Fueling Advanced Analytics with Connected Data: 3 Use Cases from Top Data Leaders

Biopharma companies are using data in new and innovative ways — especially in combination with AI — to add value to their organizations. Bristol Myers Squibb, Lundbeck, and Pfizer recently shared how they are applying Veeva Link Key People data, in both medical and commercial use cases, to help them unlock new insights, improve engagement, and increase efficiencies.

Bristol Myers Squibb: Customer engagement orchestration

At Bristol Myers Squibb (BMS), the pursuit of innovative customer engagement strategies is a top priority. Anuj Bajracharya, director of international markets predictive customer engagement, is at the forefront of these efforts, driving the integration of both internal and external data sources to enhance the quality and effectiveness of customer interactions. “We develop machine learning models to identify correlations across these diverse datasets and the impact, which allows us to generate actionable recommendations for engaging with our customers,” Bajracharya explains. By leveraging these insights, the commercial team is able to prepare more thoroughly for customer meetings, ensuring that every interaction is informed and impactful.

To maximize the value of these data-driven insights, BMS delivers information to representatives and other customer-facing teams on both a monthly and weekly basis. This regular cadence ensures that teams are equipped with the latest recommendations and can respond promptly to emerging opportunities. Bajracharya emphasizes that timely access to relevant data is essential for maintaining a competitive edge in customer engagement.

A pivotal factor in the success of this initiative has been the adoption of Link Key People, chosen for its seamless integration with BMS’s internal data. All data flows through the company’s CRM system and is consolidated within internal data warehouses, creating a unified and comprehensive view of customer activities. “This integration enables us to understand what our customers are doing beyond our internal records,” Bajracharya notes. “For instance, we can track developments in scientific publications and clinical trials, which provides valuable context for our engagement strategies.”

The team aggregates these data sources for a variety of machine learning analyses, allowing for the development of sophisticated statistical models and deeper insights into customer behavior. The ease of integration also means that successful use cases can be replicated and scaled across different markets, supporting BMS’s global commercial objectives.

Having AI-ready data has been a critical enabler for these initiatives. With structured and aggregated data, BMS can immediately feed information into advanced data science models, accelerating the pace of insight generation. “Our primary objective is to utilize these data sources in a more structured and aggregated manner across all our markets.”

Looking ahead, Bajracharya and his team are actively exploring additional applications for Link Key People data. “We are beginning to think more creatively about how to leverage information from publications and clinical trials,” he shares. “By applying cutting-edge methodologies, we can map relationships between authors, physicians, and our customers, unlocking insights that drive commercial activity and inform strategic decision-making.”

Historically, collecting and analyzing feedback regarding interactions with customers has posed significant challenges. Today, with the ability to process and analyze large volumes of unstructured data — including text and voice — the opportunities for advanced analytics are expanding rapidly. Bajracharya observes that the rapid evolution of technologies such as generative AI is further accelerating the pace at which insights can be generated and applied.

“These advancements will enable us to better understand our customers’ perspectives and needs, and to address those needs more effectively,” he says. “Leveraging cutting-edge technologies to facilitate more personalized interactions will help us reach more patients, faster.” Bajracharya also highlights the potential for these innovations to optimize resource allocation and support more effective investment decisions, ensuring that BMS continues to deliver value to both customers and patients.

“I am personally very excited about the future of our industry and the evolution of analytics,” Bajracharya concludes. “As we continue to harness the power of data and technology, we are well-positioned to transform customer engagement and make a meaningful impact on patient outcomes worldwide.”

Lundbeck: Expert tiering and territory design

What started at Lundbeck as defining and creating tier criteria to find KOLs, led to the development of a dynamic tool that not only identifies KOLs but also creates a balance between scientific and clinical experts. “We use this information to develop our MSL territories and appropriately size the team,” says Brendan Whooley, associate director, US medical affairs excellence at Lundbeck. “We also identify KOLs with the highest level of impact and visualize their collaboration networks.”

To achieve this use case, Whooley and his team integrated Link Key People and Veeva Compass Patient data. “Using these two data sets helps us understand the KOL landscape more broadly for our therapeutic area. We identify the individuals who have influence, not just at the scientific expert level, but also at the clinical expert level,” he says.

Easily integrating data sets was important to Whooley, especially for determining top experts and additional levels of expertise. “The blended data helped us find clinical experts who are part of a key scientific healthcare organization or in a healthcare organization that is associated with one of our top tier scientific experts,” he says. “We see top scientific experts there, but also see the individuals who are treating patients.”

This type of analysis continues to provide value to the team as business needs shift and they expand it into other therapeutic areas. Whooley’s team acknowledges that the structured and clean nature of the data, along with the versatile and configurable outputs they have received, make it ideal when thinking about its potential with AI. “These qualities certainly allow for this kind of data to be incorporated into future use cases within AI,” he says.

Pfizer: Horizon scanning

Brett South, lead, patient insights and modeling, PCRP, R&D, and his team launched horizon scanning at Pfizer. This use case applies bibliometric approaches to uncover insights from scientific literature, publications, and congresses — all data available via Link Key People. “We apply AI and machine learning technologies to extract information, build insights, and create more structured data,” he says.

Link Key People data enables different applications within horizon scanning. “It’s extremely versatile and useful for our indication recommendation system,” South says. He and his team have leveraged Link focus areas and the underlying annotated data that's contained in those focus areas. “This allows us to mix and match data across different applications,” says South. “For example, you can look at mono therapeutics, combinations across that, and specific molecular targets that are actually annotated in the data and repurposed.” The team has been able to modify and apply this practice to 12 different disease assets and molecular targets so far.

South highlights two important aspects of the data that make it machine ready. As part of their recommender system, they’ve developed scorecards. These include a market scorecard and a scientific scorecard that's based on confidence and rationale for scientific aspects. “In terms of AI readiness, what we see in the Veeva data is structure,” he says. “We need metrics [for these scorecards], so we count things like the number of publications, clinical trials, or different types of scientific activities,” he says. “Once we have structured data, we also need to access unstructured data that's often locked in publications, abstracts, clinical trials, and summaries.” Veeva then helps South’s team use LLMs to extract and repurpose the unstructured data. “We bring it back as structured input that we can then further refine and use for other aspects of that system,” South says.

"While there’s value in surfacing novel insights for certain assets, South encourages taking it a step further by utilizing multi-modal data sources and analytic capabilities. “Build a ranked list of diseases and potential assets, and identify what the night sky view looks like for those thousands of diseases,” he says. “From a methodology perspective, you can reapply this thinking to other types of activities like optimizing site selection.” For example, integrate Link and Compass patient data to maximize KOLs, identify clinicians who are experts in particular areas, and then force rank in the same way to make site recommendations. He notes an example of sickle cell disease site optimization. “We partnered closely with the Veeva team to optimize site recommendations and determine which ones to pursue.”

When it comes to patient insights and modeling within PCRP, South compares the underlying data to unrefined oil. “It needs to be further refined so it can become high octane gas,” he says. This means refining both structured and unstructured data and making it more usable for different types of analyses along with mitigating data harmonization challenges. “We see a big opportunity to integrate GenAI into data harmonization — resolving ambiguities across different types of concepts and adding contextual understanding,” he says.

As more companies like Bristol Myers Squibb, Lundbeck and Pfizer rely on deep data to fuel analytics and AI initiatives, fast access to quality data is a key enabler. New innovations such as the Link Direct Data API will allow high-speed access to this large dataset to continue to unlock new use cases, deepen analysis, and power innovations.