SBIR-STTR Award

Background: Preterm births, those that occur prior to 37 weeks of gestation, are the leading direct cause of neonatal mortality and morbidity. More than 1 in 9 births in the U.S. are preterm, with rates that are disproportionately high among African-Americans and families living in poverty, regardless of race. Addressing the problem of preterm birth requires both accurate identification of the factors that put women at risk, and communication of that risk to women and their healthcare providers. The reduction in preterm births is a core mission of the NICHD Pregnancy and Perinatology Branch. Study aims: 1) Apply our novel machine learning algorithm, KCI-neighbors, to a large prospective cohort study to model adverse pregnancy outcomes among women with low income and English literacy, 2) use mixed-methods research, grounded in decision science techniques, to tailor the MyHealthyPregnancy (MHP) app to the specific needs of low income and low English literacy patients, and 3) conduct small- scale usability testing of the modified technology with a sample of peripartum low-income, low-literacy patients to determine the app's acceptability and interested in targeted intervention strategies. Innovation: MHP is the first mobile health app that combines machine learning, expert models, and behavioral decision research to provide pregnant women and their providers scientifically sound, highly personalized, and actionable feedback on individual-level risk. The machine learning algorithms are designed to learn from users as the app is more widely deployed, allowing for the identification of new causal pathways linking risk factors to adverse pregnancy outcomes. MHP targets specific engagement metrics (e.g. appointment attendance) to meet health system stakeholders' goals, enabling healthcare systems to meet performance targets, as well as decrease costs through reduced adverse outcomes. Methodology and expected results: We will employ statistical machine learning to model the risk of adverse pregnancy outcomes, complemented by qualitative mixed-methods research to identify the most important measures to include in the MHP app. We anticipate that usability-testing will show an engaging app, capable of capturing and communicating risks in our target population. Potential impact: This work will advance scientific understanding of the risk factors, needs, and implementation science required to reach pregnant women who experience the most difficulty engaging in the healthcare system. Moreover, it will help improve clinical practice through development of a tool, MHP, that can detect and communicate preterm birth precursor risk to both patients and providers.

Project Terms:

Problem: Depression during pregnancy affects approximately 10% of women and is related to low birthweight and preterm birth. Similarly, up to 9% of pregnant women experience intimate partner violence (IPV) and abuse, with over 41% of assaults resulting in physical injury, and almost 30% requiring medical treatment. When untreated, these risks cost health systems at least $50B/year. Mitigation has proven difficult, where women are reluctant to disclose during clinical visits, and clinicians are unaware of resources. There are no integrated health technologies that enable timely disclosure of risks during pregnancy then aid in making decisions about risk mitigation. Fortunately, most women of reproductive age own a smartphone, and users report comfort disclosing health information to smartphones under the right conditions. Naima Health’s Proposed Solution: Naima Health is developing a digital health platform that pairs our MyHealthyPregnancy (MHP) smartphone application with an EPIC-integrated provider portal to (i) identify risks early in pregnancy, (ii) communicate those risks to women and their providers, and (iii) assist decision-making about risk mitigation. MHP identifies risk using ACOG-approved screenings, then helps patients and providers make real-time decisions about mitigation. The proposed solution aligns with the CDC’s priority of developing mobile app-based decision support systems for mental health and IPV screening, assessment, and referral. Proposed SBIR Work: In Phase I we developed expert and machine learning models to identify risks during pregnancy, then characterized issues facing Spanish-speaking women. In Phase II we extend these Phase I results using semi-structured interviews with patients and providers to understand site-specific requirements for psychosocial risk screening and referral, then update the MHP platform to meet those requirements (Aim 1). We then validate the updated platform’s performance using qualitative cognitive testing with patients and providers to ensure the platform meets site-specific requirements (Aim 2). Finally, we evaluate the platform’s acceptability and feasibility at two collaborating clinic sites, focusing on the rate of depression and IPV detected through the platform compared to historical rates, and the prevalence of risk mitigation actions measured through patient calls/click-throughs and provider referrals (Aim 3).Project Narrative Untreated psychosocial risks during pregnancy, such as depression and intimate partner violence, are significant adverse health outcomes for both parent and child, costing society more than $50 billion per year. Digital health tools for patients and providers can promote early identification of those risks to aid rapid intervention. The MyHealthyPregnancy (MHP) mobile phone application and provider portal combines decision science, medicine, machine learning, user-centered design, and app development to identify and mitigate clinical and psychosocial risks during pregnancy.