For doctors treating trauma victims, diagnosing shock and internal bleeding early is essential. A team of researchers at Dartmouth are developing a novel device to help clinicians make quick decisions on the ground to determine the condition of their patients.
Recently awarded the $3 million Precision Trauma Care Research Award from the Department of Defense’s Combat Casualty Care Research Program, the researchers will investigate methods to diagnose internal injury and shock using a combination of advanced sensors and machine learning.
“Our project is predicated on the idea that neither of these alone are actually succeeding, and that what we really need to do is combine state-of-the-art sensing with state-of-the-art artificial intelligence,” said medicine professor Norman Paradis, director of research in the section of emergency medicine at Dartmouth-Hitchcock Medical Center and a lead researcher on the project.
To bring these aspects together, the project team includes both doctors and researchers at DHMC, the Geisel School of Medicine and the Thayer School of Engineering. The joint team will investigate the creation of new, more effective sensing technology and develop a device that can be easily deployed in triage situations. The new device could be particularly effective in large-scale events with multiple patients in need of care, according to Paradis. The proposed device would then monitor the patient, setting off an alarm if signs of shock are detected.
“You have a bunch of people who don’t look externally sick, but could be internally injured, so you’d put our system on,” he said.
Engineering professor Jonathan Elliott is working on developing a new optical sensor to assist the device in detecting internal injury. Similar to the non-invasive pulse oximeter, the device will use light sensors to determine the amount of oxygen in the patient’s blood. Elliott’s new system will go a step further, using six spectrometers to monitor physiological changes connected to shock at three different locations on the patient’s body.
“We’re hoping to detect over time those subtle changes,” Elliott said.
Thayer professor of engineering and Geisel professor of surgery Ryan Halter is investigating electrical impedance sensors that could detect the precursors of shock. According to Halter, the preliminary phases of the project will be about collecting as much data as possible from a variety of sensors.
Halter said he hopes to be able to refine the amount of data collected, ultimately designing a device that will be both portable and wearable for patients.
“Ideally, we’re looking for somewhat low cost, easy-to-deploy sensors,” he said.
Engineering professor Vikrant Vaze is currently creating algorithms to process and draw conclusions from the data.
Halter, Paradis and Vaze will collaborate to produce the project’s first prototypes, using clinical research from Paradis and his team at DHMC along with data from several other institutions.
According to Paradis, the goal of the project is to determine which sensors are most effective and combine them into one device. The data received will then be fed into machine learning algorithms that evaluate the condition of a patient according to the chosen metrics.
The researchers plan to complete the project in a span of three years, developing initial algorithms this fall. The team will then create a clinical prototype next year. The last phase of the project will be collaborating with a medical technology company to create a smaller, more workable prototype for real-world application.
After clinical approval, the device will most likely be used first to treat soldiers suffering from trauma on the battlefield, according to Paradis. He added that the next phase of the project might be to seek approval from the National Institute of Health to investigate possible use of the device as a way to diagnose shock among patients with sepsis in civilian hospitals.
“Hemorrhaging in soldiers is a simple problem compared to sepsis in the U.S.,” Paradis said.
