Four Measures of Change in Physiologic State During the Feeding Period of Very Premature Infants

Abstract

Precise assessment of infants’ ability to regulate feeding challenges requires modeling of physiologic values across intervals of the feeding period without smoothing too much of the data stream. Using a subset of data from a study of physiologic stability and feeding performance in very preterm (≤30 weeks’ gestational age) infants fed in two different feeding positions, this exploratory study compared mean heart rate (HR) and its coefficient of variation sampled during the feeding period using four methods: (1) dividing the feeding period into three equal intervals; (2) extracting 2-min intervals from the initial, middle, and final third of the feeding period; (3) using successive 2-min intervals during the first 6 min of feeding; and (4) using successive 2-min intervals during the entire feeding period. Power analysis simulation was conducted with varying numbers of subjects to determine the method that requires the least number of subjects to detect a position effect for change in HR across time with 80% power. Different patterns of HR change across the feeding period were observed across the methods, which suggest potential advantages and disadvantages of each method. Among the four methods, using the first 6 min of the feeding period required the fewest subjects to detect significant differences in changes in HR across time between the feeding positions.