Virtual patients are increasingly used by medical schools to train their students
The computer programs are like flight simulators for doctors
Medical students monitor realistic patients for symptoms, interview them, order tests
Virtual training programs like i-Human could help train doctors in developing countries
Even before the examination begins, it’s clear Ann Martinez isn’t well.
Her breathing is labored. You can tell by the raspy sounds and jerky rise and fall of her chest. Clad in her underwear, she waits quietly for the doctor, letting out the occasional cough.
The physician starts by going over her history, asking a few questions and taking her vitals. Martinez, a smoker with a family history of heart disease, recently had knee-replacement surgery. She confirms she’s having trouble breathing and complains of some chest pain. While checking her pulse, the doctor notices that her heartbeat is faster than normal.
On their own, the symptoms are too common to reach any immediate conclusion. It’s possible she has a simple chest cold, but the signs could indicate something more dangerous, even fatal. More tests are needed.
There’s no risk of Martinez dying, however, because she isn’t real. She is a naturalistic, interactive virtual patient that lives on a computer screen. The simulation is part of i-Human Patients, one of a new generation of computer programs used by medical schools to train students and other professionals.
Like a flight simulator for doctors, i-Human presents cases as complicated, hands-on puzzles that require real medical skills to figure out. There is minimal guidance or hand-holding, leaving students to make hundreds of little decisions and conduct tests as they would when diagnosing a real sick person.
“A patient shows up in your office, and that’s it. That’s real life. You need to start asking questions,” said Norm Wu, CEO of i-Human.
I-Human says its program is an evolution of the popular first-generation virtual patients like InTime’s MedU, which are still used in most medical schools. Those programs and documents also ask medical students to make a diagnosis but typically with text, multimedia prompts and multiple choice options.
Medical schools also rely on a combination of actors and mannequins to help train doctors. Both have their advantages; working with an actor is great for bedside manner and interpersonal skills. But they can be expensive, and taking them home to practice isn’t really an option. With a cloud-based computer program, a medical student can practice his or her diagnoses anywhere there’s Wi-Fi. Data can be collected to let professors know how students are progressing, highlighting areas where they need improvement.
“If you’re a medical school professor, it’s very easy to test for fact-based recall,” Wu said. “How do you tell if somebody has figured out how to appropriately assess and diagnosis a patient with minimum error?”
Realistic high-tech training programs like i-Human aren’t just another helpful tool. They have the potential to address a shortage of trained doctors and nurses. The U.S. Department of Health and Human Services predicts a shortage of more than 20,000 primary care physicians by 2020, driven largely by the aging of baby boomers.
The issue is more severe in developing countries like India, where there is an average of one doctor per 1,700 people. In some Indian states the number is closer to 1 per 10,000, and hard-to-reach rural areas are far worse. Over the next 16 years, India’s government wants to train 1.6 million new physicians. Technology could play a huge role in the country, where doctors and brick-and-mortar med schools aren’t plentiful.
I-Human is working with 20 educators in India to localize the program, customizing patients and illnesses for the target market. (It has added an option for more conservative gowned avatars.)
More accessible diagnostic training for medical schools could also have an impact on the misdiagnosis rate. One out of every 20 outpatients is misdiagnosed in the U.S., according to an April study by BMJ Quality and Safety Journal – that’s 12 million cases a year.
Craig and Anne Knoche think that more realistic training could dramatically lower that number. The pair of Silicon Valley veterans launched i-Human with Corey Cerovsek in 2012 after creating standalone medical simulators for years.
Currently at use in 14 medical schools, the virtual patients are diagnosed as homework, group projects and tests and as a teaching tool in front of a class. The program has optional coaching tools like prompts, lessons and quizzes to keep beginning students on the right track, a common issue when there are thousands of possible questions to ask and hundreds of labs and tests to order.
Practicing on an avatar, no matter how realistic their gout or pneumonia, is obviously not the same as treating a real human. But a team of graphic artists has worked to make the avatars mimic real illness as much as possible. The team designs five or six new “patients” a week.
Each virtual patient has a name, a medical history, symptoms and an illness. They are a diverse selection of 3-D illustrations, with realistically rendered bodies, which makes it possible to see problems like jaundiced skin at a glance. Audio and animations tip students off to key details, like the sound of wheezing or the way a patient blinks.
Schools and other third parties can build their own cases using i-Human and share them with other customers, similar to selling apps in Apple’s App Store. The cases are peer-reviewed and subject to a review by i-Human’s two full-time staff physicians.
The virtual physical exams simulates, as much as possible on a screen, the tactical skills needed for things like measuring blood pressure and testing pupil reactions. To hear Ann Martinez’s heart, the student must know where exactly to put the stethoscope.
Eventually, if they’re on the right track, Martinez’s doctor will order a battery of tests that include a CT pulmonary angiogram. The result, a black and white image of real arteries, shows a pulmonary embolism. Pulmonary embolisms are frequently misdiagnosed and are the third most common cause of death in hospitalized patients.
If the medical student orders the correct course of treatment, Martinez will live another day and train more future doctors.