| Title | Using Neonatal Data: Signal Processing and Statistical Models |
| Author(s) | Daniel P. Lindstrom, Jorge Rojas, and Robert B. Cotton |
| Source | The Use of Computers in Perinatal Medicine, Chapter 18, Pages 327-335 |
| ISBN | 0-03-061513-5 |
| Publisher | Praeger Publishers, New York, NY |
| Publication Date | 1982 |
| Abstract | A great many potentially useful data are available in the neonatal intensive care unit (NICU) in analog form, that is, time-varying signals from pressure transducers, impedance sensors, or bioelectric potentials. Although the instruments commonly used to process these signals are becoming more sophisticated, most commercially available monitors extract only simple parameters, such as heart rate or blood pressure, and sound an alarm if preset limits are violated. We [the authors] at Vanderbilt University are attempting to use computers, on a research basis, to extract more useful information about the patient by processing the signals from relatively noninvasive physiologic sensors. We are also involved in using statistical models to characterize a given patient population as to probable course and outcome on the basis of linear discriminant function analysis of a few variables observed at specific times. |