Around 5%-10% of newborns need assistance to start breathing. Currently, there is a lack of evidence-based research, objective data collection, and opportunities for learning from real newborn resuscitation emergency events. Generating and evaluating automated newborn resuscitation algorithm activity timelines relative to the Time of Birth (ToB) offers a promising opportunity to enhance newborn care practices. Given the importance of prompt resuscitation interventions within the "golden minute" after birth, having an accurate ToB with second precision is essential for effective subsequent analysis of newborn resuscitation episodes. Instead, ToB is generally registered manually, often with minute precision, making the process inefficient and susceptible to error and imprecision. In this work, we explore the fusion of Artificial Intelligence (AI) and thermal imaging to develop the first AI-driven ToB detector. The use of temperature information offers a promising alternative to detect the newborn while respecting the privacy of healthcare providers and mothers. However, the frequent inconsistencies in measurements when using uncooled thermal sensors make normalization strategies critical. Our methodology involves a three-step process: first, we propose an adaptive normalization method based on Gaussian mixture models (GMM) to mitigate issues related to temperature variations; second, we implement an AI model to detect the presence of the newborn within the thermal video frames; and third, we evaluate and post-process the model's predictions to estimate the ToB. A precision of 88.1% and a recall of 89.3% are reported in the detection of the newborn within thermal frames during performance evaluation. Our approach achieves an absolute median deviation of 2.7 s in estimating the ToB relative to the manual annotations.

Newborn resuscitation guidelines recommend stimulation as a first step in non-breathing newborns to facilitate spontaneous breathing. We aimed to evaluate the extent of stimulation in the delivery room, heart rate (HR) responses to stimulation, and association between bradycardia (HR <100 beats per minute (bpm) ≥10 s) in the first minute after birth and the need for positive pressure ventilation (PPV) versus stimulation alone. A prospective observational study conducted at tertiary level Stavanger University Hospital, Norway, September 2022-July 2024 including newborns with gestational age ≥34 weeks. Stimulation was registered from thermal videos in the delivery room and visible light cameras over resuscitaires. HR was obtained using dry-electrode ECG. 449 included newborns had thermal videos ± visible light video and HR data. 70/449 (16%) newborns did not receive any resuscitative interventions, 345/449 (77%) received stimulation alone, 14/449 (3.1%) received continuous positive airway pressure, and 20/449 (4.4%) received PPV. We observed a HR increase of median 7 bpm among newborns with HR <140 bpm at start stimulation (p < 0.001). Newborns with bradycardia in the first minute after birth had four-fold likelihood of requiring PPV versus tactile stimulation alone (OR 4.81 95% CI 1.39-15.3, p = 0.009). Tactile stimulation is widely applied during neonatal transition. We observed an increase in HR in relation to tactile stimulation, especially among newborns with HR <140 bpm at start stimulation. Newborns with bradycardia (HR <100 bpm ≥10 s) in the first minute after birth had a four-fold likelihood of requiring PPV versus tactile stimulation alone.

Background/Objectives: Newborn heart rate is an integral part of resuscitation algorithms, but the impact of ECG monitoring on resuscitative interventions and clinical outcomes has been identified as a knowledge gap. The objective was to evaluate the impact of routine use of dry-electrode ECG in all newborns immediately after birth on time to start positive pressure ventilation (PPV) when indicated. Methods: We conducted a randomized clinical trial from June 2019 to November 2021 at Stavanger University Hospital, Norway. Dry-electrode ECG sensors were applied immediately after birth to all newborns ≥ 34 weeks’ gestation. Randomization determined whether the heart rate display was visible or masked. Time of birth was registered in an observation app. Time to start ventilation was calculated from video recordings. Results: In total, 7343 newborns ≥ 34 weeks’ gestation were enrolled, 4284 in the intervention and 3059 in the control group, and 3.7% and 3.8% received ventilation, respectively. In 171/275 (62%) of the newborns the exact time of birth and a video of the resuscitation were available, for 98 in the intervention and 73 in the control group. Ventilation was provided within 60 s to 44/98 (45%) in the intervention and 24/73 (33%) in the control group, p = 0.12. Time from birth to start of PPV was a median of 66 (44, 102) s in the intervention and 84 (49, 148) s in the control group, p = 0.058. Resuscitated newborns were apneic (74%) or breathing ineffectively (26%) at the start of PPV, and only 36% had a heart rate < 100 beats per minute. Conclusions: The use of dry-electrode ECG heart rate monitoring did not change the proportion of newborns that received ventilation within 60 s after birth, but early termination due to employee protests to video recordings rendered the trial inadequately powered to detect a difference. Breathing status was likely a more important determinant of starting ventilation than bradycardia.