| Title | http://dx.doi.org/10.1016/j.clp.2017.04.003 |
| Author(s) | Van Laere D, Meeus M, Beimnaert C, Sonck V, Laukens K, Mahieu L, Mulder A |
| Source | Clinics in Perinatology, Vol. 47, Pages 435-448 |
| DOI | doi.org/10.1016/j.clp.2020.05.002 |
| Publisher | Elsevier |
| Publication Date | 2020 |
| Abstract | In the era of increased digitalization and development of noninvasive monitoring techniques, decision making in hemodynamic care is still largely dependent on the clinician’s interpretation of monitored vital signs. The nature of neonatal intensive care makes it an interesting field for applying machine learning. During continuous monitoring of vital signs of vulnerable infants, patterns of disease progression might be captured. The awareness of clinicians to potential harm of cardiovascular treatment on newborn infants highlights the need for predictive analytics. Machine learning techniques can train models to recognize distinct patterns in monitoring data related to disease. Applying machine learning techniques to time series monitoring data requires special considerations to develop clinically useful models. |