WellSpan scales clinical AI OS systemwide after strong radiology results
WellSpan Health is rolling out Aidoc's clinical AI operating system (aiOS) across all nine hospitals and more than 250 care locations. The expansion adds 21 new AI-powered care pathways across seven service lines, building on six pathways already in use. The move follows a year of measurable gains in radiology that freed time, sped alerts and tightened care coordination.
What changed inside radiology
Using just six aiOS modules, WellSpan analyzed more than 200,000 advanced imaging studies over the past year, including brain CTs. The platform flagged more than 10,000 potentially critical findings such as pulmonary embolism and brain aneurysms, with 96% of those findings available to radiologists within three minutes.
Wait time for analysis and prioritization of outpatient cases with critical findings dropped by 65.5%. The brain aneurysm module alone helped escalate an additional 22 cases per month compared with prior run rates.
Adoption and efficiency in the workflow
Radiologist adoption has been high: 80% use the system daily. The six modules delivered more than 650 hours of read-time efficiency gains and eliminated more than 900 hours of care delays, redirecting time to patient care instead of administrative lag.
Scaling from 6 to 27 modules
Based on these results, WellSpan is expanding from six to 27 modules across the enterprise. The goal: reduce bottlenecks in urgent diagnostic workflows, improve turnaround times and standardize high-quality care at scale. In a tight labor market with radiology staffing shortages, these efficiency gains help sustain volume without adding burnout.
Can AI be trusted in high-stakes settings?
Dr. Tony Aquilina, chief physician executive at WellSpan Health, underscored that every scan is still read by a radiologist. AI supports the workflow; it doesn't replace clinical judgment. The organization's approach centers on people, process and technology, with a focus on measurable outcomes and safe deployment.
Key results at a glance
- 200,000+ imaging studies analyzed with six modules
- 10,000+ potentially critical findings flagged; 96% available to radiologists within three minutes
- 65.5% reduction in wait time for outpatient cases with critical findings
- 22 additional brain aneurysm cases escalated monthly
- 80% of radiologists engage daily
- 650+ hours of read-time efficiency gained; 900+ hours of delays removed
- Scaling from 6 to 27 modules across nine hospitals and 250+ locations
Practical takeaways for healthcare leaders
- Start where time-to-alert matters most (e.g., PE, stroke, aneurysm). Speed to review changes outcomes.
- Orchestrate AI inside existing PACS/EHR workflows; don't create parallel processes.
- Track operational and clinical metrics from day one: alert-to-review time, escalation volume, read-time savings, downstream care delays.
- Pair every algorithm with a clear escalation pathway and ownership in the ER, inpatient and ambulatory settings.
- Maintain a governance forum with radiology, ED, neurology, IT, quality and risk to review false positives, model drift and adoption.
- Use wins to counter staffing pressure and reduce burnout by freeing high-value clinical time.
To learn more about WellSpan Health, visit WellSpan.org. For guidance on safe AI use in imaging, see the American College of Radiology's Data Science Institute at ACR DSI.
If your teams need structured upskilling for AI literacy, workflow integration and governance, explore role-based programs at Complete AI Training.
Your membership also unlocks:
