Stanford freshman Yvonne Hong spent nearly two years working with hundreds of refugees fleeing Syria, Afghanistan, and South Sudan. Refugees often had medical problems, including physical injuries and psychological trauma. Despite injuries, individuals would delay or not seek out treatment because of the rising fear in authority figures. “People were afraid of exposing their identities,” says Hong.
Hong’s experiences became the inspiration for Soteria, a patient-centric telemedicine platform that connects refugees with confidential, low-cost, high-quality medical care. The app won the Stanford McCoy Family Center for Ethics in Society Most Ethically Engaged Hack at this year’s TreeHacks at Stanford. Hong, along with her teammates, Stanford students Tim Zekai Wu and Daniel Jeffrey Wu, and Justin Du, from Yale, had 36 hours to turn an idea into a real project. This year, the theme was hacking for the future: addressing deeply impactful problems and building creative, unique solutions.
“Especially at a hackathon, I feel like it’s very easy to be technically inclined. A lot of the times, the ethical implications of any project is more of an afterthought,” says Zekai Wu. “For us, it was really important to be able to consider that from the beginning, and design from that as a starting point.”
Soteria, named after the Greek goddess of salvation, aims to strike a balance between providing physicians essential information about each patient, crucial for accurate diagnoses while maintaining patient privacy such that individuals can seek medical help confidently without fear of retribution. The core platform consists of a camera and a heart rate sensor, attached to a raspberry pi. The team leveraged machine learning models to detect facial recognition and to obtain gender and age predictions. They decided to blur out people’s faces by applying a Gaussian filter over all regions of the image.
In an ideal scenario, an international organization like “Doctors without Borders” sets up a station with a pharmacy onsite at a refugee camp. Refugees open the application on a smartphone and connect immediately with a doctor, most likely based in Europe or North America. The physician then makes a diagnosis or treatment, allowing the unidentified patient to seek aid with the help of a nurse at the onsite pharmacy. The app is designed to automatically recognize basic classifiers of the patient to expedite the triage process.
From an ethical standpoint, the group contemplates the ethical pitfalls of their technology, including patient data privacy. “If someone who is showing signs of coronavirus, were to seek treatment on our platform, would we give the doctor a mechanism to unblur the image or to identify the patient? On the other hand, maybe the only reason a patient is using the platform is to protect their own identity, for fear of persecution or whatever that might be,” says Jeffrey Wu. While the group doesn’t have a concrete answer, their next step would be to design a proper patient consent form that both supports patients’ ethical rights and removes any concern about the confidentiality of the data. And if they had more time, the team believes it is necessary to take a step back and speak with groups, such as the World Health Organization, to determine what development would be most effective.
“Our goal is to provide a safe space for refugees to seek medical access. This should be as transparent as possible,” says Du.
To learn more about this project visit this page.