Improving care near the end of life. Why is it so hard?

A 56-year-old smoker with chronic obstructive pulmonary disease develops pneumonia and respiratory failure and is placed on mechanical ventilation. He dies 2 weeks later, after a stormy intensive care unit (ICU) course complicated by gastrointestinal bleeding and septic shock. The next morning, the ICU team questions how such a series of events occurred. Although each intervention could be justified as a response to a treatable complication, did the team truly consider the patient's overall prognosis or determine whether he really wanted such aggressive care? See also p 1591. The Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatments (SUPPORT) intervention described in this issue ofThe Journalattempted to prevent such high-technology deaths.1The SUPPORT investigators are to be congratulated on a rigorous, complex project. Yet, despite their best efforts in this multicenter randomized clinical trial, the intervention failed to improve any of the study outcomes.