The company VisualDx, which makes desktop and mobile apps for physicians that help them recognize skin symptom patterns in patients with various skin tones with an assist from artificial intelligence, launched a larger-scale effort in collaboration with Project Impact, the Skin of Color Society, and the New England Journal of Medicine Group. Project Impact aims to raise awareness about misdiagnosis and share resources and strategies that help physicians more accurately diagnose disease in Black and brown skin.
“We’re focused on all kinds of bias — racial and gender bias — and the reasons doctors make mistakes,” says Art Papier, MD, a dermatologist at the University of Rochester in New York and the founder of VisualDx.
“The key to good AI is good data,” says Dr. Papier. “We’ve been very purposeful since the inception of our company to collect imagery from people of all skin tones.”
Papier says he was inspired to create VisualDx after witnessing a misdiagnosis early in his career. When he was in training as a dermatology resident more than 20 years ago, he was in the ER when a patient with dark skin came into the intensive care unit with a severe rash caused by a rare condition called Stevens-Johnson syndrome.
The patient said the rash had started two weeks earlier, but other healthcare professionals hadn’t recognized its severity because it looks subtler on brown skin, Papier recalls. “That made me get committed to the idea of getting as many images as we could find in patients of color,” he says.