Advocate Health, one of the largest nonprofit health systems in the United States, announced it is partnering with AI startup Lind to automatically screen cancer patients for clinical trial eligibility across its oncology network, embedding the platform directly into its electronic health record. The move, detailed in releases from Advocate Health and coverage by Becker’s Oncology and HCI Innovation Group, targets a persistent and costly inefficiency in cancer research: nationally, only about 5% of eligible cancer patients ever end up enrolled in a clinical trial, not because they are unwilling, but because manual eligibility screening is slow, inconsistent, and easy to miss.
The bottleneck Lind is built to solve
Matching a single cancer patient to an appropriate clinical trial can require a research coordinator to manually check that patient’s chart against up to dozens of detailed eligibility criteria per trial — tumor genetics, prior treatments, organ function, comorbidities — a process that can take up to an hour per patient and simply does not scale across a busy cancer center’s full patient volume. Coordinators typically focus their limited time on patients already flagged by a treating oncologist, meaning many otherwise-eligible patients are never even considered. Lind’s platform is designed to run that screening automatically and continuously across a health system’s entire cancer population, surfacing likely matches for human review rather than requiring a manual look-up for every patient.
What the validation data shows
A retrospective study reviewing more than 1,700 individual eligibility criteria found that Lind’s automated screening achieved more than 94% agreement with human clinical reviewers, according to results cited by Advocate Health and industry trade coverage, with strong performance across multiple matching thresholds. Just as notably, the study found the AI-assisted process cut the time needed to assess a single patient’s eligibility from as long as an hour down to a matter of minutes — while preserving a “human-plus-AI” model in which a clinician still makes the final call on whether to approach a patient about a specific trial.
Why Advocate Health is moving now
Advocate Health operates one of the country’s largest nonprofit health systems, giving it a large and diverse cancer patient population against which to test trial-matching technology at scale — precisely the kind of environment where manual screening bottlenecks are most acute. By embedding Lind directly into its EHR rather than running it as a separate research tool, the health system is aiming to make automated screening part of routine oncology workflow rather than an extra step coordinators have to remember to take. Advocate’s public materials frame the partnership explicitly around equity: expanding trial access to patients who might otherwise never be flagged, rather than only the subset who happen to see an oncologist already plugged into a trial network.
The optimistic view: closing cancer research’s access gap
Clinical trial access has long skewed toward patients treated at large academic cancer centers with dedicated research infrastructure, leaving many community oncology patients unaware that a relevant trial even exists. Proponents of automated matching argue that tools like Lind could help close that gap by making systematic screening the default everywhere a health system has an EHR, not just at flagship research hospitals. Faster, broader trial enrollment also matters for drug development timelines — slow accrual is one of the most common reasons oncology trials miss their completion targets, and AI-assisted matching could meaningfully shorten that runway if it scales.
The skeptical view: agreement rates and real-world friction
A 94% agreement rate with clinical reviewers is strong, but it also means roughly one in seventeen assessments diverged from a human’s judgment — a gap that matters when the stakes involve steering seriously ill patients toward or away from experimental treatment. Reviewers of automated matching tools in oncology have also cautioned that identifying an eligibility match on paper is only the first step; patients still need to consent, may live too far from a trial site, or may not qualify once genomic testing or physical exam findings are incorporated in person. Some clinical researchers additionally point out that even a well-validated tool depends on complete and accurate EHR documentation to work, and gaps or delays in charting could cause the system to miss patients just as easily as a human reviewer might.
What to watch next
Advocate Health has not yet published enrollment figures showing whether the Lind partnership has measurably increased the number of patients actually joining trials, as opposed to simply the number flagged as eligible — a distinction that will determine whether this translates into real accrual gains. Also worth tracking: whether other large health systems follow Advocate’s lead in embedding AI-driven trial matching directly into EHR workflows, and whether pharmaceutical sponsors begin factoring AI-assisted screening capacity into where they choose to run trial sites. If the tool’s real-world performance matches its retrospective validation numbers, it could become a template for closing one of oncology’s most stubborn access gaps.
