Hi Hive

Olga Kargopolova10d ago

Dr. Ferdi, welcome to the community! The perspective you’re bringing is very different from what many of us are building around, and that’s exactly what makes a founder community more interesting and useful. Would love to see what you’re building and what you’re learning.

Stacy Wycoff10d ago

Ferdi, running a hospital while also building the tool meant to catch what the clinical team might miss is an unusual position to build from. Most people designing clinical decision support have never had to explain a missed early warning to a board or a family. Where does the resistance to trusting an early risk signal usually come from more, the doctors on the floor or the administrative side above them?

Hello Dr. Ferdi, and welcome to the community! Please keep us posted on what you're building. I am interested in learning what clinical areas you are targeting first - critical care, outpatient, or something else?

Sergey Kargopolov9d ago

Welcome to the community, Dr. Ferdi 🙋🏻‍♂️

Stacy, you've asked the question that separates real clinical AI from the hype.

Here's what I've learned from running both sides: resistance isn't really about the alert itself. It's about trajectory.

And thanks for the welcome message guys, appreciate it.

So, when you show a clinician a single data point, "Patient X is at risk", that's noise. Every patient is always at some risk. But when you show trajectory, "This patient's deterioration pattern matches documented pre-sepsis signatures from 48 hours ago. We caught it 8 hours before conventional indicators would have flagged it", that's actionable intelligence I aim for.

On the floor: clinicians become allies when they see trajectory. They stop missing the early signals that they should have caught. That's painful, but it's motivating.

On the administrative side: trajectory solves the governance problem. With a 10-year immutable record of when we knew what, you can answer the hard question: "Did acting on this signal prevent harm?" Even imperfect data gives you accountability.

Our work at Sentra centers on exactly this, building algorithms that track clinical trajectories over time, not isolated alerts. Doctors can now stop deterioration instead of managing its aftermath.

That's where the resistance dissolves, well I hope -)

Jinny Moon7d ago

Really interesting area to work in, especially the focus on clinical trajectories rather than just one-off AI recommendations. Tracking how a patient is changing over time seems much closer to how real clinical decisions actually need to be made. I also like the emphasis on keeping doctors in the loop and treating AI as decision support rather than a replacement. Looking forward to seeing what you share about the parts that work, and especially the parts that don’t.

Yes, thank you, everyone.

I truly appreciate it. That is what I hope for as well, though anyone working in the healthcare industry knows there is always another layer to test, and plenty of bureaucracy to navigate, especially in my country. Please wish me luck. I still have to work through that process and continue testing my equations.

As the pilot testing session is underway, it will be another two months before I receive my results. So far, and as a self-disclaimer, we have carried out some tests in a small primary healthcare setting, and we observed deteriorations in patients with CVA symptoms in five to six cases, which makes me very encouraged about it.