Navigating Uncertainties of Clinical Trials with Probabilistic Supply Chain Planning

A rising trend in the Supply Chain Planning industry is Probabilistic Supply Chain Planning. You may have already encountered numerous articles from leading supply chain software vendors and advisory firms like Lokad, SAP, and Kinaxis. In this article, we'll discuss the definition of Probabilistic Supply Chain Planning, its key differences from Deterministic Supply Chain Planning, and its application in the Clinical Trial Supply Chain.

Introduction

Traditionally Supply Chain Planning has been based on statistical and machine learning models created to come up with a deterministic forecast, where fixed values for forecast are received to approximate an uncertain variable - demand. Obviously the forecast will never be accurate and the companies utilized Safety Stock to compensate for the intrinsic variability in demand and lead times. However, in recent years, probabilistic planning has gained prominence due to its ability to handle uncertainty more effectively.

In the case of Clinical Trial Supply Chain Planning, Inaccuracy in Demand and Supply Planning is very high, primarily due to intrinsic uncertainties in the inputs - patient recruitment, retention probabilities, adaptive nature of clinical trail with uncertainties on dosage and titration schedules, addition of new clinical sites. These factors contribute to high operational costs, significant clinical trial wastage, and excess secondary kit inventory lying at depots and sites

Probabilistic supply chain planning offers a solution for mitigating these uncertainties in the clinical trial supply chain. By generating probabilistic forecasts, it allows clinical supply chain teams to create more realistic plans and make informed decisions.

Deterministic Supply Chain Planning: The Old Guard

Supply Chain Planning can be divided into 2 parts - Demand Planning and Supply Planning. While in Deterministic way of Demand Planning, We try to use statistical/ML methods to generate a fixed forecast number utilizing historical sales data and then adjust it in accordance with inputs from sales, marketing and finance to come up with an Constrained Demand Plan. Supply Plan is generated based Constrained Demand Plan by applying different optimization algorithms taking into account safety stock levels, lead times, lot sizes, cost; but ultimately generate a fixed value supply plan.

This works well where the demand and lead times are predictable; or there are less uncertainties. But What about Supply Chains where there is high variability, Obviously in that case either we are missing on customer orders or will be suffering from a high inventory levels.

Refer to below picture to understand this; Basically we are forecasting a sum of a number while rolling the 2 dices, obviously we can predict with a probability a the sum of 2 dices rolled but not with certainty and ideal way to demonstrate the forecast is the Binomial distribution as shown below:

When to use Deterministic Planning:

• Stable Conditions: When demand and supply patterns are relatively predictable.
• Frozen Time Periods: When operational decisions need to be made within fixed time frames (e.g., monthly production schedules).

Probabilistic Planning: Embracing Uncertainty

Probabilistic Planning adopts a different approach instead of computing a fixed value, it computes a range of values depicting multiple scenarios, each with an associated probability.

• Probabilistic Forecasting: Instead of forecasting a series of fixed values over a time period, a range of values is forecasted with confidence intervals.
• Probabilistic Inventory Planning: Probabilistic Inventory Planning takes into account probabilistic forecast generated, Variability primarily in lead times and other variances to generate an inventory plan.
• Probabilistic Supply Planning: Probabilistic Supply Plan takes Probabilistic Forecast & Inventory Plan as Input and takes into account variabilities in Sourcing/production/Transportation lead times, constraints for optimization to generate a feasible multi scenario supply plan.

Probabilistic planning prepares supply chain teams to perform their assessment and to reach at a conclusion for the frozen time period where fixed plan is required for operational execution

Application in Clinical Trial Supply Chain Planning

Clinical Trial Supply Chain Planning has inherent uncertainties and variability in the Inputs like Patient Recruitment, Dosage Schedule, Titration Schedules which makes it an ideal candidate for application of probabilistic supply chain planning.

Trends like Adaptive Clinical Trial where depending on the outcomes parameters like number of patients, dosage, titration, duration and even number of sites can change, this allows for more efficient and effective testing of new treatments, as the study can be adjusted in response to new information.

But from Supply Chain Perspective, It's a new challenge to be tackled upon. These are all the reasons

Let's see how Probabilistic Planning can be utilized for an effective supply chain planning:

1. Develop an Analytical Model for Probabilistic Demand Forecasting: An analytical model can be developed in multiple layers taking into account specific variabilities associated with each input like Patient Enrollment, Retention Schedules, Dosage Schedule, Titration Schedule to generate a Demand forecast for Secondary Kits.
2. Simulations for Multiple Scenarios: The model can then be used to run simulations that explore various possibilities. These simulations can consider different enrollment rates, patient drop-out patterns, Dosage schedules, titration, adaptive nature of clinical trials. These simulations will help Supply Chain teams to assess the scenarios they may face, and be better equip in advance from Vendor, Technology, Transportation and Infrastructure perspective.
3. Probabilistic Inventory and Supply Planning: Taking Probabilistic Demand Plan generated as Input and combining with lead time variabilities - sourcing, manufacturing and transportation, constraints like shelf life, lot sizes, capacity etc. to generate a probabilistic Inventory and supply plan.

What could be the Benefits of Probabilistic Planning:

• Reduced Clinical Trial Wastages: Since Probabilistic Supply Chain Planning gives enough time to Supply Chain teams to be ready for variability that they would be facing in medium and long term, they would be able to prepare well and be precise during short term planning which will ultimately reduce the cases of huge Clinical trial wastages that happen as a result of keeping high inventories at site. It should be noted that we are not even discussing the case of stockouts as that is out of question and buffer inventory is traditionally kept high enough to avoid any such situation which risks the clinical trial
• Reduced Cost of Clinical Trial Operations: By optimizing inventory levels, probabilistic planning can significantly reduce storage and disposal costs associated with excess drugs. It will help Pharmaceutical companies invest the additional capital in the core area of R&D which further will not only results in reduced cost but a better trials.
• Improved Trial Efficiency: With Supply Chain Operational readiness, Trials will be more efficient, leading to faster completion times and potentially earlier market access for new drugs.
• Data-Driven Decision Making: Probabilistic planning provides a data-driven approach to decision-making, allowing planners to make informed choices based on real-world uncertainty and also prepare better.

Conclusion

Deterministic supply chain planning is not suitable to tackle highly uncertain environment of clinical trials as has been proven with the fact that Pharmaceutical Industry is struggling with high costs of running clinical trials, huge clinical wastages and high environmental impact. Thus Probabilistic Planning has shown a promise a capable and robust alternative to manage and navigate around the uncertainties of Clinical Trials Supply Chain. By adopting a probabilistic approach, clinical trial supply chain teams can achieve significant improvements in operational preparedness. These advancements translate into tangible benefits, including reduced clinical trial waste, minimized operational costs, enhanced trial efficiency, and ultimately, the expedited development of life-saving treatments. As the pharmaceutical industry strives for continuous optimization and innovation, Probabilistic Planning can prove to be a strategic and a logical choice to move towards.

Readers are requested to provide their valuable feedback and incase you want to connect, discuss more on the topic, Please do reach out to me on LinkedIn and we can discuss this further.

#SupplyChainPlanning #ClinicalTrials #Pharmaceuticals #Healthcare #Logistics #InventoryManagement #Innovation #DataDriven #ProbabilisticPlanning #DemandForecasting #RiskManagement #UncertaintyManagement #ScenarioPlanning #Simulation