How to Invest in AI to Improve Clinical Trials

The dramatic rise in the adoption of artificial intelligence (AI) in clinical trials over the last decade has included a range of AI-enabled processes across the drug discovery and development value chain. Many of these innovations have been driven by an intensifying demand for machine learning (ML) algorithms that can analyze a seemingly never-ending stream of biomedical and chemistry data. Still, even while observing steady growth, many experts and thought leaders believe AI’s full, revolutionary potential to remake pharmaceutical R&D has yet to be realized. 

When it comes to investing in AI to improve clinical trials, strategic organizations stand to reap significant benefits – reducing costs, shortening development timelines, increasing productivity and boosting the overall efficiency of clinical trials.

In this article, we take a look at some specific ways current and emerging AI technologies are poised to improve clinical trials –  from processes to patient outcomes – and what to look for when choosing an AI partner or platform.

The current clinical trial landscape

Despite a slew of recent technological advancements, bringing a drug to market is still an expensive and time-intensive proposition, costing an average of $2.8 B over 10+ years with the clinical trial phase alone averaging five to seven years. Time delays in clinical trials are often attributed to a laborious, inefficient and complex data flow, which still requires a fair amount of manual labor.

But there’s good news: recent estimates predict AI-based innovations could reduce drug discovery costs by as much as 70%. Enthusiasm for AI-based improvements in clinical trials have no doubt contributed to the 22% increase in R&D investment over the last 3 years by top-15 global pharma companies. According to Deloitte’s life sciences digital innovation survey, 76% of the 150 biopharma leaders it surveyed reported they were currently investing in AI for clinical development.

Clinical trial pain points

So what are some of the primary and persistent roadblocks and challenges facing clinical trials today? 

1. Fragmented systems and knowledge silos: Trial data is often stored in disparate, decentralized locations that don’t communicate with one another; for example, from within an organization's archives, from a clinical institution, or from an external partner.
2. Reliance on legacy and manual processes: Even today, much trial data, including database transcription and creation, is still being processed manually, oftentimes using outdated systems. This means work is slow, costly and repetitive – with efforts being duplicated across trials.
3. Trouble with patient recruitment & retention: Travel to trial locations can be difficult for patients, leading to trial under-enrollment and related issues. In addition, patients who do enroll may find medical adherence and invasive monitoring to be a significant burden, leading to increased trial drop out.
4. Underrepresented populations: Trials can suffer from inaccessibility to traditionally underrepresented populations, making enrolling a diverse sample of trial participants an ongoing challenge.
5. Spiralling cost and complexity: The exorbitant cost of clinical trials is due in part to  their increasing complexity – stretching resources and necessitating the need for better, more streamlined trial design. For example, targeting smaller patient subgroups based on their individual genetic markers and/or other lifestyle/medical history factors, as is the case with precision medicine initiatives, requires a broad modernization of the traditional clinical trial ecosystem. ‍
6. Regulatory hurdles: Biopharmaceuticals is a heavily regulated industry, with many layers of oversight, requirements and governance. Meeting regulatory guidelines, therefore can present an assortment of roadblocks and challenges, from crucial delays to burdensome expenses.

How the right AI technology investments can help improve clinical trial outcomes

‍
With the right AI partner, organizations are likely to gain significant benefit. When considering investment in AI tools and tech, including platforms and services, it’s important to look for AI that addresses many of the industry’s known challenges. In general, look for AI technologies that help:

1. Optimize data collection and flow: AI-based technologies can help create better structured, standardized, and normalized data elements across a variety of inputs and sources. These tools can also help streamline the way data is integrated during a clinical trial, and generate insights from past and current trials in order to improve future trials. ML can also be used to find relationships between vast, disparate data sets – performing operations that would take large teams years to analyze.
2. Digitize legacy and manual processes: Using the power of AI, legacy clinical trial processes can be digitized and automated, helping wrap studies faster and getting medicines and life-saving treatments to patients sooner. AI technologies can also help organizations reuse their existing data when appropriate, reducing repetitive or duplicative efforts.
   
3. Improve study design: AI can also be used to create more modern, patient-centered study designs, managing complexities to reduce patient burden, increasing compliance, and creating a range of new efficiencies across study processes. For example, natural language processing (NLP) and supervised/unsupervised learning techniques can be used on clinical data to help design better study protocols. And, perhaps most critically, AI systems can be used to aid in the clinical adoption of precision medicine. Studies which leverage AI in their trial design (e.g. predicting drug efficacy based on patient variances; or identifying new drug lead candidates for patient subgroups before entering trial) stand to improve outcomes, boost likelihood of trial success, and lower costs over the long-term.
4. Streamline the patient experience: Especially with the proliferation of wearable sensor technology, today there are more ways than ever before to offer non-invasive, frictionless capture of clinical trial data. The ability to monitor patient vitals and other information remotely helps minimize the need for in-person trial sites. AI can be used alongside other wearable tech to generate real-time insights, reduce trial drop out rates and maximize patient adherence.
5. Recruit more diverse populations: The ability to conduct many parts of the clinical trial experience remotely helps enable organizations to accelerate recruitment of patients from different populations, even precision matching patients for specific trials (AI-enabled patient stratification helps guarantee the equitable allocation of participants or subgroups.) AI methods like NLP and ML can also help reduce bias and inequitable processes which can occur when recruitment is done manually. [Further reading: Integrating AI/ML Models for Patient Stratification Leveraging Omics Dataset and Clinical Biomarkers from COVID-19 Patients: A Promising Approach to Personalized Medicine]
6. Lower costs and improve speed: AI powered automation and data collection helps to generally lower cost and reduce the amount of time required to process clinical trial data. For example, ML algorithms can help predict the success of molecules that are used in clinical trials, identify new compounds, and draw novel insights. Biosimulation engines can accelerate drug development insights and de-risk R&D decisions by predicting the clinical benefit of a drug or molecule before it gets to human trial.

Bottom line: For organizations willing to make the right strategic investments, AI is poised to solve key industry pain points and meaningfully accelerate the discovery process in drug development. As the industry sees more and more clinical success of AI-driven methods, including a narrowing of the translational gap, industry-wide adoption is likely to follow – revolutionizing the industry and giving patients quicker access to better drugs.

‍