Jim Streeter, President of Technology, Chief Technology Officer at Envision Pharma Group
There’s no overstating the power of data. And when combined with AI capabilities, life sciences professionals are hungry to see what it can do for their launch plans as they march toward patient access.
As the President of Technology and Chief Technology Officer at Envision Pharma Group (Envision), I'm often asked about the intricacies of data management in the life sciences industry. Today, I'd like to delve into some key considerations we have been focusing on at Envision in our pursuit of interoperability in data systems.
One of the primary needs for medical affairs professionals and MedComms experts is using historical data to simplify their workflows and streamline their decision-making processes. This encompasses their prior engagements with various publications and medical documents, forming the foundation of their current workflow.
Furthermore, we've observed a growing interest in data sourced from generative AI and its potential to enhance their operations alongside predictive data, offering valuable insights for future endeavors. The convergence of generative AI, predictive data, and historical information empowers professionals to make well-informed decisions, significantly bolstering efficiency and effectiveness.
Humanizing Data: Ensuring Accessibility for All Stakeholders
Our extensive data repository spans over two decades. With millions of data points from 500,000 publications, it forms a robust foundation for our AI systems. We also maintain comprehensive metadata detailing the evolution of this data over the years, enabling our AI to provide precise guidance.
But this data virtually amounts to nothing more than static numbers if we don’t humanize it.
Humanizing data involves making it accessible and understandable for all stakeholders in the workflow that they use every day. It's about presenting data in a way that is intuitive and usable for the people interacting with it and not a separate tool they need to work with. Many companies simply take data and put it into an AI or analytics platform separate from their technology systems, adding more work and more complexity.
Envision focuses on integrating historical, predictive, and generative AI into a single platform, minimizing the need for life sciences professionals to switch between different systems, thereby streamlining their workflows and processes.
Mitigating Challenges of Quality, Privacy & Standardization
Addressing data quality and accuracy concerns is paramount, especially when dealing with extensive volumes of historical data.
We know that there will always be holes in data. Data sets are rarely complete and often have data in the wrong fields. To address these gaps, we use algorithms to clean the data and separate it from the rest of the high-quality, complete data we will use. We are cautious to build that into our system so that we aren’t using incomplete or inaccurate data sets, which can lead to hallucinations with AI and affect the downstream results. This diligence is increasingly vital as our customers explore the potential of new technologies and AI in expediting their route to market.
Historically, data standardization has been a challenge. Companies tend to collect data in ways that lack consistency over time. At Envision, we're actively working to standardize data, ensuring it can be effectively utilized across different systems. Master data management, including the creation of data standardization and cataloging, is instrumental in achieving this.
Privacy is a paramount concern for our customers. Many value the confidentiality of their data while still recognizing the potential benefits of shared insights. At Envision, we've devised a meticulous process of anonymizing and incorporating data into machine learning training. This guarantees that the data remains separate, secure, and confidential, never exposing it to unintended audiences. Our commitment to data privacy extends to respecting the decisions of those companies who choose not to share their data.
A More Collaborative Future
Industries are famous for creating separate, disconnected technologies independently, without looking at the bigger picture of how one technology can work with another or within an entire ecosystem.
For example, 10 years ago, analytics technologies were developed as stand-alone solutions, not incorporated into the systems and processes people used for their day-to-day jobs and workflows. The inefficiencies of disconnected technologies outweighed the applications. There are many parallels to this with AI today. CIOs and CTOs need to remember how hard it’s been with analytics and not make the same mistake again.
A major priority should be data standardization so that we can share data between systems. Someone has to stand up and ask what the standard for medical communications should be. Then, a consistent metadata framework will allow us to share between systems so we can simplify the process and day-to-day work for end users.
Next, we need to use open APIs. Our API-first strategy allows us to share data with other applications easily and simplifies workflow flows for end users. We cannot have siloed systems or data in this day and age of utilizing data for decision-making.
At Envision, we have the opportunity to lead the industry towards a more collaborative way of working. We can lead in standardizing methodologies, processes, and procedures that allow systems and ecosystems to be created in the life sciences industry, which is critical for medical communications and the entire drug development process.
In the clinical trial area, standardization is already happening with examples through CDISC with ADAM, CDASH, SDTM, PRM data standards. We need to see it more in medical communications. Our customers are asking for it so they can simplify their systems, processes, and workflow. It takes a leader to do it. I see Envision becoming that leader in helping the industry reach where it needs to be: a collaborative and AI-driven ecosystem.