Health trajectories before initiation of non-invasive ventilation for chronic obstructive pulmonary disease: a French nationwide database analysis

BACKGROUND: Chronic obstructive pulmonary disease (COPD) is the most common indication for long-term domiciliary non-invasive ventilation (NIV) but there is uncertainty in data supporting current guidelines. This study described health trajectories before initiation of at-home NIV in people with COPD, and compared mortality outcomes between groups with different pre-NIV health trajectories. METHODS: Data were from the French national health insurance reimbursement system database for individuals with COPD aged ≥40 years and ≥1 reimbursement for NIV between 1 January 2015 and 31 December 2019. Common health trajectories were determined using time sequence analysis through K-clustering (TAK analysis). FINDINGS: Data from 54,545 individuals were analysed; the population was elderly (median age 70 years) with multiple comorbidities. Four clusters were generated. Cluster 1 (n = 35,975/54,545; 66%) had NIV initiated in ambulatory settings or after the first acute event/exacerbation. Cluster 2 (6653/54,545; 12%) started NIV after ≥2 severe exacerbations in the previous 6 months. Cluster 3 (11,375/54,545; 21%) started NIV after frequent severe COPD-related exacerbations in the previous year. Cluster 4 (652/54,545; 1%) started NIV after many long-lasting severe exacerbations. The four clusters differed in age, sex, comorbidities, pre-NIV investigations, and prescriber/location of NIV initiation. Mortality differed significantly between clusters: highest in Cluster 4 and lowest in Cluster 1. INTERPRETATION: The significant heterogeneity in clinical initiation of NIV probably reflects the current lack of strong evidence and guideline recommendations. Knowledge about the characteristics and outcomes in different clusters should be used to address inequities and facilitate more consistent and personalised use domiciliary NIV in COPD. FUNDING: JLP and SB are supported by the 10.13039/501100001665French National Research Agency in the framework of the “Investissements d’avenir” program (ANR-15-IDEX-02) and the “e-health and integrated care and trajectories medicine and MIAI artificial intelligence (ANR-19-P3IA-0003)” Chairs of excellence from the 10.13039/100019979Grenoble Alpes University Foundation. This work was supported by 10.13039/100017647ResMed.