AI in Drug Development Helping Transform Clinical Trials

Clinical trials are a costly and time-consuming process, but AI can help expedite some of these challenges. This can make the process faster and more cost-effective. As AI becomes a bigger part of clinical trials, underlying intellectual property (IP) rights also become important in protecting these innovations.

Clinical trials in drug development

Clinical trials are research studies that test new drugs or treatments on people to see if they work and are safe.^[1] They are a crucial part of the drug development process since they provide a scientific basis for a drug’s safety, efficacy, and various use cases.

Clinical trials are essential for developing safe and effective treatments. They are also expensive and time-consuming. They require large cohorts of volunteers and necessitate careful scientific monitoring and testing, making them costly by nature.

Success rates in clinical trials can be quite low. Only 10% of compounds that enter clinical development progress to market entry. Moreover, clinical trials are an extensive and rigorous process and may take up to 10-15 years for both a medicine and a vaccine. Increased clinical trial complexity and larger clinical trial sizes may also contribute to the high cost in developing a medicine.^[2]

Recruiting and retaining participants in clinical trials is another major challenge, particularly since the studies must meet their target enrollment within a set timeframe. Literature indicates that if the required number of eligible participants is not reached, it can lead to delays, increased costs, or even compromise the accuracy of the study’s findings.

Moreover, managing clinical trial data can be complex in light of the large volume of information generated from multiple sources such as electronic health records, lab tests, and patient reports. Ensuring accuracy and consistency is key since inconsistencies may affect the reliability of study results.

In summary, the main challenges for running clinical trials are associated high costs, time-consuming processes, patient retention during the trial period, and data management.

How AI is changing clinical trials

Due to its massive computational power, AI is already being used to respond to some of these problems in several ways:

Reducing time to find patients for clinical trials

One of the biggest challenges in clinical trials is finding enough eligible patients. AI has the ability to quickly analyze large databases, including electronic health records (EHRs), genetic data, and even social media discussions, to identify suitable participants. For example, IBM Watson Health has developed AI tools to scan medical records and find eligible patients much faster than traditional methods. Another example is Deep 6 AI, a company that uses machine learning and claims to analyze patient data and match them with clinical trials in minutes instead of months. Using AI-based tools such as these can speed up recruitment and makes trials more efficient and targeted.

Improving trial design

AI can analyze past clinical trial data to design better trials. In a 2023 study published in Nature, Zhang et.al. noted: “In recent years, the use of AI-enabled technologies and real-world data (RWD), i.e. scientific data from a variety of sources, in healthcare has started transforming the way we approach clinical trials, which allows us to reshape key steps of clinical trial design.” The study also noted that “data sharing is challenging due to competition between institutions and data privacy laws.”

While challenges persist, advancements in AI have helped to improve trial designs, where the treatment plan can be adjusted as new data comes in. For example, Benevolent AI, a UK-based company, uses AI to analyze vast amounts of medical research and clinical trial data to predict which drug candidates have the highest chances of success using its proprietary Benevolent Platform.

Real-time patient monitoring

Traditionally, doctors monitor trial participants through periodic hospital visits. With the use of wearable devices and smartphone apps, with AI embedded, researchers are able to collect real-time health data. For instance, Varnosfaderani & Forouzanfar (2024) have stated that “AI optimizes hospital operation and management, refines medical image analysis, and revolutionizes patient care and monitoring through AI-powered wearables”, which enables remote monitoring from home instead of hospital visits.

AI algorithms can also analyze radiographic images, genetic data, and patient histories to detect early-stage cancer, while in cardiology, they can predict heart attacks and strokes by examining electrocardiogram (ECG) patterns and other vital signs. ECG is a measurement of the heart’s impulses to detect any abnormalities with the rhythm of the heartbeat.

Companies like Biofourmis use AI-powered wearables to track vital signs and detect early warning signs of side effects, allowing researchers to intervene quickly. Beyond clinical trials, AI technology has broader applications but can still enhance trial processes. For instance, during COVID-19, Biofourmis’ Biovitals® Sentinel platform was deployed by the Ministry of Health in Singapore to monitor COVID-19 patients and aid in early detection of deterioration.

In many clinical trials, a group of patients receives a placebo or an existing drug for comparison. However, AI is now being used to create a “control group” instead of enrolling real patients in control groups. According to Hutson (2024), a “start-up called Unlearn in San Francisco, California, creates digital twins of patients in clinical trials. Based on an experimental patient’s data at the start of a trial, researchers can use the twin to predict how the same patient would have progressed in the control group and compare outcomes.” Application of this or similar AI would mean that fewer patients are required for clinical trials.

Predicting success early

AI can help predict whether a drug is likely to succeed or fail much earlier in the process. Pfizer used AI in tandem with its prior work and corporate know-how to develop Paxlovid, an oral anti-viral pill used to treat COVID-19. According to the company’s website: “The company used virtual screening to help select the right molecular changes to enhance potency and then factored the data into the decisions on which compounds to make.”

Similarly, Insilico Medicine, a biotechnology company, uses AI to predict drug efficacy and has already identified several promising drug candidates using this method. AstraZeneca’s Alexion has also commenced a partnership with Verge Genomics to identify drug targets for rare neurodegenerative and neuromuscular diseases.

Role of IP in AI-driven clinical trials

The basic goal of the IP system is to encourage innovation through new technologies and creative works. This includes human-created and, in some cases, AI-created, inventions and works.

According to WIPO’s Frequently Asked Questions on AI and IP policy : “Qualifying human created works/inventions are protected by the existing IP frameworks, including patents, copyright, industrial designs, and trade secrets.” Some jurisdictions, for example in the European Union, offers database rights, which protect the investment made in compiling databases – some of which may be used by the AI.

Discussions on AI-generated inventions and IP tend to focus on whether AI can be recognized as an inventor, and how to protect AI algorithms and software. These questions are being widely debated in global health circles today. In a previous web story from 2023, WIPO’s Global Health Unit interviewed a firm that used a combination of IP to protect the algorithm underlying their patented MedTech technology.

In summary, AI’s growing role in medical innovation is supported by appropriate IP strategies. Protecting AI-driven innovations through patents, trade secrets, and clear data ownership rules will ensure that these breakthroughs continue to improve drug development while maintaining competition in the industry and supporting global health.

Footnotes

[1] Prior to clinical trials, investigators conduct preclinical research using either human cell cultures or animal models. See: Nuno Henrique Franco, Animal Experiments in Biomedical Research: A Historical Perspective, Animals (Basel, 2013).

[2] “DiMasi attributes the higher price tag largely to higher failure rates and to increasing out-of-pocket expenses, driven by factors such as increased clinical trial complexity, larger clinical trial sizes and the cost of collaborating with the medical sector”. See, Asher Mullard, New drugs cost US$2.6 billion to develop, Nature Reviews Drug Discovery (2014), making reference to a study by Joseph DiMasi of the Tufts Center for the Study of Drug Development in Boston, Massachusetts, USA, and colleagues who estimated drug development costs based on data provided by 10 drug companies for a randomly selected subset of 106 self-originated drugs that went into the clinic between 1995 and 2007.