Traceability Before Claims: An Honest First Look at Our Airway-Guidance Research

Procedural guidance will be commonplace for a majority of medical procedures in the next 5 to 7 years. The gap is not just care, it is that the quality of care rural communities experience today will keep falling further behind. I am from a rural community, which makes this something close to an obsession for me. At IntuBlade, I am beyond excited to share early research of ours in this realm.

I have not really talked publicly about this part of IntuBlade. Our intubation device is easy to explain: it is a single-use video laryngoscope that plugs into a phone. The software is harder to explain. In short, we continuously train our computer-vision models on new videos so that, one day, clinicians can perform a life-saving procedure better. It runs directly on iPhones with our disposable blades. That is it.

The Problem with "Black Box" Medical Tech

The Problem with "Black Box" Medical Tech

We recently ran 30 simulated intubations with the procedural guidance we have developed turned on. The question was basic: after the run, did we still have enough of the case to review it? Traceability is king in today's age of black-box software. Across these intubations we captured it all and, after reconciling the exports, put every field on display. That is what the paper is about.

It does not show clinical efficacy (yet). It does not validate the model. It does not say AI made anyone better at intubating (yet). It does say the system can preserve an airway attempt as something reviewable.

Why This Matters for the Future of IntuBlade

Why This Matters for the Future of IntuBlade

Airway video is not very useful if it dies as a clip on someone's outdated device or on a USB stick nobody ever reviews. To train computer vision, you need attempts tied to labels and metadata. You need the misses. You need review by operators who actually do the procedure. You need enough structure that the next model can learn from the last attempt.

This is early manikin data, now submitted for peer review. But it is the direction we are going, and I am beyond ecstatic to see it running in real time.

Awesome work from the team: Serge Klement and Benjamin Fedeles, M.D.

Partner with IntuBlade: Upcoming Fall 2026 Cadaver Studies

Partner with IntuBlade: Upcoming Fall 2026 Cadaver Studies

If you are a medical director running a cadaver lab or clinical simulation lab and training up-and-coming clinicians on intubation, or you know someone who is, please reach out. We are starting our cadaver studies this fall and would love more partners to build the future with us.