hat does the path to eradicating all human disease look like?

Almost two years ago, I asked this question in a blog titled “Accelerating the path from possibility to proof in the development of new medicines.” I wrote that leveraging our most advanced tools would drive new synergies at the intersection of safety, science, and speed and usher in a new era of drug development to save and improve countless lives. Today, we are here. Today, we stand at the threshold of a new era of scientific development fueled by artificial intelligence and machine learning (AI and ML).

“AI/ML holds unprecedented potential to help lower costs while improving speed, efficiency, and quality throughout the drug development life cycle,” said Ken Getz, executive director and professor at Tufts Center for the Study of Drug Development at Tufts University School of Medicine, in an interview for this article. “During the past two decades, we’ve witnessed a dramatic increase in the scientific and operating complexity of drug development activity coinciding with a surge in data volume. Concurrently, drug development costs and failure rates have steadily increased, while the return on drug development investment has fallen precipitously. AI/ML holds compelling promise in automating labor- and data-intensive tasks and bringing scaled and accelerated data processing and analytical capabilities.”

AI and ML might just be the most important technologies of our generation, potentially paving the way for the pharmaceutical industry to finally solve some of its most intractable challenges and harness new opportunities that could lead to cures once thought to be impossible.

Here’s a look at how AI and ML are changing clinical research in both large and small ways.

Overcoming Variability in Research

In clinical research, AI and ML can be pivotal in helping organizations overcome the workflow challenges inherent to rigid trial protocols. By necessity, clinical trials are all unique, not just because they are studying treatment response nuances in different therapeutic areas but also because of the individual characteristics of each study population and the regulatory environment. This complex situation is further compounded by the varied objectives of each trial, which can range from testing the safety and efficacy of a new treatment to the exploration of an optimal dosing regimen, to an investigation of a disease’s mechanisms, or capturing healthcare resource utilization (HCRU) data.

Such wide variability has historically presented major roadblocks to the industry’s capacity to produce new and effective treatments. On average, 38 new treatments have been produced each year (from data between 2010-2019). At this pace, it would take more than 200 years to solve every currently known disease.

Given the inherent complexities of clinical trials, which also now include the various data formats, devices, wearables, and other technologies currently being employed, investigators need sophisticated tools that can help them manage and analyze vast amounts of varied information. Yet, historically, trial investigators have been slow in taking advantage of such technologies. This must change for us to move forward in meaningful ways. We have already seen how decentralization and digitalization have greatly accelerated and democratized the clinical trial process, leading to the quicker development of new treatments without reducing confidence in safety and efficacy.

“There are so many areas that we’re excited about with AI,” said Getz. “Two that are particularly patient centric: 1) rapid identification of patients who are eligible to participate in clinical trials based on real-world data; and 2) more effective and efficient patient safety surveillance. Our largest concern at this early stage is the faulty and dangerous conclusions that might be drawn from poor-quality data.”

Improving Patient Recruitment and Screening

AI and ML offer unique opportunities for investigators to identify exactly what data to collect and from whom while being able to analyze it in real time. This can be particularly effective in the recruitment and screening of potential participants. Some companies are leading the way by making impactful use of such technology today. For example, one AI company that helps with clinical trial recruitment and screening is using ML to analyze electronic health record data to identify potential participants for clinical trials.

AI-powered capabilities will have an outsize impact in the race to cure rare diseases, where patients are more difficult to find and spread out all over the world. “These capabilities are a glimmer of hope for rare disease patients,” said Rita Naman, whose son Milo suffers from an ultra-rare mitochondrial disease, in an interview for this article. “While Milo’s disease has no cure today, these technologies have the power to help researchers bring together patients with similar conditions wherever they live and get studies underway faster. And when it comes to your own child’s life, time is everything.”

One major pharmaceutical company is using AI to identify potential participants for its clinical trials of a new cancer vaccine. The company’s AI algorithm analyzes patient data from electronic health records and social media to identify patients with certain types of cancer who may be eligible for the trial. Several global pharmaceutical companies are leveraging AI in clinical research now. One is using AI to screen participants for its clinical trials of a new drug for Alzheimer’s disease. The company’s AI algorithm analyzes medical images to identify patients with certain brain changes that are associated with Alzheimer’s disease. And another is using an AI-powered chatbot to simplify the screening process by informing patients about the trial and answering their questions, and to collect preliminary data.

Such use of this groundbreaking technology can make screening for clinical trials easier by improving the efficiency and accuracy of the overall process. AI can be used to automate tasks such as data entry and analysis, for example, as well as to identify and correct errors in the data. This can then free up researchers to focus on more important tasks, such as interacting with trial participants.

Enabling More Informed Consent

Properly educating participants about a study has presented trial investigators with challenges for generations. According to a 2023 industry survey, 35% of patients who dropped out of a study early believed it was too difficult to understand the informed consent form compared to just 16% of those who completed the trial. AI and ML are ideally suited for helping to improve the educational process by providing better and easier-to-understand educational information to participants, and to be able to adjust and improve that information based on common misunderstandings.

Once again, this technology is already being deployed in a variety of ways. Duke University is using a 24/7 AI chatbot that can be accessed from anywhere to answer participants’ questions about clinical trials and to help them understand the risks and benefits of participating. Meanwhile, Stanford University is using AI to assess its participants’ understanding of the informed consent process for its clinical trials of a new gene therapy for Parkinson’s disease. The AI algorithm analyzes participants’ responses to questions and identifies any areas where they may need additional information.

Such efficient and tailored communication and data analysis can also improve access to clinical trials for those who have often been shut out from studies in the past. AI and ML will ultimately provide investigators with better and more efficient means to find trial participants who more equitably represent the affected patient population which, in the end, results in better outcomes for all.

Automating Randomization

Ensuring that studies are unbiased and statistically valid is critical to achieving the best results. AI and ML can generate randomization schedules that are wholly objective by balancing the treatment groups with respect to known and unknown prognostic factors. This can help to ensure that the results of the clinical trial are more accurate and reliable.

AI can also be used to automate the randomization process, which can save time and reduce the risk of errors. For example, AI algorithms can be used to generate randomization schedules in real time, as participants are enrolled in the trial. This can help to ensure that participants are randomized as quickly as possible and that the treatment groups remain balanced throughout the trial.

AI can also be used to support adaptive trial designs, which are clinical trials that are modified during the trial based on the data that is collected. For example, one commercial clinical analytics company uses AI algorithms to identify participants who are at high risk of dropping out of the trial or who are not responding well to the treatment. This information can then be used to modify the trial design to improve the efficiency and effectiveness of the trial.

An AI Foundation for Growth, and Good

To reach this enormous potential, there must be widespread use of these new technologies. The most ideal way to overcome the broad and inherent challenges that are engendered within the clinical trial protocol and create scalability across an entire pipeline is with the adoption of a comprehensive AI- and ML-enhanced platform. As this technology continues to develop and evolve, such platforms will evolve along with it. They will also enable pharmaceutical companies to find commonalities more easily in conduct across their portfolios of trials and build bridges that will help create new standards that can be carried to new trials, regardless of their unique factors.

This framework will accelerate research to just one-day study start up, one-week participant enrollment, and one-year study conduct.

AI is transforming a range of industries, including clinical research. And as it continues to evolve, the potential for creating true synergy between AI, science, and compassion becomes even more tangible, enabling us to plausibly see a future where treatments and resolutions can be found for the 10,000+ remaining diseases that have burdened humanity for centuries.