It is 3 am. The tiny infant is receiving mechanical ventilation, with a central catheter, advancing on enteral feedings. She is requiring more oxygen, her abdomen is distended, and her C-reactive protein level is elevated. You increase the ventilatory support, stop her feedings, draw blood cultures, and begin therapy with broad-spectrum antibiotics. Two mornings later, the blood culture is sterile, but the infant has improved. The team asks how long we will continue antibiotic therapy. You grumble to yourself about those unreliable neonatal blood cultures and order a 10-day course of antibiotics. After all, this time is her third sepsis evaluation in 16 days, and you just want to keep the infant safe.