慢性呼吸不全患者における長期在宅非侵襲的換気における患者報告の漏れ知覚を遠隔モニタリングで予測できるか？

Home non-invasive ventilation (NIV) improves outcomes in patients with chronic respiratory failure; however, telemonitoring remains challenging due to undefined appropriate alert thresholds. Current leak thresholds lack clinical validation, potentially leading to inefficient care delivery and alert fatigue among healthcare providers.

Can telemonitoring leak data predict patient-reported perceptions of leaks in home NIV for patients with chronic respiratory failure?

Data collected between March 2019 and May 2023 from 571 patients with chronic respiratory failure receiving home NIV were analyzed retrospectively. Inclusion required at least one S3-NIV questionnaire and 28 days of preceding telemonitoring leak data. Twenty-seven leak features were extracted across 7, 14, 21, and 28-day windows. Statistical analysis included ANOVA tests, effect size calculation (Eta-squared), and comparison of predictive performance using Linear Mixed-effects regression and K-nearest neighbors algorithms.

Only S3-NIV Item 8 ("I am disturbed by leaks") showed significant associations with leak features. While no demographic or NIV setting differences were found between patients reporting leak disturbances and those who did not, disturbed patients consistently reported more NIV-related issues and respiratory symptoms. Leak features reflecting central tendency and magnitude showed the highest effect sizes (η² = 0.021-0.075). However, predictive models demonstrated limited ability to accurately predict patient-perceived leaks.

Telemonitoring leak data alone has limitations in predicting patient perception of leaks, suggesting that individual factors beyond objective measurements play a significant role. A proactive patient-centered management approach integrating telemonitoring data and patient-reported outcomes is thus needed.

Registered on Health Data Hub under number N° 23517022; URL: www.health-data-hub.fr (registration details available upon request).