Closed-Loop Control of Inspired Oxygen in Ventilated Newborn Infants

The Project

The oxygen delivery to the sick newborn infant must be adjusted very closely in order to grant adequate tissue oxygenation while minimizing possible toxic effects of supplying oxygen. Correct FiO2 settings are achieved by manual adjusting the FiO2 according to the measured SaO2. In some patients this requires frequent adjustments within short intervals. In order to reduce the strain frequent alarms and FiO2 adjustments, we often notice generous SaO2 limits resulting in an oversupply of oxygen. We therefore consider manual adjustments of the FiO2 not as an optimal solution.

The aim of our project was to develop a continuously operating automated FiO2 Controller for optimizing the oxygen delivery to the newborn infant. The automated FiO2 controller will adjust the FiO2 settings based on the continuous transcutaneous SaO2 measurements. During each control cycle, data are read from the monitor, validated, and in case preset SaO2 limits are exceeded or fall short of, the FiO2 is adjusted accordingly.

The FiO2 controller is not designed to replace personal monitoring of the newborn infant by nurses and physicians. In case of a rapid change of the SaO2 requiring immediate personal intervention, monitoring alarms are issued. The FiO2 controller is not designed for emergency interventions but for optimizing oxygen delivery during "quiet" intervals.

The Project Team

The project is a joint cooperation of the

Department of Medical Cybernetics and Artificial Intelligence (IMKAI),
Austrian Research Institute for Artificial Intelligence (OFAI),
Department of Software Technology, Vienna University of Technology (IFS),
the Department of Pedriatics of the University of Vienna (NICU),
the Department of Neonatology, University Hospital Tübingen, Germany,
the Department of Pediatrics, Hospital of Mödling (former partner),
the Department of Neonatology and Pediatric Pulmonology, School of Medicine, Hannover, Germany (former partner).

Researchers:

Werner Horn, IMKAI & ÖFAI
Harald Leitich, NICU (former member)
Silvia Miksch, IFS
Franz Paky, Mödling (former member)
Christian Poets, Tübingen
Christian Popow, NICU
Caroline Renault, ÖFAI (former member)
Andi Seyfang, IFS
Michael Urschitz, Tübingen and NICU

Publications

Urschitz M.S., Horn W., Seyfang A., Hallenberger A., Herberts T., Miksch S., Popow C., Mueller-Hansen I., Poets C.F.: Automatic Control of the Inspired Oxygen Fraction in Preterm Infants. A Randomized Cross-over Trial, Am. J. Respir. Crit. Care Med., 170(10):1095-1100, 2004.
Seyfang A., Miksch S., Horn W., Urschitz M.S., Popow C., Poets C.F.: Using Time-Oriented Data Abstraction Methods to Optimize Oxygen Supply for Neonates, in Quaglini S., et al.(eds.), Artificial Intelligence in Medicine, Proceedings of the 8th Conference on Artificial Intelligence in Medicine in Europe (AIME-2001), Springer, Berlin, pp.217-226, 2001.
Seyfang A., Miksch S., Horn W., Urschitz M., Popow C., Poets C.F.: Using Time-Oriented Data Abstraction Methods to Optimize Oxygen Supply for Neonates, ÖFAI TR-2001-03, 2001.
Miksch S., Seyfang A., Horn W., Popow C.: Abstracting Steady Qualitative Descriptions over Time from Noisy, High-Frequency Data, in Horn W. et al. (eds.), Artificial Intelligence in Medicine, Springer, Berlin, pp.281-290, 1999.

Contact

Werner Horn (email: werner@ai.univie.ac.at)

WH, 11-Dec-2004