Journal of Pediatric Psychology, Vol. 27, No. 2, 2002, pp. 209-214
© 2002 Society of Pediatric Psychology
Brief Report: Behaviors Identified by Caregivers to Detect Pain in Noncommunicating Children
Royal United Hospital
All correspondence should be sent to Paul Stallard, Department Child & Family Psychiatry, Royal United Hospital, Combe Park, Bath BA1 3NG, England. E-mail: paul.stallard{at}awp.nhs.uk .
 |
Abstract |
|---|
 Top Abstract IntroductionMethodResultsDiscussionReferences |
|---|
Objective: To develop an observational measure, based on caregiver reports, to assess chronic pain in children with significant cognitive impairment who are unable to communicate verbally. The issue of whether these children share a core set of cues to express pain was investigated.
Methods: Specific pain cues were elicited during detailed interviews with 29 female caregivers of noncommunicating children. Pain cues were categorized by a two-stage Delphi process and cues indicating severe and definite pain identified.
Results: Six cues from five different categories were used by 90% of caregivers to identify definite or severe pain in their child.
Conclusions: Although the specific expression of pain may be very individual, there is a shared set of core pain cues. The relationship between these cues and evidence of pain and distress is discussed.
Key words: pain; assessment; noncommunicating children; caregivers.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
Children with severe cognitive impairment are at significant risk of experiencing acute and chronic pain (McGrath, Rosmus, Camfield, Campbell, & Hennigar 1998
). Medical procedures and comorbid conditions
may cause pain while sedating drugs or disruptive behaviors may mask signs of
pain. The most severely impaired children will be unable to verbally
communicate, and their physical limitations may prevent them moving away from
painful stimuli.
Caregivers are well placed to assess the possibility of pain in their
children. They are sensitive to changes in their child's behavior and
demeanor, even when those changes are subtle. There is some evidence that
caregivers believe themselves to be better than health professionals at
assessing pain in their child and feel that health professionals ignore their
expertise (Carter, McArthur, &
Cunliffe, in submission;
Ljungman, Gordh, Sorensen, & Kreuger,
1999). Some studies find caregivers underestimate or undermedicate
pain (Chambers, Reid, McGrath, &
Finley, 1996; Finley, McGrath,
Forward, McNeill, & Fitzgerald, 1996). Chambers, Reid, Craig,
McGrath, and Finley (1999)
argue there is a lack of agreement between studies exploring caregivers'
ability to assess pain, with methodological problems making the findings
unclear. Most studies are based on caregivers' perception of pain in children
without learning disabilities and therefore may not reflect caregivers'
assessment of pain in their noncommunicating child.
Two groups of researchers have devised scales for assessing pain in
noncommunicating children, specifically those with cerebal palsy. Giusiano,
Jimeno, Collignon, and Chau
(1995) developed a scale
consisting of 22 items that medical staff considered to indicate pain. McGrath
et al. (1998) notes that these
items are mainly elicited in response to medical examination and so may have
only limited value in assessing pain in other situations.
McGrath et al. (1998)
devised a checklist (Noncommunicating Children's Pain Checklist) to be
completed by caregivers of noncommunicating children. Caregivers were asked to
recall pain episodes and to describe the pain behaviors they observed. In
total, 31 behaviors covering seven different domains were identified, although
most items were used by less than half the respondents. McGrath et al. note
that it may not be possible to identify a single set of items that can be used
reliably with this population, as the expression of pain is very
individualized. However, the authors argue that the widespread use among
respondents of all seven categories suggests commonality across individuals,
even though the specific nature of the child's response may vary.
The commonality of cues was explored in a subsequent study where the
checklist was completed by 33 caregivers
(Breau, Camfield, McGrath, Rosmus, &
Finley, 2001). A core set of seven items was found to be
significantly predictive of pain. The items were seeking comfort, gesturing to
part that hurts, tears, crankiness, change in eyes, less activity, and
gasping, with the first three being the strongest predictors of pain. Although
this study focused on noncommunicating children, over half of those assessed
were suffering from cerebal palsy.
The purpose of this study is to extend the work undertaken by McGrath et al. to children with conditions other than cerebral palsy. The study aims to develop an observational measure based on caregiver reports to assess chronic pain in children with significant cognitive impairment who are unable to communicate verbally. We investigated the issue of whether a core set of cues is used to express pain.
|
Method |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
Participants
Participants were the main caregivers of children in receipt of clinical services from Lifetime, a National Health Service community-based nursing and psychology service for children with nonmalignant life-limiting conditions based in Bath, United Kingdom (UK). Children meeting the project criteria of no expressive language (or alternative communication system such as sign language) and age 2 years or older were identified from the clinical database.
We contacted a total of 34 eligible caregivers, of whom four declined to participate (one was moving away from the area; one was too busy; one did not want to take part in research; and one declined to give a reason). Only one eligible participant was male; to avoid any possible gender effects, information from this participant was not entered into the analysis. This study therefore reports information from 29 female caregivers, ages 24-60 years, with a median age of 37 years. Of these, 25 were the child's birth mother, 2 were adoptive mothers and 2 were foster parents.
All 29 children studied had a chronic or lifelimiting condition. The largest group (n = 10) were those with a primary diagnosis of static central nervous system disorders such as cerebal palsy, spina bifida, and postencephalitis. Nine were diagnosed with either chromosomal disorders (e.g., partial trisomy of chromosome 13, deletion of chromosome 1) or other syndromes affecting the central nervous system (e.g., Goldenhar syndrome, Opitz C). A further five had progressive central nervous system disorders such as Retts syndrome and MPS. The remaining five children had no specific diagnosis and at the time of the study were still undergoing further investigations. The children consisted of 13 girls and 16 boys, ranging in age from 2 to 21 years, with a mean age of 7.8 years (SD = 5.28 years).
Measures
Caregivers undertook a semi-structured interview lasting approximately 1
hour. The interview focused on general and chronic pain rather than specific
(i.e., bumps and bruises occurring during play) and acute episodes. They were
asked how often they thought their child experienced pain, the duration of
these episodes, and the suspected nature of the pain (constant vs.
intermittent). The interviewer asked the caregivers to list the cues they used
to identify pain in the child. They were prompted to elaborate on their
answers to provide a full and clear picture. Specific cues thought to indicate
severe/mild and chronic/transient pain were assessed, and caregivers were
asked to identify those cues they considered to definitely reflect pain. After
spontaneously generating pain cues, the caregiver was shown a list of the pain
cues identified by McGrath et al.
(1998) and asked whether this
contained any cues that had been overlooked. Finally, the interviewer
completed a summary chart with the caregiver to check that the information had
been recorded correctly.
Procedure
Approval was obtained from the local ethics committees for all phases of
the project. Participating families received an information sheet and
completed an informed consent form.
The main caregivers of the children received a letter explaining the aims and method of the project and were followed up by telephone a few days later. Interviews were undertaken in the caregiver's home by a trained clinical psychologist with experience working with parents of severely cognitively impaired children. While undertaking this project, the psychologist also worked part-time as a member of the Lifetime clinical team.
|
Results |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
Caregiver Perceptions of Child Pain
A total of 12 (41%) caregivers assessed their child as experiencing pain almost every day, with a further five (17%) reporting their child to experience pain at least once a week. The duration of typical pain episodes was judged by nine (31%) to be less than 3 hours, while 11 (38%) assessed their child as experiencing pain for longer than 24 hours. Of the 26 caregivers who felt their child experienced regular pain, half (54%) recalled that their child was in pain within the first year of life.
All caregivers rated themselves on a four-point Likert scale as "accurate" or "very accurate" at detecting pain in their child. The majority, 25 (86%), believed they were correct nearly all of the time, with a further three believing they were right on 75% of occasions. One respondent believed she was correct only 50% of the time.
Behavioral Pain Cues Identified by Parents/Caregivers
In total, the respondents spontaneously identified 251 cues they felt
indicated pain in their child. Mean number of cues identified was 8.65, with a
range of 4-13. There was exact overlap between some of these cues, so, for
example, 19 caregivers reported that the child "cried." Identical
cues were combined, resulting in 203 separate pain cues identified.
Categorization of Pain Cues
A two-stage Delphi process (Dalkey,
1972; Jones & Hunter,
1995) was undertaken in order to sort the 203 pain cues into
common groupings. The pain cues were distributed to 10 Delphi participants,
comprising three clinical psychologists, one consultant community
pediatrician, one community children's nurse, two university researchers, two
parents of noncommunicating children (who were not involved in the study), and
one university psychology student. The raters were given the full list of pain
cues and asked to sort them in a way that made sense to the rater and to give
each category a name.
After the first round the 10 raters generated a total of 77 categories (mean = 7 categories, range = 5-11). Although some categories were used by only one rater, there was considerable overlap in the way many of the cues were grouped.
For the second stage, the 77 categories were circulated to the original raters, who were asked to consider their responses in the light of the categories identified by the other raters. Each was provided with a list of their original categories along with the categories devised by the others. In total, 16 categories were identified in the second round. Of these, 11 categories were endorsed by two or more raters, with 5 being endorsed by only one. These five categories were discarded. The data were checked to ensure that the cues included in these categories were not identified by parents as "definite signs of pain," occurring "only when the child was in pain" or indicated "severe pain." None of the five categories contained any cues that met these criteria. The remaining 11 categories identified by the Delphi process were defined as follows: vocal, facial, physical, withdrawal, seeks comfort, physiological, agitation, tense, inconsolable, pain sites, anger/irritability.
The 203 pain cues were assigned by one of the authors (LW) to the category in which they were most frequently placed during the Delphi process. The complete set of categories and cues was circulated to the original Delphi participants, who were asked to identify any cues they considered to be incorrectly allocated and to suggest a more appropriate category; 179 (88%) were judged to be correctly placed. They made a total of 28 suggestions about 24 separate cues. These cues were discussed in a meeting of four members of the project team and were allocated on the basis of consensus to one of the suggested categories. The number of cues falling within each category is summarized in Table I.
|
Identification of Core Pain Cues
During the interview, caregivers were asked to identify the cues they felt
indicated either severe pain or were considered as a definite sign of pain.
They identified 95 cues as indicating definite pain and 68 as indicating
severe pain, The 11 categories that emerged from the Delphi process were
examined to determine where these cues fell
(Table II).
|
The three most frequently endorsed categories for definite or severe pain were vocal, facial, and inconsolable. Within these categories, there was considerable overlap in the cues caregivers reported, with the following four being the most commonly identified: screams/yells; cries with or without tears; unable to be comforted; face screwed up/looks distressed. The next most frequently endorsed cues were flinches from contact (physical category) and appears tense and stiff (tense category). Cues in the remaining six categories derived from the Delphi process were less frequently endorsed as indicating either severe or definite pain. Important pain cues were therefore identified in almost half of the categories and did not cluster within a single domain.
An analysis was undertaken to determine how many caregivers endorsed one or more of the most frequently reported six core cues. A total of 26/29 (90%) caregivers identified one or more of these cues as signs of definite pain with 20/29 (69%) identifying two or more. In terms of severity, 22/29 (76%) identified one or more as indicating severe pain, with 14/29 (48%) identifying two or more.
|
Discussion |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
This study extends the work of McGrath et al. exploring the expression of pain in noncommunicating children, to those with a wider range of severe cognitive impairments. In considering these findings, the limitations of this study need to be acknowledged. In particular, the study cohort was comparatively small; although none of the children had any verbal communication, the range of primary diagnoses often resulted in only one child with a particular condition. These results should therefore be viewed as preliminary, since further research with larger, more homogeneous samples is required to substantiate these findings.
All of the parents we interviewed rated themselves as accurate or very accurate assessors of pain. Independent verification by comparing caregiver and child ratings is not of course possible with this client group, so the degree of caregiver accuracy remains unclear. However, the confidence expressed by caregivers was noticeable, with 86% reporting that they nearly always could detect pain in their child. This was further reflected during interview in their ability to identify possible pain cues. On average, each caregiver identified seven different pain cues, suggesting that they were aware of, and sensitive to, possible indicators of pain.
Our results suggest that a core set of six cues is used most frequently and
that these were used by 90% of caregivers as indicators of pain. Of these, two
cues, crying (with or without tears) and screaming/yelling, were used by 23
(79%) of caregivers. These results are consistent with the findings of McGrath
et al. (1998) and Breau,
McGrath, Camfield, Rosmus, and Finley
(2000). The cues of crying and
moaning were found to be used by 80% of the parents of children with cerebal
palsy (McGrath et al., 1988) and over 70% of the parents of cognitively
impaired children (Breau et al.,
2000) to indicate pain. In a subsequent study, Breau et al.
(2001) identified a core set of
seven pain cues that caregivers used to identify pain in their child These
findings therefore lead us to suggest that, although the specific expression
of pain may be very individual, there does appear to be a set of shared core
cues.
The six core cues used by caregivers were not focused around one domain of
behavior. The cues were spread across five of the eleven categories identified
by professionals during the Delphi process. Caregiver perception was therefore
based on vocal (screaming/yelling), visual (crying, distressed face, tense
body), and physical (difficult to comfort; flinches if touched) cues. However,
these cues are often identified during acute pain episodes, and it is unclear
whether they also reflect chronic and enduring pain. For example, during
chronic pain the cues may become more subtle and may result in decreased
activity and the person may become more withdrawn or subdued
(Chambers et al., 1996;
McGrath et al., 1998). Whether
caregivers are sensitive to and can identify these more subtle cues as
indicating possible pain in their children is unclear.
There is some similarity in the domains of the core pain cues identified by
Breau et al. (2001) and those
in this study. Although the specific terminology and exact nature of the cues
differ, both studies found that crying, comforting, and facial changes were
important cues used by caregivers to detect childhood pain. In other respects,
there were important differences. In our study, screaming and yelling were
frequently identified as pain cues, whereas Breau et al.
(2001) did not find a
significant relationship between vocal behavior and pain. This finding is
surprising since vocalization is one of the core ways in which the sensation
of pain is signalled to another. This discrepancy may be due to methodological
differences in that Breau et al.
(2001) used a checklist,
whereas our study used an interview format. Alternatively, differences in the
vocal ability of the client groups assessed or greater or lesser caregiver
sensitivity to subtle changes in vocal behavior may be pertinent. Further work
is required to establish whether vocal behavior is an important domain of the
core cues used by noncommunicating children to express pain.
Finally, it is important to acknowledge that, while we have identified a shared set of cues, these are not exclusively related to pain. Children will, for example, cry and scream for many reasons other than being in pain. The core cues we have identified may be indicators of general childhood distress rather than specific indicators of pain. Further psychometric evaluation is required to determine whether these cues can reliably distinguish between pain and nonpain episodes. A further challenge is to ensure that any scale based on these cues has clinical utility and can easily and readily be used on an everyday basis to assess and monitor long-term pain in noncommunicating children. Once a psychometrically robust and clinically useful scale has been developed, its use with other groups of noncommunicating children, such as those with pervasive developmental disorders, could be explored.
|
Acknowledgments |
|---|
This project was supported by a grant from Sport Aiding medical Research for Kids (SPARKS). Dr. McGrath is supported by a Canadian Institutes of Health Research Distinguished Scientist Award.
Received January 22, 2001; revision received March 26, 2001; accepted July 26, 2001
|
References |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
Breau, L. M., Camfield, C., McGrath, P.J., Rosmus, C., & Finley, G. A. (2001). Measuring pain accurately in children with cognitive impairments: refinement of a caregiver scale. Journal of Pediatrics, 138, 721-727.[ISI][Medline]
Breau, L. M., McGrath, P.J., Camfield, C., Rosmus, C., & Finley, G. A. (2000). Preliminary validation of an observational pain checklist for persons with cognitive impairment who are unable to verbally communicate. Developmental Medicine and Child Neurology, 42, 609-616.[ISI][Medline]
Carter, B., McArthur, E., & Cunliffe, M. (in submission). Dealing with uncertainty: Parental assessment of pain. Archives of Disease in Childhood.
Chambers, C. T., Reid, G. J., Craig, K. D., McGrath, P.J., & Finley, G. A. (1999). Agreement between child and parent reports of pain. Clinical Journal of Pain, 14, 336-342.
Chambers, C. T., Reid, G. J., McGrath, P. J., & Finley, G. A. (1996). Development and preliminary validation of post-operative pain measures for patients. Pain, 68, 307-313.[ISI][Medline]
Dalkey, N. C. (1972). DELPHI. In J. P. Martino (Ed.), An introduction to technological forecasting. London: Gordon and Breach.
Finley, G. A., McGrath, P. J., Forward, S. P., McNeill, G., & Fitzgerald, P. (1996). Parents' management of children's pain following "minor" surgery. Pain, 64, 83-87.[ISI][Medline]
Giusiano, B., Jimeno, M. T., Collignon, P., & Chau, Y. (1995). Utilization of a neural network in the elaboration of an evaluation scale for pain in cerebral palsy. Methods in Infant Medicine, 34, 498-502.
Jones, J., & Hunter, H. (1995). Consensus methods
for medical and health services research. British Medical
Journal, 311,
376-380.
Ljungman, G., Gordh, T., Sorensen, S., & Kreuger, A. (1999). Pain in paediatric oncology: Interviews with children, adolescents and their parents. Acta Paediatrica, 88(6), 623-630.[ISI][Medline]
McGrath, P. J., Rosmus, C., Camfield, C., Campbell, M. A., & Hennigar, A. (1998). Behaviours caregivers use to determine pain in non-verbal, cognitively-impaired individuals. Developmental Medicine and Child Neurology, 40(5), 340-343.[ISI][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  D. O'Rourke The Measurement of Pain in Infants, Children, and Adolescents: From Policy to Practice Physical Therapy, June 1, 2004; 84(6): 560 - 570. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||