Cardiopulmonary monitoring at home: the CHIME monitor.

A new physiologic monitor for use in the home has been developed and used for the Collaborative Home Infant Monitor Evaluation (CHIME). This monitor measures infant breathing by respiratory inductance plethysmography and transthoracic impedance; infant electrocardiogram, heart rate and R-R interval; haemoglobin O2 saturation of arterial blood at the periphery and sleep position. Monitor signals from a representative sample of 24 subjects from the CHIME database were of sufficient quality to be clinically interpreted 91.7% of the time for the respiratory inductance plethysmograph, 100% for the ECG, 99.7% for the heart rate and 87% for the 16 subjects of the 24 who used the pulse oximeter. The monitor detected breaths with a sensitivity of 96% and a specificity of 65% compared to human scorers. It detected all clinically significant bradycardias but identified an additional 737 events where a human scorer did not detect bradycardia. The monitor was considered to be superior to conventional monitors and, therefore, suitable for the successful conduct of the CHIME study.

An electronic simulator for testing infant apnoea monitors that uses actual physiologic data.

An electronic simulator of physiologic signals used in infant monitoring has been designed, constructed and applied in the Collaborative Home Infant Monitor Evaluation (CHIME). A unique feature of the simulator is that it contains actual physiologic waveforms recorded from infants rather than artificial, idealized signals. The simulator stores breathing waveforms that can be used to test transthoracic-impedance- and inductance-plethysmography-based monitors, and heart rate channels are tested by playing a neonatal QRS complex at preset fixed rates or a variable rate as determined from infant recordings. The transfer characteristics of the simulator are constant over frequencies ranging from 0.5 to 8 Hz for the respiration channels. Data stored in memory are divided into 60 second epochs that can be presented to the monitor being tested in a programmable sequence. A group of 66 CHIME monitors was tested using a simulator programmed with 17 apnoea and bradycardia waveforms. The agreement between monitors as to the duration of detected apnoea decreases as the amount of artefact in the signal increases. Discrepancies between monitors in detecting apnoea duration were found to be similar to inconsistencies between CHIME investigators manually scoring similar waveforms.