Major Pharmaceutical Companies Embrace Artificial Intelligence to Expedite Clinical Trials

Pharmaceutical giants are increasingly turning to artificial intelligence to accelerate clinical trials, Human studies have long been one of the most costly and time-consuming aspects of drug development, often taking years to recruit patients and test new medications, costing over a billion dollars from drug discovery to the finish line.

For years, pharmaceutical companies have been experimenting with artificial intelligence in hopes of letting machines discover the next blockbuster drug, while a few compounds identified by AI are now in development, these bets are still years from paying off.

However, interviews with more than a dozen drug company CEOs, regulatory officials, public health experts, and AI companies by Reuters indicate that the technology is playing a growing and significant role in human drug trials.

Companies like Amgen, Bayer, and Novartis are training AI on billions of public health records, prescription data, medical insurance claims, and their internal data to find trial patients, potentially cutting the time it takes to recruit them in half.

The US Food and Drug Administration (FDA) said it had received nearly 300 requests to incorporate AI or machine learning into drug development between 2016 and 2022, over 90% of these requests came in the past two years, with most aimed at using AI in some phase of clinical development.

Prior to AI, Amgen used to spend months sending out surveys to doctors from Johannesburg to Texas to ask if their clinic or hospital had patients with certain relevant clinical and demographic characteristics to participate in a trial.

Often, existing relationships with facilities or doctors influence which trial sites are selected.

However, Deloitte estimates that nearly 80% of studies fail to achieve their recruitment goals because clinics and hospitals overestimate the number of available patients, or there are high attrition rates, or patients don’t adhere to trial protocols.

Amgen’s AI tool, ATOMIC, now scans large sets of internal and public data to identify and rank clinics and doctors based on their past performance in recruiting patients for trials.

While Amgen previously estimated that enrolling patients in a mid-stage trial might take up to 18 months depending on the disease, ATOMIC can cut that time in half under the best scenarios.

Amgen has used ATOMIC in a handful of trials testing drugs for conditions including cardiovascular disease and cancer and plans to use it in most studies by 2024.

The company expects that by 2030, AI will have cut two years or more from the typically decade-long drug development timeline.

Overall, less than 25% of health data is available for research, according to Samir Bhatt, an AI expert at the World Health Organization.

Bayer, the German pharmaceutical firm, used AI to reduce the number of participants needed by several thousand in the late-stage trial of asondansetron, an experimental drug designed to reduce the long-term risk of stroke in adults.

Pharmaceutical companies typically seek prior approval from regulatory authorities to use an external control arm in a trial.

Bayer said it is in discussions with regulatory authorities, including the FDA, about relying on AI to create external controls for pediatric trials.

The European Medicines Agency said it has received no requests from companies seeking to use AI in this way.

Some scientists, including the head of oncology at the FDA, worry that drug companies are trying to use AI to arrive at external controls for a wide range of diseases.

Patients in clinical trials often feel they are better off than others because they believe they are getting an effective treatment and are receiving more medical attention, potentially leading to an overestimation of the drug’s success.

In light of this, regulatory authorities must remain cautious and vigilant as the use of AI in clinical trials continues to evolve and expand.