Domiciliary long-term non-invasive ventilation in COPD: should we select subgroups with a better likelihood to respond to NIV in subsequent randomised controlled trials?