A recent multicenter study in adult ICUs showed that doctors often fail to talk with patients about advance directives, prognosis, or the patients' pain. We assessed how well MDs and RNs in 2 NICUs communicate with parents.METHODS: We identified a subpopulation of high risk babies(<750g, Grade IV IVH, congenital anomaly, chronic lung disease) in 2 NICUs, and interviewed MDs, RNs, and parents biweekly. We asked parents from whom they received infor-mation, and asked MDs and RNs how often they had met parents. RESULTS: During the 9 month study, 260/1204 admissions to the NICUs met study criteria. We were able to interview parents of 147(57%)babies. UC had more babies with birthweight <750 (27 v 17% p<0.01) and with chronic lung disease (39 vs. 15% p<0.01). UA had more babies with congenital anomalies (59 v 19% p<0.001). At UC, a higher percentage of parents hadn't visited the NICU (29% v 14%, p<0.01) and there were more cases in which the MD or RN caring for the baby at the time of the interview had not met parents (42% v 16% p<0.01). Parents in both NICUs more often said that RNs, not MDs, were their primary source of information (45% vs 23% p<0.01). Discontinuity of care was the norm. Among doctors, it resulted from attendings going on [Illegible Text] off service or call. Doctors in both NICUs were responsible for >35 babies, and generally focused on the sicker ones. Among RNs, discontinuities resulted from staffing patterns. They were often assigned to different babies each day. Parents of babies who died seldom identified an MD by name. They often remembered a particular RN.CONCLUSIONS: At 2 NICUs, in different countries, serving different populations of patients, there were pervasive problems with communication and continuity of care by MDs and RNs. Similar problems are found in adult ICUs. They seem inherent to the staffing models in tertiary care academic ICUs.