Rethinking How Clinicians Interpret Patient Data with Systema Med
Team: Julio La Torre, Garman Xu (Tandon ‘27)
About the Venture: Systema Med is developing a clinical platform that interprets continuous streams of patient vital data rather than isolated measurements. The system aims to support faster and safer intervention through real-time decision support and automation.
Validating the Idea: What 50 Conversations Taught Us About Oxygen
Oxygen is one of the most widely administered therapies in medicine, yet its delivery is largely manual, inconsistent, and reactive. As, first and foremost, care providers at the hospital, we saw firsthand how often patients would get improper amounts of oxygen and recognized that this was an issue we wanted to investigate when we entered the NYU Entrepreneurs Challenge. Hyperoxia, a condition where a patient has an excess amount of oxygen, is a byproduct of outdated techniques and we believed we could build something better.
To better understand the issue, we spoke with approximately 50 clinicians including PACU nurses, anesthesiologists, pulmonologists, respiratory therapists, hospitalists, and ICU staff. Over the course of the challenge, we learned more about the clinical discourse surrounding hyperoxia and how to raise awareness for it.
Assumption #1: Clinicians Fully Appreciate the Risks of Hyperoxia
The literature around hyperoxia is clear, showing association with increased oxidative stress, lung injury, and worse outcomes in certain populations. Surprisingly, many clinicians were not fully aware of these dangers. While specialists such as pulmonologists and intensivists were familiar with the data, many frontline staff and even some physicians did not view hyperoxia as a major clinical risk. When oxygen is often perceived as inherently benign, we realized there was also an educational gap to cross where, in addition to minimizing hyperoxia through our solution, we must also spread awareness grounded in data, outcomes, and patient safety narratives.
Assumption #2: The Biggest Pain Point Is Escalation
We initially believed doctors and nurses would be most worried about not giving enough oxygen fast enough, but it turned out that taking patients off oxygen and reducing levels was a larger pain point because of the time-consuming nature of the task. Through our talks with PACU nurses, low staff bandwidth and competing priorities resulted in patients remaining on supplemental oxygen longer than necessary. It became clear that to solve oxygen titration, we must also address the operational burden around it.
Assumption #3: Automation Would Be Universally Welcomed
As an AI startup in medicine, we needed to understand whether the current landscape welcomes autonomous systems. Many clinicians were enthusiastic about automated titration due to its immediate benefits and precise control, but others expressed concerns about autonomy, trust in algorithms, and edge cases. We recognize the importance of these perspectives because they remind us that adoption in healthcare is never just about performance. As a result, we are building our product by prioritizing these concerns first, including clear override capabilities, guardrails, and a smoother integration experience to maintain transparency and retain human control.
The Path Forward
These conversations only strengthened our conviction that oxygen titration is both a clinical and operational opportunity. Oxygen is administered to tens of millions of patients annually and we truly believe that even small improvements in precision, time efficiency, and consistency could scale meaningfully. Not everyone agreed with our ideas during validation, but learning from real voices within the hospital system allowed us to take more perspectives into consideration as we actively build our vision.
