(182e) A Superstructure of Pathways for Pharmaceutical R&D and Its Use in the Optimal Planning of R&D Activities

The process research and development (R&D) of new drugs constitutes a challenging decision-making problem for major pharmaceutical companies. Multiple interdependent activities need to be carried out in order to pursue appropriate development paths for each drug in the portfolio of assets under development. Whereas the development paths need to follow closely the various phases of clinical trials that drug products must undergo before they can be launched in the market, the various assets across the portfolio compete with each other for the time of scientists, laboratory resources, manufacturing capacity, and budget allotments. Importantly, there exist multiple potential development paths for each drug asset’s Active Pharmaceutical Ingredient (API) as well as overall Drug Product development. Given all possibilities, the activity scheduling decisions need to be coupled with path selection decisions so as to achieve resource efficiency while respecting certain various deadlines for development milestones.

The complex nature of the drug product development process makes the portfolio-wide R&D activity planning very challenging. Researchers in the Process System Engineering community have long been interested in this problem (see [1,2] for comprehensive reviews of applying optimization approaches in this context). Past efforts have proposed relevant ontologies [3,4,5] as well as optimization frameworks [6,7,8,9,10] that can generate efficient scheduling decisions; however, the important decisions to select pathways of development have not received much attention. In this work, we construct a superstructure of the drug development process that captures all possible development paths, including technology switching between phases of development, chemical route re-identification as well as possible acceleration of development phases. Based on this superstructure representation, a new mixed integer linear optimization model is formulated to solve the portfolio-wide activity planning and resource allocation problem for pharmaceutical R&D. The model is tested on portfolio-wide instances using data inspired from real-life operations at a major pharmaceutical company. This study reveals interesting trade-offs between characteristics of the problem instances and their tractability. We also show how the latter can be further improved by applying a number of model reformulations and practical heuristic techniques to expedite solution times. Finally, we discuss our experience with deploying a decision support tool based on our optimization approach for the systematic and largely automated derivation of development paths and activity schedules in the real-life setting.

References:

[1] Laínez, José Miguel, E. Schaefer, and Gintaras V. Reklaitis. "Challenges and opportunities in enterprise-wide optimization in the pharmaceutical industry." Computers & chemical engineering 47 (2012): 19-28.

[2] Marques, Catarina M., Samuel Moniz, Jorge Pinho de Sousa, Ana Paula Barbosa-Povoa, and Gintaras Reklaitis. "Decision-support challenges in the chemical-pharmaceutical industry: Findings and future research directions." Computers & Chemical Engineering 134 (2020): 106672.

[3] Venkatasubramanian, Venkat, Chunhua Zhao, Girish Joglekar, Ankur Jain, Leaelaf Hailemariam, Pradeep Suresh, Pavankumar Akkisetty, Ken Morris, and Gintaras V. Reklaitis. "Ontological informatics infrastructure for pharmaceutical product development and manufacturing." Computers & chemical engineering 30, no. 10-12 (2006): 1482-1496.

[4] Moniz, Samuel, Ana Paula Barbosa-Póvoa, Jorge Pinho de Sousa, and Pedro Duarte. "Solution methodology for scheduling problems in batch plants." Industrial & engineering chemistry research 53, no. 49 (2014): 19265-19281.

[5] Viswanath, Shekhar, Steve Guntz, Jon Dieringer, Shankarraman Vaidyaraman, Hua Wang, and Chrysanthos Gounaris. "An Ontology to Describe Small Molecule Pharmaceutical Product Development and Methodology for Optimal Activity Scheduling." Journal of Pharmaceutical Innovation (2020): 1-15.

[6] Maravelias, Christos T., and Ignacio E. Grossmann. "Simultaneous planning for new product development and batch manufacturing facilities." Industrial & engineering chemistry research 40, no. 26 (2001): 6147-6164.

[7] Varma, Vishal A., Joseph F. Pekny, Gary E. Blau, and Gintaras V. Reklaitis. "A framework for addressing stochastic and combinatorial aspects of scheduling and resource allocation in pharmaceutical R&D pipelines." Computers & Chemical Engineering 32, no. 4-5 (2008): 1000-1015.

[8] Colvin, Matthew, and Christos T. Maravelias. "Modeling methods and a branch and cut algorithm for pharmaceutical clinical trial planning using stochastic programming." European Journal of Operational Research 203, no. 1 (2010): 205-215.

[9] Colvin, Matthew, and Christos T. Maravelias. "R&D pipeline management: Task interdependencies and risk management." European Journal of Operational Research 215, no. 3 (2011): 616-628.

[10] Wang, Hua, Jon Dieringer, Steve Guntz, Shankarraman Vaidyaraman, Shekhar Viswanath, Nikolaos H. Lappas, Sal Garcia-Munoz, and Chrysanthos E. Gounaris. "Portfolio-wide optimization of pharmaceutical R&D activities using mathematical programming." INFORMS Journal on Applied Analytics (2021). Accepted for Publication.