Effectively incorporating AI into drug research, development

Key Challenges in Research and Development

Acknowledging the ongoing challenges in this field is important to understanding the opportunities for AI to improve research and development. The drug research and development landscape presents many difficulties. Although the complexity and magnitude of these challenges change based on the company and environment, Swarna offered some insight into the significant challenges faced by this industry.

Scientific Growth

“The biggest challenge is both an opportunity and a challenge. We see that there've been tremendous advances in biology and basic disease biology. We understand human disease biology better than we've ever done as a global scientific community working to address unmet needs in disease areas,” he stated, suggesting that one challenge is keeping up with scientific growth.

“Science has grown dramatically in terms of its impact on patients and the benefit for patients, but we've not been able to keep pace with the growth in science in terms of executing on that scientific progress and being able to shorten the time and the cost associated with bringing therapeutics and devices to patients.”

Swarna referred to the mechanics of clinical trials as “fraught,” meaning data management challenges and other complexities littered throughout the trial process. For example, clinical trials often take a long time to execute. While significant resources across the industry have been allocated to shorten the time it takes to get a drug to market, it can still take over a decade and billions of dollars.

Macroeconomics

In addition to keeping up with advancing science, drug research and development firms are also challenged by macroeconomic conditions, including reimbursement challenges and the Inflation Reduction Act in the US.

“It's changing how companies think about their portfolios and the molecules as well as the disease areas they will focus on,” Swarna emphasized. “There's a whole retooling of the industry in terms of thinking about the impact of the macroeconomic conditions, reimbursement, and the economics of bringing drugs to market. That's a second driver.”

Technology Optimization

The final challenge is optimizing technology in research and development. Many companies have invested in various technologies across the industry.

“Technology investments that companies have made have been very important and very useful investments in individual areas, but bringing them all together and systematically utilizing data from early research through to late stage development and regulatory submission [is a challenge], he explained. “There's a lot that can be done. There's a lot of opportunity to unify all of that.”

Reinventing Research and Development with AI

The report focuses on the idea that technological intervention can reinvent and improve the research and development pipeline.

“Unlike other previous technology movements, this is not about an individual technology. Generative AI and analytics are not just about one technology implementation. It is about a systematic rethinking of the processes and the data flow and the investments that companies make in technology to really go end-to-end early research, early discovery through to late-stage development and beyond,” he said.

He reiterated that technology reinvention is about enterprise-wide strategy and implementation.

Responsible AI

However, reinvention requires companies to consider the potential challenges that AI may present. “We have a very robust framework at Accenture that we call our responsible AI framework,” noted Swarna.

Responsible AI is an industry-wide term that considers how AI can be used effectively and how certain challenges that are inevitable with AI, like bias and security, can be addressed.

Accenture considers multiple factors regarding bias, which means choosing the right patient populations to study and understanding how bias might have impacted existing data that inform ongoing research.

In addition, the company considers security issues, including intellectual property protection and patient privacy, in its responsible AI framework.

“Protecting patient privacy and being compliant with all the regulations around patient privacy around the world is really important for us to think about [to use] technology in the right way inside an organization because this is a very powerful technology, and [without] guardrails around how it's used, we could run into data loss or data breaches.”

Measuring ROI

Swarna explained how companies can ensure that the technologies they deploy effectively and positively impact their research and development life cycle.

“We've developed a rubric for being able to do that,” he noted. “The one thing that is potentially different in our industry from other industries is that we are a long cycle industry.”

“We've tried to come up with an objective measure in terms of the impact and the return on investment on these technologies in terms of changing or bending the curve in terms of both cost and time. So we have a framework that we use internally and with our clients.”

Additionally, companies have developed quantitative measures to act as interim milestones. For example, companies can assess their recruitment cost per patient, recruitment velocity, and other factors.

“Those are all measures that we can observe in smaller chunks of time as opposed to waiting to see the impact across the entire value chain from end to end. So those are some of the measures we can take,” he noted. “The one key thing would be it'll vary by company and the stage in which they are implementing. Some companies have decided to lead with this in terms of making this part of the core strategy. Others are taking a more conservative view based on where they are in their own internal business cycle.”

Benefits of Using AI in Clinical Trials

Theoretically, incorporating AI into clinical trials can help reduce research times and spending. While Swarna and Accenture cannot speak to the greater implications of policy on drug pricing, he offered a theoretical outcome that underscores the benefits of using AI in clinical research.

“An individual drug can become more profitable for a company because the actual development costs, in theory, could be lower.”

Swarna explained that, theoretically, a cost reduction achieved through optimizing the right technologies could result in lower development costs, which could trickle down to make medicine more affordable and accessible globally.

He also explained that reducing research and development spending could provide more effective drugs to broader populations without excessively costing the overall healthcare system.