Indicators of persistent pain in preterm neonates are poorly defined. In the setting of a double blind, placebo-controlled trial investigating morphine use in ventilated preterm infants (NEOPAIN Trial) we aimed to identify factors that may be useful in assessing persistent pain. Twenty-two babies (morphine 12; placebo 10) were assessed for comfort, pain or distress and clinical staff described the factors they had considered. This assessment was performed during the first period of duty with the baby. Based on this, they stated which study drug they believed the infant was receiving. Eighty-nine assessments were made in total (1-14 per baby). The drug was correctly identified on 71% of occasions. Staff considered one or more of the following factors: infant activity; response to routine care; known pain-related behaviours; posture/quality of movements; respiratory effort; synchrony with ventilator; blood pressure and heart rate. Four factors most frequently identified babies receiving placebo: facial expressions of pain, high activity levels, poor response to handling and poor synchrony with ventilation. Absence of pain-related behaviour was less discriminating. Observation of a good response to handling, good synchrony with ventilation, a "settled" baby, normal blood pressure and heart rate were poor discriminators. Hypotension and poor respiratory drive were noted exclusively in babies receiving morphine infusions. Facial expressions of pain, high activity levels, poor response to routine care, and poor ventilator synchrony were associated with placebo versus morphine therapy, and may be considered useful markers for persistent pain in preterm infants.