The effect on participation rates of including focused spirometry information in a health check invitation: a cluster-randomised trial in Denmark