Accurate assessment is essential for appropriate and successful management of pain in children with cancer. Too often, however, the practice of assessing pain is forgotten, ignored, or conducted haphazardly. The inattention to assessment, in part, accounts for the undertreatment and inappropriate management of pain in children with cancer.
The state of the science offers minimal help to health care providers wishing to include pain assessment in their practice with children who have cancer. Recent reviews of the literature suggest that few studies on children with cancer have addressed pain assessment. Sutters and Miaskowski (98) identified only four studies that focused exclusively on assessment of cancer pain. Three studies (6,8,56,63) addressed pain associated with invasive procedures (i.e., lumbar punctures and bone marrow aspirations); one study (32) attended to pain associated with the disease. Also, other studies focused on estimating prevalence have included assessment of the intensity and source of pain (29,80,81). After a comprehensive review of child pain literature for the Cancer Pain Management Guidelines Panel for the Agency for Health Care Policy and Research, Hester1 concurred with Sutters and Miaskowski (98) that few studies have focused on assessing pain.
Other studies not focused exclusively on cancer have included children with cancer in their samples (e.g., ref. 51). However, these studies have addressed primarily acute pain related to invasive procedures and surgery. With so little research on pain in children with cancer, assessment approaches must be borrowed from other pain models, such as postoperative pain. Inherent in borrowing from other pain models are concerns of whether the approaches are appropriate for assessing cancer pain and whether they capture the complexity of the pain experience in children with cancer. These concerns are foremost for future research.
Contended here is that the deliberate and systematic use of a comprehensive
assessment approach can capture the pain experience of children with cancer
and can improve pain management. Proposed in this chapter is
a comprehensive assessment approach that emphasizes open communication
about pain. Open communication is best accomplished when partnership among
health care providers, the child, and family exists (48).
The use of a comprehensive approach to pain assessment will help attain the goals for caring for children with pain. These goals are to (a) enhance the child's comfort, (b) promote recovery when possible, (c) improve functional status, and (d) prevent detrimental effects from unrelieved pain.
The proposed pain assessment approach draws from guidelines previously developed for the assessment and management of pain (5,71,88,96,112) and includes the systematic use of a pain history, self-report and proxy measures, behavioral/observation approaches, and physiological measures. This information is integrated with knowledge about the child's disease, child and family factors, concurrent symptomatology, and unstructured observations to yield an individual judgment about the pain experienced by a particular child. A review of this approach yields information on whether research has provided sufficient and adequate tools for measuring pain in children with cancer.
Although research suggests that disease-related pain is low for childhood cancers, prevalence and incidence documentation is available in few studies. In a retrospective chart review of notations regarding pain for the decade beginning in 1973, Cornaglia et al. (22) reported that the occurrence of pain varied by diagnosis. For example, the incidence was 89% for Ewing's sarcoma; 58% for acute lymphocytic leukemia, and 38% for Wilms' tumor. These findings, however, may not be applicable to the current era, considering that diagnostic approaches and treatment modalities for cancer and pain have advanced since the conduct of the study.
A more recent study revealed that, for most children, pain is a presenting
symptom when cancer is diagnosed. Miser (81)
reported that 52% to 100% (M = 78%) of children and young
adults (less than 24 years) presented with pain. Prior to cancer therapy,
pain was present for a median of 74 days, ranging from 3 to 821 days. It
decreased after the initiation of treatment, persisting for a median of
Pain sources prior to diagnosis are disease-related; but, according to limited research, disease pain often remits rapidly once effective treatment begins (81). After diagnosis and therapy initiation, pain from invasive procedures (venipunctures, lumbar punctures, and bone marrow aspirations) (113) and therapy appear to predominate. For example, Elliot and colleagues (29) reported a low prevalence for pain after treatment initiation. The pain, according to Elliot, generally was associated with therapy side effects. Further findings from Miser (79) and Miser and colleagues (80) suggested that occurrence of chronic disease pain was low in children.
Therapy-related pain includes chemotherapy- and radiation-induced mucositis, postoperative pain, neuropathies from sources such as vincristine or amputation, abdominal pain from chemotherapy-induced nausea, radiation dermatitis, and infections (80). For children requiring surgical amputation, phantom limb pain becomes a problem. Diagnostic and monitoring procedures (venipunctures, lumbar punctures, and bone marrow aspirations) cause extensive pain and distress in children. Children frequently report pain from the invasiveness of these procedures as the worst hurt they could have (28,31,43). Weekes and Savedra (109) reported that adolescents associated cancer pain with needles. In addition to pain during procedures, pain also occurs following procedures. Hester and colleagues (51) found that children with recently inserted central lines reported small amounts of pain (1 to 2 on a 0-to-4 scale) a few days after the insertion, suggesting pain related to the inflammatory process. Pain also occurs after other procedures; one example is the prolonged, often intense, headaches that follow lumbar punctures (80). ``While not all cancer pain is severe, the nature and etiology of the pain in these children (e.g., `phantom limb' pain, radiation-induced mucositis) can create complex management problems'' (114).
Adequate identification and management of pain requires an understanding of the child's disease, its progression, its treatment, and the related diagnostic and monitoring procedures. An understanding of the disease will help the health care provider anticipate pain so that it can be prevented or promptly treated. The health care provider must be aware that children with cancer also experience pain from noncancer sources (e.g., headaches and abdominal pain), as do children without cancer. These pain experiences also deserve provider attention.
Unfortunately, the literature is bereft of research regarding an average course of pain for a child with different types of cancer. Especially needed is information on the occurrence and intensity of pain related to disease stages and on other common noncancer pains experienced by children.
Family factors include the meaning of pain; cultural background; health and illness of family members; socioeconomic situation; and experience with and expectations of the illness, hospitalization, and death. Little research has focused on these issues. Pfefferbaum et al. (86) reported that, although cultural differences affected the reliability of behavioral measures, findings in self-report and observed response between Anglo and Hispanic children were similar. A vivid example of how a father's illness affected a child's pain response is illustrated in this quote from a 10-year-old girl who was hospitalized for a leg fracture: ``When I have pain, it's hard on my mom because my dad has MS. She has to take care of him. ... So when I get hurt, I feel guilty sometimes because that puts more pressure on my mom and she has to work a lot harder to take care of both of us'' (40). A 7-year-old girl, hospitalized for diabetes, expressed her concern about the family financial situation: ``I don't like ... them to think that I'm really, really sick and spend all their money'' (49). Although these children did not have cancer, these quotes exemplify how family issues may affect the child's pain experience.
Little research addresses the effects of child and family factors on pain. Health care providers, however, must be cognizant that such factors may increase, decrease, mask, or nullify the pain experience. Astuteness is necessary so that these underlying factors are identified and treated by providers or consulting services whenever possible.
Deliberate and systematic assessments are comprehensive, integrating
structured and unstructured information. Structured information results
through the use of psychometrically sound tools, whereas unstructured information
occurs through more casual and less focused observations. Integration of
the structured and unstructured information forms the basis for clinical
judgments about pain.
The Pain Experience History by Hester and Barcus (41), recommended in the acute pain management clinical practice guidelines (5), structures this pain history discussion. The Pain Experience History, best used at an initial clinic visit or at admission to the hospital, has two forms, one for the child and the other for the parent. The child form is for verbal children, generally at least 4 years of age. The health care provider asks the child the questions and records the child's responses. The parent form contains parallel questions. The parent can either fill out the form or respond to the questions orally (see Table 1).
The first question on the Pain Experience History addresses whether
the child understands the word pain. This question evolved
from two studies (42,43) documenting that hospitalized
children often do not understand what pain is. The preferred word for many
children under 12 years is hurt (49).
The child's response to ``Tell me what pain is'' provides the health care
provider with the word to use when discussing pain with the child. For
example, a 12-year-old boy with Burkitt's lymphoma said pain was ``real
hard throbbing'' (52). This child seemed to understand
the word pain. In contrast, a 7-year-old boy with more than
20 previous surgeries denied knowing what the word pain
meant (42). He did, however, understand the word
Chapter 14 table 1: Pain experience history
|Child form||Parent Form|
|Tell me what pain is.||What word(s) does your child use in regard to pain?|
|Tell me about the hurt you have had before.||Describe the pain experiences your child has had before.|
|Do you tell others when you hurt? If yes, who?||Does your child tell you or others when he or she is hurting?|
|What do you do for yourself when you are hurting?||How do you know when your child is in pain?|
|What do you want others to do for you when you hurt?||How does your child usually react to pain?|
|What don't you want others to do for you when you hurt?||What do you do for your child when he or she is hurting?|
|What helps the most to take your hurt away?||What does your child do for him- or herself when he or she is hurting?|
|Is there anything special that you want me to know about you when you hurt? (If yes, have child describe.)||What works best to decrease or take away your child's pain?|
|Is there anything special that you would like me to know about your child and pain? (If yes, describe.)|
A 9-year-old girl with osteosarcoma and severe mucositis said she told others when she hurt, but with the mucositis, she had to communicate by writing (52). She wanted to hold her mother's hand when she hurt and described back rubs, conversation, quiet distraction, and a quiet room as helpful in relieving her pain. She did not want others to surprise her by ``moving her sore leg.'' This child's responses would be extremely helpful in planning care to identify, prevent, and treat pain.
Information from parents, especially important for the preverbal child, augments information from the verbal child. For example, one mother described her 9-year-old daughter with osteosarcoma as behaving in the following way when she was hurting: ``Tense knot in forehead. Facial expression changes. Lies on one side and curls up, head down.'' According to the mother, her daughter tried to relieve pain using several approaches: ``Curls up on right side. Wants Mom to hold hand. Braces self together'' (52). Her mother also used multiple strategies to assist her daughter: raised the head of her bed, talked to her daughter, held her hand, and told her stories about an aunt in the Orient. Her mother felt that the pain was best relieved by drawing, closing her eyes and focusing on other things, and thinking on her own about things such as a ``little person'' in her head. Through the pain history, this mother and her daughter provided the health care providers with important information about pain expression and preferred relief strategies. The mother commented on the usefulness of the dialogue generated through the Pain Experience History for her and her daughter. Nurses who participated in this study thought that the use of this tool helped them better understand this child's pain and empowered parents to be more involved in relieving their child's pain.
For children with repeated hospitalizations and health care provider
visits, previous hospital and clinic records may augment information obtained
through the pain history. Too often, however, little information on pain
is documented. The chart provides data on the prescription and administration
of analgesics and adjuvant drugs; however, documentation of pain episode
characteristics, pharmacological effectiveness, and the use and effectiveness
of nonpharmacological approaches are often lacking or too inadequate to
be of use.
Although a pain history can provide valuable information for caring
for the child in pain, incorporating the information into practice may
be difficult, especially if the pain history is not readily accessible
to health care providers. At Harbor-UCLA Medical Center, Hall et al.2
developed an innovative strategy to ensure that pain history information
was readily available to health care providers on the pediatric
unit. A form entitled ``Feel Better Plan'' was placed in a convenient but
visible place in a child's room, e.g., on the bulletin board at the head
of the bed or taped on the head or foot of the bed. This plan, written
in big letters, included the child's name and age, the Poker Chip Tool
instructions (discussed later), and the statement ``To help the hurt feel
better, try.'' Blank lines followed for writing in child-preferred actions.
Innovations such as this are needed to facilitate use in practice.
Although most self-report approaches were developed for children with pain secondary to diagnostic, monitoring, and surgical procedures, self-report tools with adequate psychometric evidence appear to capture appropriately the pain intensity for children with cancer. Further psychometric work, however, is necessary to defend or refute this contention. Often approaches with adequate psychometric evidence lack attributes necessary for clinical applicability. Recent work by Hester and colleagues with nurses on the pediatric oncology unit (52) and a general pediatric unit (53) suggests that tools need to be easy to understand, learn, teach, use, and carry. A simple, easily portable format with few instructions is preferred.
Self-report approaches are most suitable for children at least 4 years of age (5,70). Hester and colleagues found that some 4-year-old hospitalized children had difficulty using a simple self-report tool (i.e., the Poker Chip Tool) (51) and recommended a minimum age of 4.5 years for its use. Some researchers, however, contend that self-report approaches are appropriate for children as young as 3 years of age (15,71). Based on a critical review of data within the studies with 3-year-old children, self-report use with children under 4 years yields inconsistent findings. Use with children under 4 years is highly individual and should be done with some skepticism and the use of adjunct approaches (i.e., proxy, behavioral, and physiological) is advocated.
Most tools measure intensity of pain. One of the earliest tools developed, the Poker Chip Tool (39), measures intensity only. Extensive psychometric work has documented strong evidence for generalizability; content, convergent, discriminant, and construct validity; and sensitivity (11,16,17,39,46,47,50,51). Although research studies have focused on use with 4- through 13-year-old children, anecdotal information has suggested that older adolescents have used it successfully in the clinical setting (M. Jordan-Marsh, March 1991, personal communication ).
The Poker Chip Tool consists of four red poker chips that represent
pieces of hurt. Table 2 provides the
instructions in both English and Spanish. Children learn to use this tool
rapidly. They can either state how many pieces of hurt they have, pick
up the chips, or point to the chip in the position representing the number
of pieces of hurt, as in
Chapter 14 table 2 Poker chip tool instructions
Another tool focused on intensity only, the Oucher (11,14-17), uses a vertical format for the presentation of two simultaneously presented scales: a photographic scale and a 0 to 100 (presented by tens) numeric rating scale. To determine which scale to use, the child is asked to count from 1 to 100 by ones. If successful, he or she uses the numeric scale. Unclear, however, is the rationale for counting by ones to 100 when the numeric format is presented by tens. The photographic scale, reserved for children unable to count to 100, consists of six photographs of a Caucasian boy with facial expressions purported to depict differing pain levels. Lines between the two scales connect the photographs with specific numeric ratings. The correspondence between the facial expressions and designated numeric values have yet to be validated, however, and so may be misleading to the child and the provider interpreting the response. Although research on the Oucher separates analysis by scale (photographic and numeric), the presentation of both formats simultaneously precludes clarity on whether the child used only one format or was informed by the other in making the pain level decision. Recent work on Hispanic and African-American versions of the Oucher have addressed the problem of simultaneous presentation of the two tools by (a) eliminating the lines between the photographs and the numbers and (b) placing the photographic format on one side of the poster and the numeric format on the other (108).
Cartoon-face scales (19,66,68,111) provide another approach to eliciting the child's report of pain. Generally, these tools consist of five to six cartoon faces, beginning with a neutral face and advancing to a crying face (68,111). Unfortunately, few studies have focused on the psychometric properties of these tools and whether the faces actually depict pain. Hence, little is known about reliability, generalizability, validity, and sensitivity. Of concern is whether the cartoon faces actually depict pain, a validity issue. The use of the same tool to measure pain and anxiety confounds psychometric analyses (68). Conceptually, faces representing pain and anxiety represent different emotional states, and using the same faces to measure the same concept is confusing and poses a major threat to construct validity. P.|A. McGrath (70) has addressed this issue by using a series of nine cartoon faces that measure pain affect, i.e. the unpleasantness of pain. She uses the pain affect scale in conjunction with a visual analogue scale that addresses intensity.
Clearly, more work is needed on the validity of scales using faces. Validity could be addressed by examining the faces with the Maximally Discriminate Facial Coding System and the Facial Action Coding System to differentiate between pain and anxiety. Neuman et al. (83) used this process to examine the content validity of facial photographs included in the three versions of the Oucher but results were not provided in the abstract.
The Eland Color Tool (27,28) focuses on both
pain intensity and location. This tool involves the child's development
of a color scale. After selecting four colors to represent no hurt, little
hurt, moderate hurt, and the most hurt, the child chooses the color representing
his or her hurt and marks where the pain is on a clothed body map. For
example, a 12-year-old boy who used this tool after a lumbar puncture colored
a red line along his spinal cord and then made a zigzag line on top of
it. He commented that the zigzag line was ``because it hurt so bad'' (43).
Although the tool purportedly measures intensity, some investigators (43;
Falco, 1985, personal communication ) who have examined
the relationship between other intensity measures and color selection have
found nonsignificant correlations of small magnitudes. Hence, they have
questioned whether the tool measures intensity or another qualitative dimension
2R. D. Hall, M. Jordan-Marsh, G. McFarlane-Sosa, and R. Watson developed this innovation for the Harbor-UCLA Medical Center Humor Project for Children, Harbor-UCLA Medical Center, Los Angeles, 1991.
The Adolescent Pediatric Pain Tool (92-94) is similar to the McGill Pain Questionnaire in that it measures intensity, location, and quality. Extensive psychometric work since 1982 has documented its reliability, validity, and sensitivity (90,93,94,103-105). This tool is recommended for children from 8 to 17 years of age. A body map is on the first page of the tool. In contrast to the Eland Color Tool, which uses a clothed body map, the Adolescent Pediatric Pain Tool uses a nude body map. No genitalia are included, thus allowing one form of the tool for both boys and girls (see Figure 2 ). Even though school-age children and adolescents could point to the pain location on their own bodies, the inclusion of the body map has some advantages. Children 8 years and older are able to be very specific in locating pain. Instead of putting one mark above the knee, a 15-year-old boy with a limb salvage identified three specific locations of pain above the knee. The body map can become part of the child's record, so that location patterns can be examined over time. According to Foster (June 1992, personal communication ), the body map is particularly useful for children and adolescents who have had amputations to depict the presence of phantom limb pain.
Pain intensity is measured using a word-graphic rating scale. Selection of this horizontal scale resulted from a series of studies involving approximately 1,000 children and adolescents (103). These studies focused on comparing five different rating scales: a color scale, a visual analogue scale, a graded-graphic scale, a word-graphic rating scale, and a magnitude estimation scale. The scales yielded comparable data from the children and adolescents. This finding is not surprising since the scales (except the magnitude estimation scale) are similar in format, 100 mm in length, and the children completed all of them for each pain stimulus. Given the comparability of results, child and adolescent preference was the principal criterion for scale selection. Interestingly, although several authors recommend the visual analogue scale for children as young as 5 (70), 6 (88), 7 (73,89) and 8 (95) years, only 3% of the children preferred the visual analogue scale. Overall, the visual analogue scale was the least preferred of the five scales (103).
Children and adolescents can mark anywhere on the 100-mm word-graphic rating scale (see Figure 3 ). To determine the pain score, the scale must be measured from the no-pain end to the mark. Normally the requirement to measure a line to score a scale is a clinical limitation. Savedra and colleagues (92) tried to overcome this limitation by making a ruler at the bottom of page 1 of the tool (see Figure 2 ) that can be folded up to measure the word-graphic rating scale on page 2 of the tool (see Figure 3 ).
Included on page 2 of the tool (see Figure 3 ) is a list of words that describe pain. Children and adolescents circle the words they feel capture the quality of their pain. They also may add words not provided on the list in the blank spaces. The criteria for scoring and interpreting word selections are not yet available; hence, at most, the words provide a picture of how the child or adolescent perceives his or her pain.
An aspect of self-report tools that has received little attention is
their word anchors. For example, the word anchors for the Poker Chip Tool
are ``a little hurt'' and ``the most hurt you could ever have.'' Hester
and Foster (46,47), questioning the appropriateness
of the upper anchor, conducted two studies, one with 60 well children and
the other with 120 hospitalized children. Two groups of children, 4 through
13 years of age, rated pain depicted in a series of rabbit cartoon pictures
using the Poker Chip Tool either with ``most hurt you could ever have''
or ``a lot of hurt.'' Pain ratings did not differ significantly by upper
anchor group in either the well-child study or the hospitalized-child study.
Thus these preliminary analyses suggest that the upper anchors function
Tesler and colleagues (103) raised an issue about a middle anchor. The word moderate is often used, but after pilot testing the word-graphic rating scale, they found that medium was the word children and adolescents preferred. How use of different words affects child use of scales is unknown.
Children need a tool format that is simple and precise. A recent study took issue with the Pain Ladder (Hay, 1984, unpublished research report ). Hester and colleagues (50,51) found the Pain Ladder to be too complex for children from 4 through 13 years. The ladder has 10 rungs, 11 spaces, and an ambiguous zero point. Hospitalized children often used only the lower end of the scale, reserving the upper end for very severe pain. Essentially, children reduced the scale complexity by constructing their own scale within the Pain Ladder.
The score ranges on tools for use with children vary considerably. The Eland Color Tool has a range of 3 (0 to 3); the Poker Chip Tool, 4 (0 to 4); and the photographic Oucher, 5 (0 to 5). In contrast, the numeric Oucher and the word-graphic rating scale each have a range of 100 (0 to 100). Supposedly, tools with larger ranges have greater sensitivity from a psychometric perspective; however, Jensen (60) noted that, for adult pain assessment tools, evidence does not support the notion that sensitivity is compromised for scales with fewer response categories. He suggested that 21 points was optimal for adults and he cautioned that sensitivity of scales with ranges of 4 or less may be less in some situations. If the sensitivity is not compromised, what is the practical and clinical value of using tools with larger ranges? What is the practical and clinical significance of knowing that pain is 95 versus 97, 56 versus 65, and 29 versus 38? Would interventions differ? Interestingly, researchers often group values on tools with large ranges into fewer categories for descriptive purposes. For example, Miser and colleagues (80) aggregated data generated with a 100-mm visual analogue scale into five groups: 0 mm, no pain; 1-35 mm, mild pain; 36-50 mm, mild to moderate pain; 51-75 mm, moderate pain; and 76-100 mm, severe pain. Criteria for groupings such as these are generally not reported. Further psychometric work on scaling related to ranges and sensitivity are needed for self-report pain assessment tools used with children.
These studies (2,66,67,106) focused only on one pain measurement by each rater. Other studies have examined the correspondence across multiple measurement occasions. Moinpour et al. (82) reported strong concordance between parent and child visual analogue scale ratings for mouth pain following bone marrow transplants. LeBaron and Zeltzer (68) reported low to moderate correlations (0.39, 0.50, and 0.11) on three occasions between a child and a trained observer for pain associated with a bone marrow aspiration measured with the Faces Scale. Although all of the correlations are positive, the magnitude varies widely and unpredictably without a pattern. Hester and colleagues (51) examined the relationships among pain ratings obtained from the child, parent, and nurse. The children in this study, 4 to 13 years, were hospitalized for a variety of reasons including cancer. Randomly assigned to either the Pain Ladder Study or the Poker Chip Tool Study, children, their parents, and the nurses caring for the children measured the children's pain with the tool corresponding to the study assignment on four occasions within each nurse's shift. The correlations among the raters are presented in Table 3. As with the face scale, all correlations were positive, but an interesting pattern occurred with the Pain Ladder: the correlations among the raters decreased from moderate to low across the measurement occasions. This finding suggests that as the individuals became more familiar with the Pain Ladder their correspondence decreased. Interestingly, the highest correlations occurred between the adult raters. Perhaps adults used the Pain Ladder differently from children. In contrast, correlations among the raters using the Poker Chip Tool increased in magnitude across measurement occasions. This pattern is more consistent with expectations concerning learning: as the raters become more familiar with a tool, correspondence will increase (see Table 3).
Hester and colleagues speculated that the simplicity of the Poker Chip
Tool contributed to the high correspondence among the child and proxy raters.
Tools that demonstrate higher correspondence between adult raters than
between child and adult (e.g., the Pain Ladder and the visual analogue
scale) suggest that tools are used differently by the child and adult raters;
potentially, complexity or abstractness accounts for these findings. More
research, however, is needed to substantiate these conclusions.
Capter 14 table 3 Correlation coefficients among raters using the pain ladder and the poker chip tool
|Pain ladder||Poker chip tool|
Few behavioral/observation methods are available (see Table
4). Examination of the behavioral assessment tools for children demonstrates
the inadequacy of the science in providing psychometrically sound behavioral
tools to measure pain. Several issues pertain to the development of behavioral
measures. Foremost is whether the measures capture behaviors that represent
pain. Confounding the development of measures are behaviors that signal
pain intensity but also prevent or alleviate pain. This paradoxical nature
presents validity problems in the selection of behaviors for behavioral/observation
Another issue is the lack of attention to developmental differences in behavioral manifestations for infants, children, and adolescents. Findings from studies demonstrate that age or developmental level affects the overt manifestation of some behaviors, e.g., vocalizations and large motor movements. Foster (30a), Craig et al. (25), and LeBaron and Zeltzer (68) concluded that younger children undergoing invasive procedures exhibited more crying behavior than did older children. A rationale for these findings is that physiological competency for expression of specific behaviors (e.g., crying or movement) is tempered by cognitive interpretations of sociocultural appropriateness. For example, whereas children from birth through adolescence are physiologically capable of crying, screaming, and flailing extremities, these behaviors are not socially appropriate for school-age children and adolescents. Thus, an older child or adolescent who feels like crying may suppress the behavior because ``crying is for babies'' and because of what his or her peers might think of such behavior. This example suggests that different tools are needed to account for developmental differences. Jay et al. (58) emphasized that current tools may be not useful for assessing pain in older adolescents, who are less likely to exhibit obvious behavioral responses. The Procedure Behavior Rating Scale (63), the Procedure Behavior Checklist (68), and the Observational Scale of Behavioral Distress (56,57,59) all demonstrate this issue. Research on these tools show age differences: older children and adolescents exhibit fewer behaviors than younger children. This finding raises a validity issue regarding the use of the same behavioral tool for children across developmental levels.
Mapping developmentally appropriate behaviors indicative of pain is a major undertaking. The task is further confounded when the type of pain is considered. Behavioral expression differs with the type of pain and noxious stimuli. For example, a child's response to an invasive procedure differs from his or her response during the immediate postanesthetic period, the postoperative period, episodic disease pain, and chronic pain. In the development of behavioral/observation methods, researchers and clinicians often have assumed the following behaviors to be indicative of pain: vocal (e.g., crying and whining), verbal (i.e., intelligible statements regarding pain), facial expression (e.g., grimacing), and large motor movements (e.g., flailing of arms and legs) (49). Research on these behaviors has focused primarily on pain associated with diagnostic and monitoring procedures (e.g., injections, venipunctures, lumbar punctures, and bone marrow aspirations) but findings are not directly applicable to children with acute and chronic pain.
Hospitalized for medical or surgical reasons, older children and adolescents often do not cry when in pain (51), but young children who undergo invasive diagnostic procedures often do cry. Inhibition of motor responses is characteristic of children, especially postsurgically (18). Taylor's comment regarding children postsurgically is particularly poignant: as children ``became more awake and alert, they seemed to realize that movements increased their pain'' (102).
Issues such as these provide a rationale for the paucity of available reliable and valid behavioral/observation tools to measure pain. Craig (23) emphasized the principal issue in developing behavioral tools: ``The claim that pain behaviors can be easily recorded and quantified through direct observation underestimates the difficulties involved.'' He also underscored that the ``reason for heavy emphasis on self-report is its methodological convenience.''
Most of the behavioral tools (91%) listed in Table 4 focus on behavioral responses to invasive procedures and to surgery. Only one tool focuses on disease-related pain, i.e., cancer. Further, most of the tools (73%) focus on the verbal population. These findings extend Craig's commentary on methodological convenience to behavioral tools. Emphasis in the research community is on the development of behavioral tools for the easiest pain to measure, i.e., visible pain (e.g., invasive procedures and surgery), and for the easiest population of children to access, those who are verbal. Of concern is the limited emphasis on accessing the more difficult pain (e.g., pain from disease and pain with unknown etiology) and the more difficult populations (e.g., preverbal and nonverbal children). The travesty is the emphasis on developing behavioral tools for children who could use self-report without attending to the populations for whom the primary access to pain is through behaviors.
Chapter 14 table 5Guidelines for critical review of pain measurement tools
|Is pain or a related concept being measured by this tool?|
|Is the type of pain (e.g., acute, procedural, chronic malignant) being measured similar to the type of pain I plan to measure?|
|Was the tool designed to use with patients like mine?|
|Was the tool designed to use in a setting like mine?|
|Is the format of the tool appropriate for my patients?|
|Is the length of the tool appropriate for my patients?|
|Is the readability of the tool appropriate for my patients?|
|Is the tool developmentally appropriate for my patients?|
|Is the amount of practice time required by the patients appropriate for my patients?|
|Would my patients like this tool?|
|Is the tool affordable?|
|Is the tool readily available?|
|Is the tool easily reproducible?|
|Is the tool easily transported?|
|Is the tool easily disinfected?|
|Is the tool easy to score?|
|Is the scoring easily interpretable?|
|Is the tool appropriate for patients from different cultures?|
|Is the estimated reliability adequate?|
|Is the estimated validity adequate?|
|Are the estimated reliability and validity appropriate for the tool and its intended uses?|
|Is the tool sensitive enough to determine differences in pain levels?|
|Has additional research been done using this tool?|
|Does the additional research support the reliability, validity, and sensitivity of the tool?|
|Are other clinical settings using this tool?|
|After addressing these criteria, decide whether the tool
is adequate for use in your study.
Adequate: Proceed with clinical application
Questionable: Use clinically with caution; pilot-test the tool before adopting it; contact the author and those who have used it clinically or in research to discuss your concerns
Inadequate: Do not use
Approaches other than self-report are necessary for preverbal and nonverbal children. Needed are reliable and valid approaches to assessing pain through observations, behaviors, and physiological parameters. These approaches must be examined for efficacy with the preverbal, verbal, and nonverbal populations of children who have cancer. The use of multiple approaches, including proxy measures from parents and looking for changes in behavior, especially facial (24), temperament, and physiological parameters, should be valuable in assessing pain in these special populations. Another approach is to validate a suspicion of pain by giving the child a trial dose of analgesics (5) and observing for subsequent changes.
Field T, Goldson E. Pacifying effects of nonnutritive sucking on term and preterm neonates during heelstick procedures. Pediatrics 1984;74:1012-1015.
Molsberry DMJ. Young children's subjective quantifications of pain following surgery. Unpublished Master's Thesis. University of Iowa, 1979.
National Center for Nursing Research. National nursing research agenda: vol IV: symptom management. NIH publication no. 93-2420. Bethesda, MD: NIH, U.S. Public Health Service, U.S. Department of Health and Human Services (in press).
Tyler DC, Krane EJ, eds. Pediatric Pain. New York: Raven Press, 1990.