Identifying Risk of Stillbirth Using Machine Learning

Overview

Journal Am J Obstet Gynecol

Publisher Elsevier

Specialty Gynecology & Obstetrics

Date 2023 Jun 14

PMID 37315754

Authors

Tess E K Cersonsky

Nina K Ayala

Halit Pinar

Donald J Dudley

George R Saade

Robert M Silver

Adam K Lewkowitz

Affiliations

Soon will be listed here.

Abstract

Background: Previous predictive models using logistic regression for stillbirth do not leverage the advanced and nuanced techniques involved in sophisticated machine learning methods, such as modeling nonlinear relationships between outcomes.

Objective: This study aimed to create and refine machine learning models for predicting stillbirth using data available before viability (22-24 weeks) and throughout pregnancy, as well as demographic, medical, and prenatal visit data, including ultrasound and fetal genetics.

Study Design: This is a secondary analysis of the Stillbirth Collaborative Research Network, which included data from pregnancies resulting in stillborn and live-born infants delivered at 59 hospitals in 5 diverse regions across the United States from 2006 to 2009. The primary aim was the creation of a model for predicting stillbirth using data available before viability. Secondary aims included refining models with variables available throughout pregnancy and determining variable importance.

Results: Among 3000 live births and 982 stillbirths, 101 variables of interest were identified. Of the models incorporating data available before viability, the random forests model had 85.1% accuracy (area under the curve) and high sensitivity (88.6%), specificity (85.3%), positive predictive value (85.3%), and negative predictive value (84.8%). A random forests model using data collected throughout pregnancy resulted in accuracy of 85.0%; this model had 92.2% sensitivity, 77.9% specificity, 84.7% positive predictive value, and 88.3% negative predictive value. Important variables in the previability model included previous stillbirth, minority race, gestational age at the earliest prenatal visit and ultrasound, and second-trimester serum screening.

Conclusion: Applying advanced machine learning techniques to a comprehensive database of stillbirths and live births with unique and clinically relevant variables resulted in an algorithm that could accurately identify 85% of pregnancies that would result in stillbirth, before they reached viability. Once validated in representative databases reflective of the US birthing population and then prospectively, these models may provide effective risk stratification and clinical decision-making support to better identify and monitor those at risk of stillbirth.

References

Willinger M, Ko C, Reddy U . Racial disparities in stillbirth risk across gestation in the United States. Am J Obstet Gynecol. 2009; 201(5):469.e1-8. PMC: 2788431. DOI: 10.1016/j.ajog.2009.06.057. View

Khatibi T, Hanifi E, Sepehri M, Allahqoli L . Proposing a machine-learning based method to predict stillbirth before and during delivery and ranking the features: nationwide retrospective cross-sectional study. BMC Pregnancy Childbirth. 2021; 21(1):202. PMC: 7953639. DOI: 10.1186/s12884-021-03658-z. View

Sidey-Gibbons J, Sidey-Gibbons C . Machine learning in medicine: a practical introduction. BMC Med Res Methodol. 2019; 19(1):64. PMC: 6425557. DOI: 10.1186/s12874-019-0681-4. View

Fiaschi L, Nelson-Piercy C, Gibson J, Szatkowski L, Tata L . Adverse Maternal and Birth Outcomes in Women Admitted to Hospital for Hyperemesis Gravidarum: a Population-Based Cohort Study. Paediatr Perinat Epidemiol. 2017; 32(1):40-51. DOI: 10.1111/ppe.12416. View

Hall K, Dalton V, Zochowski M, Johnson T, Harris L . Stressful Life Events Around the Time of Unplanned Pregnancy and Women's Health: Exploratory Findings from a National Sample. Matern Child Health J. 2017; 21(6):1336-1348. PMC: 5444959. DOI: 10.1007/s10995-016-2238-z. View

Smith G . Predicting antepartum stillbirth. Curr Opin Obstet Gynecol. 2006; 18(6):625-30. DOI: 10.1097/GCO.0b013e32801062ff. View

Zeisler H, Llurba E, Chantraine F, Vatish M, Staff A, Sennstrom M . Predictive Value of the sFlt-1:PlGF Ratio in Women with Suspected Preeclampsia. N Engl J Med. 2016; 374(1):13-22. DOI: 10.1056/NEJMoa1414838. View

Trudell A, Tuuli M, Colditz G, Macones G, Odibo A . A stillbirth calculator: Development and internal validation of a clinical prediction model to quantify stillbirth risk. PLoS One. 2017; 12(3):e0173461. PMC: 5340400. DOI: 10.1371/journal.pone.0173461. View

. Screening for Fetal Chromosomal Abnormalities: ACOG Practice Bulletin, Number 226. Obstet Gynecol. 2020; 136(4):e48-e69. DOI: 10.1097/AOG.0000000000004084. View

10.

Thornton P, Liese K, Adlam K, Erbe K, McFarlin B . Calculators Estimating the Likelihood of Vaginal Birth After Cesarean: Uses and Perceptions. J Midwifery Womens Health. 2020; 65(5):621-626. PMC: 7606451. DOI: 10.1111/jmwh.13141. View

11.

Williams A, Wallace M, Nobles C, Mendola P . Racial residential segregation and racial disparities in stillbirth in the United States. Health Place. 2018; 51:208-216. PMC: 6287738. DOI: 10.1016/j.healthplace.2018.04.005. View

12.

Townsend R, Manji A, Allotey J, Heazell A, Jorgensen L, Magee L . Can risk prediction models help us individualise stillbirth prevention? A systematic review and critical appraisal of published risk models. BJOG. 2020; 128(2):214-224. DOI: 10.1111/1471-0528.16487. View

13.

Rudin C . Stop Explaining Black Box Machine Learning Models for High Stakes Decisions and Use Interpretable Models Instead. Nat Mach Intell. 2022; 1(5):206-215. PMC: 9122117. DOI: 10.1038/s42256-019-0048-x. View

14.

Berhie K, Gebresilassie H . Logistic regression analysis on the determinants of stillbirth in Ethiopia. Matern Health Neonatol Perinatol. 2016; 2:10. PMC: 5025573. DOI: 10.1186/s40748-016-0038-5. View

15.

Lee T, Shah N, Haack A, Baxter S . Clinical Implementation of Predictive Models Embedded within Electronic Health Record Systems: A Systematic Review. Informatics (MDPI). 2020; 7(3). PMC: 7710328. DOI: 10.3390/informatics7030025. View

16.

Reddy U . Management of pregnancy after stillbirth. Clin Obstet Gynecol. 2010; 53(3):700-9. DOI: 10.1097/GRF.0b013e3181eba25e. View

17.

Partridge S, Balayla J, Holcroft C, Abenhaim H . Inadequate prenatal care utilization and risks of infant mortality and poor birth outcome: a retrospective analysis of 28,729,765 U.S. deliveries over 8 years. Am J Perinatol. 2012; 29(10):787-93. DOI: 10.1055/s-0032-1316439. View

18.

Saha S, Sandha S, Srivastava M . Machine Learning for Microcontroller-Class Hardware: A Review. IEEE Sens J. 2022; 22(22):21362-21390. PMC: 9683383. DOI: 10.1109/jsen.2022.3210773. View

19.

Mboya I, Mahande M, Mohammed M, Obure J, Mwambi H . Prediction of perinatal death using machine learning models: a birth registry-based cohort study in northern Tanzania. BMJ Open. 2020; 10(10):e040132. PMC: 7574940. DOI: 10.1136/bmjopen-2020-040132. View

20.

Goy J, Dodds L, Rosenberg M, King W . Health-risk behaviours: examining social disparities in the occurrence of stillbirth. Paediatr Perinat Epidemiol. 2008; 22(4):314-20. DOI: 10.1111/j.1365-3016.2008.00947.x. View