Mobile and intelligent – emergency blood analysis

Things need to move fast in an emergency. Making the right call in this setting can be a challenge for emergency medical services – especially when symptoms are ambiguous, which is the case if a patient has difficulty breathing or exhibits a cardiovascular or poisoning emergency. A blood analysis is paramount to deliver a fast and accurate diagnosis. This is where mobOx comes in.

In this MEDICA-tradefair.com interview, Prof. Stefan Müller talks about the benefits of a point-of-care blood analysis, explains the role artificial intelligence (AI) plays in this setting, and reveals his take on the future of emergency medicine.

Professor Müller, why did you develop mobOx?

Prof. Stefan Müller: Back in 2015, we began to develop an optical sensor system for hemoglobin measurement (CO-oximetry) as part of a project funded by the German Federal Ministry for Economic Affairs and Energy. The sensor was initially intended for a stationary blood gas analyzer. One stipulation was that the blood sample should not be pretreated prior to the analysis. This resulted in a complex laboratory setup with which we examined the optical behavior of numerous blood samples. For experimental purposes, we set multiple clinically relevant parameters to simulate diseases and patterns. We analyzed a multitude of different algorithms with the help of the resulting database. An AI-based approach had shown particular promise, and we decided to further develop it. The final tests revealed that apart from its high accuracy, it also proved very robust in an environment with fluctuating temperatures or vibrations. This is when we decided that our approach might also be ideal for mobile scenarios.

How does AI assist the analysis?

Müller: Our method detects changes in the color spectrum of the blood sample. AI-based algorithms help recognize patterns in the data and correlate various causes. Instrument drift refers to errors caused by temperature fluctuations and vibrations, for example, and is a major source of error that leads to inaccuracies in the measurement of blood parameters. Our algorithms are not susceptible to these influences. They also make it possible to accurately model complex, highly non-linear relationships in the data.

How can mobOx improve emergency medicine?

Müller: In many cases, a rapid, easy, and reliable point-of-care blood analysis can facilitate immediate targeted treatment and subsequently minimize any catastrophic long-term impact for the patient. For example, this enables health professionals to adjust the administration of medication during transport or decide to head for a specialized hospital clinic. MobOx gives emergency medical personnel an added reliable decision-making tool.

The device is still in development and very compact and robust thanks to the optical, AI-based measurement concept. It will approximately have the dimensions of a smartphone and its operation will resemble a blood glucose meter. It features a wide operating temperature range and is not susceptible to vibration. The requisite test strips have a long shelf life because they forego enzymatic components.

What are your plans for the device? What are some conceivable modifications?

Müller: We see mobOx as a mobile blood analysis platform for emergency medical services. The robust optical measurement principle combined with the AI-based algorithms make the system very flexible and easily expandable. In the medium term, we intend to add additional conventional blood gas analysis parameters to our system. In the coming months we plan to expand by including the pH and pCO2 parameters, which are frequently required in emergency medicine. The latter pertains to the partial pressure of carbon dioxide, which provides information about the lung and heart function. We will also study the interaction of cyanides as components of exhaust gases with hemoglobin in terms of a potential diagnostic approach.

What will the future of emergency medicine look like?

Müller: As in many other areas, I see an unstoppable digitization trend. This goes hand in hand with medical device networking and streamlined data management. In the future, I think it will be a matter of course that the emergency health care team will have received all relevant diagnostic data in digital format prior to the patient's arrival. AI-based assistance systems are another trend from my perspective as they assist staff in making a timely diagnosis and improve decision-making about the path of care.