OBJECTIVES. Because the concept of brain death is difficult to define and to apply, we hypothesized that significant variability exists in pediatric brain death determination and documentation. METHODS. Children (0–18 years of age) for whom death was determined with neurologic criteria between January 2000 and December 2004, in southern California, were included. Medical charts were reviewed for documented performance of 14 specific elements derived from the 1987 brain death guidelines and confirmatory testing. RESULTS. A total of 51.2% of children (142 of 277 children) referred to OneLegacy became organ donors. Care locations varied, including PICUs (68%), adult ICUs (29%), and other (3%). One patient was <7 days, 6 were 7 days to 2 months, 22 were 2 months to 1 year, and 113 were >1 year of age. The number of brain death examinations performed was 0 (4 patients), 2 (122 patients), 3 (14 patients), or 4 (2 patients). Recommended intervals between examinations were followed for 18% of patients >1 year of age and for no younger patients. A mean of only 5.5 of 14 examination elements were completed by neurologists and pediatric intensivists and 5.8 by neurosurgeons. No apnea testing was recorded in 60% of cases, and inadequate Paco2 increase occurred in more than one half. Cerebral blood flow determination was performed as a confirmatory test 74% of the time (83 of 112 cases), compared with 26% (29 of 112 cases) for electroencephalography alone. CONCLUSIONS. Children suffering brain death are cared for in various locations by a diverse group of specialists. Clinical practice varies greatly from established guidelines, and documentation is incomplete for most patients. Physicians rely on cerebral blood flow measurements more than electroencephalography for confirmatory testing. Codifying clinical and testing criteria into a checklist could lend uniformity and enhance the quality and rigor of this crucial determination.