In this clinical audit, we compared the guidelines on cardiopulmonary resuscitation in our hospital with those from the Royal College of Physicians (London). Forty-six of 72 cases designated as "do not resuscitate" at the Prince of Wales Hospital and transferred to Shatin Hospital were re-categorised to receive active resuscitation. Documentation and the reasons for the "do not resuscitate" order were not made in fewer than 7% of patients. The decision was made with patients/carers in 61% of patients designated "do not resuscitate" at Shatin Hospital. Of those 72 patients, 68% were eventually discharged and only 16% died in the subsequent nine months. Among those discharged, functional levels were not worse compared with their premorbid state. Hence, our policy has not led to undue prolongation of life and disability with poor quality of life. We recommend more training in patient communication skills and studies on attitudes of patients towards the decision not to resuscitate.