Variability in infant acute pain responding meaningfully obscured by averaging pain responses

Summary Stable variability in infant pain responses increases with age, challenging the validity of using a mean infant pain score in pain research. Abstract Given the inherent variability in pain responding, using an “average” pain score may pose serious threats to internal and external validity. Using growth mixture modeling (GMM), this article first examines whether infants can be differentiated into stable groups based on their pain response patterns over a 2‐minute post‐needle period. Secondary analyses, to specifically address the issue of averaging pain scores to represent a sample, qualitatively described clinically meaningful differences between pain scores of the discerned groups and the overall mean (irrespective of groups). Infants were part of Canadian longitudinal cohort naturalistically observed during their 2‐, 4‐, 6‐, and/or 12‐month immunization appointments (N = 458 to 574) at 3 pediatrician clinics between 2007 and 2012. At every age, GMM analyses discerned distinct groups of infants with significantly variable patterns of pain responding over the 2 minutes post‐needle. Our secondary suggested that the overall mean pain score immediately post‐needle reflected most groups well at every age. However, for older infants (6 and 12 months, especially), the overall mean pain responses at 1 and 2 minutes post‐needle significantly over or underestimated groups that contained 48% to 100% of the sample. These results combined highlight the significant variability of infant pain responding patterns between groups of infants and furthermore, calls into question the validity of using an overall mean in research with older infants during the regulatory phase post‐needle.