Some hospitals are trying software that searches patient records to find people who might qualify for clinical trials. A 2026 study summary from a cancer program reports that it can save research staff time. But saving time doesn't answer a separate question: is the hospital allowed to use and share patient information this way? Before turning it on, write down where the data goes, which approvals already cover it, and who checks the results.
How this software works
Research teams usually find trial patients by reading charts by hand, which is slow. This software searches the electronic health record (EHR), where patient charts are kept, and suggests patients who might fit a trial. Research staff then review those suggestions before anyone contacts a patient. This step is called trial prescreening.
What the study found
A 2026 study summary, or abstract, in JCO Clinical Cancer Informatics describes one such system. It was first tested against manual chart review in one trial for advanced (metastatic) colorectal cancer, then used across 29 cancer trials.
In that first test, the abstract reports that the system picked the right patients 94 percent of the time when checked against past records, and 88 percent of the time with new patients. It doesn't say how many patients those numbers are based on.
Across the 29 trials, it reports that staff had 10 times less chart review to do, but it doesn't explain how that was measured or what it was compared with. The abstract reports screening time falling from 3.1 to 1.8 minutes per reviewed chart, but does not clearly identify the comparison underlying that reduction or whether it applies to the broader rollout.
What these numbers do and don't tell you
The results are encouraging and worth testing. But they come from one cancer program, and the summary leaves out details needed to judge them fully. We'd treat them as a reason to test the tool on your own patients, not as results you can expect.
We'd also keep the two time-saving numbers apart. One is about how many charts staff need to review. The other is about how long each review takes. Combining them into one "time saved" figure would go beyond what the study shows.