We have been amazed at the complexity of the journey of clinical supplies to investigator sites is and delays can be extremely costly for sponsors. In fact, supply chain logistics can account for up to 25% of total annual pharmaceutical R&D costs. With stakes this high, it’s important to understand the strengths and weaknesses of tools uses to plan such critical projects.

Spreadsheet Models

The most common tool used for planning and managing Clinical Supply Chains are spreadsheet models. While this approach has some attractive qualities (simplicity, familiarity, and ease of sharing results) they suffer from several drawbacks.

For basic or static analysis, spreadsheets are often the right tool. However, supply chains are not static. They have parameters, resource constraints, queues, and demand curves that will change over time. This is where spreadsheet models show their limitations in capturing real-world situations. Although they will provide an answer, if you only use averages as your inputs, you will always get the average result, and not the range of real-world possibilities.

Simulation

Due to the drawbacks of spreadsheet models, approximately 10% of pharma companies have tried to solve these shortcomings by utilizing more sophisticated, dynamic tools like simulation. However, while simulation solves many of the shortcomings of spreadsheet models, they also introduce new problems; namely complexity and lack of transparency. Developing realistic simulations of supply chains is not an easy task because the software is often proprietary and the learning curve for programming them can be extremely high unless you are a data scientist or software developer.

Due to these barriers, the management of clinical trial supply chains and approaches to minimize costs has received little attention in formal research [Chen, Mockus, Orcun, & Reklaitis, 2012]. Much of the research that has been done does not consider the inherent randomness found in real-world clinical supply chain systems. This makes much of the research less applicable to pharma companies because uncertainty and variability is inherent to patient enrollment, shipment lead times, and process delays [Chen et al., 2012].

Alternatives

Given the challenges with both spreadsheet models and simulation, what are clinical trial supply chain decision makers to do? There are two possible solutions:

1) Develop highly intricate and complex simulation models using Discrete Event Simulation (DES) software like ExtendSim (https://goo.gl/VFSHak) or Python SimPy in hopes of capturing all the aspects of a complex supply chain; or

2) Simply eliminate the need and cost of the supply chain by providing subjects with a clinical study pharmacy card.

Some pharma companies take the first route and try to model complex clinical trial supply chains in hopes of maximizing efficiency and expediting clinical trials. Although not the easiest route, companies that wish to pursue this path would benefit from developing an understanding of dynamic programming and simulation optimization strategies. A good starting point would be a paper by Jung, Blau, Pekny, Reklaitis, and Eversdyk (2004), covering a simulation optimization approach to solve a generalized supply chain problem under demand uncertainty.

Using the RxStudyCard Instead of trying to build a large-scale simulation and then run hundreds of scenarios in hopes of identifying efficiencies, decision makers can simply eliminate much of it using a clinical study pharmacy card which leverages a network of over 60,000 retail pharmacies to deliver medicine and supplies to subjects participating in clinical studies. This type of program provides a safe and efficient method of dispensing unblinded medicines and supplies to study subjects that saves administrative effort, wasted supply overage, and time as well as reduces the amount of capital expense required by 30 to 60%, depending on the size of the study.

Chen, Ye, Linas Mockus, Seza Orcun, and Gintaras V. Reklaitis. “Simulation-Optimization Approach to Clinical Trial Supply Chain Management with Demand Scenario Forecast.” Computers & Chemical Engineering 40, no. Supplement C (May 11, 2012): 82–96. https://doi.org/10.1016/j.compchemeng.2012.01.007.

Jung, June Young, Gary Blau, Joseph F. Pekny, Gintaras V. Reklaitis, and David Eversdyk. “A Simulation Based Optimization Approach to Supply Chain Management under Demand Uncertainty.” Computers & Chemical Engineering, Special Issue for Professor Arthur W. Westerberg, 28, no. 10 (September 15, 2004): 2087–2106. https://doi.org/10.1016/j.compchemeng.2004.06.006.

Gerald Klein, MD