Earlier this month, we sat down with Dr. Peter Weyhrauch, VP, Human-Centered Intelligent Systems and Principal Scientist at Charles River Analytics to learn more about how AI and robotics are supporting medical training, decision support for clinical providers, battlefield medicine and cancer research to improve health outcomes for cancer survivors. Below are highlights from our conversation.
Tell us about Charles River Analytics healthcare initiatives.
Since 2012, we have been active in healthcare research applying AI to military medical training. Military medical care providers are under constant strain to learn new medical devices, procedures, and skills to ensure the best possible patient care. At Charles River Analytics, we have developed adaptive training systems that tailor themselves to individual provider needs that can reduce this strain.
Some of our work includes a game-based rehab app using augmented reality for people with balance issues, a VR headset with multiple sensors to aid diagnosis of traumatic brain injury, tech supporting cancer survivors, and a VR training system that will help first responders learn new decontamination protocols to minimize the risk of COVID-19 or other airborne virus transmission.
Tell us about your recently awarded contract from the National Cancer Institute?
We were recently awarded a contract by NCI to develop a privacy-preserving Wearable Health Informatics Platform for Extensible Research (WHISPER) in response to the technology and healthcare trends that we’re seeing. On the tech side, with mobile phones, wearables, connected devices, and IoT — people have become increasingly connected.
On the healthcare side, there is a trend toward out-of-clinic research and patient assessment using passive, continuously measured data streams to measure cancer and symptom progression. Through WHISPER, we are bringing data to researchers so they can streamline their population health studies and synthesize data across devices, participants, and studies.
What impact do you expect your WHISPER technology to have across both patient & provider uses?
Ultimately, we see this technology being very valuable in community clinics. Firstly, it can be a big part of creating care equity. For those that live in rural areas, for example, there’s no need to go to the clinic on a regular basis, the data can go straight to the caregivers who need to see it.
Secondly, it can give researchers and clinicians the ability to align objective markers with patients’ perception of symptoms, which we know can improve health outcomes in cancer populations. Allowing patients to control their health records and volunteer their data to research will open up new opportunities in population-level cancer research. Our collective challenge going forward will be to seize new opportunities while staying vigilant about maintaining the security and privacy of sensitive health information or unnecessarily adding to the burden of already overburdened health care professionals.
What type of work are you doing on AI and robotics that is impacting the patient/provider relationship?
Right now, some of our most interesting work involves robotics and AI to support battlefield medicine. This includes developing tech to support robots that evacuate casualties, and an augmented reality app designed to help medics monitor patients, diagnose injuries, and provide treatment in areas that are isolated or difficult to access. Our parallel work on medical training and our resulting knowledgebase is allowing AI to guide people making decisions in the field.
Can you share some of the latest healthcare tech advances Charles River Analytics is leading?
Unmanned systems can have big benefits in dangerous situations, like extracting injured warfighters from an active battlefield. We are using state-of-the-art perception technologies to identify casualties in an image and model their position and body shape in 3D so that a team of small autonomous robots can remove them from harm.
Our tech is giving these robots the ability to maneuver and manipulate objects under all sorts of operating conditions using dynamic world modeling, path planning, and localization onboard each vehicle. It also lets them share information with other manned and unmanned teammates so they can work together.
What plans do you have to further the effectiveness of AI and robotics on healthcare training and research?
Our ability to codify the reasons and procedures used for treatment in the field means we’ve got this data in a form that our algorithms can use to adapt and customize training in a medical simulation.Our algorithms accurately classify an individual’s skill and continuously adapt training content to meet their learning needs – maximizing both skill improvement and sustainment while reducing training time.
As we look toward the future, we’re interested in charting a course to deliver intelligent, adaptive healthcare based on the deep knowledge we have developed and encoded for medical training, decision assistance, and robotics. This will include investigating how robotics and vision and pose estimation algorithms can help guide treatment via telemedicine. This is an exciting time to be in the industry!