Delayed diagnosis inspires student to create AI tool for mental health
Each year, Loc Phan, a physics major in the Eberly College of Science, would travel from his urban home in Ho Chi Minh City to the rural Vietnamese province of Vinh Long and visit family members. One summer, Phan said he observed something different about one of his relatives. The events that followed ignited a journey of seeking change that brought him to the Nittany AI Challenge and gave him the chance to create a tool for mental health in rural communities.
“I noticed my uncle had very unusual behavior,” Phan said. “He kept talking to someone, he said that he heard stuff from people, but nobody really knew whom he was talking to.”
Phan said the family grew worried when his uncle’s symptoms persisted, and they did not know what was causing his alarming mental state.
“The people of the rural countryside had speculations. Their spiritual-based theories explained his behavior as conversations with ghosts or dark spirits,” Phan said. “Unfortunately, we did not receive the medical advice or treatment options desired. Instead, the doctor attributed his behavior to overwork and recommended rest. The doctor sent him home.”
It was not until weeks later that Phan’s uncle was eventually able to be brought to a bigger hospital in the city. There, doctors gave him the diagnosis of schizophrenia. After that summer, Phan said he knew there was a big problem that he needed to solve and wondered what could be done to help solve the lack of diagnosis his uncle received.
“The local doctor sent him home because he said he was overworked,” Phan said. “But what if that doctor had a tool, maybe an AI assistant that could tell him, ‘Your patient has a serious problem, possibly schizophrenia, send him to the state hospital immediately.’ That is how I came up with the solution.”
Four years after his uncle was diagnosed with schizophrenia, Phan saw a flyer in the HUB-Robeson Center for the Nittany AI Challenge.
“I saw that every team had to come up with an artificial intelligence (AI) solution that tackles [a real-world] problem,” Phan recalled. “This was the opportunity for me to make something that is meaningful to help with mental health and my rural area community in Vietnam but also in developing countries in general.”
Phan teamed up with nine other University Park students to form SchizophrenAI. Together, the group has worked on creating an AI-based platform that calculates quantifiable data to help undertrained medical workers reach diagnoses. Parker Sell, a computational data science major in the College of Engineering and SchizophrenAI team lead, describes how it is important to have objective measurements when dealing with mental illness.
“Currently schizophrenia is diagnosed with subjective measures,” Sell said. “There are a bunch of subjective markers that only very trained professionals can recognize. So, we are hoping to provide a test and then use AI to automate the test to allow for more objective measures.”
Phan said that while the experienced coders on the team worked on the AI product, he took on an ambassador role and facilitated conversations with doctors and healthcare professionals in Vietnam about the necessity of a product like SchizophrenAI to improve mental health care for people in rural areas. The conversations confirmed to Phan and his team that a solution like their AI product has yet to be implemented in the country.
“What we learned is that nobody has implemented this AI solution before, as far as I know, in the south of Vietnam,” Phan stated. “If our team can deliver an application to a mental hospital or rural area hospital, we could be among the first to run these tests on patients.”
The professionals that Phan spoke to shared potential challenges facing the team as they pursue introducing their application within the Vietnamese healthcare system.
“First, they said they liked it, but they also said they think it will be hard to make it happen,” Phan said. “We have to talk to lawyers, hospitals, government, and the people who regulate laws before we can implement our solution.”
The team’s short-term goals are to complete the final phase of the Nittany AI Challenge and finish their minimum viable product (MVP) by the end of August. They will present their MVP to a panel of judges who will reveal the winners who will receive a portion of the remaining funds from a $50,000 prize pool, during the “AI For Good Expo” at the HUB-Robeson Center on Sept. 8.
Sell described the team’s future goals for growing into a startup company and distributing their product to the areas of the world that need it the most.
“Our goal is to win the next phase of the Nittany AI Challenge and receive the required funding to continue working,” Sell said. “If we do, we want to be able to put this product into place and connect it with the right people within a rural community or busy clinic in Vietnam.”
Phan said his goal is to share the AI solution and help prevent future hardship for rural families like his own.
“I hope my story can help raise awareness and action toward better levels of care for those struggling with mental health in underdeveloped parts of the countries around the world,” Phan said. “I want people to acknowledge the current state of mental health and understand the awareness in developing countries is still lacking and there is a lot of space for improvement in funding for the future; a lot of people need help.”
To learn more about the Nittany AI Challenge and “AI for Good Expo” visit the Nittany AI Alliance, a service of Penn State Outreach.