The primary outcome will be the number of COPD exacerbation days/patient/year, which will be assessed using Poisson regression with a scale-parameter to correct for over dispersion. Secondary analyses will include between-group differences
in the proportion of patients who have a total number of exacerbation days above the median, 75th, and 90th percentile of the population as a whole. Between-group differences of secondary continuous variables (e.g. CRQ, HADS) over the 12-month period will be assessed using a multi-level model to incorporate the repeated measurements over time. The total number of health care contacts and courses of corticosteroids/
antibiotics will be analysed using Poisson regression