Informatics Ph.D. candidate Mayara Costa Figueiredo has been selected to participate in the Doctoral Consortium at the ACM CHI Conference on Human Factors in Computing Systems (CHI 2020) in Honolulu next April. Her dissertation work, “Self-Tracking for Fertility Care: A Holistic Approach,” was one of 20 papers selected from 84 submissions.
“CHI is one of the top venues in my field, and the Doctoral Consortium is usually very competitive, so I was very happy when I got the news,” says Figueiredo. This will be her first time attending the conference, which attracts thousands of international attendees annually. “It will be really nice to attend the conference and explore all the different kinds of work that will be there.” She is also excited about the location. “The other part that is really nice is that it’s going to be in Hawaii,” says the native of Brazil, “and I’ve never been.”
Of course, she is mostly focused on the consortium, where a panel of distinguished researchers will provide feedback on her dissertation. “I’m expecting to have discussion about my research and learn about other people’s research, and I think it will be very productive,” she says.
Her research interest is patient-generated health data in general and fertility tracking in particular. Working with her adviser, Associate Professor of Informatics Professor Yunan Chen, Figueiredo has analyzed data from online forums as well as 31 popular fertility apps. “We found that the apps aim to offer support for multiple goals — for example, trying to conceive, avoiding conception or tracking your menstrual cycle — but often the focus is on conceiving, and people complain about that,” she says. The main complaint has been the lack of support when transitioning between goals or when experiencing an unexpected event, such as a miscarriage. “It can be emotionally difficult if you just lost a baby, but the app keeps sending messages about the pregnancy,” explains Figueiredo.
She has also noted that while the apps offer opportunities to track everything from symptoms and mood to water intake and exercise, the visualizations are usually limited to information related to conception. Most have a calendar as well graphs to track temperature or ovulation, but not everybody who is tracking their cycle wants to get pregnant. “You may just want to see if your migraines match with your cycle,” says Figueiredo. So, she is exploring different designs and hopes to test some prototypes with focus groups.
She also considering how best to design for transparency, especially given the increase in apps claiming to use artificial intelligence. “The algorithms are always proprietary, so we don’t know how the algorithms make predictions or what data they use,” she says. “We want people to trust the recommendations, but when it’s something as delicate as fertility, which can have serious consequences, like an unplanned pregnancy, we want some sort of balanced trust.” The goal is to design the tools in such a way as to offer support to users while still making them aware of any limitations.
Hopefully, researchers at the Doctoral Consortium will help Figueiredo fine-tune her work, and Honolulu will be one of the final stops on her journey to becoming a professor and reaching her goal of using computer science to help the broader public.
— Shani Murray