Machine assessment of neonatal facial expressions of acute pain
OBJECTIVE To evaluate a method of pain assessment to be used for hospitalized infants requiring blood test by a heel lance procedure. DESIGN Observational study evaluating pain measurement and interrater reliability of pain measurement in hospitalized infants. SETTING A Level III neonatal unit and a cardiac surgical unit at a major pediatric teaching hospital. PARTICIPANTS 20 infants whose gestational age ranged from 28 weeks to full-term. MAIN OUTCOME MEASURES Observations included behavioral measurements (facial expressions, body movements, and crying characteristics) and the physiologic measure of heart rate. Interrater reliability and the feasibility of using the procedural pain assessment method at the bedside were considered. RESULTS The three behavioral measurements and heart rate were responsive to the heel lance. Interrater reliability was high for facial expressions and crying scores but was low for body movements. CONCLUSIONS Modifications have been made to the method of procedural pain assessment to be used in a subsequent study. The modified method is expected to be a reliable measure of procedural pain caused by a heel lance and can easily be used at the bedside during the course of further research.