Assessment of paediatric pain: a critical review

Introduction

The clinical management and assessment of pain associated with care of newborns and young infants is on the threshold of dramatic change. Paediatric pain assessment, an essential component of paediatric care, has received much attention over the past decade. The International Association for the Study of Pain has developed a standard definition of pain, i.e. “Pain is an unpleasant sensory and emotional experience, associated with actual or potential tissue damage”. The elements of actual or potential tissue damage are valuable in paediatric pain management as their assessment is difficult. In 2001, the American Academy of Pediatrics and the American Pain Society issued a statement to ensure humane and competent treatment of pain and suffering in all children and adolescents in order to focus the attention on an interdisciplinary therapeutic approach, including pharmacologic, cognitive-behavioural, psychologic, and physical treatments [1]. Children’s pain perception should be an integral part of their pain assessment, which supports Meinhart and McCaffery’s [2] statement that pain is whatever the person says it is. Thus, it is important to listen to and understand the child in order to create a “child-centred focus” in managing paediatric pain [2, 3].

Erroneous assumptions in earlier investigations have hampered efforts to assess and manage pain effectively in children [4]. A prevalent belief is that children do not effectively experience pain, on the assumption that incomplete myelination of nerve fibres precludes the transmission of pain impulses [5]. Prevalent myths substantiate the potency of this assumption: (a) Young children are unable to sense pain neurologically; (b) they do not interpret noxious stimulus as pain; and (c) they do not experience the deleterious consequences of severe pain in the same way as adults. However, research on animal models and clinical research has challenged these myths. Neuroscience research has documented the intactness and functional ability of the neuroanatomical apparatus and the neurochemical systems involved in pain transmission and modulation in children. Clinical research has shown that preterm neonates can mount stress response to noxious stimuli, and these responses can be blocked with analgesics [6, 7]. The argument that preverbal children do experience pain is further corroborated by the fact that research has demonstrated the presence of intact neural pathways of pain transmission and modulation [8].

For the preverbal population, the American Society for Pain Management Nursing has recommended using multidimensional aspects that are characterised by both behavioural and physiological indicators of pain [9]. Many previous studies have suggested the need for a pain assessment tool that uses multimodal variables [10–13]. Different techniques do already exist to assess pain in children. In infants, the pain assessment methods/measures used are cry, movement, facial expressions, and physiological changes [14–19]. It has been reported that facial activity is strikingly visible in infants reacting to noxious events [15]. The Neonatal Facial Coding System (NFCS) and the Facial Action Coding System (FACS) are the two measures used to examine facial expressions in infants [15]. The NFCS focuses upon a limited subset of all possible facial actions that had been identified previously as responsive to painful events, whereas the FACS is a comprehensive system that is inclusive of all facial actions [15]. Another neonatal pain assessment tool is the CRIES [18]. It is an acronym of five physiological and behavioural variables previously shown to be associated with neonatal pain: C – crying; R – requires increased oxygen administration; I – increased vital signs; E – expression; S – sleeplessness. This tool has been reported to be valid, reliable, and well accepted by neonatal nurses [18]. In a recent study, the diagnostic performances of the Face, Legs, Activity, Cry, and Consolability Scale and Neonatal Infant Pain Scale were reported to have excellent sensitivity and specificity for pain assessment in infants [19].

In children, an earlier study has provided a preliminary indication that the Child Facial Coding System (CFCS) can be reduced to components that reflect several aspects of children’s acute pain experience and predict self-reports and observer reports of children’s pain [14]. The CFCS comprises 13 discrete facial actions, used to code each second of five 10-s phases from videotape: baseline, preneedle, needle, postneedle, and posthandling. The children provided a self-report using a Faces Pain Scale, and parents and nurses rated the children’s pain and anxiety using visual analog scales [14]. Recently, a smartphone application called Panda has been developed that contains an electronic version of the Faces Pain Scale-Revised (FPS-R) and the Color Analog Scale (CAS) for postoperative pain assessment in children [20]. This application can be used in lieu of the original FPS-R and CAS for assessment of pain in children, and the children’s preference for the application may translate to improved cooperation with self-report of pain [20].

In toddlers, several previous studies have explored behavioural and verbal expressions [21–24]. An earlier study has reported that among the five pain scales, i.e. word-graphic rating scale, a visual analogue scale, a graded-graphic rating scale, a magnitude estimation scale (0–10), and a colour scale, the word-graphic rating scale was the most supported for measuring pain intensity in paediatric populations [23]. Recently, the COMFORT behavioural scale has been reported valid to assess sedation, pain, and distress in toddlers unable to report pain [24].

With preschoolers, interval rating scales exist, while tools for older children include the visual analog scale, body outline, projective techniques, etc. [25]. The children clearly describe pain, and there are no appreciable differences by age groups, but children who are hospitalised describe pain differently from children who are not [25]. Despite these, a definite paediatric pain assessment tool still eludes us.

Nociception and the neural system in children

While the neonate’s pain transmission and modulation is intact and functional, research has revealed some pertinent development issues. Myelination of nerve fibres is incomplete in preterm and term neonates, and also perhaps in young infants [5]. The primary afferent axons in the peripheral nerves vary in size and degree of myelination and include large myelinated fibres, small myelinated fibres, and small unmyelinated fibres. The myelinated fibres are not myelinated at birth, and myelination occurs during the early months of infancy [26]. Incomplete myelination, however, does not alter pain perception because of lack of myelination. It merely implies slower conduction of pain in neonatal nerves or central nerve tracts, and this is offset completely by the shorter interneuron and neuromuscular distances travelled by the impulse. Hence, despite incomplete myelination, young infants appear to have the appropriate peripheral and central neural mechanisms to encode pain. The C fibres are mature from an anatomical and neurochemical perspective, but their full physiological potential develops slowly during the postnatal period. Sensitivity to pain increases as the physiological functions of the C fibres reach maturity, and the neonatal period is critical for this maturation to occur [27, 28].

Endogenous pain control mechanisms that dampen pain impulses appear to be lacking or are immature in the young infant [29]. Serotonin levels in the young infant are low and may limit the effectiveness of the endogenous pain control mechanisms [27]. These findings led Fitzgerald [27] to speculate that the established pain in the neonate could be worse than the observed pain response. Neuroscience research primarily using animal models suggests that neural pathways for pain in preterm and term neonates are neuroanatomically intact [27]. Of importance are the findings that preterm and term neonates (a) exhibit unpredictable, imprecise, disorganised, and weak behavioural responses to pain secondary to the continuing postnatal development of the neurological system; (b) may lack the ability to modulate pain effectively through endogenous mechanisms; and (c) may experience long-term structural changes in the neurological system secondary to untreated pain. With maturing of neural pathways, older infants and toddlers (a) exhibit more precise, more organised, and stronger behavioural responses to pain and (b) may have the ability to modulate pain more effectively [8, 27].

Lacunae in paediatric pain assessment

Like an adult, a child’s pain is also plastic and complex. Environmental and internal factors do modify the pain experience [30]. Children are known to evaluate their pain in relation to a particular frame of reference, i.e. the number, diversity, and strength of previous pains experienced [31]. Episodes of pain could register in their memory and in turn influence a wide range of their sensory, affective, and behavioural reactions. In a study of how attitudes about pain change as children develop, they were asked, What is pain? Provide examples of smallest and strongest pains. Most children provided different answers when asked the same questions 2 years later. This was probably because they had experienced a wide variety of pains during this period [32–34].

Optimal pain management in children requires an assessment of the factors that influence their pain. This normally involves determination of a qualitative (burning, aching, dull, sharp, etc.) or a quantitative scale (numerical values of pain) to provide meaningful information about the pain experience. Reliability and validity of children’s pain scales cannot be assessed in the same manner as adults. Adults participate in laboratory studies in which they rate how much pain is produced by a well-controlled noxious stimulus, when all other factors are controlled. With children, however, it is not possible to use experimental pain to determine a reliable and valid pain measure, since the responses vary. Instead, it could be possible to determine the reliability and validity in the natural settings in which the child experiences pain (Table 1), but this is difficult to achieve [35].

Measurement and assessment of pain in infants and children remains a challenge. Many studies in this regard have been conducted to validate and use multicomponent behavioural measures in prematures and neonates [16, 36, 37]. Grunau and Craig [16] provided additional evidence that coding of facial actions is more accurate with tissue-injuring procedures than with factors that increase arousal per se, that these measures change appropriately with analgesic treatments, and that they correlate with physiologic variables in restricted settings. The utility of the NFCS and the FACS has been supported by Craig and co-workers [15]. Physiologic pain measures in the neonate continue to be studied but with little evidence of specificity across a range of physiologic circumstances. For example, it has been shown that there is some correlation between salivary pH and certain behavioural autonomic multi-parameter scores. However, since arterial and tissue pH are known to vary with a range of respiratory and metabolic circumstances unrelated to pain, it is unlikely that this measure will show any specificity for pain [38].

Measurement refers to application of some metric to a specific quality of pain (usually intensity), while assessment is a broader endeavor, which encompasses the measurement of interplay of various factors on the total pain experience. While measurement has become increasingly sophisticated, assessment has lagged behind. Various assessment packages for paediatric pain have been developed. Though useful, they have not been guided by an explicit conceptual model [39–41].

Pain assessment should be in accordance with the type of pain and the age of the child [42]. There have been several theoretical arguments favouring the presence in the neonates of both acute sensitivity to pain and effective communication systems for signalling painful distress [16]. Research has concluded that all sensory systems including nociception are functional and active in the newborn, but the cognitive processes necessary to attach a meaning and understand the nature of painful experience remain undeveloped. Perceptual processes in the newborn serve to organise the substantial inflow of information and provide the mechanism whereby the earliest experiences register and slowly acquire significance [43].

Behavioural and physiological responses to noxious stimuli have been monitored in infants to identify pain responses [44]. Paradoxically, we have better behavioural pain measures for use with newborns than we do for 2–4 year olds, in part because available measures in toddlers and preschool children are so confounded by fear and anxiety. Goodenough and coworkers [45] analysed self-report scales using a range of facial expression scales and concluded that there were limitations of each and no reason to pick one as the “best” of these sorts of scales.

Behavioural responses in infants include overt body movements (withdrawal, rigidity in the limbs, torso motion, etc.), cry/cry patterns, and specific facial expressions [19, 20]. Behavioural methods infer an infant’s pain from the presence/quality of certain overt distress responses. An earlier study has shown that the infant’s distress responses during noxious stimulation are not consistent with the tissue-damaging stimulus, and instead they represent complex responses dependent on the age and behaviour state of the child [16]. Though there are better pain assessment measures in older children, no single behaviour constitutes an unequivocal measure of an infant’s pain. However, characteristic patterns of distress behaviour are now emerging after their analysis. Further studies are needed to evaluate how distress indices vary in relation to the quality, intensity, and duration of pain so as to understand how an infant’s particular behaviour reflects his pain. As yet, we have no simple, direct, and practical behavioural measures to precisely identify an infant’s pain [36, 37].

Physiological responses that have been monitored in infants include heart rate, respiratory rate, palmar sweating, transcutaneous oxygen levels, blood pressure, and cortisol levels [46, 47]. Till date, no single physiological measure has been identified that constitutes a pure pain measure in infants. Physiological responses are actually complex body responses to stress situations and not specific pain responses [48]. It is an interesting observation that, though the physiological state of the critically ill infant is monitored continuously, little attention has been focused on establishing criteria for an infant’s pain status. Records of physiological health indices during an infant’s recovery can yield important clues to which physiological systems are involved in the propagation of pain. These studies should be complemented by research on infant’s responses to novel or stressful non-pain-producing stimuli in order to determine the physiological responses specific to noxious stimuli. Several studies have reported that physiological pain responses are not ideal to assess pain and suggest that the pain assessment tools be based on behavioural pain indicators [10, 49, 50]. Further research is required to determine how responses change according to development level, health status, and the nature of noxious stimulus.

Physiological and behavioural distress measures as potential pain measures have also been investigated in older children. Children’s physiological responses mirror their overall distress state in a painful situation [51, 52]. There is insufficient evidence to conclude that physiological responses correlate directly with the pain experience. In fact, some physiological changes occur in direct response to the quality, intensity, and duration of the noxious stimulus, whereas other changes occur in response to the child’s stress and reactions to the stimulus [40].

Behavioural methods for assessing pain in older children generally include observations in which independent raters record children’s behaviours when they are in pain. An objective evaluation of the type and frequency of pain behaviours could provide an accurate estimate of the child’s pain experience. Behavioural indices are actually an integrated index of the child’s anxiety, fear, distress, and pain. These indices are not specific pain indices but reliable measurements of a child’s overt distress [51, 52]. Children’s behaviour in response to noxious stimuli is actually a passive reflection of their pain. Some children behave stoically and still experience pain, while others exhibit distress responses even before the pain stimulus has been given. Some distress behaviours may indicate pain in one situation, but this behaviour may not be appropriate in another. Behavioural measures in children are still qualitative and do not provide sufficient details to understand a child’s pain experience [40].

Attention has also been focused on the use of self-report measures in which verbal children directly describe the quality and intensity of their pain [53, 54]. A child’s understanding of pain and the ability to communicate his understanding is dependent on their developmental level and the nature and diversity of their pain experience [25]. Although the language they use, ingenuity of their pain descriptions, and quality and diversity of their plans are different, it is clear that these children should understand the concept and the multidimensional nature of the pain experience. Children’s verbal descriptions of pain do help facilitate medical diagnosis by providing information about the quality, severity, location, and duration of pain. They also provide information, which helps identify the environmental and internal factors that exacerbate their pain. An array of projective methods in which the children’s pain attitudes are inferred from their colour selections, drawings, etc. have been used as self-report methods [40, 55, 56]. However, we really do not know as to how qualitative differences in children’s responses reflect quantitative differences in their pain perceptions. Direct scaling techniques have also provided the most flexible and comprehensive method of pain assessment in older children. However, factors like cognitive level, previous pain experiences, and age must always be kept in mind while evaluating these techniques [57].

Thus, a wide variety of pain measures and different tools are available to assess paediatric pain. Sometimes, these measures are integrated to form a comprehensive package for evaluating all aspects of a child’s pain. However, there are many lacunae and we still lack a specific objective pain measure in children.

Management

It appears reasonable to attribute a capacity to perceive pain to the neonate, but just what the neonate actually experiences remains an open question. Neonatal pain is qualitatively very different from the pain experienced later in life, when it gets transformed through biological maturation [58]. We normally assume that the neonatal pain experiences comprise the same sensory, affective, and cognitive components as adults, but they are at the rudimentary level. Thus, in infants, we should not expect finely differentiated reactions as observed in adults. It can be speculated that absence of finely differentiated sensory information could lead to an even more intense reaction to pain in infants. We have now moved from the position that only undifferentiated precursors of emotions exist in the form of distress and non-distress states to a position that infants are born biologically capable of experiencing and communicating complex emotional states. Infants have been observed to be capable of experiencing emotional reactions whose assessment is qualitative in nature [34]. The most confusing is the cognitive domain due to the absence of history of direct experiences of pain to give a meaning to it. Previous literature suggests that it is inappropriate to conclude that children’s pain perception is diffuse and amorphous [59]. A great degree of behavioural organisation has been demonstrated [59]. It has also been indicated that children pass through distinct stages of understanding pain [25, 31]. The thoughts and perception of pain develop in an orderly and behavioural sequence of stages, each characterised by a distinct structure of thought that parallels the development of other health-related concepts [3]. Theoretically, the newborn may remember painful experiences and adjust their behaviour when confronted with similar noxious events [60]. However, it is unlikely that the effects of the painful experiences on the memory of the newborn are explicit and subject to recall. Amnesia for infant experiences in the adults is well documented [61].

The infant is predisposed to adapt and learn through experience. Painful events induce a state of arousal, which optimises learning [31]. While evidence supports the theory that painful experiences have an impact on the subsequent behaviour of the newborn, little can be said about the duration and retention of the impact [31]. At some stage of development, painful experiences do produce memory traces, and these provide the basis of cumulative patterns of response to pain and have a long-term impact on the child’s behaviour. Memory for painful events is readily demonstrated through habituation research [31, 62]. Habituation to non-noxious sensory stimuli in the newborns is known, but it is still uncertain how responses to noxious events would change with repeated exposure to the same event. In clinical situations, infants are sometimes exposed to identical invasive procedures at brief intervals. Franck [63] reported more rapid cry and reflex withdrawal to a second heel-stick stimulus, but his statistical analysis was lacking. Several investigators have discussed long-term consequences of untreated pain in the newborn. Taddio and colleagues [64, 65] added to their provocative and fascinating finding that newborns circumcised without anaesthesia, when compared to those circumcised with a eutectic mixture of local anaesthetics or to uncircumcised neonates, showed more intense and prolonged behavioural responses to the pain of immunisations at 4 and 6 months of age. The implications of Taddio’s work and related studies of long-term consequences of pain are extremely noteworthy. They provided concrete evidence to refute the view that untreated suffering is of no lasting effect on newborn. These observations substantiate that the pain experience of the newborn has an important role on his subsequent reactions to noxious stimuli, which lack uniformity [38, 63, 66].

Though many behavioural, physiological, and psychological methods are available for the investigation of paediatric pain experience, all are qualitative and inadequate and classify well as measures of overt distress rather than pain. These measures are not versatile and do not provide quantitative estimates of different dimensions of the childhood pain experience; they only give us an indirect inference of a child’s pain. Hence, no measure is available that constitutes a pure pain measure [67]. Self-report measures do provide a natural format for obtaining information about the pain experience, but they are restricted to children who can communicate. Direct scaling measures also provide quantitative estimates of a child’s pain experience, but they are possible only in children who are able to make proportional judgements about their pain perceptions. Though we do have a whole lot of qualitative measures for clinical appraisal of pain in children who can communicate, we do not have a definite measure that provides us with quantifiable information of their pain experience [32, 68].

Current and future trends

Though today’s assessment of pain in children is qualitative, research is now being directed towards development of an objective tool for assessment. There is a challenging need for an accurate measure of childhood pain for research and clinical purposes. Studies have demonstrated usefulness of somatosensory evoked potentials as potential tools for assessing pain in children. Monitoring such as online spectrographic analysis of EEG may demonstrate specific features, which may have potential application as objective pain measures [34, 69]. It is now argued that the foetus and the newborn can perceive and respond to pain. Evidence that the newborn have the anatomical and functional components required for pain perception has been established with evaluation of techniques like EEG and somatosensory evoked potential changes to noxious stimuli [34]. It has been established that acute pain triggers a stress arousal response stemming from increased sympathetic activity, and this response is absent in infants receiving opioids [31]. Several potential ratio-level measures of speech seem to be influenced by stress arousal. These ratio-level scales have been found to have a true zero point, and so data could be calibrated on this scale. Stress arousal increases striated muscle tension, rate and depth of breathing, and the airway diameter, which thus elevates the voice pitch. These could be promising indicators since cry is thought to involve complex neurophysiological mechanisms that provide information about changes in the infant’s biological status. Guided by this conceptual framework, researchers are now involved in using these ratio-level acoustic variables either singly or in combination with other pain assessment scales, to find a potential, reliable, objective, and quantitative pain scale in children. These techniques are still in research stage, and their clinical application could be a distant possibility [70, 71].

Some of the most fundamental types of perceptual research still remain to be attempted. Sensitivity and the pattern of response to different magnitudes of noxious stimuli are still largely unknown, and there is a need to examine these. Till date, the emphasis has been on phasic rather than long-term pain in infants. The pattern of response to persistent pain needs to be examined. The database for developmental changes in sensitivity is also inadequate, and we are still uncertain whether the earlier studies are methodologically flawed or valid. The special challenges of assessing pain in the acute care settings have been inadequately addressed. Usually, the equipment needed for acoustical analysis, pschophysiological measures, etc., are often not available and also impractical since immediate decisions are made there. Further, intensive care nurseries may be silent because the babies are mostly intubated so that they cannot cry, making it even more difficult to assess their pain. Facial expression may be difficult to discern because of the equipment obscuring their face. Studies of risk-benefit ratios of different analgesic infusions in neonates are necessary and should focus both on short-term end-points, including pain scores, and on longer-term medical and neurobehavioural consequences [34, 38].

Conclusions

Several studies have led to an awareness of a substantial sensitivity to painful events in children. However, there is a striking inability of the medical community to recognise and quantify paediatric pain. Clear demonstrations of investigating paediatric pain are not available. Misperceptions of children’s pain are mainly attributed to pain assessment being still qualitative in nature.

The challenge to the medical fraternity involves deciphering nonverbal clues in preverbal children. Information is embedded in the complex, ongoing activity of the infant in distress, but careful attention is needed if pain is to be differentiated from other aversive states. Early investigations of neonatal pain were inadequate in their use of systematic observation, methodological controls, and quantitative analysis, leading to inaccuracies. Fortunately, now sophisticated measurement strategies are being developed that attend specifically to infants’ vocal and expressive behaviour. Advances in the study of cry, facial activity, limb and bodily movement, various physiological responses, and other activity have contributed to a better understanding of the neonatal pain experience. However, most of these assessment strategies are not quantifiable, difficult to apply during clinical development, and focus primarily on the initial response to the acute events, thereby leaving us uncertain about the perception of persistent pain. Crying serves as the infant’s attention to temporal characteristics; amplitude and other acoustical qualities may provide information to ascertain a child’s pain experience. Failure to recognise pain in infants could lead to lasting effects, and the episodes of pain could register in their memory and influence a range of sensory, affective, and behavioural reactions to a painful situation. The study of the persisting effects of pain in infants still remains open in the field of pain perception research.

Though assessment of pain in verbal children appears simpler, because of the indirectness and developmental complexity of the pain experience in children, the interpretation is difficult. Pain may be expressed differently even in older children, making pain assessment sometimes quite confusing.

Despite considerable progress, there is a critical need for a more accurate paediatric pain measurement tool for both research and clinical purposes. Though the future may hold considerable challenges, we may perhaps never have an ideal measure for paediatric pain of different age groups. Longer-lasting pain may pose a greater challenge than short and sharp pain in the future. Research in future may be directed at the mechanisms underlying long-term cognitive, emotional, and behavioural effects of pain in children, which in turn could help investigators look for an accurate or near-accurate paediatric pain assessment tool.