Validation of a new domiciliary diagnosis device for automatic diagnosis of patients with clinical suspicion of OSA

Abstract

Background and objective: Obstructive sleep apnoea is a prevalent and considerably underdiagnosed disease. The development of cost-effective, home-based, automatic diagnostic devices to improve the diagnosis accessibility is therefore essential. Methods: In this study, a new portable polygraph (BTI-APNiA) was used to validate automatic scoring. This five-channel device records respiratory flow, oxygen saturation, heart rate, body position and snoring. The validation was performed in two phases. In the first phase, manual and automatic scorings of a new respiratory polygraphy (RP) device (BTI-APNiA) were compared. In the second phase, automatic analysis performed with BTI-APNiA was compared with manual scoring of a validated RP device (Embletta Gold). Results: Phase I was completed by 424 patients (50.5% males, 52.2 ± 12.4 years, BMI of 25.4 ± 4.8 kg/m2 and Epworth Sleepiness Scale score of 8.0 ± 4.0). Manual and automatic analysis resulted in an apnoea–hypopnoea index (AHI) of 13.7 ± 12.7 and 14.0 ± 12.5 (P > 0.05), respectively. The interclass correlation coefficient (ICC) was 0.99 (P < 0.001). During Phase II, 28 patients were evaluated (72.0% men, 49.1 ± 10.9 years, BMI of 27.1 ± 4.2 kg/m2 and Epworth Sleepiness Scale score of 7.5 ± 4.2). Manual analysis of Embletta Gold recordings indicated an AHI of 12.3 ± 14.0, while automatic analysis of the BTI-APNiA was 13.4 ± 14.7 (P > 0.05). The ICC was 0.68 (P < 0.01). Conclusion: The automatic analysis of the BTI-APNiA is as accurate as manual analysis of AHI. This automatic analysis compared well with the manual analysis of a validated RP device (Embletta Gold).