AstraZeneca and Oracle sign non-binding deal to test AI for trial recruitment

AstraZeneca and Oracle have signed a non-binding agreement to use Oracle technology to help AstraZeneca teams find patients for trials in oncology; cardiovascular, renal and metabolic diseases; respiratory and immunology; and rare disease, the companies said. The plan is to apply an AI model to de-identified data from clinical research and routine care to inform the design of future studies, safety monitoring and patient care measures.

Oracle’s technology will be used in several clinical programs in Europe, the US and the Middle East. The first is a collaboration with M42, an Abu Dhabi-headquartered healthcare firm, focused on hypertension. The AstraZeneca-M42 clinical team will use the technology to identify patients with uncontrolled hypertension for planned clinical studies, AstraZeneca said.

The partners plan to set shared governance, reusable technical components and agreed outcome measures to assess results and decide whether successful approaches can extend across disease areas and geographies, the companies said.

AstraZeneca CEO Pascal Soriot said:

“Combining AstraZeneca’s world-class science with Oracle’s deep data and technology expertise, we can help clinicians identify patients sooner, optimize clinical trials, and generate evidence continuously across the lifecycle of a medicine.”

Oracle's Mike Sicilia said:

“Some of the most important opportunities for AI are in healthcare, where better use of data can help turn scientific innovation into better outcomes for patients. We have an opportunity to help accelerate research, connect more patients to clinical trials, and ultimately bring relevant insights into care faster.”

Why recruitment matters

Finding patients is one of the slowest parts of drug development. When trials sign up people slowly, they run longer and cost more, The Next Web reported. Sponsors and agencies are also bringing recruitment strategy into protocol design and site selection instead of treating it as last-minute advertising, according to MM+M.

AI can screen far more patients, but it cannot fix restrictive eligibility criteria, missing health data or geographic hurdles, MM+M reported. That limits what any recruitment tool can deliver on its own.

Evidence and limits

Research on AI eligibility screening is early. Jin and colleagues reported in 2024 that TrialGPT judged eligibility with 87.3% accuracy across 1,015 patient-criterion pairs and cut screening time 42.6% in user validation. Unlu and colleagues randomised 4,476 heart failure patients in 2025; eligibility was 20.4% with language-model support versus 12.7% by hand, and 35 patients enrolled versus 19. The TrialGPT test used synthetic data, and the heart failure comparison covered one site and one trial.

The companies did not give a start date for recruitment or a patient target for the M42 hypertension work.