The promise of AI!
Its indeed romantic attribution to the “artificial intelligence” that skims through the “doctors notes” and processes the information to suggest the patients best suitable for the trials.
The devil lies in the details. Even if you OCR the notes, it is still a challenge to ensure accuracy for the analogue to digital conversion and the accuracy claims are usually off the mark. What is required is a structured template with a UI/UX layer to capture the data without investing in $$ to attempt to rewrite everything. It is difficult, if not impossible, to improve upon the promised accuracy since it is limited by efficiency gains.
Clinical trials themselves could change significantly thanks to AI advances. The technology has the potential to improve logistics such as patient enrolment by quickly scanning electronic medical records and clinical trial eligibility criteria.
For instance, 86 per cent of trials do not meet enrolment deadlines, and nearly a third of Phase III trials fail because of inability to recruit patients. AI could improve the selection of participants by sifting through electronic medical records, genetic information and unstructured data, including doctors’ notes, to find the best fits between proposed therapies and disease signatures.
These statements represent an extreme case of “techno-optimism”.