Heart rate variability: impact of differences in outlier identification and management strategies on common measures in three clinical populations.

Heart rate variability (HRV) is reported increasingly in pediatric research, but different strategies used to identify and manage potential outlier beats impact HRV parameter values in adults and animals. Do they in pediatrics? To compare the impact of different strategies to identifying and managing outliers, we used interbeat interval (IBI) data from three different populations: 10 stable premature infants, 33 stable pediatric oncology patients, and 15 healthy adults. Five commonly reported HRV parameters were compared using three identification and two management strategies to filter potential outliers. The three populations had different resting heart rates: 155 +/- 9 beats per minute (bpm) in infants, 105 +/- 17 bpm in children, and 87 +/- 12 bpm in adults. All three identification strategies flagged fewer than 2% of IBIs; the threshold identification strategy, excluding IBIs denoting heart rates <30 or >300 bpm, identified significantly fewer outliers than the other two strategies and generated higher HRV parameters in all populations (p < 0.001). There were no significant differences in HRV parameters calculated by managing identified outliers by "tossing" them versus "interpolating" values. Different strategies for identifying potential outliers are associated with significant differences in HRV parameters. Pediatric researchers who report HRV should detail their outlier filtering strategies.