For pharma teams, generative AI may ultimately be able to answer pressing questions like Is my medical strategy working or not? and What should I be doing differently? But current discussions about AI are happening on a more tactical level, including practical examples of how the technology can transform market behavior analysis and HCP insights. When it comes to getting the most from technology, should teams expect to use AI for real-time insights?
In a conversation presented by Reuters Events Pharma, moderator Paul Tunnah, Within3 CEO Lance Hill, and experts from AstraZeneca, GSK, and Novartis discussed the opportunities and challenges of AI – and the difference between real-time and right-time insights. Panelists in the discussion were:
- Carlos Eid, Executive Medical Director, Cardiovascular, Novartis
- Lance Hill, Founder & CEO, Within3
- Nathan Lear, Director of Commercial Data Science & AI, US Oncology, AstraZeneca
- Neeraj Goel Mittal, Global Head of Data, Analytics & AI, GSK
- Paul Tunnah, moderator on behalf of Reuters Events
The biggest challenge in AI for pharma insights: data quality and availability
“I truly believe the biggest challenge [in using AI for real-time insights] is data quality and data availability,” said Neeraj Goel Mittal, Global Head of Data, Analytics and AI for GSK. “We forget that insights are based on data, and the principles of data lineage in terms of data ingestion, data quality, data transformation, and data to action still apply.” These practicalities should temper unrealistic expectations about what AI in pharma can do and at what speed. “Sometimes, because of the hype, leadership expects the insights to be available [quickly],” said Mittal. “When the data on which those insights depend might not be available in the format we need it.”
“AI is not a strategy. It’s how you do things. The strategy is to effect some business result or change that you are dealing with depending on where you are within the organization and understanding the latest ways that companies are trying to solve that. – Lance Hill, Within3 CEO
AI: more about right time than real-time
“Instead of real-time, think about it in terms of ‘right time,’” said Within3 CEO Lance Hill. “If you’re talking about understanding over time, is there an impact on strategy?” For most cases in pharma, real-time – or instantaneously, as many people understand it – doesn’t necessarily apply.
“Does knowing something 30 seconds after a patient leaves a doctor’s office change our world in each occurrence? Probably not. But if you’re in the middle of a launch or you’re looking at information daily, weekly, or biweekly…real time might actually mean real-time for the decisions you want to make.”
In the context of real-time versus right-time, it’s important to understand the cadence of the problem and the ultimate objective. If data, analysis, and insights are available precisely when you need them, that is of high value. If data takes longer to aggregate and analyze for impactful action than the timeframe requires, that can lead to frustration.
For medical affairs professionals seeking impact from advisory boards or at medical congresses, the right time can approach real-time given AI’s current capabilities to aggregate, analyze, and report on large volumes of data. There should be no need to wait until April for insights generated from conversations in January. Insights around product launches – including uptake and adoption or physician and patient sentiment – could also be directionally relevant in shorter timelines.
Data and associated insights can be scarce in other examples, such as rare disease work. In those cases, a three-to-five-month timeline may be needed for sufficient impact – and as close to real-time as possible.
“For me, a timely insight is really getting live chatter at the right moment and not having a delay of four or five months until you get the analysis you need to make a positive impact,” said Carlos Eid of Novartis.
It’s important to consider the trade-offs or distinctions between data and insights for immediate action and those same data and insights being used for prediction. “A lot of times, we’ll be pushed to deliver something that is just as accurate as concrete data with some amount of a prediction looking forward two, three, four months or beyond,” said AstraZeneca’s Nathan Lear. “You have to assess and weigh the risks of predicting further ahead to get a timely insight versus having something that’s more concrete but perhaps is more retrospective.”
Watch the full Reuters webinar on demand. Or, grab a new white paper co-authored by Reuters, Within3, and medical affairs leaders: you can download “Elevating Medical Affairs Insight Management” for a deep dive into our recent industry survey.
