Abstract
The advent of artificial intelligence/machine learning (AI/ML) in medicine has opened opportunities for harnessing the power of ML to predict patient outcomes based on diverse data contained in the electronic medical records (EMR). Taken by themselves singular features represented by such diverse data are often not clearly predictive but combined in an ML modeling framework such so-called weak learners have the power to yield highly predictive models of health outcomes. We hypothesized that sociodemographic and basic, easily obtainable health characteristics of pregnant mothers strongly predict the risk of rehospitalization. Using the nuMoM2b dataset from NICHD and reproducible, open access AI/ML techniques we tested and validated this hypothesis. Among a large (n=10,038) and geographically diverse cohort of nulliparous women with singleton gestations, the present findings show that while some of these characteristics highlight the known disparities in health outcomes (race/ethnicity), most of them are modifiable through social policies and health counseling. These include access to and support during education, comprehensive health insurance, alleviating coming out of poverty, smoking, and unhealthy lifestyle habits. Across the four ML models deployed, the results are mutually reinforcing yielding a clear cohesive picture. This solution can be used to develop a prospective risk prediction for this important health outcome in pregnant mothers. The strengths of the presented solution are its focus on preventable maternal morbidities and risk factors, demonstration of the impact on adversely affected populations, and reproducible step-by-step executable, documented open-access code with relatively low computational requirements. This enables easy adoption and further development on this and other similar datasets to study the contributions and predictive power of various factors to maternal morbidities.
Competing Interest Statement
MGF holds patents on fetal monitoring. MGF advises and holds stock in Delfina, a maternal-fetal health AI company.
Clinical Protocols
https://dash.nichd.nih.gov/study/226675
Funding Statement
This study did not receive any funding.
Author Declarations
I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained.
Yes
The details of the IRB/oversight body that provided approval or exemption for the research described are given below:
ClinicalTrials.gov Identifier: NCT01322529 IRB of Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) gave ethical approval for this work. The data was collected originally by NICHD Division/Branch/Center: DER - Pregnancy and Perinatology Branch (PPB) with NICHD Research Networks and Initiatives: Nulliparous Pregnancy Outcomes Study: Monitoring Mothers-to-Be (nuMoM2b) Study Description: NuMoM2b studied underlying, interrelated, mechanisms of adverse pregnancy outcomes (e.g., preterm birth, preeclampsia, fetal growth restriction) in pregnant women with no previous pregnancy lasting 20 weeks-0 days or more estimated gestational age (nulliparas). The network included 8 clinical sites, subsites, NICHD, and a data center. Women enrolled early in pregnancy and were followed through delivery, with study visits at 6 weeks-0 days to 13 weeks-6 days, 16 weeks-0 days to 21 weeks-6 days, 22 weeks-0 days to 29 weeks-6 days gestation, and delivery. Data collected through interviews, self-administered questionnaires, clinical measurements, ultrasounds, and medical records review included demographic, psychosocial, dietary, physiologic, health, and pregnancy outcome information. Subsets of participants enrolled in sleep disordered breathing, sleep patterns and quality, and fetal adrenal gland substudies. Source Repository: DASH https://dash.nichd.nih.gov/study/226675
I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals.
Yes
I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance).
Yes
I have followed all appropriate research reporting guidelines and uploaded the relevant EQUATOR Network research reporting checklist(s) and other pertinent material as supplementary files, if applicable.
Yes
Data Availability
The code is available as follows: https://doi.org/10.5281/zenodo.5785159 https://martinfrasch.github.io/maternal_nu_data.nb.html The raw data can be obtained directly from NICHD via the NICHD DASH portal (https://dash.nichd.nih.gov/study/226675).