O026 Concordance between a home sleep apnoea test based on peripheral artery tonometry and laboratory polysomnography

Abstract

Abstract Introduction Limited channel home sleep apnoea test (HSAT) devices are an emerging alternative to laboratory-based PSG. Despite increasingly commercially availability, there is limited external validation of their diagnostic performance. We recently independently validated the Night Owl Mini HSAT (NOM) against in-lab PSG and demonstrated a satisfactory level of agreement (LOA) of 63.8% overall. However, lower rates of agreement were observed for severe OSA. We speculated that signal artefact may have contributed to reduced concordance. Methods Data was drawn from our previously published prospective cohort of 100 participants undergoing PSG for suspected OSA, simultaneously fitted with the NOM. PSG and NOM studies were re-processed, manually remarking artefact. LOA was recalculated comparing only temporally aligned data segments for which both devices provided intact signals, this allowed direct comparison of pulse oximetry data and software algorithms that generated the ODI3%. Two new datasets were generated for PSG and NOM for which combined artefact was deleted (COM-PSG, COM-NOM). Results To date, datasets from 12 participants have been assessed. The mean difference in ODI3% between PSG and NOM was 5.25/hr (95% limits of agreement -0.73 – 11.23), while the mean difference in ODI3% between COM-PSG and COM-NOM was 2.53/hr (95% limits of agreement -1.77 – 6.83). In our previous study, the overall mean difference in ODI3% was –0.21/hr (95% limits of agreement -18.1 – 17.7). Conclusion Preliminary results may indicate signal artefact contributes to reduced level of agreement between PSG and NOM.