From enabling the discovery of life-changing new drugs and therapies, to ensuring the consistent supply of essential medication to diverse populations around the world, data truly is the lifeblood of the life sciences industry. And today, it’s both more valuable, and more essential than ever.
The Life Sciences industry is on the cusp of a revolution. Organizations that have historically focused primarily on drug discovery, research and development, or drug manufacturing are actively expanding their capabilities across the life sciences value chain, in a push to become better integrated, and adopt new business models.
The life sciences leaders of tomorrow will be masters of their entire value chain. From the earliest stages of drug discovery, right through to the creation and execution of innovative commercial strategies, they’ll harness data to accelerate and amplify value creation, driving every stage of the process themselves.
To help them get into that position, some are expanding through acquisition — merging with other organizations to build broader enterprises with operations that span the entire value chain. Others are pursuing organic growth, establishing new business units to help them survive and thrive in this new industry paradigm.
However organizations choose to gain the necessary capabilities, there’s two things they’ll all need — flexible, adaptive data architectures combined with new operating models that will enable them to operate at the scale and speed of their aspirations. That’s where data mesh comes in.
Today’s top data demand: Interoperability
Right now, the biggest barrier standing between life sciences organizations and their vision of deeply connected, intelligent operations that span the value chain is data interoperability. Interoperability is the ultimate goal and spans both domains within the organization and external partners including contract research and manufacturing organizations, and syndicated data providers.
A lot of the data used by teams across life sciences organizations is relevant and valuable to other internal and external teams too. For example, data generated during research and development directly feeds into clinical trials, and contains vital insights for the commercial teams that will market and sell the drugs.
A key example of the need for interoperability is the new set of standards coming out from the FDA and EMA around the Identification of Medicinal Products (IDMP). These new regulatory standards involve the unique identification of medicinal products in the context of pharmacovigilance and the safety of medications. This includes data from across the value chain and ultimately requires companies to be able to identify and describe medicinal products with consistent documentation and terminologies. With over 1500 attributes required for compliance (most of which are contained in unstructured data sets), this presents a major opportunity for pharmaceutical and biotech companies to embark on their journey towards interoperability.
But right now, interoperability is certainly not the standard across the industry. Often, each of the teams across the life sciences value chain has their own systems and storage mechanisms, which leads to data either getting permanently trapped in domain-specific silos or ends up stored in formats that aren’t interoperable with the systems used by other teams. Organizations understand the value that’s trapped within that data, but without the right architecture and operating model, they can’t unlock it. In order to achieve interoperability, there is a trade off. Many parts of the organization (e.g., R&D, Clinical) may lose some autonomy, but the entire organization gains interoperability.
That’s where the domain-oriented design of the data mesh becomes extremely valuable for life sciences organizations. Within a data mesh, each domain creates and curates its own data products that meet its unique needs. But crucially, everyone else across the organization is also free to access those data products, and repurpose them to meet their own unique requirements.
Data mesh, especially the principle of federated computation governance, enables the seamless interoperability of all kinds of data sets. Each data product is owned by the domain team closest to it, which puts them in total control of how it’s created, while simultaneously giving every other domain access to it. It enables interoperability, without demanding compromise from any specific domain.
Data mesh in action: Early drug discovery
By creating an interoperable data foundation, data mesh can remove many of the barriers standing between life sciences organizations and their vision for hyper-connected intelligent operations. But ultimately, what they do with that foundation remains in their hands.
One area that many organizations are focusing on is early drug discovery. Discovering new drugs and therapies and bringing them to market quickly has always been a key success factor for pharmaceutical and biotechnology companies. However, the drug discovery landscape has recently started to transform significantly.
Early drug discovery now goes beyond the basic identification of molecules. It’s expanded to include chemical modalities such as RNA therapeutics, cyclopeptides, antibody drug conjugates, and gene therapy. For pharmaceutical and biotechnology companies, the practical upshot of that is that there’s a lot more data to crawl through, and less time to do it in.
The shift is creating a new data-driven early drug discovery paradigm — one where the capabilities companies develop to accelerate discovery are the blockbusters, rather than the drugs and therapies themselves. The organization that can build the strongest discovery capabilities through proprietary systems will write the rules of the new playing field.
Those proprietary systems typically use capabilities such as AI, deep learning, neural networks and other advanced analytics techniques to process, understand, and extract insights from huge, diverse discovery data sets. For example, data sets coming from partnerships with academic and/or contract research organizations are prodigious. They can be so large and diverse that it would take substantial man hours to begin to understand, analyze, and apply it all. The ability to quickly ingest this data and begin analyzing it is key.
The data mesh supports those capabilities in two main ways:
The interoperable, domain-orientated data foundation created by the data mesh makes it easy to access diverse data sets from across and outside of the organization, and create data products that automatically pull data from sources elsewhere in the organization and externally to augment and support discovery processes.
Data mesh enables the creation of bespoke data products, designed to make tasks like training AI models or building clear pipelines to feed deep learning models easy. Data mesh facilitates faster time to insights and actions by providing clean, ready to consume data which reduces (or nearly eliminates) the time spent identifying issues with the data and cleansing it. Having a Product Manager that monitors and extends these data products ensures that they remain production-grade throughout their lifecycle.
By making data easier to access and translate into vital discovery insights at speed, data mesh architectures are enabling organizations to build leading discovery capabilities that will help them create value and bring new drugs to market quickly, for years to come.
Data mesh in action: Commercial transformation
Personalized medicine and data-driven delivery have transformed the commercial landscape across the life sciences industry. Today, organizations are tasked with managing deep individual relationships with Healthcare Providers (HCPs) and Healthcare Organizations (HCOs) — using data to track, understand, and meet their rapidly-changing needs.
For commercial teams, that means mastering a hybrid personalization approach for HCP engagement across in-person, virtual, and non-personal channels. Practically, they need to build the capabilities to translate diverse insights from HCPs and HCOs into new go-to-market strategies at speed.
Within a data mesh, commercial teams can build their own data products to help them do exactly that. They can define diverse sources of HCP, HCO, and syndicated data, and bring them together within data products that can help them quickly understand what that data is telling them, and convert it into value-generating commercial decisions.
Crucially, the data mesh approach also enables other teams to act on those insights. For example, whatever a diagnostics commercial team learns, the pharmaceutical commercial team can act on too, at a similar pace. Plus, upstream teams can benefit from practitioner-level insights into drug performance and uptake, using those insights to make better manufacturing, research, and discovery decisions.
Data mesh isn’t a ‘cure all’, but it’s a well-targeted solution to the symptoms life sciences companies are facing today
As the life sciences industry evolves, every domain — from early drug discovery and research, to supply chain and commercial teams — needs access to complete, relevant data, and the power to operationalize it at speed.
That’s what makes data mesh such a clear fit for life sciences organizations. It’s an architectural and operational paradigm that empowers diverse domains to take charge of their data and apply it to solve their domain-specific challenges, while simultaneously improving the accessibility and interoperability of data across the enterprise.
Early adopters are already seeing significant success with data mesh, and blazing a trail that will help them succeed in their rapidly-transforming industry. To learn more about the approach, and discover how it could help you overcome the diverse data challenges faced in across the life sciences industry today, visit our data mesh hub.