Identifying saltatory growth patterns in infancy: A comparison of results based on measurement protocol.

Abstract

The present analyses address the question of how frequently time-intensive growth measurements need to be taken in order to provide sufficient data for the statistical identification of pulsatile saltatory growth patterns during infancy. The daily serial growth measurements of three infants during 4 months are analyzed in seven subsets, which include progressively fewer data points from daily to weekly intervals. Saltatory growth pulse identification and growth pattern analyses are compared between the temporally distinctive data sets by the saltatory algorithm and continuous curvilinear models. Statistically significant pulse identification and saltatory growth pattern resolution decrease with time intervals > 2-5 days in these data sets. These results suggest that the convenience of a Monday, Wednesday, Friday protocol may not be sufficient to characterize saltatory growth in individual infants, and the dependability factor is a significant source of error that may contribute to growth pattern misperception in data collected by infrequent protocols. Am. J. Hum. Biol. 9:343-355, 1997. © 1997 Wiley-Liss, Inc.