87% of Biopharma Respondents Use AI/ML for Post-Market Activities, But Key Missed Opportunities Remain
In the survey, which was conducted among 150 participants whose companies develop products for rare diseases and/or specialty therapies, 87% of respondents said their companies use AI or ML tools during the commercialization stage of the biopharma lifecycle. The most-cited use cases of automation included finding new patients, identifying key opinion leaders other than treating physicians, and assessing medical journeys.
The findings point to bigger trends occurring at the crossroads of automation and biopharma at large, but also align directly to challenges in patient identification for orphan and specialty products, says Dan Fisher, managing director of IPM.ai.
“The value of using artificial intelligence and the different disciplines that sit within AI has been realized across the industry,” Fisher said in a recent survey report summarizing the survey’s findings. “It’s certainly been effective in larger-prevalence disease states for segmenting or clustering patients and reaching them with the right level of engagement. But it’s also been tremendously effective for rare disease companies at identifying potential patients.”
Missed Opportunities in AI/ML Applications
Despite high awareness of AI and ML platforms, the survey suggests that some of the biggest promises of automation have yet to be embraced.
For example, less than half — 40% — of biopharma respondents who are using AI/ML technologies are using them to find new prescribers, even though the treating physician landscape plays a large role in new patient identification for rare disease and specialty products.
Others may be missing growth opportunities by focusing their AI/ML-supported patient finding efforts too narrowly on already-diagnosed patients. Just 34% of respondents said they’re currently using AI/ML to find potentially undiagnosed or misdiagnosed patients, even though individuals with rare diseases frequently navigate multiple misdiagnoses before receiving the correct one.
Fisher says automated tools can help manufacturers gain the early insights needed to fuel successful commercialization strategies. To realize these potential benefits, organizations must expand their use cases for AI/ML – especially when it comes to use cases that incorporate predictive modeling.
“The adage that we use most often is ‘if you’ve seen one patient, you’ve seen one patient,’” Fisher said. “But in the data, there are a lot of interconnected steps that happen for each rare disease patient. We call it the constellation of events that occurred over their history. With machine learning and the ability to view the population at scale, we can see others who fit the same pattern but may not have yet reached an initial diagnosis.”
Concerns of Privacy, Security, Accuracy and Budget
Possibilities of AI notwithstanding, biopharma leaders still have concerns about AI/ML, including a lack of confidence in its accuracy, as well as questions about privacy and security.
“Those are absolutely legitimate concerns that anybody should have when determining whether to use AI or which vendor to use,” Fisher said. “And they all underscore the importance of working with a partner that has privacy built into the foundation of their business, as well as uncompromised compliance.”
He adds that overcoming these challenges may require that manufacturers not only prioritize data quality and access, but also scrutinize their selected vendors’ data sharing: 17% of survey respondents said that limited access to patient and prescriber data is the top restrictive factor for commercial success.
Additionally, 29% of respondents said that shrinking budgets that restrict the ability to invest in technology is the industry trend that has the most limiting effect on the overall commercial success of their product — emphasizing what Fisher considers a need to reimagine how organizations assess the ROI of automation.
“Think about AI as an enabler more than anything else,” he said. “Technology that works at scale and that helps overcome the modern challenges of finding and reaching prescribers and patients offers tremendous — and maybe underestimated — value. With the velocity of change we’ve seen in AI/ML tools, you’re now able to do far more by spending far less.”
For full insights from the 2023 IPM.ai and BioPharma Dive survey, download the report here.
IPM.ai, part of Real Chemistry (www.ipm.ai), transforms real-world data into real-world insights that uncover the ideal patient and their healthcare ecosystem in specialty and rare diseases. IPM.ai’s Insights as a Service (IaaS) platform optimizes drug development, clinical study, product launch and commercial operations by utilizing longitudinal analytics, artificial intelligence and machine learning in conjunction with a real-world data universe of over 300 million de-identified patient journeys and 65 billion anonymized social determinants of health signals.