OBJECTIVE: To define the core data elements of a neonatal intensive care unit (NICU) handoff compare NICU residents' written and verbal handoff data with real-time, auto-populated data and identify the epidemiology of handoff errors. STUDY DESIGN: We defined nine core data elements for a NICU patient handoff. We then compared residents' written and verbal handoffs against real-time, auto-populated data for each core element. RESULT: A total of 101 NICU patient handoffs (31 unique patients) were analyzed. Per patient, residents made more written errors for infants in critical-care beds than for infants in step-down beds (2.33 vs 1.67, P=0.04). Replacing residents' written handoffs with the gold-standard, auto-populated data would have prevented 92% of written errors. CONCLUSION: NICU infants are subjected to many handoff errors. Sicker infants are at higher risk for error. Auto-population can reduce written handoff errors and allow residents more time for training and educational opportunities.