Digital Twins Tentatively Finding Traction in the Pharmaceutical Industry

While asset and process twins have seen significant interest and adoption in the manufacturing space, the pharmaceutical industry is one where uptake has been much slower. The technologies are, however, increasingly seeing development by digital twin technology suppliers; as accuracy and capabilities mature, digital twins will drive reams of new applications in the pharmaceutical space.

Early research projects for digital twins in the pharmaceutical industry have looked at creating models of patients’ bodies. Certara’s Virtual Twin solution (2018) models individual patients based on physical and genetic attributes to help predict the impact of certain drugs and dosages on a case-by-case basis. Dassault’s Living Heart (2014) and Living Brain (2019) projects create digital models of the named organs for testing and studying drugs or medical interventions. Siemens Healthineers, Hewlett Packard, GE, and Philips are among other companies working on digital models of organs for similar purposes.

These early research-stage applications have more recently been complemented by an interest in digital twins at the drug manufacturing level.  In 2019, Infosys worked with a European pharmaceutical to predict and monitor the cell culture process for vaccine development. In February 2020, Honeywell and Bigfinite partnered to enhance pharmaceutical companies’ manufacturing process operations through respective automation and analytics specialisms. In May 2020 Siemens and Atos went a step further with an expansion of their long-standing partnership (since 2011) by bringing digital simulations and predictive models to the pharmaceutical manufacturing process.

A 2019 study by the Tufts Center for the Study of Drug Development (Tufts CSDD) revealed a number of challenges facing the pharmaceutical industry. The study estimated the cost of a new successful prescription drug development to be nearly US$2.6 billion—up from US$802 million in 2003—and the success rate of these drugs through clinical trials to be only 12%—down from 21.5% in 2003. High Research and Development (R&D) costs are often cited by pharmaceutical companies as the principal reason behind high drug prices, supported to a great extent by a complex regulatory landscape.

Siemens and Atos’ partnership is the latest development in established automation and digital twin suppliers creating solutions specific to the pharmaceutical industry. Their Process Digital Twin aims to address the above challenges by creating a digital replica of the drug manufacturing process—enabling simulation of outcomes with reduced physical experimentation time and material, reducing the costs and Time to Market of drugs, and enabling replicability of the process at scale. While the pharmaceutical and healthcare industries were early adopters of digitalization using sensors, this has largely remained confined to equipment and patient monitoring. The development of process twins, beyond its impact on R&D and product development and manufacture, will drive an acceleration of the shift toward product design and patient modelling, enabling low-cost “personalized” medicine. The partnership between engineering giant Siemens and digital transformation giant Atos puts them at the forefront of such innovations.

Leading digital twin suppliers for the manufacturing industry include Honeywell (Force), Rockwell Automation, PTC, Siemens, GE, Bentley, Rolls Royce, and Dassault Systèmes, with several of these in the early stages of bringing to market solutions specifically for the pharmaceutical industry. As more of these companies expand their focus to the pharmaceutical space, competition and real-life applications—as opposed to research projects, where the market is currently focused—will begin to transform the industry. With pharmaceuticals applying these solutions, and with the expansion of the quantity of personalized medical data—driven by increasingly intelligent wearables—a “federated” ecosystem of digital models will enable faster and safer clinical trials and treatments with greater flexibility and direct interplay between manufacturing firms and end patients.

From the industry’s perspective, there are challenges to overcome before widespread adoption is seen. Firstly, pharmaceutical processes are exponentially more difficult to model and simulate than mechanical and physical models, where digital twins are currently used. While open-source software may be sufficient to build a predictive model of an engine’s or a machine’s lifecycle—in part due to the technology’s relative maturity in these applications—simulating and predicting changes brought about by modifications in the pharmaceutical manufacturing process requires a more individualized and specialized approach to ensure that digital simulations are an accurate reflection of the physical process. Secondly, and on a related note, the complexity of the modelling process and the consequent limitations of accuracy are a hard sell where existing processes have been refined over time to address the highly regulated landscape, with extensive data and model training necessary before twins reach an acceptable degree of accuracy. Thirdly, while simulating discrete events is currently the principal focus of digital twin technology, scaling this to a lengthy process chain presents a number of additional challenges—already an advanced use of the Internet of Things (IoT), digital twins will see slower uptake in industries where more complex hardware, software, and integration requirements are necessary.

Overcoming these challenges requires partnerships offering expertise in different fields and combining players across the value-chain—including Original Equipment Manufacturers (OEMs), sensor device makers, and software and cloud service providers offering joint Go-to-Market (GTM) solutions for pharmaceutical companies. Ecosystem development through partnerships and acquisitions has already been seen and leveraged in the manufacturing industry, with PTC and Rockwell examples of companies vigorously vying for position, helping to drive competition and democratization of solutions in this space as discussed in the ABI Insight A US$1 Billion Industrial IoT Partnership: Who Wins? (IN-5175). As sensor technical capabilities and advanced data analytics systems mature, the pharmaceutical industry becomes another high-value addressable market. To gain a foothold in the market, digital twin solution providers must prioritize partnerships and the value of a first-move advantage; this will not only help pharmaceutical companies gain a competitive edge in their existing drug development and manufacturing domain but also open up new downstream opportunities for them all the way to the point-of-care market.