, “without data you are just another person with an opinion.”

Modeling and Visualization – the Essence of Supply Chain Planning

As drug supply professionals manage increasingly dynamic, complex programs and trials, accurate modeling and visualization of potential issues becomes a critically important element in keeping ‘the trains on the tracks.’ Most companies develop two types of mechanisms to help with this. The most widespread way is to develop models (or implement systems) that provide dashboards, reports, or other visual means of very clearly highlighting potential issues. This is the ‘pull information’ approach.  Alternatively, ‘pushes’ of information can be developed, in order to notify and alert supply managers that an issue requires attention.These can be as simple as delayed shipment notifications, or as complex as push updates when enrollment goes beyond certain parameters.A wise supply chain organization seeks to enable both types of mechanisms. Some general advice:

A). MS Excel can be beautiful in its simplicity as well as its complexity.Nothing can match its flexibility to perform quantitative modeling. As well, Visual Basic for Applications (VBA) with Excel opens up immense possibilities for modeling and visualization (but requires some specialized skill within your organization).All of this said, Excel has major drawbacks for planning if overused:

i). Lack of standardization – Standardization between spreadsheets (models) of different authors is difficult to enforce. With any attrition, the transition of protocol responsibility to another plannercan become difficult with any medium to large sized study.

ii). Prone to errors – It’s extremely easy to get one formula wrong, with disastrous results.  Or to lose a file, or accidentally overwrite a key model input.Accuracy is critical when the welfare of patients is at stake.

iii). Excel is not good at processing non-standard inputs – One example of this would be inventory lots of various expiries. The difficulty of dealing with this using standard types of formulas can cause oversimplification in models, leading to divergence from what can be expected to happen in the real world.

Because of these drawbacks, most mature clinical organizations eventually seek to employ software specifically tailored to planning clinical demand versus supply. This software can help manage the particular complexities of our industry, be it expiry, import licensing, inventory pooling, etc.

B). Perhaps it goes without saying, but ensure your systems and models describe ‘real life’ to the extent possible. Whether it’s demand-side or supply-side information, ensure those inputs remain as accurate as possible. Agree on cadence for updating and verifying information.  Ideally, some inputs can be automatically fed into your systems and models to avoid errors in data entry.

C). Simulation and optimization have begun to find their place in the world of IMP. These types of systems can become very powerful if they can be prescriptive; telling you what you should be doing – say, in terms of sub-depot ordering, or setting of certain IRT thresholds to minimize waste over time. Once again, it’s critical to believe in the forecast being simulated, in order to believe in the direction the system provides you with.

D). “Fire drill” your models. What happens if you lose a batch of API, or drug product? What happens if those final patients who were supposed to transition onward now need to stay on their current treatment for an additional six months? Brainstorming these types of scenarios needs to be a natural activity for a prescient supply planner.

The Promise of Actuals-Based Forecasting

Once solid planning systems and models are in place, efficiency can be driven by improving the way information is updated. “Actuals-based forecasting” is the name de jour for simplified processing of data coming in from your IxRS systems. Standardizing an output specification, for this information to automatically be sent from various IxRS platforms, is usually the first step toward cross-portfolio, “big data” style analysis and the possibilities that open up for your group. After that, getting a pulse on enrollment at sites and inventory levels at depots are low hanging fruit. Here are some additional ideas, but the possibilities for innovation are endless:

  • How much inventory is actually out at sites, which will need to be replaced at various points in time with fresher supply? How does this match up with supply at depots?
  • Which sites are getting more shipments per month than you would like? Are there ways to minimize these?
  • What is the expected lead time for sub-depot shipments, and how are actual shipment times comparing?
  • Which sites take longer than they should to receive inventory in IxRS?
  • Which studies are less ‘healthy’ in terms of inventory coverage than others? Which ones are carrying too much inventory?

Execution and Delivery in Supply Chain

In the Plan, Source, Make, and Deliver landscape, it’s of primary importance that all functions are marching to the beat of the same drum. If plans for manufacturing and distribution are the crown jewels of any supply chain organization, the system they are housed in, for all to see and work toward, should be considered the vault. Ensure this information is backed up regularly and fully supported by IT. Make it easy to understand readiness for each upcoming manufacturing activity, and meet regularly to ensure alignment.Understand time horizons for locking down plans (usually referred to as the “frozen” period), but also fight for flexibility to deal with the tricky situations that will invariably come up.Actively measure bottleneck readiness tasks, since they are on the critical path.

Where delivery is concerned, import license (IL) tracking continues to be a sore spot for the industry.  Countries provide companies with different types of licenses, some based on expiry, some quantity based, and some even constraining shelf life allowed or formatting of inventory coding (e.g. must align with what’s on application). Importers of record can be either CRO or CMO associated, which means that neither is likely to go above and beyond in helping with IL tracking responsibilities (unless IMP planning is subcontracted out to them). Therefore, hold your own organization accountable for planning around import license intricacies. Keep requirements visible – and check in with partners often. Know when they will need to reapply for a license based on anticipated lead times, and hold them responsible for doing so. This is critical to your success.

A great number of dynamic pitfalls and opportunities present themselves to any clinical drug supply organization. Identifying and managing these is the great challenge of our work, as we seek to serve our patients and impact their lives in meaningful ways.


Barrett Glasser

Associate Director, Clinical Drug Supply



Post a comment

Comments may be moderated for spam, obscenities or defamation.