Selja Health

In trauma, the transfer decision often comes too late.

We're building clinical decision support for physicians, with physicians, to solve that problem.

Request a briefing See the approach

Retrospective validation across seven trauma centers · 1M+ records under analysis

A patient just arrived.

  • Stabilize the patient. Done.
  • Assess the patient. Done.
  • Maybe order imaging, maybe do labs. Done.
  • Talk to the family. Done.
  • Decide and coordinate on transfer. Still waiting on results above.

And many other patients at the same time.

The transfer decision, made earlier.

Today the transfer call is placed late, after a long sequence of evaluation, much of it spent waiting on information that would not change the call. Moving the decision earlier in the sequence is what we are studying.

Two illustrative timelines, not drawn to minutes, each from the patient arriving to the patient leaving. Today, the transfer decision comes about three-fifths of the way through, after stabilization, evaluation, imaging and the wait on labs; coordination follows, including the wait for EMS, and the patient leaves at the end. With Selja, the decision comes about a fifth of the way through, after stabilization and a basic evaluation; coordination follows, including the wait for EMS, the patient leaves about three-fifths of the way through, and the rest of the timeline is time saved.
Today
Patient arrives Patient leaves
  • Active clinical work : stabilize · evaluate · image · wait on labs
  • Coordination : phone calls · partial briefings · waiting for EMS
transfer call earlier
With Selja
Patient arrives Patient leaves
  • Active clinical work : stabilize · basic evaluation
  • Coordination : recommendation · structured packet (in build) · waiting for EMS
  • Time saved

Active clinical work

Placed earlier, the call waits only for the work that must happen before the patient can move.

Coordination

With the call made earlier, the calls to the receiving hospital and the wait for EMS start sooner.

How Selja reaches a recommendation.

120 injured patients arrive at a small local hospital.

24 clearly qualify for transfer under established guidelines.

For the rest, Selja estimates what specialist care each one needs.

Selja suggests transfer when this hospital can't provide that care, and the doctor decides.

A figure following 120 patients through four stages. All 120 arrive at a community emergency department, undecided. 24 meet documented criteria for immediate transfer, which decides them outright and estimates nothing. Of the remaining 96, Selja estimates that 48 are likely to need one of 12 specialty services (trauma surgery, orthopaedic surgery and neurosurgery among them) and that 48 need none. The facility in this example offers none of those specialties, so the recommendation is transfer for 72 patients and stay for 48.

The methods come first.

Nothing goes into clinical use before validation across hospital sites is complete, and no result is claimed before it has been reviewed. Methods and results will be reported through the channels clinicians already trust.

retrospective trauma records under analysis
1,000,000+
patients with detailed data collected first-hand
5,000+
Level I trauma centers
7

Cohorts under analysis in an ongoing academic research program; figures are current counts, not results.

Built by clinicians and scientists.

Trauma surgery, biomedical engineering, and machine learning research, with the clinicians who make this call advising the work.

Sambit Panda, PhD

Co-founder & CEO

PhD, Biomedical Engineering, Johns Hopkins. Builds the system and leads commercial strategy.

Amina Qutub, PhD

Co-founder & CSO

Burzik Endowed Professor, UT San Antonio. Research in computational modeling for medicine. Leads the scientific program.

COL (Ret) Brian Eastridge, MD, FACS

Senior Clinical Advisor

Served as the first Director of the U.S. Army's Joint Theater Trauma System. Advises on clinical protocol design, the surgeon-in-the-loop interaction model, and trauma system integration.

Clinical advisory board

Ronald Stewart, MD, FACS

Chair of surgery at UT Health San Antonio and STRAC board chair since 1996.

Alan Cook, MD, MS, FACS

Professor of surgery at UT Tyler and medical director of its trauma research center.

Elizabeth Scherer, MD

Trauma and critical care surgeon at UT Health San Antonio.

Mark Goldberg, MD

LeWinn Endowed Professor of Neurology at UT Health San Antonio and stroke specialist.

David Limon, MD

Trauma surgery resident at UT Health San Antonio.

Rachelle Jonas

Research operations manager at UT Health San Antonio.

Lauran Barry

Clinical research project manager at UT Health San Antonio.

Technical advisory board

Chris Rathbone, PhD

Associate dean and professor of biomedical engineering at UT San Antonio.

Joseph Houpt, PhD

Professor of psychology at UT San Antonio.

Jayanta Dey, PhD

AI researcher with the MATRIX AI Consortium at UT San Antonio.

Christian Cruz

AI engineer with the MATRIX AI Consortium at UT San Antonio.

Gloria Manavi

Researcher with the MATRIX AI Consortium at UT San Antonio.

Institutional affiliations are stated for identification only. They describe where a person works and do not imply that the institution endorses, sponsors, or is affiliated with Selja. Full statement.

What we don't do.

Three constraints we hold ourselves to, even when overclaiming would be easier.

We don't replace clinicians.

Selja supports the physician's judgment and never decides on its own.

We don't add to the cognitive load.

A trauma forces parallel streams of clinical decisions: ATLS, stabilization, family communication, transfer logistics. Selja puts what the transfer decision needs in one place.

We don't deploy models without validation.

The stakes are too high to deploy untested. Validation across multiple trauma centers comes before clinical use.

If you make this call, or answer it.

Sending-facility physicians, trauma medical directors, health systems, and research partners. If your work touches the trauma transfer decision, we would like to hear from you.

Email the founders

A call with the founders, usually within a few business days. By contacting us you agree to our Terms of use and Privacy policy.