Neonatologie in tijden van big data, machine learning en artificiële intelligentie: potentiële toepassingen gebruikmakend van continu gemonitorde vitale parameters. Een systematische review.

Abstract

Neonatology in times of big data, machine learning and artificial intelligence: potential applications using continuously monitored vital signs. A systematic review Background: At the neonatal intensive care unit vital signs are continuously monitored, which yields a huge amount of values. At the moment, these are mainly used in a real-time bedside manner when values cross fixed boundaries triggering an alarm, also general trends are followed. Subtle patterns which may be a first sign of disease or clinical condition can be missed. By using scientific revolutions as big data, machine learning and artificial intelligence it may be possible to extract much more information from this amount of data. Objective: We make an overview of possible applications more thoroughly using continuously monitored vital signs in the NICU, developed by using scientific evolutions as big data, machine learning or artificial intelligence. Methods: A systematic search is carried out in four databases, PRISMA-guidelines are followed. The retained articles are analysed regarding the topic, the used methods, the results and possible implementation. Results: We included 35 articles. The investigated topics and used methods were very diverse. Only a minority of possible applications are already studied enough to enable use in daily practice. Conclusion: Applications using continuously monitored vital signs, developed by using scientific evolutions as big data, machine learning and artificial intelligence are promising and a lot of research is being carried out. Nevertheless, implementation is still very limited.