A novel classification of obstructive sleep apnoea enabling precision medicine to improve adherence to continuous positive airway pressure: A growth mixture modelling analysis

ABSTRACT

Introduction: High non-adherence rates to CPAP remain a major obstacle to good outcomes in OSA. In trials, 29%-83% of patients do not adhere to CPAP. CPAP adherence in clinical practice, and the effect of clinical pathways and interventions, remain unknown because of incomplete datasets and use of non-clinically relevant criteria for adherence in previous studies. Patients are reported to become adherent or non-adherent to CPAP from treatment onset, forming the basis of current clinical practice, but the studies have been small. We addressed these evidence gaps using a large, UK multicentre clinical dataset, using changes to sleep centres' treatment pathways during the COVID-19 pandemic as a natural experiment. Method(s) Five sleep centres that telemonitored patient data in 2019 and 2020 were recruited. Using a 18% difference in CPAP adherence between years (Philips Respironics data), 80% power, alpha < 0.05, n = 92 was required. Objective CPAP-usage data over the first three months of treatment was collected from 100 patients who started CPAP prepandemic (April 2019) and 100 patients post-start of pandemic (September 2020), per centre. CPAP adherence criteria Mean CPAP use >=4 h/night for >=70% of nights (for Night 1-3 period, median CPAP use used, as data non-normally distributed). Growth mixture modelling (GMM) and logistic regression were performed using all centres' data (1000 patients). Result(s) Three months after treatment started, only 34% of patients were treatment-adherent in 2019 and 42% in 2020 (p = 0.24). GMM identified six distinct, CPAP-usage behaviours over the first month, each with a different likelihood of CPAP non-adherence at three months. Four behaviours consisted of changing (increasing or decreasing) CPAP use (54% of patients), two behaviours consisted of consistent good or no use (remaining 46%). Treatment pathway determined prevalence of behaviours and CPAP adherence at three months;OSA severity was a weaker determinant of CPAP adherence at three months. Conclusion(s) CPAP use at treatment onset does not predict long-term adherence in most patients. This can explain why current practice is ineffective, and may even be detrimental, as the changing users are inappropriately managed as consistent users . Our data supports precision medicine tailored to specific behaviour from Week 2 of treatment.