Full Text Available
Note: Clicking the button above will open the full text document at the original institutional repository in a new window.
The development of a neonatal vital signs database
Modern intelligent monitoring systems use digital computer technology to analyze and evaluate physiological vital signs. This analytical and evaluative process is performed by algorithms developed for this purpose. The degree of 'intelligence' of the monitoring system is dependent on the 'sensitivit...
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis |
| Language: | English |
| Published: |
Department of Human Biology
2017
|
| Subjects: | |
| Tags: |
No Tags, Be the first to tag this record!
|
| Summary: | Modern intelligent monitoring systems use digital computer technology to analyze and evaluate physiological vital signs. This analytical and evaluative process is performed by algorithms developed for this purpose. The degree of 'intelligence' of the monitoring system is dependent on the 'sensitivity' and 'specificity' of these algorithms. In order to develop robust and clinically valid algorithms, a database of representative waveforms is required. The aim of this thesis was to create a neonatal vital signs database to be used for this purpose, by means of a computer-based central station. The computer was interfaced to a number of neonatal monitors (Neonatal ICU, Groote Schuur Hospital). The monitors were interrogated to obtain patient condition, ECG waveforms and respiration waveforms using the impedance technique. When possible, percentage oxygen saturation was also captured. The database contains 509 documented clinical records obtained from 35 patients and 20 records containing examples of technical alarm conditions and high frequency noise. Additional patient record data is included. Clinical events recorded include apnoea, bradycardia, periodic breathing tachycardia, tachypnoea and normal traces. These events were recorded against a variety of signal quality conditions that have been characterized in Appendix C. A prototype rate detection algorithm was checked using samples from the database. |
|---|