How girls and women can help lead the effort to leverage artificial intelligence (AI) to combat preterm birth and related disparities

Dr. Jelliffe-Pawlowski is an Associate Professor of Epidemiology & Biostatistics in the UCSF School of Medicine and is Director of Discovery and  Precision Health with the UCSF California Preterm Birth Initiative


Today is the International Day of Women and Girls in Science”. In honor of this day I was asked to write a story on how girls and women might contribute to the field of artificial intelligence (AI) and pregnancy by the conveners of the upcoming symposium on “Maternal Health and Artificial Intelligence” (to be held this Thursday and Friday, February 13th and 14th in Boston, Massachusetts). I am thrilled to be included in such a meeting and especially grateful to see a number of other women researchers beginning to focus in this space.

This request got me thinking a lot about my work as a researcher in the field of preterm birth and how girls and women might take more of a leadership role in the growing AI space. We currently find ourselves at a crossroads in the field of preterm birth given the deep, deep need to figure out some way to turn the curve on rising rates and disparities. For example, in California Black women experience preterm birth at a rate that is 65% higher than that of White women. Preterm birth remains a leading cause of death and disability in infants, children, adolescents, and adults worldwide. As women and mothers, many of us feel that we have a profound personal stake in this field.

 

Rates of preterm birth are going up in California in all groups except in White women.

Rates of preterm birth are going up in California in all groups except in White women.

 

I am not only a researcher in this field, but I am also the mom of a daughter who was born preterm after I was diagnosed with preeclampsia and had to be induced. In many ways, when my own professional journey collided with my personal health, my work got more focused and the dedication to my research deepened, working to turn discoveries into interventions. As we encourage more girls and women to enter this field it will be essential that we encourage them to bring both their professional and personal passions into their work. This, I would argue, is where AI has the potential to help transform the field by encouraging the examination of different types of data and differing perspectives contributed by individuals with widely differing professional and personal experiences, preferences and ideas.

Given the role that technology now plays in all of our lives, one can argue that AI is more approachable than ever before. In medical research, most AI work falls into groupings that include, for example, machine learning, probabilistic reasoning and neural networks. It can be argued that all of these analytic approaches are much more familiar to girls and women of younger generations than those of us who grew up in say the 1980s, given at least a familiarity with functionality. After all, pop-up ads that result from tracked spending patterns are closely tied to machine learning and probabilistic reasoning and facial recognition software that opens many of our smartphones often uses neural network approaches (for more information about AI in medical research a good resource is the Elsevier report on AI in medical research published in 2018).

As AI has become more approachable, more people have been and will continue to be drawn to the field including girls and women. The potential this continuing trend holds for a field like preterm birth is that as more girls and women see their own personal links and experiences in the work, the more drive and determination there will be to find answers and to develop new interventions. For Black and Brown girls and women, this impetus has the potential to be especially profound given the disproportionate and devastating burden of preterm birth in families and communities of color. The involvement of girls and women of color in this growing push to understand preterm birth using the best available technology has particular urgency given the risk that AI poses in terms of increasing bias in medicine rather than on addressing and decreasing it if studies and data are not approached with these populations in mind.

Our group, the UCSF California Preterm Birth Initiative (PTBI-CA), and other groups focused on preterm birth and other pregnancy outcomes, are starting to leverage AI in order to bring multiple types of data together and to look for novel patterns that we have never been able to examine or test due to the hundreds of thousands of comparisons and computations that such an effort would require. I would argue that this approach, in particular, holds promise for the field given that more girls and women may see their own interests and expertise reflected in the questions being asked and the kinds of technology be used to ask and answer questions. I am particularly proud that in partnership with the UCSF Newborn Brain Research Institute, the UCSF/NIH Eureka platform, and Oda (omics data automation), the PTBi-CA is working to launch a research platform that will include embedded AI in order to accelerate research on preterm birth and associated complications. The platform is called “HOPE”, which stands for “Healthy Outcomes of Pregnancy for Everyone”).

HOPE Graphic

The UCSF HOPE platform will be comprised of three specific components: a data collection and upload arm, a data storage and sharing arm, and a data analysis and visualization arm that includes AI. A focus on resilience as well as risk is also a key element of the HOPE platform in that we believe there is a great deal to learn about preterm birth, complications of prematurity and other obstetric, newborn and pediatric outcomes by also focusing on healthy outcomes in high-risk women and infants. From the perspective of encouraging more girls and women to contribute to our field, this perspective of also focusing on health expands opportunities for collaboration even further. It is clear that we need to understand health in pregnancy better and health in preterm babies better if we are to identify in-roads to preventing preterm birth and related complications. While AI is arguably an essential component if we are to understand these patterns, so is the energy and perspective of interested and excited investigators – particularly girls and women, whose contributions may be key to truly unleashing progress and solutions.