Overall, the professionals in this study did not rate the pain of neonates differently when provided with information indicating those infants had mild, moderate or severe risk for neurological impairment. The professionals' perception of the infants' level of risk also did not affect their ratings of the infants' distress, or their belief in how long the infant would take to calm after pain without intervention. Professionals did perceive that cuddling would be significantly less effective for infants at high risk for neurological impairment than for infants with mild or moderate impairment. However, this effect was not large, and, although it was statistically significant, it may be spurious. Further research should examine whether beliefs regarding pain experience in this group and beliefs regarding the effectiveness of cuddling and other nonpharmacological interventions are truly independent. These results are inconsistent with the results of our previous questionnaire study indicating professionals, with similar levels of experience in neonatal intensive care settings, perceive the pain experience of infants as reduced as their level of risk for neurological impairment increases . There are several possible reasons for these discrepant results.
The professionals who participated in this study were asked to rate the risk for neurological impairment of each infant they viewed on videotape. Asking them to do this may have alerted them to the purpose of the study and elicited efforts on their behalf to provide ratings that were unbiased. However, their ratings of the perceived effectiveness of cuddling did vary by level of risk for impairment, suggesting attempts to appear unbiased do not fully explain the results found.
In our previous studies, questionnaires elicited beliefs about the pain experience of infants and children with varying levels of risk relative to the pain experience of those without risk [10–12]. In contrast, no infants in the current study were described as having no risk for neurological impairment. This was because the infants' appearance made it apparent that they were not healthy full-term infants. It may be that the comparative nature of the questions in the previous studies made the possibility of differences in pain experience due to neurological risk more salient to participants. Thus, the pain ratings provided here did not differ among levels of risk, but had ratings of healthy infants been included in the task, they may have differed significantly from them.
It is also possible that the beliefs expressed by professionals in our previous study  do not moderate professionals' behaviour in relation to pain assessment for specific infants, as was found here. A discordance between expressed beliefs and behaviour, in regard to pediatric pain management, has been reported elsewhere [20, 21]. Thus, the professionals here may hold similar beliefs to the professionals in our previous study, but these beliefs did not alter their behaviour when asked to judge pain in a specific infant based on observable behaviour. This interpretation is supported by the current results because no differences were found due to level of risk for ratings that the professionals could base on behaviour they observed on the video clips: pain, distress, time to calm. In contrast, professionals' judgments of the effectiveness of cuddling were influenced by the descriptions of the infants' level of risk for neurological impairment. This may be because there was no visual information to base this rating upon, so professionals used the descriptions of risk provided, presumably in light of their previous experience with these groups in the neonatal setting.
The finding that pain ratings did not vary due to level of risk for neurological impairment raises questions about our previous study that revealed infants at risk for neurological impairment receive less pain treatment in the NICU . When a group is provided less medication for pain, it is typically assumed that this is because their pain was judged as less. However, it is possible that professionals hold beliefs about pain treatment that directly impact upon treatment decisions, irregardless of pain assessment. For example, they may hold beliefs about the appropriateness of medication for specific groups that are unrelated to beliefs about the amount of pain that group experiences. In support of this perspective, research indicates that nurses hold negative attitudes towards pharmacological treatment for pain  and that steps to improve pain assessment do not necessarily result in changes in pain management .
Further research is needed to reconcile the current results with beliefs that risk for neurological impairment does affect pain experience expressed by a similar group of professionals in our previous survey  and the results of our study indicating procedural pain is not treated as frequently for infants in the NICU who have greater risk for neurological impairment . If this reflects a disconnect between pain beliefs related to assessment and those related to treatment for infants at risk for neurological impairment, then educational interventions aimed at improving care through changes in pain assessment may be ineffective. In that case, other avenues to changing professionals' pain management for this group should be explored.
Another finding in this study warrants discussion. Professionals' judgments of the effectiveness of cuddling decreased with increasing risk for neurological impairment, despite their having judged pain as similar in intensity. This result is similar to a finding by Fanurik et al. . They found nurses, but not physicians, responding to vignettes of children undergoing painful procedures, indicated nonpharmacological interventions would be less appropriate as level of cognitive impairment increased. The same professionals' ratings of the pain intensity experienced by the children in that study did not differ due to perceived level of cognitive impairment.
The current results, along with those of Fanurik's group , raise the question of whether professionals perceive the pain experienced by those at risk for or with neurological impairment as similar in intensity, but differing in quality from those at lesser risk. Because the current study elicited ratings only of the intensity of pain and distress and professionals were not asked about the nature of the pain the infants experienced, the results cannot confirm this possible explanation, as data regarding pain quality was not collected. However, professionals in our survey study differentiated between physiological aspects of pain and internal and external responses to pain, such as emotional reaction, behavioural reaction and communication of pain . They also believed the experience of infants at greater risk was more reduced along the latter aspects that are more psychological in nature. Caregivers' have expressed similar beliefs, and also perceived the behaviour of children with more severe impairment is more closely related to their physiological pain experience . From this finding, we could suggest that there is a belief, on the part of professionals and caregivers, that the pain behaviour of those at greater risk for, or with, neurological impairment is more reflexive in nature. We could further speculate that the underlying rationale may be that they are seen as less able to interpret their pain, both cognitively and emotionally, due to their neurological impairment. However, we would need to conduct further research to substantiate this rationale.
If professionals and caregivers do believe pain behaviour is more reflexive, and that pain experience is more physiologically based when a child has neurological impairment, it could explain the current results regarding the effectiveness of cuddling. Professionals viewing the video clips may have perceived the behavioural responses of the infants with different levels of risk for impairment as being similar in intensity. Nonetheless, they may have interpreted the behaviour of those with more risk as more of a reflexive response to a physiological insult, while they saw the behaviour of those with lesser risk as reflecting a more multidimensional pain experience incorporating both physical and psychological suffering. Thus, we could again speculate that they may have felt cuddling, an intervention that would address physical and psychological aspects of pain, would be more effective for the less impaired groups. This phenomenon would not be novel or unique. For most of recorded history, there has been a belief that cognitive interpretation of pain was necessary for pain to result in long-term negative consequences. This belief was often the justification for poorer pain management for both children and infants . Although this belief is fading in regard to children and infants in general, it is still held in relation to those who are most severely at risk for, or have neurological impairment, and are perceived as least capable of interpreting their pain. Alternatively, this belief may be based on the actual experience of professionals in this study, that it is more difficult to calm an infant at risk for neurological impairment. This experience may also be an accurate perception of the difficulty infants at greater risk for impairment may have in responding to behavioural interventions because of their reduced ability to organize behavioural state and biobehavioural responses. Further research should examine these areas of speculation to specifically determine whether the perception that a behavioural intervention will be less effective for infants at greater risk for neurological impairment does reflect professionals' direct experience with this group or their understanding of how the pain experience may be affected by neurological impairment that may affect pain interpretation.
The current study has several limitations. Professionals were asked to rate the pain experience of infants receiving heel sticks from videotape. Although this may approximate the real situation in a NICU setting, it is not identical. In a NICU setting, professionals would have rich information from the environment, previous contact with an infant, physiological data, and medical records that guide their assessment of pain. They would also view this infant within the context of all other infants in the unit. Professionals here were also asked only to provide ratings of pain intensity. As the results suggest, this is only one dimension of pain and may not be the dimension that plays the largest role in their judgments regarding pain in a clinical setting. The professionals here were experienced in the types of pain experienced in the NICU and may have held a priori beliefs about the painfulness of this procedure that moderated their judgments. Research suggests professionals' beliefs regarding the painfulness of a procedure play a large role in their assessments of children's pain [7, 9, 25].