Introducing HEA3RT

November 17 marks the launch of the Stanford Healthcare AI Applied Research Team (HEA₃RT) and Simulation Lab on the Redwood City campus.

Back in the 1960s, health care professionals and members of the general public could learn how to do cardiopulmonary resuscitation using Resusci Annie, a dummy torso of a human on which to practice chest compressions and rescue breathing. Often learning under the watchful eye of Red Cross experts, regular people gained confidence in their ability to ‘do it right’ and left with a signed certificate. This was many people’s introduction to medical simulation.

We’ve come a long way.

November 17, 2020, marks the official opening of the Simulation Lab of the Stanford Healthcare AI Applied Research Team (HEA₃RT) on the Redwood City campus.

Steven Lin, MD, Executive Director of HEA₃RT and clinical associate professor, primary care and population health, explains that the simulation lab “is a space where we work with tech partners to do research and development for ideas and devices. Our simulation lab is uniquely equipped to allow extensive learning before live clinical testing. We want to go from code to the bedside. Right now, it involves Stanford partners and industry partners with 20 projects in process or completed.”

Lin explains the phrase ‘from code to the bedside’ in these words: “We use a mixed methods approach and work with clinical, operational, and technical teams to advance the development and validation of clinically relevant models, leveraging quality improvement, implementation science, design thinking, and traditional research methods. Workflow integration and scalability are at the forefront from inception to project completion.”

The lab has several spaces outfitted as clinic rooms or hospital bedrooms monitored by video and sensor technology, so that techniques and skills can be learned in a safe environment before going live in another setting. An open collaboration space, as described on the website, “can be used to build other clinical settings, simulate workflows, evaluate the utility of artificial intelligence (AI) technologies, and rapidly test and scale AI solutions for healthcare.”

Although he has yet to see the simulation lab firsthand, Department of Medicine Chair Bob Harrington, MD, knows a lot about it: “What an awesome idea! It’s learning how to do something in the environment in which you’re ultimately going to do it but recognizing that you have to acquire a set of skills. The first time you do CPR you’d rather it not be on a real person, so the Resuscitation Annie worked well. Today we think that some digital tools or advanced computational work like machine learning is going to be helpful, but we won’t know until we put it into the wild, if you will. On the way to the wild we can stop in the simulation lab and test to see what might need to be done to operationalize it most effectively.”

Harrington describes a situation in which the simulation lab could avert a disaster: “Say you have this idea that sounds fantastic, but once it gets applied in practice it’s a disaster. It slows everyone down and within a half-hour of opening the clinic, you’re already two hours behind. Having systems tested in the sim lab to consider that particular clinic and that particular doctor’s unique way of working could be quite valuable.”

There are ways to enhance collaboration among specialties, too. Preventive cardiology, for example, places a lot of the focus on blood pressure management, glucose management, activity encouragement, and lipid management. Many patients who need preventive cardiology are first identified in the primary care setting. Might there be digital tools to scan a medical record in real-time to inform the clinician that this patient has an atherosclerotic cardiovascular disease risk of 15%? Might that scan of the health record be integrated seamlessly into the workflow so the primary care doctor could loop in a colleague from preventive cardiology to start thinking jointly about how best to take care of that patient while the patient was still in the room?

We can stop in the sim lab and test to see what might need to be done to operationalize it most effectively.

Maybe members of the cardiologist’s team get involved to provide some diabetes or blood pressure education. Perhaps that can take place by video while the primary care doctor goes on to the next patient.

Quality improvement is yet another aspect that HEA₃RT will be well equipped to address. In an age of guideline-directed medicine, most of the needed elements of creating guidelines have been tested and proven. Investigators know how to formulate a question, answer the question through a clinical trial, and get the clinical trial data into a guideline. “Where we’re not so good,” says Harrington, “is how do we better implement guidelines in a typical practice. I think the sim lab could help us with that.”

There’s a bright new AI-inspired world waiting in Redwood City to help Stanford clinicians take even better care of their patients.