Assessment of pain in infants has remained somewhat of an enigma over the past two decades. Although a plethora of infant pain measures has been developed , they are infrequently and inconsistently used in clinical practice. This paucity of assessment is particularly evident in infants who are the most vulnerable; such as those at risk for neurological impairment (NI). Infants at varying levels of risk for NI are exposed to multiple painful procedures during their initial days in the NICU . Stevens et al , found that during the first day of life, neonates at highest risk for NI experienced the greatest number of painful procedures (e.g., suctioning, heel lances, intravenous starts) compared to lower risk groups and were administered the least amount of opiods . To appropriately manage these at-risk infants, a reliable and valid approach to assessing pain is required.
To address this issue, a comprehensive approach to measurement development taking into account the perspectives of multiple stakeholders such as health care providers from different disciplines, parents and family members needs to be undertaken. Eliciting the perceived importance (in accurately and consistently identifying pain) and clinical usefulness (how feasible and useful the measure is for making decisions concerning pain management) of physiologic, behavioural and contextual indicators that comprise pain measures by experts in infant pain research is an appropriate starting place.
Only a few infant pain measures have taken contextual factors, which consist of any factor either known or thought to influence the infant's pain response (such as gestational age, sleep/wake status, severity of illness, that would assist to describe an infant's pain response within a particular context [3–5]) into consideration. As no measures have incorporated risk for neurological impairment (NI) as a contextual factor, the results of this study will contribute to our understanding as to whether there is a difference in response based on risk for NI or whether this risk in and of itself should be considered a contextual factor. There is inconsistent evidence on the differences in behavioural and physiological indicators when comparing infants with and without NI. Stevens et al  reported there were differences in facial activity and heart rate variability with the most at risk for NI infants demonstrating the least response following heel lance. Conversely, Oberlander et al  found similar facial and heart rate variability responses to heel lance between groups of preterm infants with and without parenchymal brain injury.
Generally, infants with NI may show fewer and less clear emotional responses [8, 9]. For example, infants with Down's syndrome or asphyxia exhibited cries that were less frequent, less variable in intensity and fundamental frequency (pitch) [10–14] and of longer latency from the painful stimulus than cries in infants with no disabilities . Furthermore, there is speculation that differences in facial musculature, hypotonia, and aberrant neural information programming may affect facial pain responses [15, 16]. The multidimensional pain response in infants at risk for NI has not been consistently or comprehensively described, and no measure to assess pain in this population has been validated.
Using the Delphi method, the aim of this study was to establish consensus amongst infant pain research experts about which behavioural, physiologic and contextual indicators characterized pain in infants at high, moderate and low risk for NI. The ultimate goal is the development of a new measure or validation of an existing measure for acute pain assessment in this population.