Background: While most people die in the hospital or a nursing home, surveys indicate that more than 70% of people would prefer to die at home. Expert panel recommendations have called for epidemiologic studies to document the nature of dying in America. Objective: To determine if the experience of dying differed among settings in New Hampshire. Design: A voluntary, statewide medical record audit of adult deaths in hospitals, nursing homes and home care/hospice agencies was conducted for February and March 2002. Measurements: Records were examined for place of death, patient decision-making capacity and advance directives (ADs). Information was collected on demographic characteristics, primary and secondary diagnoses, presence of a "values history" (documented discussion with patient about their values and end-of-life care) and whether emotional and spiritual support was provided to patients and their families. Medical chart notes for the 48 hours preceding death were reviewed for "pain" and "other symptoms routinely assessed, treated and documented," and for whether the patient had undergone any of the following "treatments": surgery, ventilator, cardiopulmonary resuscitation, or extubation. Results: Nearly one third (32%) of health care organizations in the state reported on 782 deaths (424 hospital, 148 nursing home, and 210 home care/hospice) representing 44% of the adult deaths during this period. Significant differences among settings (p < 0.001) were found for mean age, gender, marital status, primary insurance, diagnosis, ADs, symptom assessment, and provision of emotional and spiritual support for patients and families. Among hospital decedents, 56% were in acute care beds, 30% were in intensive care units, and 4% were in palliative care beds. Nineteen percent of decedents received interventions such as extubation, placed on a ventilator or surgery in the 48 hours preceding death. Over 80% had a do-not-resuscitate (DNR) order, 45% had a Durable Power of Attorney for Health Care, and 37% had a Living Will. Age and setting were significant factors in the presence of ADs. Conclusions: This information provides a benchmark for different care systems to identify areas for improvements in end-of-life care. © Mary Ann Liebert, Inc.