P074 Three models of pulse oximeters are within specifications when assessed using a benchtop pulse oximeter simulator

To assess possible SpO2 accuracy differences between 3 different (oximeter & software) testing techniques. Hypothesised that differences in SpO2 would be within the oximeter manufacturer’s specification. A synthetic digit, OX-1 OxSim® Optical SpO2 Pulse Oximeter Simulator (Pronk Technologies Inc., Sun Valley, CA) can provide simulated SpO2 & HR values : 85% (80BPM), 98% (80BPM), 98% (140BPM), 99% (80BPM) at low perfusion and 95% (40BPM) low HR. Simulated SpO2 delivered for 5 minutes, average calculated from signal recorded with corresponding download software: Rad-5 (Masimo, Irvine, CA, USA, 2sec averaging time) - Visi download (Stowood Scientific, Oxford, UK), the Rad-7 (Masimo, Irvine, CA, USA, 2-4sec averaging time) - PSG 4 (Compumedics, Melbourne, Australia), Wrist-Ox2® (Nonin Medical Inc, Plymouth, MN, USA, 3sec averaging time) - Noxturnal software (Nox Medical Inc, Reykjavik, Iceland). Four devices for each model Rad-5, Rad-7 and Nox Wrist-Ox2® were examined at 5 SpO2 & HR settings. The SpO2 downloaded from each device was compared with the OxSim® values. Oximeter reading v OxSim® : at 85%, Rad-5 (y = -0.146x + 14.375, R² = 0.9627 bias 0.606), Rad-7 (y = -0.0897x + 8.737, R² = 0.5295 bias 0.295), Wrist-Ox2® (y = -0.0161x + 1.2261, R² = 0.0481 bias -0.281). At all other comparison values, including low perfusion and low HR, all values fell within +/- 2% SpO2. Small differences in the measured SpO2 for the 3 oximeter models & paired software were noted, however all devices gave SpO2 values within 2% of simulated SpO2 values as per manufacturer specifications.