Kolstad Vilde, García-Torres Jorge, Brunner Sara, Johannessen Anders, Foglia Elizabeth, Ersdal Hege, Meinich-Bache Øyvind, Rettedal Siren
Department of Simulation-Based Learning, Stavanger University Hospital, Stavanger, Norway.
Department of Electrical Engineering and Computer Science, University of Stavanger, Stavanger, Norway.
Front Pediatr. 2024 Aug 14;12:1342415. doi: 10.3389/fped.2024.1342415. eCollection 2024.
Newborn resuscitation algorithms emphasize that resuscitation is time-critical, and all algorithm steps are related to the time of birth. Infrared thermal video has the potential to capture events in the delivery room, such as birth, cord clamping, and resuscitative interventions, while upholding the privacy of patients and healthcare providers.
The objectives of this concept study were to (i) investigate the technical feasibility of using thermal video in the delivery room to detect birth and cord clamping, and (ii) evaluate the accuracy of manual real-time registrations of the time of birth and cord clamping by comparing it with the accuracy of registrations abstracted from thermal videos.
An observational study with data collected at Stavanger University Hospital, Norway, from September 2022 to August 2023. The time of birth and cord clamping were manually registered on a portable tablet by healthcare providers. Thermal cameras were placed in the delivery rooms and operating theatre to capture births. Videos were retrospectively reviewed to determine the time of birth and cord clamping.
Participation consent was obtained from 306 mothers, of which 195 births occurred in delivery rooms or an operating theatre with a thermal camera installed. We excluded 12 videos in which no births occurred. Births were detectable in all 183 (100%) thermal videos evaluated. There was a median (quartiles) of 1.8 (0.7, 5.4) s deviation in the manual registrations of the times of births relative to those abstracted from thermal videos. Cord clamping was detectable in 173 of the 183 (95%) thermal videos, with a median of 18.3 (3.3, 108) s deviation in the manual registrations of the times of cord clampings relative to those abstracted from thermal videos.
Recognizing the time of birth and cord clamping from thermal videos is technically feasible and provides a method for determining when resuscitative events occur.
新生儿复苏算法强调复苏对时间要求严格，且所有算法步骤都与出生时间相关。红外热成像视频有潜力在维护患者和医护人员隐私的同时，捕捉产房内的事件，如出生、脐带结扎和复苏干预。
本概念研究的目的是：（i）调查在产房使用热成像视频检测出生和脐带结扎的技术可行性；（ii）通过将手动实时记录的出生时间和脐带结扎时间的准确性与从热成像视频中提取的记录准确性进行比较，评估其准确性。
在挪威斯塔万格大学医院进行的一项观察性研究，收集2022年9月至2023年8月的数据。医护人员在便携式平板电脑上手动记录出生时间和脐带结扎时间。在产房和手术室放置热成像摄像头以捕捉分娩过程。对视频进行回顾性审查以确定出生时间和脐带结扎时间。
获得了306名母亲的参与同意，其中195例分娩发生在安装了热成像摄像头的产房或手术室。我们排除了12个未发生分娩的视频。在评估的所有183个（100%）热成像视频中都能检测到分娩。与从热成像视频中提取的出生时间相比，手动记录的出生时间中位数（四分位数）偏差为1.8（0.7，5.4）秒。在183个热成像视频中的173个（95%）中能检测到脐带结扎，与从热成像视频中提取的脐带结扎时间相比，手动记录的脐带结扎时间中位数偏差为18.3（3.3，108）秒。
从热成像视频中识别出生时间和脐带结扎时间在技术上是可行的，并提供了一种确定复苏事件发生时间的方法。
