Life science companies' supply chains extend from drug research to production and customer acquisition

No industry is more regulated and resource intensive than pharmaceuticals. The opportunity for optimization are presented the most at the R&D stage where tremendous resources are required to evaluate many molecules and therapies via myriad studies and tests. Companies try to differentiate by smartly deploying their limited resources. Resource allocation is complicated by many uncertainties and by the long horizon of resource planning (5 – 10 years) required.

Operations research and machine learning can benefit pharmaceuticals in many ways.

Drug research and manufacturing decision support

Resource management in drug safety assessment
Clinical trial supply chain optimization
Drug portfolio optimization
Long range forecasting and planning
Optimal planning and scheduling of drug manufacturing
Process Analytical Technology (PAT) for manufacturing

Patient care decision support systems

Decision support for physicians
Analysis and predictive models using EMR data
Scheduling or Operating Rooms
Scheduling of nurses
Hospital capacity planning
Optimal routing of hospice care personnel

Demand forecasting and promotions optimization

Drug launch forecasting
Promotions optimization
Analysis of competing events in the marketplace
Disease progression modeling (Epidemiology)
Pricing optimization under regulationary constraints

The role of predictive and prescriptive analytics in pharma

Drug uptake and resource usage forecasting are key requirements for pharma. Drug demand forecasting can be leveraged for designing appropriate marketing (launching promotions, for example) and business development (licensing-in competing drugs/molecules, for example) activities. Modeling of disease progression (with or without intervention) can be achieved using rigorous modeling paradigms (like Markov chain approach). Such models can be used to study the evolution of markets and to justify R&D investments and for identification of new disease states to address, etc.

Typical approach to forecasting in pharma is to use Spreadsheet-based models. Such models can be bulky, slow, inaccurate, and hard to maintain as employees change roles or quit the company.

Merck & Co. Inc.

Optimal Solutions (OSI) has helped Merck build a strong analytical foundation and improved its capability in resource and operational planning by providing excellent services in business modeling. Compared to other consulting companies in its specialization, OSI has provided exceptional services, and delivered on their promises. I am impressed with their consultants’ dedication and their well-rounded experiences and skills ranging from general technology support, to business model development, and to advanced theoretical application. I am glad that Merck made the choice to go with Optimal Solutions, Inc.

– Nakin Sriobchoey, Manager RDM, Merck.

The role of machine learning in pharmaceuticals

Many tough problems in pharma can be addressed using machine learning. Raw material identification and drug counterfeit identification are prime examples. Pharma adopts FDA’s Process Analytical Technology (PAT) which hinges on monitoring manufacturing processes in real-time. Cutting-edge infrared sensors are available now which, in conjunction with machine learning models, can be used for PAT.

Research is underway in our group to prove these advanced concepts and initial results look very promising. We are collaborating with large life science companies, sensor makers, and academia in this DOE-funded project.

Interested in a deeper conversation?

We can expose you to what predictive and prescriptive analytics can do for your business. We can showcase our experience in machine learning and PAT so you can leverage our experience.