The National Institutes of Health (NIH) will award up to $400,000 to individuals or groups who design an effective method for analysing a large data set of first-time pregnancies. The innovative methods also need to identify risk factors for adverse outcomes, such as hypertensive disorders, diabetes and infection. A total of $50,000 will be awarded to each of the seven winners designing the most effective means to analyse the data. An additional $10,000 will be awarded to the top five winners whose methods identify risk factors in disadvantaged populations.
The Decoding Maternal Morbidity Data Challenge will be administered by NIH’s Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD). Using computational analysis, data mining, Artificial Intelligence (AI) and other methods, winning entrants will need to devise ways for analysing the vast store of participant data from the Nulliparous Pregnancy Outcomes Study: Monitoring Mothers-to-Be.
The study consists of a racially, ethnically and geographically diverse sample of people beginning in the sixth week of pregnancy and continuing through delivery. The study was established in 2010 and has compiled data on more than 10,000 pregnant women. The data were collected from interviews, questionnaires, clinical measurements, patient charts and biological specimens. The study aims to identify pregnancy risks for women who have not given birth previously.
NICHD Pregnancy and Perinatology Branch stated that without a prior pregnancy for comparison, identifying risks for adverse pregnancy outcomes is difficult. The study has provided important information on the health needs of this unique population, and they look to the Decoding Maternal Morbidity Data Challenge to identify even more effective ways to harness these powerful data.
Pregnancy complications, or morbidity, may result from conditions women have before pregnancy or develop during pregnancy. It is difficult to estimate the effects of pregnancy complications on maternal and newborn outcomes because they encompass a broad range of conditions that vary in severity. According to the U.S. Centers for Disease Control and Prevention, severe maternal morbidity includes unexpected outcomes of labour and delivery that result in significant short- or long-term consequences to a woman’s health.
The Decoding Maternal Morbidity Data Challenge seeks to address maternal mortality and morbidity by focusing on its underlying causes, such as obesity, mental health issues and substance abuse disorders. Limited access to health insurance and health care are also contributing factors. Because maternal mortality and morbidity affect Black and Indigenous/Alaskan Native women at a much higher rate than other groups, applicants are encouraged to develop methods addressing the needs of these communities.
A panel of federal employees serving as judges will review submissions based on the following criteria:
- Proposal quality: Is the proposal complete? Does it address the challenge? Is the information presented clearly?
- Problem comprehension and approach: Does the team have a clear understanding of the problem? Is there a clear explanation of how their solution addresses the challenge? Does the solution clearly address an issue related to maternal morbidity?
- Innovation: Is the solution unexpected or out of the box? Does it have the potential to move science beyond what was previously considered? Did they approach the data in a new or unique way?
- Potential Impact: Does the solution have the potential to impact large numbers of pregnant women? Does the solution have the potential to have a profound impact on a subset of the population? Does the solution address what could be a big problem?
- Holistic view of proposal: Sense of the overall viability, innovation, thoroughness, potential to complete a winning prototype.
- Applicability: Can the solution be extended or applied to other areas beyond maternal morbidity?
- Disadvantaged racial and ethnic communities: How does the solution impact adversely impacted populations? Is the impact clearly described and demonstrated? Do they demonstrate an understanding of issues around maternal morbidities within in that population?
- Robustness: Is the solution robust enough to stand up to issues such as noisy or incomplete data? How large a data set is required to function correctly? Was the solution adequately and suitably tested? Did the solution perform well under the tests? Did they adequately describe their results and their rationale?
- Efficient/Ease of Use: How efficient is the implementation of the model? Is the data solution clear? Not overly complicated? Would it be straightforward for a researcher to move forward with the solution?