Duly Health and Care, a Chicago‑based, physician‑owned multi‑specialty group, said it used a unified data platform and artificial intelligence tools to tackle three operational challenges that threatened both financial performance and patient care: providers referring patients outside the organization, high readmission rates, and limited access to actionable business and clinical data.
Context: two payment models, heightened financial risk
Duly operates under two distinct reimbursement approaches: a traditional fee‑for‑service model, and a value‑based care arrangement under Medicare where the organization receives a fixed payment per patient based on risk. In the value‑based model, the group effectively assumes financial responsibility for downstream care costs when patients are hospitalized. That structure made reductions in readmissions and in‑network retention priorities for both clinical quality and financial stability.
What the organization built and why it mattered
To keep care within its system and to give clinicians and business teams easier access to data, Duly developed a set of tools on the Databricks platform. Key efforts included:
- A searchable "Physician Finder" that lets referring providers locate in‑network colleagues by specialty, procedure, language, gender and other filters.
- A conversational AI interface so clinicians can make natural‑language queries from mobile or desktop (examples reported include queries such as "find me a gastroenterologist taking new patients in Elgin").
- Predictive modeling aimed at identifying patients at high risk of readmission so care teams could intervene earlier.
Measured outcome
After rolling out the Databricks platform and associated tools, Duly reported a notable improvement in a key quality metric: a 29% reduction in readmission rates. The organization also cited improved ability to keep referrals inside its network and better access to actionable data for business teams.
| Metric | Reported change |
|---|---|
| Readmission rate | −29% |
Why this matters beyond one health system
Health systems nationwide are under pressure to control costs, improve outcomes and demonstrate value — particularly those participating in Medicare value‑based arrangements where financial risk is shared. Duly’s experience highlights how integrated data platforms and AI interfaces can support clinicians in finding in‑network care, enable real‑time decision support, and target interventions to reduce costly readmissions.
At the same time, reported gains like a 29% reduction in readmissions raise important follow‑up questions that health leaders and policymakers should examine: sustained results over time, the specifics of model development and validation, data governance and equity implications, clinician workflow impacts, and patient outcomes beyond readmission counts. The available reporting describes the technical approach and the headline improvement but does not detail baseline rates, absolute numbers, or how gains were distributed across patient populations.
The Duly case adds to a growing body of examples showing how health systems pair clinical expertise with data engineering to pursue both better outcomes and financial sustainability. As more providers experiment with similar tools, careful evaluation and transparent reporting will be important to separate durable advances from pilot‑specific effects.
