We reviewed the records of 274 consecutive deaths at the Milton S. Hershey Medical Center, Hershey, Pa, occurring through May 1988 to examine the approach of physicians, patients, and families in making the decision to invoke the "do-not-resuscitate" order. Of these 274 patients who died, 171 (62%) had do-not-resuscitate orders. Of these 171 patients, 86 (50%) were judged fully mentally competent on admission to the hospital; 44 (51%) of these 86 fully competent patients were included in the decision to withhold resuscitative efforts. In the remainder, the family was usually involved in the decision without input from the patient. Only 6 patients (4%) were admitted to the hospital with a preexisting do-not-resuscitate order. For the remainder, the do-not-resuscitate order was written a mean of 8.5 days following admission and 3.3 days before death. Documentation of this order with a specific progress note was universal. The principle reason cited for a do-not-resuscitate order was the presence of irreversible terminal disease in 52% and an unacceptable quality of life in 33%. When considered separately, patients with a principle diagnosis of malignant neoplasm had a do-not-resuscitate order written 80% of the time. Of 88 such patients, 48 (55%) were fully competent at admission. In turn, 36 (75%) of these patients participated in the do-not-resuscitate decision. Nursing activities were quantified for the 24 hours preceding and the 24 hours following the do-not-resuscitate order. No difference could be found comparing these two periods whether the comparison was made on the general hospital ward or in the intensive care unit.(ABSTRACT TRUNCATED AT 250 WORDS)