AI, Big Data, and Digital Disruption

Technological innovations and digital disruptions have a long history in clinical research. We often push boundaries of what’s possible to enable better, faster drug development to bring therapies to patients. However, this innovation push is pulled back by a necessary caution because, at the end of the day, patient safety, data integrity and regulatory compliance are paramount to our operations. 

Icon’s digital disruption survey and the publication of a related white paper (Digital Disruption. Surveying the Industry’s Evolving Landscape) is a follow-on to a similar survey we conducted in 2019, when ideas such as blockchain and organ-on-a-chip were buzzing in the industry. 

In the time since, we have seen an organic shift in the hype cycle, with some technologies seeing more significant uptake than others. AI, for example, has become more pervasive as a solution to other technological capabilities, such as remote monitoring, and enabled the industry’s current data abundance. 

We were also interested to see how pandemic-era digital adoptions impacted the overall trajectory of technology in clinical research. Now that we are post-pandemic, this survey gave us a snapshot of the changing attitudes and challenges for digital innovation in clinical research. 

Main findings and biggest surprises
 

Key findings from this survey showed that attitudes towards the potential impact of digital tools and AI solutions remain relatively positive. Sponsors see the value of new technology, and optimism about its potential to improve R&D productivity has grown by 5 percent since 2019 – though with more modest expectations overall as the hype of innovation meets the reality of implementation. Improving returns on R&D is still seen as a leading area of impact, followed by improved safety and efficacy. 

Given our level of engagement with sponsors across a wide range of therapeutic areas, we were less surprised by the survey results because they echoed a lot of what we have encountered anecdotally. Compared to expectations set in our 2019 survey, however, the rate of digital adoption has been slower than expected.

But it hasn’t stalled. We see a slow and steady increase in implementation. For example, the 2024 survey shows a 10 percent increase in the number of respondents that indicated their organizations were using AI and big data analytics in their programs, totalling 49 percent. We also see an increase in the number of organizations piloting these technologies, growing from 29 percent in 2019 to 35 percent in 2024. Considering the length of an innovation hype cycle, where this process takes years to settle into a significant adoption pattern, the progress made has occurred over a relatively short period of time. 

The majority of respondents (88 percent) plan to increase investment in digital-enabled technologies in the next one or two years. Many organizations are reluctant to be early adopters of unproven digital innovations – especially when the success of important therapies and costly trials could be jeopardised. External factors such as a perceived lack of regulatory support or clarity also impacts technological uptake, as do the rising costs associated with licensing models and keeping digital infrastructure secure from cyberattacks.

Depending on the existing infrastructure, internal digital transformations require a significant investment of time and resources that can be difficult to undertake while maintaining a favorable return on investment. In our survey, 70 percent of respondents were either piloting or selectively using AI in clinical development, but only 13 percent had a comprehensive program in place. 

Ensuring that technology is scalable and adaptable to various trial designs and therapeutics is another consideration that may be impacting the slower pace of comprehensive programs. In other words, companies may be struggling to design their programs so that their tools offer interoperability to allow different configurations within the ecosystem of digital solutions. 

Improving clinical trials with AI
 

As a CRO, we see a lot of potential around AI for improving R&D. The clinical trial phase of drug development is the longest, riskiest and most expensive. Integrating digital solutions across clinical phases through to postmarketing can streamline trial management, reduce manual workloads and enhance data quality to ensure efficient, compliant trials. 

When developed and implemented strategically, AI tools can make informed decision-making processes faster, accelerate timelines, improve control, forecast risks and requirements, and create more responsive operational efficiencies. Eliminating inefficiency and centralizing certain processes through connected AI and data systems can have big impacts on R&D. As just one example; AI tools can be used to identify the right clinical trial sites, first time, by parsing massive amounts of real-world data, evaluating connections and ranking results for best-fit sites. 

AI is the natural next step in making the data we capture more usable, faster, which is why it’s more prevalent in clinical development. The probabilistic, data-driven models used in AI are increasingly allowing scientists to better model and understand complex systems. AI can help approach problems from the bottom up instead of the top down – measuring lots of data first and using algorithms to come up with suggested rules and patterns to assist our overall understanding to challenges.

The future of AI in pharma
 

Our 2024 survey affirmed an attitude shift we had already noticed: hype around AI is settling to a more realistic range of optimism. We expect this to continue as we move farther away from AI for AI sake, and as solutions evolve into more thoughtful, interconnected, and interoperable systems. 

There are exciting applications for AI to have a wider positive impact. For example, the technology can support more sustainable operations. AI tools bring efficiencies that help to reduce resource waste, greenhouse gas emissions, and carbon footprints. For example, ICON is piloting a proprietary AI forecasting system based on our own datasets that is improving kit resourcing processes during clinical trials and reducing kit waste, which is a significant issue industry wide.

Proponents of AI claim that it has the potential to make drug discovery and development more efficient and cost effective. Today, it takes on average 8-10 years and costs more than $2 billion to develop a drug – and less than 10 percent of drugs that enter clinical trials are approved by regulators. The opportunity for AI to bridge this gap is immense. AI facilitates the understanding of diseases and what drugs may be most effective by analyzing large quantities of disparate data at faster-than-human rates. Software can pinpoint promising molecules and fine-tune their structures to boost their chance of success in trials. Generative AI also has the potential to go a step further by suggesting new molecules to be tested. 

^{Image Credit: Original image sourced from Adobestock.com}