Interobserver variability of the neurological optimality score

Objective: To assess the interobserver reliability of the neurological optimality score. Study design: The neurological optimality scare of 21 full term healthy, neurologically normal newborn infants was determined by two well trained observers. Results: The interclass correlation coefficient was 0.31. Kappa for optimality (score of 58 or higher) was 0.19. A systematic difference of 1.3 points between the two observers was present. Conclusions: The interobserver variability of the neurological optimality score of the newborn infant is substantial. The subtle judgement of elicited responses as optimal or non-optimal proved to be especially critical in this concordance study. A difference of at least two points in the score is considered as a valid endpoint for comparitive studies. If two or more observers are involved in the neurological examination of the newborn infant in a study to assess influences on perinatal morbidity, frequent re-instruction sessions are recommended. (C) 1999 Elsevier Science Ireland Ltd. All rights reserved.