Predicting labor onset relative to the estimated date of delivery using smart ring physiological data

Area of Science:

• Reproductive Endocrinology
• Physiological Monitoring
• Maternal-Fetal Medicine

Background:

• Parturition is a complex physiological process involving maternal, fetal, and placental tissues.
• Maternal physiological metrics may reflect the transition from pregnancy to parturition.
• Understanding pregnancy duration and predicting labor onset are crucial for maternal and infant care.

Purpose of the Study:

• To investigate the association between maternal physiological metrics, collected via smart ring, and gestational age.
• To determine if physiological data can predict whether labor begins before or after the estimated delivery date (EDD).
• To explore the relationship between self-reported labor symptoms and gestational age or labor timing.

Main Methods:

• Enrolled 118 pregnant participants in their third trimester.
• Utilized continuously worn smart ring devices to monitor heart rate, heart rate variability, skin temperature, sleep, and physical activity.
• Collected weekly survey data on labor symptoms, pain, fatigue, and mood.
• Employed mixed-effects regression to analyze associations between metrics and gestational age/labor timing.
• Trained a boosted random forest model to predict labor onset relative to EDD.

Main Results:

• Many sleep, activity, pain, fatigue, and labor symptom metrics correlated with gestational age.
• Advancing gestational age was associated with decreased resting heart rate and increased heart rate variability.
• Physical activity and sleep durations decreased with advancing gestational age.
• The random forest model predicted pregnancies passing EDD with 0.71 area under the receiver operating curve (AUC).
• Self-reported labor symptoms correlated with gestational age but not with labor timing relative to EDD.

Conclusions:

• Maternal physiological data from smart rings in the third trimester shows correlations with gestational age.
• Smart ring data shows potential for improving predictions of pregnancy duration relative to the EDD.
• Self-reported symptoms are less predictive of labor timing than objective physiological metrics.