Project Spotlight: MyHealthyPregnancy

The costs of preterm births to the healthcare system and society are exceedingly large, and the incidence of preterm births is on the rise despite medical advances. This has created a need to provide reliable, personalized information to pregnant women about how they can reduce their risk of premature births. In response, researchers have created the “MyHealthyPregnancy” (MHP) smartphone app, which applies statistical machine learning techniques to comprehensive pregnancy data sets to improve the app’s patient-specific risk predictions of adverse pregnancy outcomes. Tamar Krishnamurti, PhD, CMU, is a leading researcher of this project along with Hyagriv Simhan, MD, of Magee-Women’s Hospital, and Alexander Davis, PhD, of Carnegie Mellon University. Tamar shares more about her experiences and the MyHealthyPregnancy app below.

Available for download in the App Store and on Google Play.

 

Please share a little about your background and your research experiences.

I am a behavioral scientist with a background in human decision making. I work as an Assistant Professor of Medicine and Clinical and Translational science at the University of Pittsburgh and maintain an adjunct position in the Department of Engineering and Public Policy at Carnegie Mellon. I use and develop methods in the social and decision sciences to examine problems at the intersection of health, risk, technology, and the environment. A lot of my current research focuses on risk perception and communication and the development of strategies to identify and intervene on maternal health risks.

What led you to the PHDA?

At the time of applying for PHDA funds, I was faculty at Carnegie Mellon, working with a cross-institutional team of decision scientists, physician-scientists, and applied statisticians. I was very interested in understanding how to leverage findings from big data analysis to provide a form of personalized medicine that could deliver actionable feedback to people based on their individual social, psychological, and clinical risk factors. I believe that well-designed responsive mobile health (mhealth) tools have great potential for promoting health equity around critical peripartum risks. Smartphone use is fairly ubiquitous, allowing for more equitable access. Additionally, embedding algorithms for risk identification can serve as a bias-minimizing approach to monitoring and communicating a woman’s symptoms to her healthcare provider. These goals seemed very aligned with the mission of the PHDA and so I applied for funding to further develop the MyHealthyPregnancy mobile health platform through the Center for Machine Learning and Health at CMU.

Walk us through your project.

Preterm birth, where delivery occurs before 37 weeks of gestation, affects 1 out of every 10 babies born in the US and is the leading cause of infant morbidity and mortality. It disproportionately affects black women in the US, making it a health equity issue, as well as a clinical issue. Preterm labor is challenging to predict, though, because there are multiple risk factors (maternal medical history, demographics, social determinants) and several clinical precursors, such as gestational diabetes and preeclampsia.

MyHealthyPregnancy (MHP) is a maternal mobile health platform, integrated with the health system, that identifies and communicates actionable risks, from a variety of clinical and psychosocial risk factors, enabling real-time interventions with the goal of improving patient outcomes and reducing healthcare costs. Our approach allows us to model each individual user’s risk of prematurity, among other pregnancy-related complications. It offers actionable feedback to women using it, connecting them to local resources and supports when a specific risk is identified. It also simultaneously shares relevant data with providers to support clinical decision making.

How do you and your project partners’ strengths complement each other?

Effectively modeling, communicating, and intervening on pregnancy-related risk requires statistical expertise to build models, clinical expertise to make them interpretable, and behavioral expertise to make them actionable. My colleague at CMU, Dr. Alex Davis, brings the machine learning experience needed to create risk models. Dr. Hyagriv Simhan, at Magee-Womens Research Institute, is one of the leading experts in prematurity. As a behavioral scientist, I have worked closely with pregnant and postpartum women to ensure that the way we communicate risk information is accessible. We wouldn’t have been able to produce the MHP platform without the combination of these different strengths. Our partnership with UPMC and UPMC Enterprises has also been key in being able to integrate MHP into the healthcare system, actually getting it into the hands of the women we aim to serve.

What are your project’s next steps?

We licensed our IP from CMU and Pitt and formed a spinout company, Naima Health, which is located in East Liberty. We are currently beta-testing our app with pregnant patients at UPMC. We have just finished a project for the Centers for Disease Control to create a culturally sensitive Spanish language version of MHP and are excited for our platform to reach patients in other health systems as well.

Please note that UPMC has a financial interest in Naima Health.