Researchers in Penn State's Colleges of Engineering and of Medicine aim to reduce error rates in central-line placement by designing and implementing a novel robotic training program that uses life-like manikins and real-time feedback to simulate patient scenarios. Their recent study, published in the Journal of Surgical Education, demonstrated a significant reduction in placement errors and in complications, including infections and blood clots.
Currently implemented at the Penn State Milton S. Hershey Medical Center and at Cedars-Sinai Medical Center in Los Angeles, with additional partner agreements planned, the team has trained more than 700 physicians and is currently adding to that total with 200 new physicians trained each year. The team is also expanding the medical simulation training at participating hospitals to include colonoscopies and laparoscopies.
Professor Scarlett Miller, Professor Jason Moore, and Jessica González Vargas explain how a novel robotic training program designed by researchers at Penn State uses life-like manikins and augmented reality to train the next generation of physicians. Students engage in a high-fidelity environment to simulate real-world scenarios for common procedures including central line placement, laparoscopic surgery, and colonoscopies. Using AI and machine learning, the medical simulator provides automated feedback in real-time.
Expand
Re-engineering medical training
In 2013, Scarlett Miller and Jason Moore, both professors of mechanical engineering, partnered to investigate how they could apply their individual expertise in simulation, robotics and engineering design to solve problems in healthcare.
While sourcing data and speaking with clinicians, the topic of reducing complications for central venous catheterizations (CVC) procedures surfaced.
"I was shocked to learn that new physicians who have performed CVC procedures fewer than 50 times were more than twice as likely to commit an error," said Miller, who is also a professor of industrial engineering.
A human factors and design engineer, Miller's research focuses on designing machines that incorporate user needs to reduce human errors, particularly within medical applications. When she and Moore learned that the source of these complications was a lack of practice and feedback for new physicians, the team set out to engineer a device that would simulate the CVC procedure and provide automated feedback.
"Simulation-based medical training utilizing manikins exists for certain procedures, such as central venous catheterization and colonoscopy, but current models tend to be restricted to specific scenarios and lack adequate feedback mechanisms," Moore said. "Training systems with physical tools, such as rubber skin, are not resilient against hundreds of residents practicing thousands of times. These training systems fall short."
Moore and Miller's development, funded by the U.S. National Institutes of Health, combines a manikin that can simulate multiple scenarios, provide real-time feedback, and withstand the repetitive training needed to reduce complication rates.
Residents can use ultrasound to monitor their progress as they thread a needle into the target vein. The system provides immediate alerts to correct positioning and other potential errors. Critically, the researchers said, the training system can also simulate multiple patient types by adjusting parameters related to weight, height, sex and age.
"All of our devices provide a diversity of anatomy," Moore said. "If we're able to simulate lots of different types of anatomies and individuals, that wealth of knowledge will translate much better into the clinic by exposing physicians to the variety of patients they will encounter."
Standardizing education when patients are non-standard
Recalling his time as a resident, Sanjib D. Adhikary, professor of anesthesiology and vice chair for research and innovation in the Department of Anesthesia at Penn State Health, witnessed a central line lodged in a patient's airway. A cardiac surgeon was nearby and available to help him safely remove the catheter with no lasting harm to the patient, but the experience left a mark on Adhikary.
"There are devastating complications associated with central lines," Adhikary said, explaining that several vital structures are at risk during central line placement, from blood vessels to lungs to the airway. "Residents may feel confident from learning in class or observing, but that confidence can falter when faced with specific or compounding factors in a critical situation."
Miller and Moore approached Adhikary with their early robotic simulation ideas, and he knew it was the best pathway for residents to build confidence and comfort with procedures - without risking lives. He explained that residents' training is typically in high-pressure situations, with a lot of opportunity for miscommunication.
"When you're doing a procedure and directing a new resident, you may say go left, and they have to figure out if they mean your left or their left or the patient's left - it's easier to navigate in a controlled environment," Adhikary said. "That's why simulation is the way to go. With medical simulation, education can be propagated more easily and safely before they try it on a patient."
He also noted that simulation training in health care is extremely cost-effective when considering the resources involved in teaching, and it could provide a better teaching experience. A single supervising physician may be able to train about 10 people a year, while the simulation program can train more and provide objective feedback. Residents are also able to train on their own time and even record their practice for their supervisors to provide personal feedback.
"We used to teach that you see one, do one, and teach one - that was the method because there was nothing like this available," Adhikary said. "The training ensures that everybody is talking in the same language, that they have the same training. By using the simulator consistently in a standardized way, we are increasing the value of education and of medicine."
Expand the system, expand the skills
Based on their technology, the researchers have been granted two patents, with three patents pending and more in development. In 2018, Miller and Moore co-founded Medulate, a start-up company aimed at increasing the efficiency and effectiveness of medical education by commercializing the Penn State-generated intellectual property around medical simulation.
Since founding the company, they have continued to develop and assess novel medical training simulators that provide validated patient benefits. Miller was approached by doctors at Cedars-Sinai Medical Center in California interested in integrating the central line training in their residency program after noticing the same preventable errors that the simulator targets among their trainees. Over the last five years, Cedars-Sinai has implemented the new resident training program, demonstrating its scalability and application beyond Penn State.
Now, the team is working to broaden the impact of the training system to include experience with sterile technique and error prevention and recovery. With the use of artificial intelligence, the system can monitor sensitivities such as the sterility of the surgical field - the area surrounding the site of the procedure.
"Ultimately, we need to scale. We need to get this out there to the broader medical community - we have the opportunity to impact so many lives in a positive way," Miller said.
This is just the beginning for Miller and Moore. The team's research and training program is expanding to include colonoscopies, using augmented and virtual reality simulation. They recently received $1.7 million from the National Institutes of Health to develop, deploy and assess a colonoscopy training system.
"If colon cancer is caught early, there's a much better patient prognosis," Miller said, noting that 15 million colonoscopies are performed each year in the United States. "But research shows that it takes completing the procedure 200 times for a physician to become proficient. In residency training, physicians are only required to complete 50 colonoscopies. We're trying to reduce that learning curve so that every clinician doesn't need to practice - and potentially miss early cancer signs - on 150 patients."
With the medical simulation training program, medical students, residents and other practitioners can practice these delicate procedures countless times in a risk-free environment. Their trainees enter the medical field confident in the procedures, and, more importantly, more experienced, resulting in better patient outcomes.
"Historically, there's been a gap between education and clinical practice - and we're closing it," Miller said. "In terms of clinical significance, we can show that we are saving lives."
Professor of industrial engineering and of mechanical engineering and director of the Center for Research in Design and Innovation, Penn State College of Engineering
- Ergonomic product design, design cognition and human-computer interaction
Professor of mechanical engineering, Penn State College of Engineering
- Minimally invasive medical instruments
