UTMB cardiologists review an AI pre-screened list; as many as a third of heart failure patients flagged
Qualified Health's platform reads fragmented clinical records at the University of Texas Medical Branch and surfaces patients who may qualify for guideline-based care. Clinicians review every output.
- Organization
- Qualified Health and University of Texas Medical Branch
- Industry
- Healthcare
- When
- 2026
- Evidence
- Vendor case study
- Agents and tools
- Claude Sonnet 4.5
Reported outcome
In the first month, the platform showed that as many as a third of heart failure patients had opportunities for better guideline-directed medical therapy.
- as many as a thirdHeart failure patients with an opportunity
- dozensPatients in the initial wave of review
Caveats
Published by the company that sells the agent. Treat numbers as marketing until confirmed elsewhere.
- Published by Anthropic, which sells the model, with statements from Qualified Health and UTMB.
- 'As many as a third' and 'dozens' are loose figures. The page gives no patient count, no false-positive rate and no health outcomes.
- An opportunity found is not a treatment changed. The page does not say how many patients had their therapy changed.
- The page shows no publication date. It describes plans for the end of 2026.
- The estimate of 4–6 million unidentified Texas patients a year is context from the page, not a result.
What the agent does
Qualified Health built a platform on Claude Sonnet 4.5. It brings together notes, lab results, imaging and procedure records, and applies guideline-based clinical criteria to find patients who may qualify for proven treatments.
At UTMB's Sealy Heart and Vascular Institute, cardiologists log in and see a list of patients the AI has pre-screened, in areas like heart failure and valve disease. For each patient the platform shows the relevant details from the chart and traces each finding back to the source data.
Why this works
- The page quotes the founder: "There are no fully automated clinical decisions being made." Clinicians review every output against source data.
- Each deployment is staged: back-testing on historical data, automated evaluations, structured physician review, then a controlled rollout.
- Cardiology was chosen first because it has clear criteria, clear guidelines and proven interventions.
What to copy
- Use the model for search and retrieval across records. Leave the decision to the clinician.
- Make every finding point to its source record.
- Start in a specialty with clear written criteria.
Starter prompt
Written by us, not by Qualified Health or Anthropic. Not for clinical use without validation, approval and clinician review.
Guideline criteria: {criteria}
Patient record extracts: {records}
For each criterion, say met, not met, or not enough data.
Quote the exact record line behind each answer, with its date and source.
Do not recommend a treatment. Write "For clinician review" at the top.
Sources
- Qualified Health customer story — Anthropic, 2026
Added Nov 16, 2026 · Sources last checked Nov 16, 2026.
Get new use cases weekly
One short email. Sources included.
The newsletter is paused. Past issues stay in the archive.
Related use cases
Works Human Intelligence cut its HR browser agent's cost per process by up to 97%
AWS reports the browser operation agent's cost per process fell by up to 97% after prompt caching, shorter sub-agent prompts and a switch from Claude Sonnet 4.5 to Haiku 4.5.
Cohere Health's policy agent cut the time to digitize a clinical policy from 2 h 15 min to 1 h 35 min
Overall time spent on policy digitization fell 30 percent, from 2 hours 15 minutes to 1 hour 35 minutes per policy. Agent deployments went from 3–4 months to 2–6 weeks.
Bluesight's Prism Assistant: 20 health systems, recurring drug diversion reports in 15 minutes, not six hours
By Bluesight's internal measurements across 20 health systems: recurring reports 96 percent faster (about six hours to 15 minutes) and triage time down 90 percent.