Table 1. Summary of Children's Scores on Dependent and Independent Variables
Correlates of Pain-Related Responses to Venipunctures in School-Age Children
Effects of Distraction on Children's Pain and Distress During Medical Procedures: A Meta-Analysis
Correlates of Pain-Related Responses to Venipunctures in School-Age Children
Outline
Graphics
Guided by the Roy Adaptation Model of Nursing, the relationship of children's age, gender, exposure to past painful experiences, temperament, fears, and child-rearing practices to their pain responses to a venipuncture was examined. A sample of 94 children aged 8 to 12 years and their female caregivers were recruited from three outpatient clinics. During the venipuncture, children's behavioral and heart rate responses were monitored; immediately after, their subjective responses were recorded. Canonical correlation revealed two variates. In the first, age and threshold(temperamental dimension) correlated with pain quality, behavioral responses, and heart rate responses, explaining 12% of the variance. In the second, age, the temperamental dimensions of distractibility and threshold, and medical fears explained only 5.7% of the variance in pain quality and heart rate magnitude. Significant correlations between pain intensity, quality, behavioral responses, and heart rate responses support the multidimensionality of pain.
Factors that account for variability in children's responses to venipunctures have not been fully identified. The purpose of this study, therefore, was to examine the relationship of a set of correlates, including age, gender, past painful experiences, temperament, general and medical fears, and child-rearing practices, on school-age children's subjective, behavioral, and heart rate responses to a venipuncture.
Background
Conceptual Framework: The Roy Adaptation Model of Nursing (RAM)(Roy and Andrews, 1991) and the research literature on children's responses to painful procedures guided this study. According to the RAM, the goal of nursing is to facilitate adaptation between the person and the environment through the management of stimuli (Roy& Corliss 1993). One of the assumptions of adaptation-level theory is that the person's 'adaptive behavior is a function of the stimulus and adaptation level, that is, the pooled effects of the focal, contextual, and residual stimuli' (Roy & Corliss, 1993, p. 217). The focal stimulus immediately confronts the person; the contextual stimuli contribute to the effect of the focal stimulus; and residual stimuli are factors whose effects on the person's adaptation have not been clearly determined.
The focal and contextual stimuli are processed through coping mechanisms, such as the regulator and cognator subsystems. The regulator subsystem induces physiological responses through neural, chemical, and endocrine processes. The cognator subsystem elicits responses through perceptual/information processing, learning, judgment, and emotion processes (Andrews & Roy, 1991a). Outcomes of these coping mechanisms are the person's responses in four modes: self-concept, role function, interdependence, and physiological. The self-concept mode relates to feelings about one's personal and physical self. The role-function mode is associated with the need for social integrity based on roles assumed within society. The interdependence mode focuses on interactions that fulfill the need for affectional adequacy and support. The physiological mode is described as the person's physical responses to stimuli from the environment. In accordance with the proposition that stimuli serve as inputs to the person to elicit a response, children's pain-related responses were tested as a function of the pooled effects of the focal stimulus, the venipuncture, and the contributory effects of the contextual stimuli congruent with the RAM and the empiric pediatric pain literature. Contextual stimuli relate to culture, family, developmental stage, and cognator effectiveness, as well as those factors that have an effect on the person's adaptive responses in any of the modes (Andrews & Roy, 1991b). The pooled effects of a venipuncture, children's age, gender, exposure to past painful experiences, temperament, fears, and child-rearing practices were studied in relation to children's pain-related responses to the venipuncture.
Roy (1991) defined pain within the physiological mode, as a sensory experience of acute or chronic nature, coded into the somatosensory pathways. Acute pain refers to 'discomfort which is intense but relatively short lived and reversible' (p. 166). Using principles from neurophysiology, Roy stated that a sensory experience such as pain involves the transmission of neural activity through specialized receptors, and transmission of information from sensory pathways to the cerebral cortex. A sensation results from receptors' activity and is converted into perceptual activity involving mental representations and interpretations (Roy). Thus, pain can be understood to be both a sensory and perceptual experience.
As a manifestation of the regulator subsystem activity, the sensory dimension was represented in this study by children's behavioral and heart rate responses. The perceptual dimension of pain, which is the response to the cognator subsystem activity, was portrayed by children's subjective responses about the location, intensity, and quality of pain.
Review of the Literature: Empirical evidence supports a negative relationship between the contextual stimulus of children's chronological age and pain intensity (Fradet et al. 1990; Lander & Fowler-Kerry, 1991; Manne et al., 1992) and behavioral responses to venipunctures (Fradet et al., 1990; Humphrey, Boon, van Linden van den Heuvell, van de Wiel, 1992; Jacobsen et al., 1990). Mean pain intensity responses to a venipuncture or intravenous cannulation were found to be greater in preschoolers than in schoolage children; however, it is not clear whether this difference was statistically significant (Van Cleve et al., 1996). Gender has been believed to be a mediator in pain experiences (Katz, Kellerman, & Siegel, 1980). However, researchers have shown that gender has no effect on pain intensity responses (Fowler-Kerry & Lander, 1991; Fradet et al., 1990; Manne et al., 1992) and behavioral responses to venipunctures (Fradet et al.; Humphrey et al.; Jacobsen et al.). In view of the influence of socialization on schoolage children's sex-role stereotypes, it was relevant to reexamine the effects of gender on pain responses.
Several researchers have reported that children's exposure to past painful procedures is inversely related to their behavioral responses to a venipuncture (Fradet et al., 1990; Jacobsen et al., 1990). However, in another study, this relationship did not reach statistical significance (Manne et al., 1992). The effects of past exposure to venipunctures on children's subjective, behavioral, and heart rate responses to a venipuncture remain unclear.
The contextual stimulus of temperament may explain the individual variability of responses across situations. According to Thomas and Chess (1977), temperament is the result of an interactive process between child and parent and consists of nine temperamental dimensions: activity, rhythmicity, approach/withdrawal, sensory threshold, intensity of reaction, quality of mood, distractibility, and attention span/persistence, and adaptability. Although the predictive effects of temperament have been studied in children's adaptation to chronic illness (Garrison, Biggs, & Williams, 1990; Wallander, Hubert,& Varni, 1988), the relationship between temperament and pain related to invasive procedures has only been examined in preschool children.Young and Fu (1988) found that a child's rhythmicity had a small effect on pain intensity and that approach accounted for 7% of children's behavioral responses to venipunctures. In view of the brevity of the venipuncture, the theoretical relevancy of each temperamental dimension, and school-age children's increased behavioral mastery, only sensory threshold, intensity, and distractibility were judged important in this study. Sensory threshold reflects sensitivity to stimulation and may be important in self-regulation and defense mechanisms (Rothbart & Derryberry, 1981). Intensity is the energy level of a response irrespective of the direction (Thomas & Chess, 1977). Distractibility, a measure of how sensitive one's attention is to environmental stimulation, may be relevant in children's coping mechanisms and adaptive responses.
Children fearful of medical procedures report higher pain intensity to venipunctures (Broome, Bates, Lillis, Wilson, & McGahee, 1990) and display more behavioral responses (Jacobsen et al., 1990). Nevertheless, investigators have not examined the contextual stimulus of children's medical fears in relation to a multidimensional view of the pain experience or accounted for another contextual stimulus, children's general fears. While fear is an immediate response to a threatening situation, general fears may serve as a context for the development of medical fears.
The family's influence on children's behaviors through the provision of structure and discipline is relevant in the study of children's responses to painful situations (Melamed & Bush, 1985). Parental restrictiveness and nurturance toward a child's behavioral expressiveness may help in understanding children's responses to pain. The contextual stimulus of child-rearing practices was examined in the situation of immunizations (Broome & Endsley, 1989). Children of authoritative(high-control, high-warmth) parents exhibited significantly fewer behavioral responses than those of authoritarian (high-control, low-warmth), permissive(low-control, high-warmth), and unresponsive (low-control, low-warmth) parents. This study extended examination of the influence of child-rearing practices to children's pain responses to venipunctures.
Previous studies of the effects of multiple variables on children's pain intensity and behavioral responses to venipunctures have been informative. However, children's pain has not been measured in a comprehensive fashion. The relationship between children's behavioral display and reported pain intensity has been found to be moderate in school-age children during venipunctures (Fradet et al., 1990; Humphrey et al., 1992; Manne et al., 1992) and during bone marrow aspirations (Jay & Elliott, 1984; Jay, Ozolins, Elliot, & Caldwell, 1983). Inasmuch as pain intensity represents one aspect of the pain experience, children's behavioral responses needed to be examined in relation to a global assessment of pain. Though not specific to pain, physiological responses have been described in acute pain experiences. However, studies involving cardiac rate during painful procedures have yielded equivocal results. Broome and Endsley (1987) found no relationship between preschoolers' behavioral responses and heart rate during a finger stick procedure, whereas Jay and Elliott (1984) reported a moderate correlation between behaviors and heart rate of school-age children and adolescents during bone marrow aspirations. These findings support the need to reexamine the role of sympathetic responses of heart rate combined with other pain measures during an invasive procedure. Based on the RAM and the empirical pediatric pain literature, the hypothesis tested was that there is a relationship between the set of independent variables age, gender, past painful experiences, temperament, medical fears, general fears, and child-rearing practices and the set of school-age children's responses to venipuncture: pain location, pain intensity, pain quality, observed behaviors, and heart rate.
Method
Sample: The sample consisted of 94 children and their female caregivers recruited from three outpatient clinics (gastroenterology, nephrology, and preoperative) at a large pediatric hospital in a mid-Atlantic state. An initial power analysis for multiple regression with 9 independent variables, a power of .80, a medium effect size, and an alpha level of .05 revealed that 119 subjects were necessary (Cohen, 1988). However, the actual obtained power of the main analysis performed to test the study hypothesis was so low that an increase in sample size would not have yielded more meaningful results. Consequently, the study sample was judged sufficient.
Children between 8 and 12 years of age, cognitively normal for their school grade, accompanied by a female caregiver, and expected to receive a venipuncture during their clinic visit were asked to participate. Of the 121 subjects who were approached, 10 refused (8%) for lack of time or personal reasons. Of those who agreed, 17 (15%) were excluded for various reasons; 4 children had cognitive deficits, 2 were not accompanied by a female caregiver, 7 did not require a venipuncture, and 4 caregivers did not have time to complete the questionnaires. Children's mean age was 10.3 (SD = 1.4). The majority of the children were female (54.3%) and white (86.2%).
Instruments: The Child Information Sheet (CHILDIS) was used to record information about the child's gender, age, school grade, number of past hospitalizations, and number of past venipunctures and other painful procedures. The Caregiver Information Sheet (CIS) was used to record demographic information about the caregiver, including age, gender, ethnicity, marital status, education level, employment status, and family income. Caregivers' perceptions about the child's experiences with past hospitalizations and painful procedures were also requested.
The Middle Childhood Temperament Questionnaire (MCTQ)(Hegvik, McDevitt, & Carey, 1982) was used to measure temperament of children 8 to 12 years old. The MCTQ is a 99-item parent report using a 6-point scale from 1 (almost always) to 6 (almost never) for each of nine temperamental dimensions. Three dimensions, distractibility, intensity, and threshold, were included in this study. Higher scores are indicative of higher distractibility and intensity but lower sensory threshold. Satisfactory criterion-related validity was evidenced in comparisons of children's temperament at ages 7 and 12 (Maziade, Câté, Boudreault, Thivierge, & Boutin, 1986). For this study, the Cronbach's alpha internal consistency reliability coefficients were threshold, .68, distractibility, .75, and intensity, .81.
The Child Medical Fears Scale (CMFS) (Broome, Hellier, Wilson, Dale, & Glanville, 1988) is used to measure children's levels of reported fears related to medical personnel and diagnostic or therapeutic procedures. Children rate their level of fear for 17 items on a scale of 1 (not at all), 2 (a little), and 3 = (a lot afraid), with higher scores indicating greater fear. Total scores range from 17 to 51. The content validity index for the CMFS is 78% (Broome et al.). Criterion validity was established with the original Fear Survey Schedule (Scherer & Nakamura, 1968), with a correlation of.71 (Broome et al.). The Cronbach's alpha internal consistency reliability coefficient was .87 for this study sample.
The Revised Fear Survey Schedule for Children (R-FSSC)(Ollendick, 1983) is an 80-item questionnaire used to measure children's general fears. Children rate their level of fear to the unknown, supernatural events, bodily injury, small animals, and death on a 3-point scale of 1 (none), 2 (some), and 3 (a lot). Total scores range from 80 to 240, with higher scores indicating greater fear. Construct validity was established by discriminating fears of phobic and normal children (Ollendick) and by supporting a decline in children's fears with age (King, Gullone, & Ollendick, 1991). The generalizability of a five-factor structure has been shown across cultures (Ollendick & Yule, 1990). The Cronbach's alpha internal consistency coefficient for the present study was .95.
The Modified Child-Rearing Practices Report (M-CRPR) (Dekovic, Janssens, & Gerris, 1991; Rickel & Biasatti, 1982) was used to measure two parental attitudes toward child rearing: parental restrictiveness, characterized by a high degree of control and endorsement of strict rules and restrictions, and parental nurturance, characterized by the willingness of parents to share feelings with their children and to show responsiveness to the child's needs. The M-CRPR consists of 40 statements with a 6-point response format, from 1 (not at all descriptive of me) to 6 (highly descriptive of me). Two scores are obtained, with lower scores indicative of low restrictiveness (range 0 to 22) and low nurturance (range 0 to 18). The validity of the M-CRPR was supported by discriminating the childrearing practices of parents of rejected and highly sociable children (Dekovic et al.) and by factor-structuring the scale (Dekovic et al. ; Rickel & Biasatti). In this study, the Cronbach's alphas for the restrictiveness and nurturance subscales were.90 and .92, respectively.
The Adolescent Pediatric Pain Tool (APPT) (Savedra, Tesler, Holzemer, & Ward, 1992) is a self-report measure of location, intensity, and quality of pain in children aged 8 to 17 years. Pain location is measured using a body outline figure on which children are instructed to mark the location(s) of their current pain. The number of locations is summed, with scores ranging from 0 to 43 (Savedra, Tesler, Holzemer, Wilkie, & Ward, 1989). Criterion-related validity of the pain location scale was documented when children's markings and investigators' observations reached an agreement of at least 80% (Savedra, Tesler, Holzemer, Wilkie, & Ward, 1990). A postoperative decrease in pediatric surgical patients' pain sites was found, thus supporting construct validity (Savedra et al., 1990). Intrarater reliability estimates of the agreement between subjects' markings and pointings ranged from 83% to 94%(Savedra et al., 1989).
Pain intensity was measured using a 100-mm wordgraphic rating scale, with scores ranging from 0 to 100. A decline in postoperative pain intensity scores supported construct validity (Savedra et al., 1990). Criterion-related validity was established through correlations with four other pain intensity scales (Tesler et al., 1991). Test-retest reliability was .91 (Tesler et al., 1991).
Pain quality was measured using a list of 67 words that relate to the sensory, affective, evaluative, and temporal experiences of pain (Tesler, Savedra, Ward, Holzemer, & Wilkie, 1988). Based on a factor analysis that confirmed only three factors (Wilkie et al., 1990), scores, which range from 0 to 56, are reported only for sensory, affective, and evaluative subscales. Criterion-related validity was evidenced by correlations with pain intensity scores (Savedra et al., 1990). A significant decrease in the number of words used by recovering pediatric surgical patients (Savedra et al., 1990; Savedra, Holzemer, Tesler, & Wilkie, 1993) supported construct validity. Test-retest reliability of total, sensory, affective, and evaluative scores of surgical patients revealed high correlations, .95, .91, .97, and .78, respectively (Tesler et al., 1991).
The Observed Child Distress (OCDS) (Jacobsen et al., 1990; Manne et al., 1992) was used to measure six behavioral responses to venipunctures: pain verbalizations, cry/scream, request for termination of procedure, refusing to assume body position, muscular rigidity, and requiring physical restraint. These behaviors are observed during three phases of the venipuncture (phase 1, from sitting in the chair until the tourniquet is applied; phase 2, from tourniquet application until needle is to be inserted; and phase 3, from piercing the skin to bandage application). They are rated for their presence (1) or absence (0), for a total score ranging from 0 to 18. Construct validity of the OCDS was supported by a positive correlation between behavioral scores and self-reports of pain intensity (Manne et al., 1992). Cronbach's alpha internal consistency coefficient for the study sample was.83.
A Nellcor electronic pulse oximeter (N-10; Haywood, CA) was used to measure heart rate during the venipuncture through a taped sensor to a finger. Heart rate was measured at rest and monitored every 10 seconds throughout the venipuncture. The magnitude of heart rate change (highest heart rate during phase 3 relative to baseline heart rate) was calculated for each child.
Procedure: All subjects were recruited before their clinic appointments. Participation in the study was voluntary, and informed consents from female caregivers and children's assents were obtained in accordance with the institution's Committee for Protection of Human Subjects.
During waiting periods, caregivers completed the CIS, the MCTQ, and the M-CRPR. In the presence of their caregivers, children were asked to answer verbally to the CHILDIS, R-FSSC, and the CMFS. A baseline heart rate was obtained after each child had rested for 15 minutes in a sitting position. Throughout all phases of the procedure, children's behavioral responses were measured using the OCDS, and heart rate was monitored with a pulse oximeter. Data obtained during the third phase of the venipuncture, which is associated with the experience of pain, is reported in this study. Only one caregiver was absent during the venipuncture. Within 10 minutes following the procedure, children completed the APPT.
Results
Examination of the distribution of the pain-related variables (Table 1) led to the exclusion of pain location scores from further statistical analyses due to limited variability. Since the distributions of pain intensity and quality scores were skewed, square root transformations were performed. Though the majority of children reported minimal pain intensity, 20% of the children regarded venipunctures as very painful procedures. Most children (98%) described their pain experiences using sensory descriptors, 75% chose evaluative words, and 40% selected affective qualities. Children's behavioral responses associated with the insertion of the needle were minimal; however, heart rate changes were of greater magnitude. For most children (96.7%), magnitude in heart rate change was within two standard deviations. For 3.3% of the sample, important changes in heart rate were recorded (< or > 3 SD).
Table 1. Summary of Children's Scores on Dependent and Independent Variables
Correlations were computed between all independent and dependent variables. Threshold was correlated with pain quality (r(94) =.25,p <.05), that is, low sensory threshold was associated with more pain descriptors. Age, r(93) = -.48, p <.001, and threshold, r(93) =.24, p < .05, were correlated with behavioral responses, suggesting that with age, children manifest fewer behavioral responses, and that low sensory threshold is associated with more behavioral responses. Distractibility, r(90) =.33, p<.05, threshold, r(90) =.23, p = .03, general fears,r(87) =.27, p <.05, and medical fears,r(86) =.26, p < .02, were correlated with magnitude in heart rate change. Children with high distractibility, low threshold, and high general and medical fears had greater changes in heart rate. None of the independent variables were related to pain intensity. Consequently, pain intensity was excluded from the main analysis.
Examination of the correlations between the independent variables revealed multicollinearity for general and medical fears, r(87) =.83,p <.001. Consequently, general fears were excluded from the main analysis.
The correlation matrix for the dependent variables revealed low to moderate correlations between pain quality, intensity, behavioral responses, and magnitude in heart rate change. Specifically, pain quality correlated with pain intensity ( r(94)=0.59, p < 0.001), behavioral responses (r(93)=0.41, p < 0.001), and magnitude in heart rate change (r(90)=32, p < 0.05). Lower correlations were found between pain intensity and behavioral responses(r (93)=0.26, p < 0.05) and magnitude in heart rate change (r(90)=0.22, p < 0.05). As expected, behavioral responses were correlated with magnitude in heart rate change(r (90)=0.31, p < 0.01). Together, these findings suggest that as children select more pain descriptors, they report higher pain intensity and exhibit more behavioral responses and greater heart rate responses.
Of the initial variables, only the independent and dependent variables that correlated significantly were retained for the canonical analysis. They were age, distractibility, threshold, medical fears, pain quality, behavioral responses, and magnitude of heart rate change. As can be seen in Table 2, two canonical variates were found to be significant. The first canonical variate (.526) was found for age and threshold, and correlated with pain quality, behavioral responses, and magnitude of heart rate change, explaining 12% of the variance. The second canonical variate (.411) revealed that age, medical fears, distractibility, and threshold correlated with pain quality and magnitude of heart rate change, explaining 5.7% of the variance. Overall, 17.7% of the variance was accounted for. Inasmuch as all variables did not enter the analysis, the study hypothesis was not supported.
Table 2. Canonical Correlation Analysis Summary Table Between Age, Distractibility, Threshold, and Medical Fears (Set 1) and Pain Quality, Behavioral Responses, and Heart Rate Magnitude (Set 2)
Additional findings from t-test and chi-square analyses showed no differences on most independent and dependent variables between subjects from the preoperative clinic and those from the gastroenterology and nephrology clinics. However, children from the preoperative group had less experience with past venipunctures, [chi]2 (1, N = 84) = 4.24, p < .05, and reported higher pain quality, M = 2.7, SD = 1.1, T(93) = -2.1, p < .05. In the total sample, girls had higher general fears than boys, T(90)= -2.0, p < .001, but no difference was found with regard to medical fears. Finally, girls had higher temperamental intensity than boys(girls, M = 3.9, SD = 0.8 vs. boys, M = 3.5,SD = 0.9; T(94) = -2.05, p < .05).
On all dependent variables, children were found to be homogeneous except that girls cried significantly more than boys during the venipuncture,[chi]2 (1, N = 93) = 4.22, p < .05. No differences were noted in children experienced and inexperienced with venipunctures with regard to temperament, general and medical fears, pain intensity and pain quality scores, behavioral responses, and heart rate magnitude. Regardless of children's health problems, family income, and race, there were no differences in children's responses to pain.
Discussion
Although the study hypothesis was not supported, the results from the canonical correlation revealed several important relationships. Data from the first variate showed that with increasing age and in children with high sensory threshold, fewer words are used to describe pain, fewer behavioral responses are manifested, and lower magnitude of change in heart rate is observed. With increasing age, children are more emotionally and behaviorally organized (Maccoby, 1983). Moreover, school-age children's greater understanding of the procedure, increasing awareness of socially acceptable behaviors, and competency in controlling behaviors may account for the restricted body movements. Finally, children less sensitive to sensory stimuli were less upset by the venipuncture and showed fewer behavioral responses and changes in heart rate.
Findings from the second variate suggest that younger, highly fearful, distractible, and sensitive children report higher pain quality and have higher heart rate reactivity. Lack of familiarity combined with a limited repertoire of coping skills may account for the younger child's increased vulnerability to stressful events. These findings support the need for the implementation of interventions for young children before and even during such relatively brief and simple medical procedures as venipunctures.
Correlations between the dependent variables as shown in the correlational matrix and in the first canonical variate support a relationship between the perceptual and sensory dimensions of pain. However, the relationship is low in magnitude. Gross motor responses may be less relevant in school-age children, suggesting that a focus on muscular rigidity and/or facial activity might be more appropriate for this group. In this study, no relationship was found between the independent variables and pain intensity. It may be that the venipuncture did not evoke enough variability in children's pain intensity or that no relationship can be established with pain intensity since it is essentially a subjective and unpredictable characteristic of pain.
The study results provided limited support for the Roy Adaptation Model. Based on the proposition that focal and contextual stimuli influence responses, empirical support was found for the contextual stimuli of age, medical fears, and the temperamental dimensions of distractibility and threshold. Only the contextual stimuli that affect developmental stage (age), self-concept (medical fears), and interdependence between parent and child(temperamental dimensions of distractibility and sensory threshold) were supported. The lack of a relationship between gender and subjective pain responses, though unexpected, is consistent with prior work on gender and pain intensity (Fowler-Kerry & Lander, 1991; Fradet et al., 1990; Manne et al., 1992). Limited support was found for the relationship between gender and behavioral responses, in that girls cried more than boys. The influence of parental child-rearing styles on responses to venipunctures was not noted in this study.
Contrary to prior research (Fradet et al., 1990; Jacobsen et al., 1990) but consistent with the work of Manne et al. (1992), the findings showed no relationship between experience with venipunctures and children's pain-related responses. This suggests that experienced children did not habituate to the procedure. It may also be that frequency of exposure is not sufficient information for understanding children's responses to pain. Rather, as the RAM suggests, children's coping abilities with procedures need to be taken into account.
Most importantly, findings from the correlational and canonical analyses support the multidimensionality of pain as conceptualized by the RAM. This empirical evidence is consistent with the need for clinicians and researchers to use a comprehensive approach to assess pain by integrating valid and reliable subjective, behavioral, and physiological measures. Such a global approach to understanding pain is in accordance with the RAM.
Several instruments used in the study need further evaluation. The Revised Fear Survey Schedule for Children should be revised to be more sociohistorically appropriate for children. Low scores obtained by the Child Medical Fears Scale suggest a need to reexamine the relevancy of several items with a school-age population. In view of the fact that there was no relationship between behavioral responses and pain intensity, more attention needs to be paid to the meaning of behavioral responses. For example, it is important to understand which behavioral responses are reflective of the pain experience in different age groups. This information is particularly critical in the care of children unable to express their needs verbally.
This study supports the need to assess children's pain and to identify the factors that may aggravate the pain experience. Research on helping children cope with aversive medical procedures has produced somewhat equivocal results (Dahlquist, 1992). While certain strategies may be helpful to some children, others may have no or negative effects. In order to individualize the care of children undergoing procedures, future research may be directed toward matching interventions with children's age, fears, and temperament. The findings of this research have contributed to the extension of the knowledge base on school-age children's pain to venipunctures.
References
Andrews, H. A., & Roy, C. (1991a). Essentials of the Roy adaptation model. In C. Roy & H. A. Andrews (Eds.), The Roy Adaptation Model: The definitive statement (pp. 325). Norwalk, CT: Appleton & Lange. [Content Link]
Andrews, H. A., & Roy,. C. (1991b). The nursing process according to the Roy adaptation model. In C. Roy & H. A. Andrews (Eds.),The Roy Adaptation Model: The definitive statement (pp. 27-54). Norwalk, CT: Appleton & Lange. [Content Link]
Broome, M. E., Bates, T. A., Lillis, P. P., & Wilson McGahee, T. (1990). Children's medical fears, coping behaviors, and pain perceptions during a lumbar puncture. Oncology Nursing Forum,17, 361-367. [Content Link]
Broome, M. E., & Endsley, R. C. (1989). Maternal presence, childrearing practices, and children's response to an injection.Research in Nursing and Health, 12, 229-235. [Content Link]
Broome, M. E., Hellier, A., Wilson, T., Dale, S., & Glanville, C. (1988). Measuring children's fears of medical experiences. In C. F. Waltz & O. L. Strickland (Eds.), Measurement of Nursing Outcomes (Vol. 1, pp. 201-214). New York: Springer. [Content Link]
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum. [Content Link]
Dahlquist, L. M. (1992). Coping with aversive medical treatments. In A. M. La Greca, L. J. Siegel, J. L. Wallander, & C. E. Walker (Eds.), In stress and coping with child health (pp. 345-376). New York: Guilford Press. [Content Link]
Dekovic, M., Janssens, J. M. A. M., & Gerris, J. R. M. (1991). Factor structure validity of the Block child rearing practices report (CRPR). Psychological Assessment: A Journal of Consulting and Clinical Psychology, 3, 182-187. [Content Link]
Fowler-Kerry, S. (1990). Adolescent oncology survivors' recollection of pain. In D. C. Tyler & E. J. Krane (Eds.),Advances in pain research therapy (Vol. 15, pp. 365-371). New York: Raven Press. [Content Link]
Fowler-Kerry, S., & Lander, J. (1991). Assessment of sex differences in children's and adolescents self-reported from venipunctures. Journal of Pediatric Psychology, 16, 783-793. [Content Link]
Fradet, C., McGrath, P. J., Kay, J., Adams, S., & Luke, B. (1990). A prospective survey of reactions to blood tests by children and adolescents. Pain, 40, 53-60. [Content Link]
Garrison, W. T., Biggs, D., & Williams, K. (1990). Temperament characteristics and clinical outcomes in young children with diabetes mellitus. Child Psychology and Psychiatry, 31, 1079-1088. [Content Link]
Harrison, A. (1991). Preparing children for venous sampling. Pain, 45, 299-306. [Content Link]
Hegvik, R. L., McDevitt, S. C., & Carey, W. B. (1982). The middle childhood temperament questionnaire. Developmental and Behavioral Pediatrics, 3, 197-200. [Content Link]
Humphrey, G. B., Boon, C. M. J., van Linden van den Heuvell, G. F. E. C., & van de Wiel, H. B. M. (1992). The occurrence of high levels of acute behavioral distress in children and adolescents undergoing routine venipunctures. Pediatrics, 90, 87-91. [Content Link]
Jacobsen, P. B., Manne, S. L., Gorfinkle, K., Schorr, O., Rapkin, B., & Redd, W. H. (1990). Analysis of child and parent behavior during painful medical procedures. Health Psychology,9, 559-576. [Content Link]
Jay, S.M., & Elliot, C. (1984). Behavioral observation scale for measuring children's distress: The effects of increased methodological rigor. Journal of Consulting and Clinical Psychology, 52, 1106-1107. [Content Link]
Jay, S.M., Ozolins, M., Elliot, C.H., & Caldwell, S. (1983). Assessment of children's distress during painful medical procedures. Health Psychology, 2, 133-147. [Content Link]
Katz, E. R., Kellerman, J., & Siegel, S. (1980). Behavioral distress in children with cancer undergoing medical procedures: Developmental considerations. Journal of Consulting and Clinical Psychology, 48, 356-365. [Content Link]
King, N. J., Gullone, E., & Ollendick, T. H. (1991). Manifest anxiety and fearfulness in children and adolescents.The Journal of Genetic Psychology, 153, 63-73. [Content Link]
Lander, J., & Fowler-Kerry, (1991). Age differences in children's pain. Perceptual and Motor Skills, 73, 415-418. [Content Link]
Maccoby, E. E. (1983). Social-emotional development and response to stressor. In N. Garmezy & M. Rutter (Eds.), Stress, coping, and development in children (pp. 217-234). New York: McGraw-Hill. [Content Link]
Manne, S. L., Jacobsen, P. B. & Redd, W. H. (1992). Assessment of acute pediatric pain: Do child self-report, parent ratings and nurse ratings measure the same phenomenon? Pain, 48, 45-52. [Content Link]
Maziade, M., Coté, R., Boudreault, M., Thivierge, J.,& Boutin, P. (1986). Family correlates of temperament continuity and change across middle childhood. American Journal of Orthopsychiatry,56, 195-203. [Content Link]
Melamed, B. G., & Bush, J. P. (1985). Family factors in children with acute illness. In D. Turk & R. Kerns (Eds.),Health, illness and families: A life-span perspective(pp. 183-219). New York: Wiley. [Content Link]
Ollendick, T. H. (1983). Reliability and validity of the revised fear survey schedule for children (FSSC-R). Behavior Research Therapy, 21, 685-692. [Content Link]
Ollendick, T. H., & Yule, W. (1990). Depression in British and American children and its relation to anxiety and fear.Journal of Consulting and Clinical Psychology, 58, 126-129. [Content Link]
Rickel, A. U., & Biasatti, L. L. (1982). Modification of the Block child rearing practice report. Journal of Clinical Psychology,38, 129-134. [Content Link]
Rothbart, M. K., & Derryberry, D. (1981). Development of individual differences in temperament. In M. E. Lamb & A. L. Brown (Eds.),Advances in developmental psychology: Vol. (pp. 37-86). Hillsdale, NJ: Erlbaum. [Content Link]
Roy, C. (1991). Senses. In C. Roy & H. A. Andrews(Eds.), The Roy Adaptation Model: The definitive statement (pp. 164-189). Norwalk, CT: Appleton & Lange. [Content Link]
Roy, C., & Andrews, H. A. (1991). The Roy Adaptation Model: The definitive statement. Norwalk, CT: Appleton & Lange. [Content Link]
Roy, C., & Corliss, C. P. (1993). The Roy Adaptation Model: Theoretical update and knowledge for practice. In M. E. Parker (Ed).Patterns of nursing theories in practice (pp. 215-229). New York: National League for Nursing. [Content Link]
Savedra, M. C., Holzemer, W. L., Tesler, M. D., & Wilkie, D. J. (1993). Assessment of postoperative pain in children and adolescents using the adolescent pediatric pain tool. Nursing Research, 42, 5-9. [Content Link]
Savedra, M. C., Tesler, M. D., Holzemer, W. L., & Ward, J. A. (1992). Adolescent Pediatric Pain Tool (APPT): User's manual. San Francisco: University of California, School of Nursing. [Content Link]
Savedra, M. C., Tesler, M. D., Holzemer, W. L., Wilkie, D. J., & Ward, J. A. (1989). Pain location: Validity and reliability of body outline markings by hospitalized children and adolescents.Research in Nursing and Health, 12, 307-314. [Content Link]
Savedra, M. C., Tesler, M. D., Holzemer, W. L., Wilkie, D. J., & Ward, J. A. (1990). Testing a tool to assess postoperative pediatric and adolescent pain. In D. C. Tyler & E. J. Krane (Eds.),Advances in Pain Therapy (Vol. 15, pp. 85 - 93), New York: Raven Press. [Content Link]
Scherer, M. W., & Nakamura, C. Y. (1968). A fear survey for children (FSS-FC): A factor analytic comparison with manifest anxiety (CMAS). Behavior Research and Therapy, 6, 173 - 182. [Content Link]
Tesler, M., Savedra, M., Ward, J. A., Holzemer, W. L.,& Wilkie, D. (1988). Children's language of pain. In R. Dubner, G. F. Gebhart, & M. R. Bond (Eds.), Pain research and clinical management,Proceedings of the 5th World Congress on Pain (vol. 3, pp. 348 - 352). Amsterdam: Elsevier. [Content Link]
Tesler, M. D., Savedra, M. C., Holzemer, W. L., Wilkie, D. J., Ward, J. A., & Paul, S. M. (1991). The word-graphic rating scale as a measure of children's and adolescents' pain intensity. Research in Nursing and Health, 14, 361-371. [Content Link]
Thomas, A., & Chess, S. (1977). Temperament and development. New York: Brunner and Mazel. [Content Link]
Van Cleve, L., Johnson, L., & Pothier, P. (1996). Pain responses of hospitalized infants and children to venipuncture and intravenous cannulation. Journal of Pediatric Nursing,11, 169-174. [Content Link]
Wallander, J. L., Hubert, N. C. & Varni, J. W. (1988). Child and maternal temperament characteristics, goodness of fit, and adjustment in physically handicapped children. Journal of Clinical Child Psychology, 17, 336-344. [Content Link]
Wilkie, D.J., Holzemer, W.L., Tesler, M.D., Ward, J.A., Paul, S.M., & Savedra, M.C. (1990). Measuring pain quality: Validity and reliability of children's and adolescents' pain language. Pain,41, 151-159. [Content Link]
Wong, D. L., & Baker, C. M. (1988). Pain in children: Comparison of assessment scales. Pediatric Nursing,14, 9-17. [Content Link]
Young, M. R., & Fu, V. R. (1988). Influence of play and temperament on the young child's response to pain. Children's Health Care, 16, 209-215. [Content Link]
Cognitive-behavioral interventions, typically effective in reducing anxiety and pain, have not been applied to adolescents undergoing major orthopaedic surgery.
To determine the effectiveness of three cognitive-behavioral interventions for reducing adolescents' postoperative anxiety and pain following spinal fusion surgery for scoliosis, and whether effectiveness depended on preoperative anxiety and age.
A randomized controlled trial with four groups receiving a videotape intervention (information only, coping only, information plus coping, or control) used a convenience sample of 109 adolescents (88 female, 93 White), 11-18 years of age ( M = 14). Speilberger's (1983) State Anxiety scale assessed anxiety preoperatively and postoperatively on Day 2. A visual analogue scale assessed pain postoperatively on Days 2 and 4.
Information plus coping was most effective for reducing postoperative anxiety in adolescents with high preoperative anxiety. Coping instruction led to less postoperative anxiety and pain for adolescents ages 13 and younger. The control group reported the highest levels of pain on Day 4.
Cognitive-behavioral interventions designed to prepare adolescents for surgery should be tailored to individual factors and developmental needs, especially the adolescents' preoperative anxiety level and age.
Accepted publication February 4, 2003.
Parts of this study were funded by NIH-National Institute of Nursing Research, grant #RO1NR02673 to Lynda LaMontagne.
Corresponding author: Lynda L. LaMontagne, RN, DNSc, School of Nursing, Vanderbilt University 412 Godchaux Hall, 461 21 st Avenue South, Nashville, TN 37240-0008 (e-mail: [email protected]).
Lynda L. LaMontagne, RN, DNSc, is Professor; Joseph T. Hepworth, PhD, is Research Associate Professor; and Michele H. Salisbury, RNC, PhD, is Assistant Professor, Vanderbilt University School of Nursing, Nashville, Tennessee.
Frances Cohen, PhD, is Associate Professor, School of Medicine, University of California, San Francisco.
*If there are more than two groups to be examined using the Johnson-Neyman technique, an alternative strategy has been suggested by Huitema (1980) and used by Klein and Ross (1993) . Rather than combining groups with similar slopes to create two comparison groups to use in analyses, all possible pair-wise comparisons among groups are assessed. These pair-wise comparisons were also made and the results were essentially the same as reported here.
Idiopathic scoliosis is the most prevalent deforming orthopaedic condition that affects children during their transition into adolescence, creating significant disabilities if the spinal curve progresses beyond 45 ° . Surgical correction is needed when a back brace fails to correct the curve. The primary problems an adolescent faces immediately following major spinal surgery are controlling anxiety and managing intense pain. Surgery to correct idiopathic scoliosis is both long (6-10 hours) and physically demanding. Because of the complexity of the surgical procedure and resulting postoperative pain and distress, spinal fusion surgery is one of the most invasive orthopaedic surgeries performed on children ( Kotzer, 2000 ). The purpose of this study is to determine the effectiveness of three cognitive-behavioral interventions for reducing adolescents' postoperative anxiety and pain following major spinal surgery and whether preoperative anxiety and age influenced the effectiveness of the interventions.
Cognitive-behavioral strategies are an important component of preparation for children and adolescents scheduled for medical procedures or surgery. Cognitive-behavioral interventions (techniques intended to alter the patient's emotional distress) change the individual's thoughts through attention processes, images, and positive statements. The goal is to alter the individuals' thoughts and interpretation of the stressors encountered and to lessen their distress by enhancing their sense of control or confidence in coping with the experience. These techniques include presentations about the procedure or operation and sensations that are to be expected ( Johnson, 1973 ; Johnson, Fieler, Wlasowicz, Mitchell, & Jones, 1997 ), and instruction presentations in coping using strategies (e.g., guided imagery, relaxation exercises, positive self-talk statements) ( Broome, Rehwaldt, & Foley, 1995 ; Powers, 1999 ).
Most studies testing cognitive-behavioral interventions for decreasing children's and adolescents' procedure-related distress and pain have been conducted with individuals undergoing invasive treatment (e.g., bone marrow aspiration for childhood cancer) ( DuHamel, Redd, & Vickberg, 1999 ; Powers, 1999 ). Children implementing cognitive-behavioral strategies show less anxiety or distress during the procedure and self report less pain after the procedure. While most studies report reductions in children's distress, not all children are responsive to the interventions. Children who are highly anxious may benefit less from the interventions because of their inability to concentrate and/or practice the techniques ( Broome et al., 1995 ).
Psychological preparation of children for surgery using behavioral strategies (e.g., relaxation and training in coping skills) has been beneficial in reducing postoperative anxiety and distress and generally improving psychological adjustment ( Kain, Caldwell-Andrews, & Wang, 2002 ). Cognitivebehavioral strategies, tested with younger children (2-12 years of age) undergoing minor surgery, indicate that strategies such as (a) relaxation ( Kain et al., 2002 ); (b) role play ( Hatava, Olsson, & Lagerkranser, 2000 ; Kain, Mayes, & Caramico, 1996 ); (c) film modeling ( Pinto & Hollandsworth, 1989 ; Melamed & Siegel, 1975 ); and (d) training in coping skills ( Kain et al., 1998 ) are effective for reducing preoperative fear, anxiety and distress. One study showed that children who received a combined cognitive-behavioral intervention (information modeling and coping) fared better than the children who received only one of the intervention modalities ( Kain et al., 1998 ). In this study the combined intervention was only effective for reducing anxiety and fear preoperatively. These positive effects did not extend to the postoperative recovery period. No studies systematically tested cognitive-behavioral strategies with adolescents undergoing major surgery.
The efficacy of cognitive-behavioral interventions may also vary depending upon the child's age. Because older school-age children and adolescents are more advanced in their cognitive development (e.g., the ability to consider more than one variable at one time) than younger children, they are typically more active in their coping. Older children can more ably seek and use information for managing a stressful event. Providing clear, concise information about what to expect and instruction on how to implement coping behavior during treatment helped older children feel less distress following a medical procedure ( LaMontagne, 1993 ). Thus, informational interventions may be more influential on older school-age children and adolescents in handling the stresses of surgery.
The literature suggests that cognitive-behavioral interventions for children undergoing surgery are effective for reducing anxiety and distress and enhancing coping. However, the interventions' content differs widely. Not only are some interventions more effective than others, but also different combinations of interventions may vary in their efficacy. More recent studies suggest that information interventions and coping instruction interventions may both enhance recovery for most patients, and that a combined approach may be most effective ( Kain et al., 1998 ). However, the generalizability of these findings to children awaiting major surgery has not been demonstrated. Children undergoing minor surgery have fewer physical and psychological demands to face because they typically spend only 1 or 2 days in the hospital and have a shorter recovery period than children undergoing major surgery. The stressors of major orthopaedic surgery are more complex and demanding ( LaMontagne, Hepworth, & Cohen, 2000 ).
The current study investigates the effectiveness of cognitive-behavioral information and coping instruction interventions on adolescents' postoperative anxiety and pain following major spinal surgery. Three cognitive-behavioral intervention modalities were tested: (a) a concrete-objective information-only intervention which described the upcoming surgery including the sensations the patient would likely experience; (b) a coping-only instruction intervention which instructed the adolescents in what strategies they could realistically implement postoperatively; and (c) a combined information plus coping intervention. The first set of questions addressed whether there were differences among the three modalities (i.e., did the interventions have an effect on postoperative anxiety and pain?). Because an adolescent's preoperative anxiety level or age might influence her/his ability to use the cognitive-behavioral information effectively, a second set of questions examined whether preoperative anxiety level or age influenced the effectiveness of the interventions.
The design was a randomized control trial with four groups. Participants were randomly assigned (according to a computer-generated list of random numbers) to one of three intervention groups or a control group. The three interventions consisted of concrete-objective information plus coping instruction, concrete-objective information only, or coping instruction only. Anxiety was assessed preoperatively and again on the second postoperative day. Pain was assessed on the second and fourth days following surgery.
The effect sizes which could be detected using the various analyses ( N = 100, alpha level = .05, power = .80) to be employed in the study were calculated. The effect size, which could be detected for the main and interaction effects ( N = 100 [25 per cell], alpha = .05, power = .80), was an f of .29 ( Cohen, 1988 , p. 312). Thus, effect sizes slightly larger than Cohen's medium effect size would be detected. In order for an intervention to be of clinical significance, an effect size at least in this range is desirable.
The concrete-objective information plus coping instruction intervention included a combination of procedural and sensory information and instruction in specific coping behaviors that children could realistically perform during recovery. Information focused on specific details about the surgical procedure such as: (a) the length of surgery; (b) the insertion of spinal instrumentation; and (c) postsurgery expectations (e.g., Foley catheter, device for patient control of pain). Information was also presented regarding (a) normal sensation expectations; (b) the length of these sensations; (c) the time that sitting up and ambulation would occur; and (d) the type of activity restrictions that would be imposed. Coping behaviors (i.e., relaxation, deep breathing, positive thinking) were presented as strategies for children to manage postoperative stressors such as pain.
The concrete-objective information only intervention included only procedural and sensory information identical to the concrete-objective information presented in the combined intervention; the coping instruction was omitted. The coping-only intervention included only the information related to coping behaviors identical to the coping information presented in the combined intervention. No concrete-objective information was given.
The control intervention provided standard information about the surgery experience such as (a) information about the length of surgery; (b) postoperative routines (e.g., turning in bed, I.V. medication); (c) the hospital environment (e.g., recovery room, patient rooms); (d) the usual hospital and nursing routines; (e) family visitation; and (f) postoperative clinic visits.
In order to enhance the consistency of the delivery of the intervention among participants, all interventions were delivered by videotape. Each videotape was 8-10 minutes in length. The intervention videotapes were developed and filmed by a Child Life Specialist who delivered the content by Teleprompter in order to look directly toward the person viewing the video. The Child Life Specialist delivered all videotaped interventions (including control) to the participants.
Adolescents were asked to participate in the study if they met the following criteria: (a) scheduled for major orthopaedic surgery for repair of idiopathic scoliosis that included the insertion of spinal instrumentation; (b) 11-18 years of age; (c) no previous major orthopaedic surgery for repair of idiopathic scoliosis; (d) no learning or developmental problems (as noted in the adolescent's medical record); and (e) English-speaking child and parent. All participants and their parents read and signed an informed consent document, which, along with the study protocol, had been approved by the medical center Institutional Review Board.
Anxiety was measured using the child/adolescent version of Spielberger's (1983) State-Trait Anxiety Inventory; the 20-item State Anxiety Scale was used. The age appropriate state scale, 'How I Feel Questionnaire,' is widely used to measure anxiety in children and adolescents because it has good internal consistency and evidence of construct validity ( Spielberger, 1983 ). Scores range from 20-60, with higher scores indicating higher anxiety. In previous research, coefficient alphas were .92 and .93 for children's preoperative and postoperative anxiety, respectively ( LaMontagne, Hepworth, Johnson, & Cohen, 1996 ).
A Visual Analogue Scale (VAS) was used to measure pain. For this study, a graduated 10-centimeter line with anchors at the endpoints of 0 and 10 was used. This allowed the participants to rate their level of pain intensity on a continuous scale from 0 (no pain at all) to 10 (worst pain ever). Various studies have demonstrated that school-age children and adolescents can accurately describe their own pain using self-report measures ( Abu-Saad & Holzemer, 1981 ; Buchanan, Voigtman, & Mills, 1997 ). In a meta-analysis of pain interventions with children, highly significant relationships between pain management and behavioral self-report and distress responses of children were found ( Broome, Lillis, & Smith, 1989 ). The validity of using a VAS to measure children's postoperative pain is further supported by studies of children who reported less pain after the administration of pain medication ( Tyler, Tu, Douthit, & Chapman, 1993 ), and a decrease in pain intensity at different times over the postoperative course ( McGrath et al., 1996 ; Savedra, Holzemer, Tesler, & Wilkie, 1993 ). Participants practiced using the VAS preoperatively and were told that the scale would be used to assess their pain following surgery.
On the day before surgery, adolescents and their parents were approached by the researcher during their preoperative orthopaedic clinic visit and asked to participate in the study. After reading and signing the informed consent document, those who agreed to participate were randomly assigned to one of the four groups. The adolescents and their parents were interviewed separately by one of the researchers in a private room. The adolescents completed the anxiety scale and the parents provided demographic information verbally to the researcher. Then, the adolescents and their parents jointly viewed the assigned videotaped intervention with the researcher. Each intervention video was approximately 8-10 minutes long. Participants in the groups receiving coping instruction interventions practiced the coping skills taught on the videotape with the researcher. This occurred after they viewed the videotape; parents also participated.
On the second postoperative day, participants completed both the VAS to assess their pain and the anxiety scale usually between 12:00 noon and 3:00 PM because patients were typically awake at this time. Both scales were verbally administered by a researcher. On the fourth postoperative day, participants again completed the VAS. All patients had the same method of pain control (i.e., patient-controlled analgesia for the first 3 postoperative days) as managed by the pediatric pain service of the medical center.
Equivalence of the intervention groups was tested using analysis of variance (ANOVA) for continuous variables and chi-square for discrete variables. Results showed there were no differences in the groups on age, ethnic background, sex, socioeconomic status, or preoperative anxiety indicating the randomization procedure for the study was successfully implemented. Having all data collected in an interview format minimized missing data. To maximize the power of the statistical analyses, all participants with valid data for each analysis were included.
In order to determine the effects of the interventions on postoperative anxiety, the data were analyzed using ANOVA. Analyses of covariance (ANCOVA) models were then used to address our questions more precisely and determine whether the differences depended on the participants' level of preoperative anxiety or age. The covariate by intervention interaction term in these models assessed the homogeneity of regression assumption. When a significant covariate by intervention interaction was found, this meant there was heterogeneity of slopes and, thus, the assumptions of the ANCOVA analysis were not met. In this case, the Johnson-Neyman technique is an appropriate analysis strategy ( Dorsey & Soeken, 1996 ; Huitema, 1980 ; Johnson & Neyman, 1936 ; Pedhazur & Schmelkin, 1991 ). The technique specifies the values of a covariate for which the dependent variable is significantly different between the two groups. To carry out the analysis, the slope for each individual intervention group was examined and intervention groups with similar slopes were combined for subsequent analyses employing the Johnson-Neyman technique ( Johnson & Neyman, 1936 ; Pedhazur & Schmelkin, 1991 ). The analysis uses the entire undivided sample and determines for which values of the covariate the two groups differ significantly on the dependent variable.
To assess the effects of the interventions on postoperative pain, repeated measures ANOVA was used with the repeated measure being Day 2 and Day 4 postoperative pain. Repeated measures ANCOVA was also used controlling for participants' age, using the same modeling strategy as described above for anxiety (i.e., assessing the homogeneity of regression assumption, using the Johnson-Neyman technique if there is a significant covariate by intervention interaction).
Of the 117 patients identified from the preadmission fact sheet (e.g., patient's age, diagnosis, surgical procedure) of three orthopaedic surgeons in a Southeastern Medical Center, four families refused to participate for personal reasons, resulting in a convenience sample of 113 adolescents. Two participants died from surgical complications on the first day after surgery and two participants withdrew on Day 2 because they felt too distressed to answer questions. This left 109 participants available for analyses of postoperative anxiety. However, four additional participants withdrew on Day 4 and scheduling problems developed over collecting the pain data; 89 participants were available for the pain analyses.
The participants were predominantly female ( n = 88, 81%), and White ( n = 93, 85%); 32 of the participants (all girls) were 11 or 12 years of age. These numbers are consistent with the incidence of idiopathic scoliosis, which is typically diagnosed in young female adolescents during their growth spurt; incidence is 7:1 females to males and is more common in White adolescents ( Goldberg, Mayo, Poitras, Scott, & Hanley, 1994 ). The mean age was 13.9 years of age at enrollment into the study with a range of 11-18 years. According to the Hollingshead Four Factor Index of Socioeconomic Position (1975) , families ranged from Class 1 (unskilled laborers, 3.7%) to Class 5 (professional, 17.4%), with the largest percent being in Class 4 (semiprofessional, 36.7%).
The means for anxiety, pain, and age are presented in Table 1 . Postoperative anxiety was significantly correlated with both postoperative Day 2 pain ( r = .55, p < .01) and postoperative Day 4 pain ( r = .37, p < .01). Postoperative Day 2 pain was significantly correlated with postoperative Day 4 pain ( r = .52, p < .01).
The first analysis examined whether there were overall differences in postoperative anxiety among the three intervention groups: information group ( n = 27); coping group ( n = 27); information plus coping group ( n = 30); and control ( n = 25). The ANOVA analysis showed no differences among groups (F[3,105] = .92, p = .44).
The next analysis investigated whether the level of preoperative anxiety influenced the effectiveness of the interventions. Results of an ANCOVA analysis controlling for preoperative anxiety showed a significant covariate (preoperative anxiety) by intervention group interaction ( p < .01). Regressing postoperative anxiety on preoperative anxiety for the four groups indicated that the regression lines did not all show similar slopes ( Figure 1 ). The regression for the information plus coping group was markedly different from those of the other three groups. The intercept for the information plus coping group was more than twice the size of the other groups (63.5) and it had a negative slope (-.78) indicating that higher levels of preoperative anxiety were associated with lower levels of postoperative anxiety. The regressions for the other three groups were all quite similar, with positive slopes ranging from .23-.27 and positive intercepts ranging from 27.51-31.46. When the information only, coping only, and control groups were analyzed together, there was no covariate by group interaction effect (F[2,73] = 0.01, p = .99) and no group effect (F(2,73) = 1.17, p = .32), indicating that the slopes and intercepts were not significantly different among the three groups. Based on the homogeneity of the three groups, the two single modality intervention groups (information only, coping only) and the control group were combined into a single redefined group (referred to hereafter as redefined group) and compared with the information plus coping group in the subsequent analyses. *
|
FIGURE 1.
Within Group Regression Results of Postoperative Anxiety Regressed on Preoperative Anxiety. [solid grey line] Concrete-objective information + coping; [broken dark line] Concrete-objective information only; [solid dark line] Coping only; [broken grey line] Control. |
Analysis using the two comparison groups (the information plus coping group and the redefined group) yielded a significant interaction effect (F[1,105] = 13.28, p < .01) indicating that the regression of postoperative anxiety on preoperative anxiety differed for the two groups. The Johnson-Neyman technique was employed and the analysis revealed that for preoperative anxiety scores greater than 37 (i.e., higher anxiety), the information plus coping intervention was significantly more effective for decreasing postoperative anxiety than in the redefined group. For preoperative anxiety scores less than 26 (i.e., lower anxiety), the redefined group had lower postoperative anxiety than the information plus coping group.
Similar analyses were used to determine if age influenced the effectiveness of the interventions. Results of an ANCOVA controlling for age showed that there was a significant covariate by intervention group interaction ( p = .03). The relationship between age and postoperative anxiety differed depending on group. Regression analyses revealed a positive relationship between age and postoperative anxiety for the two interventions that included coping (information plus coping and coping only). When analyzing these two groups together there was no covariate by group interaction effect (F (1, 53) = 0.91, p = .34), and no group effect (F (1, 53) = 0.00, p = .95), indicating that the slopes and intercepts for these two groups were not significantly different. Regression analyses revealed a negative relationship between age and postoperative anxiety for the two interventions that did not include coping instruction (information only and control). When analyzing these two groups together there was no covariate by group interaction effect (F (1, 48) = 0.22, p = .64) and no group effect (F (1, 48) = 0.04, p = .84) indicating that the slopes and intercepts for these groups also were not significantly different. Therefore, based on the homogeneity of regressions within each of these pairs, the four groups were redefined into two groups, coping group and no coping group.
Analysis comparing coping group vs. no coping group yielded a significant interaction effect (F[1,105] = 8.48, p < .01) indicating that the regression of postoperative anxiety on age differed for the two groups. The Johnson-Neyman technique revealed that for adolescents younger than 13.25 years, the interventions including coping were significantly more effective for reducing postoperative anxiety than the interventions that did not include coping. The difference between the interventions was not significant for older adolescents.
A 2 × 4 repeated measures ANOVA was used to assess the effects of the interventions on postoperative pain measured at Day 2 and Day 4 ( Table 1 ). The significant decrease in pain from Day 2 (M = 6.38) to Day 4 (M = 5.12), a 20% reduction, is interpreted in light of the significant time by group interaction. The simple effects for time revealed significant reductions in pain from Day 2 to Day 4 for the information only group, the information plus coping group, and the control group. There was no significant difference for the coping only group ( Table 2 ).
Repeated measures ANCOVA analyses were used to determine if age influenced the effectiveness of the interventions. When controlling for age, the intervention group main effect ( p = .01) and the intervention group by age interaction ( p = .02) were significant. Although the significant group main effect indicated that there were differences among the four groups on postoperative pain, Tukey's post hoc tests revealed that none of the pair-wise comparisons of the four groups were significantly different. The group effect should be interpreted in light of the significant group by age interaction.
This interaction was probed by regressing the average least squares pain mean score on age for each group. The relationship between pain and age was negative for the information only group (b = -.58) and the control group (b = -.23) and it was positive for the information plus coping group (b = .19) and the coping only group (B = .36). Therefore, two redefined groups were formed: coping group and no coping group.
Analysis comparing the coping group vs. no coping group yielded a significant interaction effect (F(1,85) = 10.50, p < .01) indicating that the regression of postoperative pain on age differed for the two groups. The Johnson-Neyman technique revealed that for adolescents 13.46 years old or younger, the interventions that included coping were significantly more effective for reducing average levels of postoperative pain than the interventions that did not include coping. The difference between the interventions was not significant for older adolescents.
In this randomized control trial the concrete-objective information plus coping intervention provided the most benefit in anxiety reduction when preoperative anxiety was high. It is possible that learning specific coping behaviors enhanced the effectiveness of the information provided because adolescents could anticipate recovery and use techniques to help them manage their emotional distress. The combination of accurate expectations and knowledge in dealing with the stressors may have offered strategies needed for optimal preparation for surgery (e.g., Johnson, 1973 ).
When preoperative anxiety was low, the single interventions (concrete-objective information only; coping only) or the control (i.e., standard information) were more effective in reducing postoperative anxiety than the combined (concrete-objective information plus coping) intervention. Most studies have shown that cognitive-behavioral techniques, such as providing information about the surgery or instruction in coping techniques, help to reduce children's anxiety. This study confirms that these single modality approaches (e.g., information only or coping only) are useful in reducing postoperative anxiety for low anxiety adolescents undergoing major orthopaedic surgery. Individuals with low preoperative anxiety may benefit from the information or strategies provided by a single modality; a combined approach may raise their anxiety level by offering more information and advice than they can assimilate.
Pain levels significantly decreased from Day 2 to Day 4. The information interventions were the most effective for reducing postoperative pain. Adolescents in the information only and information plus coping group, and in the control group (standard information) had significant pain reduction from Day 2 to Day 4. The combined information plus coping group had the most pain reduction (35%). This finding suggests that providing specific information (even standard information) about postoperative pain may be necessary to facilitate adolescents' recognition of the importance of pain control and identification of when it is needed. Perhaps the adolescents in the coping only group needed more time to recover from the surgical procedure itself before they could effectively use the coping techniques in regard to pain.
The fact that postoperative pain was not influenced by the coping intervention may be a reflection of the nature of the recovery process. Patients are physically exhausted after undergoing spinal surgery and they are normally under high doses of narcotic pain medication during the first few days after surgery. Under these circumstances it may be unrealistic to expect patients to be able to fully use the coping instruction provided by the interventions. It is possible that coaching by nurses and parents could have helped the patients implement the coping strategies. This statement poses an important issue for future research.
When controlling for age, the pattern of results for anxiety and pain revealed important insights. For younger adolescents (11-13 years), the two coping interventions (coping only; information plus coping) were the most effective interventions in decreasing postoperative anxiety and pain whereas no differences were found among the intervention groups for older adolescents (ages 14-18). The cognitive-behavioral interventions used in this study were based on the theoretical assumption that children in the age range studied (11-18) are able to explore several alternatives and demonstrate an age-related ability to use cognitive coping techniques ( Crain, 1992 ). Yet only younger adolescents (11-13 years) benefited from the two interventions that included coping. It is possible that the characteristic developmental tasks of the three distinctive adolescent phases ( Hamburg, 1985 )-early (11-13 years), mid (14-16 years) and late (17-18 years)-influenced the behavioral outcomes.
For example, although adolescents in each developmental phase may be similar in cognitive ability, the psychosocial changes that typically occur during each phase can influence how information is used. In early adolescence there is concern over body image ( Hamburg, 1985 ). Most adolescents are deeply concerned about physical attractiveness and vulnerable to real or imagined assaults on body integrity. Increased concern about physical appearance among the younger adolescents may have engendered receptivity to strategies for dealing with surgical recovery (e.g., focusing on the positive benefits of surgical repair) and resulted in a greater benefit from the coping interventions.
In contrast, adolescents in the later phases of development are less susceptible to outside influences and are typically more secure about their own judgments and abilities to handle difficult situations. Older adolescents have developed a larger repertoire of useful coping strategies and may not need nor want more instruction in coping techniques. Older adolescents tend to rely more on friends and siblings for information. Videotaped interventions presented by an older adolescent who also had gone through the experience of coping with spinal surgery might have had a strong impact.
The results of this study have important implications. Cognitive-behavioral interventions designed to prepare adolescents for surgery that are tailored to individual factors (e.g., level of preoperative anxiety) and developmental needs may be needed. When preoperative anxiety is high, a concrete-objective information plus coping intervention may be effective for reducing anxiety because it offers details about the surgery, and provides information about how to cope with distress. Interventions that include coping may be especially effective in reducing anxiety and pain for younger adolescents because they may be more open to advice and instruction in coping techniques.
Effects of Distraction on Children's Pain and Distress During Medical Procedures: A Meta-Analysis Outline
Graphics
Background: It is difficult to determine the usefulness of distraction to decrease children's distress behavior and pain during medical procedures because many studies use very small samples and report inconsistent findings.
Objectives: To investigate the mean effect sizes across studies for the effects of distraction on young children's distress behavior and self-reported pain during medical procedures.
Method: Hunter and Schmidt's (1990) procedures were used to analyze 16 studies (total n = 491) on children's distress behavior and 10 studies (total n = 535) on children's pain.
Results: For distress behavior, the mean effect size was 0.33 (±0.17), with 74% of the variance accounted for by sampling and measurement error. For pain, the mean effect size was 0.62 (±0.42) with 35% of the variance accounted for. Analysis of studies on pain that limited the sample to children 7 years of age or younger (total n = 286) increased the amount of explained variance to 60%.
Conclusions: Distraction had a positive effect on children's distress behavior across the populations represented in this study. The effect of distraction on children's self-reported pain is influenced by moderator variables. Controlling for age and type of painful procedure significantly increased the amount of explained variance, but there are other unidentified moderators at work.
Relevant Literature
Distraction is a class of cognitive coping strategies that divert attention from a noxious stimulus through passively redirecting the subject's attention or by actively involving the subject in the performance of a distractor task (Fernandez, 1986). Thus, distraction involves the cognition, expectancies, or appraisals of an individual, and results in a modification of the individual's behavior. According to McCaul and Malott (1984), distraction affects pain perception because: (a) pain perception is partially a cognitively controlled process; and (b) distraction consumes part of an individual's finite attentional capabilities, leaving less attention or focus available to perceive pain.
McGrath (1991) suggests that distraction affects the perception of pain because it directly interferes with neuronal activity associated with pain. The gate control theory of pain (Melzack & Wall, 1965; Wall, 1978) suggests that pain is modulated by a gating mechanism that opens and closes nerve impulses to the brain. The gating mechanism is influenced by cognitive processes, such as attention to the noxious stimuli. The relationship between cognitive processes and the perception of pain has been questioned by others, however. Willis & Coggeshall (1991) suggest that there may be dual neuronal pathways for the sensation of pain and the reaction to pain, and cite studies reporting that patients who have had frontal lobotomy feel pain but are not 'concerned' by it, and that patients who have had spinal transection do not feel pain but continue to have a vigorous flexor response to painful stimuli. In a descriptive study of pain after major surgery, Beyer, McGrath, and Berde (1990) found discordance between the intensity of children's perceived pain and behavioral reactions to pain. The relationships between pain stimulation, pain-related behavior, and perceived pain are probably more complex than previously thought.
Some researchers use distress behaviors (e.g., crying, moaning, fighting, and verbal resistance) as a proxy for pain. However, behavior can be influenced by many things other than pain (e.g., fear, anxiety, and temperament). Therefore, the research on the use of distraction with children during medical procedures should be explored and judged for its separate effects on pain and on behavioral distress, as put forth in this analysis.
The purpose of this analysis is to quantitatively estimate the effect of distraction interventions on young children's perceived pain and observed distress behavior during medical procedures. The research questions are: (a) What are the average effect sizes for pain and distress behavior among the studies included in the analysis?; (b) What is the variability among the effect sizes?; and (c) How much of the observed variance is due to sampling and measurement error?
Method
Sample: Standard search procedures were used to locate published and unpublished studies. Electronic databases searched were Cancerlit (1992-August 1996), Healthstar (1994-September, 1996), Medline (1966-October 1996), and CINAHL (1982-August 1996), using the key terms pain, distraction, imagery, and attention, and was limited to the age group infant to 12 years and to the English language. A hand search of the CINAHL database from 1970 to 1982 was conducted using the key terms pain and children. Additional sources included the Psychology Database at the University of Iowa libraries (containing journal articles from 1967 to October 1996 and book chapters from 1967 to 1980 and 1987 to 1996); the Periodical Abstracts Database (indexing articles from 1,600 common journals from 1986 to 1996); ERIC (an index of education-related literature); and Wilson Database (WLS). The search strategy for those databases included the keywords imagery or imagination, attention, distraction, pain, and child age group. Unpublished studies were discovered by searching the Dissertation Abstracts database for 1961 to 1996 using an intentionally broad search strategy with the keywords child and pain or distraction or distress. Although this strategy captured a large number of studies, very few of them dealt with medical procedures. Reference lists accompanying research and review articles on children and pain were scanned for any studies missed through the database searches. The searches resulted in a large number of citations for initial screening: 192 from Medline; 784 from Dissertation Abstracts; and 136 from the psychological databases.
Study Selection: The inclusion criteria were: (a) mean age of subjects was under 12 years; (b) study designs were randomized clinical trials (RCT) or repeated measures (RM) designs; (c) means and standard deviations (SDs) for both control and experimental conditions were available for the outcome variables self-reported pain or observed behavioral distress; and (d) distraction was the intervention. For the purposes of this analysis, distraction was defined as any intervention intended to focus the subject's attention away from pain or discomfort.
Nineteen studies (see Table 1) met all of the inclusion criteria. Seventeen studies included published means and SDs and two authors (S. Arts, personal communication, May 12, 1997; R. Blount, personal communication, June 2, 1997) provided the unpublished statistics. An additional eight authors, whose published works were missing group means and SDs, either did not have the information available or did not respond to inquiries.
TABLE 1. Effects of Distraction on Children's Self-reported Pain and Observed Distress Behavior
Study Descriptions: The studies used in the meta-analyses are summarized in Table 1. The mean age for subjects was 6.6 years, with a range of 3 to 15 years. The medical procedures varied in complexity and painfulness. The rationale for accepting studies with varying procedures was that this was an effort to sample the universe of painful experiences that children encounter. Therefore, the intensity of the medical procedure experience was not restricted through the choice of studies included in the analysis.
The distraction interventions (independent variable) varied in complexity, from simple things that could be manipulated, such as a kaleidoscope, to a package of distraction techniques that included concentrating on breathing, imagery, and nonprocedural talk. The consistent intent of the distraction interventions was to divert the child's attention away from the medical procedure.
The self-reported pain measures used in these studies are generally visual one-item scales. Most of the behavioral distress scales used in these studies are adaptations of the Behavior Rating Profile Scale created by Melamed (Melamed, Hawes, Heiby, & Glick, 1975). Behaviors such as crying, whining, grimacing, and thrashing were coded as either present or absent at varying intervals throughout the procedure and a final total score was given for behavioral distress.
Meta-Analysis Procedure: An effect size statistic (d) for each study was computed directly from means and SDs. Cohen's (1977) formula for d ([mean of experimental condition - mean of control condition)/pooled group SD] was used for RCT design studies. For RM studies, the denominator of the equation was the control condition SD (Hunter & Schmidt, 1990, p. 352). Reported means and SDs were always used for calculating d when available, which was the case for 14 studies. One study reported results in logged form; the numbers were anti-logged for use in this analysis. Four studies reported results in graphic form. The authors of this review independently transformed graphed data into numeric data. One hundred percent agreement was reached on the transformed numbers. Means and SDs were then calculated for each study.
Hunter and Schmidt's (1990) meta-analytic technique was used to calculate the mean effect size weighted by sample size Equation 1 the observed variance of the d values Equation 2 and sampling error. Equation 3 For studies with RM dependent group designs, a different formula Equation 4 was used to calculate sampling error because it is affected by the correlation between pretest and post-test scores (Becker, 1988). When the size of the correlation between the scores is unknown, the relationship can be estimated from data given in the studies (Dunlap, Cortina, Vaslow, & Burke, 1996). Seventeen sets of baseline and control condition scores for 17 children reported in the studies by Elliott and Olson (1983); Jay, Elliot, Ozolins, Olson, and Pruitt (1985); Powers, Blount, Bachanas, Cotter, and Swan (1993); and Stark et al. (1989) were used to calculate a correlation coefficient (r = .55).
Results
The results are reported in Table 2. The effect sizes for both pain and behavioral distress had negative signs, meaning that distraction decreased self-reported pain and observed distress behavior. Because readers are used to thinking of beneficial effects as positive, the absolute values of the effect sizes are used in this report.
TABLE 2. Meta-analysis Results of the Effects of Distraction on Pain and Observed Distress
As stated previously, younger children tend to report more pain with medical procedures. Therefore, a subanalysis was conducted on the three studies with subjects exclusively between the ages of 3 and 7 years (Blount et al., 1992; Fowler-Kerry and Lander, 1987; Gonzalez, Routh, & Armstrong, 1993). Because all of the subjects had injections as part of well-child care, this subanalysis also controlled for the type of medical procedure. For the 268 children in these three randomized clinical trials, the average effect size was .47 (±0.26). Sampling and measurement error accounted for 60% of the variance in the effect size. Thus, it appears that moderator variables other than age and procedure type influence the effectiveness of distraction.
Distress: The analysis for behavioral distress shows an average effect size (absolute value) of 0.33 (±0.17) for the 491 children in the sample studies. The credibility interval for d¯ does not include zero, indicating distraction has a positive effect on children's distress behavior. Sampling error and measurement error accounted for 73.44% and .4% of the observed variance, respectively. Hunter and Schmidt (1990) suggest that when approximately 75% of the observed variance is accounted for, the remaining variance is likely to be caused by a combination of other statistical artifacts, namely reliability of the independent variables, study differences in range restriction, and instrument validity.
Discussion
This meta-analysis found a moderate effect size for the influence of distraction on observed distress behavior in children. Sampling and measurement error accounted for almost 74% of the observed variance among the studies, indicating that the positive effect of distraction on distress behavior is seen across the populations sampled by the studies.
For the effect of distraction on self-reported pain, the 90% credibility interval was wide and encompassed zero. Credibility intervals are calculated with the corrected standard deviation of the mean effect size and are used in meta-analysis to determine whether or not moderator variables are operating (Whitener, 1990). When the interval includes zero, variability in the effect size might be due to remaining statistical artifacts or to moderator variables. Because only about 35% of the observed variance was accounted for by sampling error and measurement error, a subanalysis that controlled for age and type of medical procedure was completed. Although substantially more of the observed variance was accounted for in this subanalysis, there is still some variability left to explain. Possible moderator variables that come to mind are inconsistencies in the distraction interventions and variations in the characteristics of the children.
Research has shown that the child's innate temperament influences distress behavior during painful procedures (Corbo-Richert, 1994; Lee & White-Traut, 1996; Schechter, Bernstein, Beck, Hart, & Scherzer, 1991; Young & Fu, 1998). However, the relationships between child temperament and perceived pain have not been explored. The child's history of bad experiences with procedures is another possible moderator. In a study of preparation methods for medical procedures, Dahlquist et al. (1986) found a significant main effect for quality of previous medical experience on the behavioral distress of children during throat culture. The amount of past exposure to medical procedures was not related to behavioral distress. Thus, it appears that it is the quality of the child's experiences and not the quantity that matters. Systematic inquiry is needed to explore the influences of temperament and prior experience on children's responses during medical procedures, and to search for additional moderator variables.
Using distraction with children during medical procedures will reduce the amount of observed distress behavior for most children. The magnitude of the benefit will vary from child to child. Distraction is a low-cost intervention that has no risk to the patient and has a measurable benefit.
REFERENCES
References marked with an asterisk (*) indicate studies included in the metaanalysis.
*Arts, S. E., Abu-Saad, H. H., Champion, G. D., Crawford, M. R., Fisher, R. J., Juniper, K. H., & Ziegler, J. B. (1994). Age-related response to lidocaine-prilocaine (EMLA) emulsion and effect of music distraction on the pain of intravenous cannulation. Pediatrics, 93(5), 797-801.
Becker, B. J. (1988). Synthesizing standardized mean-change measures. British Journal of Mathematical and Statistical Psychology, 41, 257-278. [Content Link]
Beyer, J. E., McGrath, P. J., & Berde, C. B. (1990). Discordance between self-report and behavioral pain measures in children aged 3-7 years after surgery. Journal of Pain and Symptom Management, 5, 350-356. [Content Link]
*Blount, R. L., Bachanas, P. J., Powers, S. W., Cotter, M. C., Franklin, A., Chaplin, W., Mayfield, J., Henderson, M., & Blount, S. D. (1992). Training children to cope and parents to coach them during routine immunizations: Effects on child, parent, and staff behaviors. Behavior Therapy, 23, 689-705. [Content Link]
*Broome, M. E., Lillis, P. P, McGahee, T. W., & Bates, T. (1992). The use of distraction and imagery with children during painful procedures. Oncology Nursing Forum, 19, 499-502.
Cohen, J. (1977). Statistical power analysis for the behavioral sciences (revised ed.). New York: Academic Press. [Content Link]
Corbo-Richert, B. H. (1994). Coping behaviors of young children during a chest tube procedure in the pediatric intensive care unit. Maternal-Child Nursing Journal, 22, 134-146. [Content Link]
Dahlquist, L. M., Gil, K. M., Armstrong, F. D., DeLawyer, D. D., Greene, P., & Wuori, D. (1986). Preparing children for medical examinations: The importance of previous medical experience. Health Psychology, 5, 249-259. [Content Link]
Dahlquist, L. M., Power, T., Cox, C., & Fernbach, D. (1994). Parenting and child distress during cancer procedures: A multi-dimensional assessment. Children's Health Care, 23, 149-166. [Content Link]
Dunlap, W. P., Cortina, J. M., Vaslow, J. B., & Burke M. J. (1996). Meta-analysis of experiments with matched groups or repeated measures designs. Psychological Methods, 1, 170-177. [Content Link]
*Elliott, C. H., & Olson, R. A. (1983). The management of children's distress in response to painful medical treatment for burn injuries. Behaviour Research and Therapy, 21(6), 675-683. [Content Link]
Fernandez, E. (1986). A classification system of cognitive coping strategies for pain. Pain, 26, 141-151. [Content Link]
*Foertsch, C. E. (1993). Investigation of an imagery-based treatment for the control of children's behavioral distress during burn dressing changes. Unpublished doctoral dissertation, University of Iowa.
*Fowler-Kerry, S., & Lander, J. R. (1987). Management of injection pain on children. Pain, 30, 169-175. [Content Link]
*Gonzalez, J. C., Routh, D. K., & Armstrong, F. D. (1993). Effects of maternal distraction versus reassurance on children's reactions to injections. Journal of Pediatric Psychology, 18(5), 593-604. [Content Link]
Hester, N. O. (1993). Pain in children. In J. Fitzpatrick & J. Stevenson (Eds.), Annual review of nursing research: Vol. 11 (pp. 105-142). New York: Springer. [Content Link]
Humphrey, G. B., Boon, C. M., van Linden van den Heuvell, G. F., & van de Wiel, H. B. (1992). The occurrence of high levels of acute behavioral distress in children and adolescents undergoing routine venipunctures. Pediatrics, 90, 8-91. [Content Link]
Hunter, J. E., & Schmidt, F. L. (1990). Methods of meta-analysis: Correcting error and bias in research findings. Newbury Park, CA: Sage. [Content Link]
Jay, S. M., Elliott, C. H., Katz, E., & Siegel, S. E. (1987). Cognitive-behavioral and pharmacologic interventions for children's distress during painful medical procedures. Journal of Consulting and Clinical Psychology, 55(6), 860-865. [Content Link]
*Jay, S. M., Elliott, C. H., Ozolins, M., Olson, R. A., & Pruitt, S. D. (1985). Behavioral management of children's distress during painful medical procedures. Behaviour Research and Therapy, 23, 513-520. [Content Link]
Jay, S. M., Ozolins, M., Elliott, C. H., & Caldwell, S. (1983). Assessment of children's distress during painful medical procedures. Health Psychology, 2, 133-147. [Content Link]
Katz, E. R., Kellerman, J., & Siegel, S. (1980). Behavioral distress in children with cancer undergoing medical procedures: Developmental considerations. Journal of Consulting and Clinical Psychology, 48, 356-365. [Content Link]
Keck, J. F., Gerkensmeyer, J. E., Joyce, B. A., & Schade, J. G. (1996). Reliability and validity of the Faces and Word Descriptor Scales to measure procedural pain. Journal of Pediatric Nursing, 11, 368-374. [Content Link]
Lander, J., & Fowler-Kerry, S. (1991). Age differences in children's pain. Perceptual and Motor Skills, 73, 415-418. [Content Link]
Lee, L. W., & White-Traut, R. C. (1996). The role of temperament in pediatric pain response. Issues in Comprehensive Pediatric Nursing, 19, 49-63. [Content Link]
*Manne, S. L., Redd, W. H., Jacobsen, P. B., Gorfinkle, K., Schorr, O., & Rapkin, B. (1990). Behavioral intervention to reduce child and parent distress during venipuncture. Journal of Consulting and Clinical Psychology, 58, 565-572. [Content Link]
McCarthy, A. M., Cool, V. A., Petersen, M., & Bruene, D. A. (1996). Cognitive behavioral pain and anxiety interventions in pediatric oncology centers and bone marrow transplant units. Journal of Pediatric Oncology Nursing, 13, 3-12. [Content Link]
McCaul, K. D., & Malott, J. M. (1984). Distraction and coping with pain. Psychological Bulletin, 95, 516-533. [Content Link]
McGrath, P. A. (1991). Intervention and management. In J. P. Bush & S. W. Harkins (Eds.), Children in pain. New York: Springer-Verlag. [Content Link]
*McGrath, P. A., & deVeber, L. L. (1986). The management of acute pain evoked by medical procedures in children with cancer. Journal of Pain and Symptom Management, 1(3), 145-150.
Melamed, B., Hawes, R., Heiby, E., & Glick, J. (1975). The use of film modeling to reduce uncooperative behavior of children during dental treatment. Journal of Dental Research, 54, 797-801. [Content Link]
Melzack, R., & Wall, P. D. (1965). Pain mechanisms: A new theory. Science, 150, 971-979. [Content Link]
*Olsen, B. R. (1991). Brief interventions for routine use with children in a phlebotomy laboratory. Unpublished doctoral dissertation, West Virginia University.
*Powers, S. W., Blount, R. L., Bachanas, P. J., Cotter, M. W., & Swan, S. C. (1993). Helping preschool leukemia patients and their parents cope during injections. Journal of Pediatric Psychology, 18(6), 681-695. [Content Link]
Schechter, N. L., Bernstein, B. A., Beck, A., Hart, L., & Scherzer, L. (1991). Individual differences in children's response to pain: Role of temperament and parental characteristics. Pediatrics, 87, 171-177. [Content Link]
*Schur, J. M. (1986). Alleviating behavioral distress with music or Lamaze pant-blow breathing in children undergoing bone marrow aspirations and lumbar punctures. Unpublished doctoral dissertation, University of Texas Southwestern Medical Center at Dallas.
*Smith, J. T., Barabasz, A., & Barabasz, M. (1996). Comparison of hypnosis and distraction in severely ill children undergoing painful medical procedures. Journal of Counseling Psychology, 43(2), 187-195.
*Smith, K. E., Ackerson, J. D., & Blotcky, A. D. (1989). Reducing distress during invasive medical procedures: Relating behavioral interventions to preferred coping style in pediatric cancer patients. Journal of Pediatric Psychology, 14(3), 405-419.
*Stark, L. J., Allen, K. D., Hurst, M., Nash, D. A., Rigney, B., & Stokes, T. F. (1989). Distraction: Its utilization and efficacy with children undergoing dental treatment. Journal of Applied Behavior Analysis, 22(3), 297-307. [Content Link]
*Vessey, J. A., Carlson, K. L., & McGill, J. (1994). Use of distraction with children during an acute pain experience. Nursing Research, 43, 369-372. [Content Link]
*Winborn, M. D. (1987). Associative and dissociative preparatory strategies for children undergoing dental treatment. Unpublished doctoral dissertation, Memphis State University.
Young, M. R., & Fu, V. R. (1988). Influence of play and temperament on the young child's response to pain. Children's Health Care, 18, 209-217. [Content Link]
Wall, P. D. (1978). The gate control theory of pain mechanisms. A re-examination and re-statement. Brain, 101, 1-18. [Content Link]
Whitener, E. (1990). Confusion of confidence intervals and credibility intervals in meta-analysis. Journal of Applied Psychology, 75, 315-321. [Content Link]
Willis, W. E., & Coggeshall, R. E. (Eds.). (1991). Sensory mechanisms of the spinal cord (2nd ed.). New York: Plenum Press. [Content Link]
*Zabin, M. A. (1982). The modification of children's behavior during blood work procedures. Unpublished doctoral dissertation, West Virginia University.
Key words: children; distraction; distress; pain
Clinical Outcomes of Educating Nurses About Pediatric Pain
OutlineThis pilot study examined the influence of a mandatory hospital in-service pediatric pain management program on nurses' administration of analgesics. Chart audits were conducted using a convenience sample of all children who had undergone tonsillectomies during a 2-week period before and after the in-service. In contrast to what was expected, after the educational program, there was an increased length of time before the first dose of analgesic was given and between doses of analgesics given to patients.
Although knowledge about assessment of children's pain and management of pediatric pain has increased,1-3 studies suggest that children still are seriously undertreated for pain.4,5 Research has demonstrated consistently that children receive fewer and less frequent doses of opioids postoperatively and fewer analgesic medications than adults after the same type of surgery.6-10 Undermedication results in detrimental physiological responses to pain.11-13
Schechter 8 attributed undertreatment of children in pain to three factors: incorrect assumptions, prevailing attitudes, and the complexity of pain assessment. Pharmacologic interventions for pediatric pain relief have been hampered by incorrect beliefs about the effects of analgesics on children and the risks involved.14 Misconceptions may influence nurses in their decisions not to medicate pediatric patients.15,16
Burokas 7 found that what influenced nurses the most in medicating pediatric patients after surgery was the nurses' goals of pain relief. Only 16% of the nurses reported complete pain relief as their goal, with 61% hoping to relieve as much pain as possible, and 23% intending to relieve enough pain to render the patient functional.
Purpose
Adequate pain management has been lacking in our health care system. Educational programs on children's pain have demonstrated some success in increasing nurses' knowledge, assessment, and management of children's pain.17,18 The purpose of this pilot study was to examine the outcomes of a pediatric pain measurement program on nurses' practice.
Methods
A convenience sample of all pediatric nurses (n = 52) employed in a 75-bed children's hospital attended a mandatory pain management in-service program. All 52 nurses completed a questionnaire adapted with permission from Burokas 7 that consisted of multiple-choice questions relating to nurses' goals of pain relief, comfort in administering analgesic medication, and demographic items. Burokas 7 did not report reliability. Content validity of the revised questionnaire was established for this pilot study by two expert pediatric nurses. Questionnaires were completed and returned anonymously as a group before the start of the in-service program. Chart review was reported by code number to maintain confidentiality.
Objectives of the 3-hour in-service program were to assist the participant in understanding myths and misconceptions about pediatric pain management, increase knowledge of medications used in the treatment of pediatric pain, and inform the participants of pediatric pain assessment tools and documentation. The Agency for Health Care Policy and Research (AHCPR) quick reference guide for clinicians on acute pain management 1 was used as the basis for the content presented. Participants received individual copies of the guide for personal use.
The content of the in-service program included history and research on children's pain management, myths and misconceptions, assessment tools, nonpharmacologic and pharmacologic interventions, benefits of pain relief, ethical considerations, and parental involvement. Additionally, a pediatric board-certified physician discussed concerns about underusage of pain analgesia in pediatric patients. In the concluding open forum, participants asked many pertinent questions and analyzed examples of clinical situations and interventions for treating children's pain.
Charts using a convenience sample of all children admitted to the general pediatric unit after tonsillectomy/adenoidectomy (T&A) were audited during a 2-week period before (n = 22) and after (n = 20) the in-service program.
Results
Background of Nurses
The nurses were female: 2 (3.8%) licensed practical nurses, 18 (34.6%) with associate degrees, 12 (23%) with diplomas, and 20 (38.5%) with baccalaureate preparation. Five (9.6%) nurses had worked in pediatrics less than 2 years, 19 (36.5%) 2 to 5 years, 11 (21%) 6 to 15 years, and 17 (32.7%) more than 15 years. Most (n = 31; 59.6%) worked primarily days. Thirty (57.7%) of the 52 nurses surveyed had never taken a class on pain management.
Most of the nurses reported that their knowledge about pain had come from clinical experience (n = 40; 76.9%), basic nursing education (n = 14; 26.9%), personal pain experiences (n = 13; 25%), or continuing education classes (n = 11; 21.2%). Forty-two (80.8%) nurses reported that they had personally experienced pain, and 20 (38.5%) nurses recounted that their own children had endured severe pain in the past.
Chart audits of 22 post-T&A patients (10 male; 12 female) before the in-service program revealed an age range of 4 to 18 years (mean, 8.9 years). The chart audit of 20 post-T&A patients (10 male; 10 female) after the in-service program revealed an age range of 3 to 15 years (mean, 6.6 years).
Effect of In-service Program on Pain Management
The effect of the in-service program was examined by the amount of time that elapsed before the first dose of pain medication was given to the post-T&A pediatric patients before and after the program occurred. In contrast to what was expected, the elapsed time before the first dose of analgesic was given was shorter before the pain management in-service program than after. The standard of care was to give the first dose of analgesic medication within the first 2 hours of admission to the pediatric unit after the T&A.
Before the in-service program, the first dose of analgesic was received by 30% of the patients within the first hour, by 58% within 2 hours, and by 100% within 5 hours after T&A. After the in-service program, however, the first dose of analgesics was received by 16% of the patients in the first hour, by 35% within 2 hours, and by 100% within 11 hours after T&A. Analysis of variance (ANOVA) indicated significant differences in the time until the first dose of analgesic was given before and after the in-service program (F[1,16] = 32.82; p <= 0.001).
The difference in the amount of time between doses of pain medication given to the post-T&A patients also was examined. The length of time that elapsed between doses of analgesic medication was shorter before the in-service program than after. The mean time between doses before the in-service program was 5.01 (SD = 2.41) hours; afterward, it was 6.30 (SD = 2.70) hours. A t test indicated a significant difference between these findings (p < 0.001). The pre-in-service group (n = 22) received a total of 33 doses of analgesics. The post-in-service group (n = 20) received 12 doses, with some receiving no analgesics at all before discharge.
Discussion
The 3-hour mandatory in-service program on management of children's pain unexpectedly did not increase the amount of analgesics given to children after T&A. It is possible that the information given during the in-service was repetitious or overly familiar for some nurses and new material for other nurses. A needs assessment of the nurses' previous experiences and knowledge might have helped to individualize the in-service program to meet the needs of everyone more fully.
The in-service program included a variety of teaching methods: lecture, media, reference guides, and discussion. Perhaps other interactive teaching methods such as games, simulations, videotapes, and computer programs may be more effective. The in-service program focused on knowledge of pediatric pain management, but the problem may lie in changing nurses' values. Teaching strategies such as values clarification, role play, and simulations to increase awareness of pain may be more effective in altering values associated with assessing and managing children's pain. Because most of the nurses had learned about pediatric pain management on the job, perhaps having nurses work closely on a one-to-one basis with an expert may help them realize why pain management with children is critical.19,20
The patients in the post-in-service group may have had a higher tolerance for pain than those in the pre-in-service group because a convenience sample of charts was used for the audit. The small sample size was another limitation. Perhaps the nurses incorporated in their practice more of the nonpharmacologic pain-relieving measures learned during the in-service program. However, this was not assessed as part of the study.
Removing biases and examining myths and misconceptions about pain may need multiple and frequent reinforcement and positive mentoring before a change in practice is seen. Perhaps more time should have elapsed between the educational program and post-in-service chart audit to give the nurses more time to incorporate new knowledge into their practice.
Furthermore, the highest level of educational preparation was the baccalaureate degree. Margolius et al.15 reported that master's-prepared nurses demonstrated the greatest insight into children's pain management, whereas nurses with less education who were providing direct patient care had more misconception and were least likely to examine the effectiveness of their interventions in reducing pain.
Another possible reason for the lack of positive results from the in-service program is that nurses may have believed that they were already meeting their patient's pain needs and did not need further education on the topic. Suggesting a change in practice may have been interpreted by the nurses as criticism of their care for patients in the past.
Further research on management of pediatric pain is indicated.21 Studies of nurses' attitudes and values regarding pediatric pain management, surveys of patient satisfaction with pain management, and studies of contextual factors influencing the management of children's pain are important areas for future investigations.
References
1. Acute pain management guideline panel. Acute pain management in infants, children, and adolescents: operative and medical procedures. Quick reference guide for clinicians (AHCPR Pub. No. 92-0020). Rockville, MD: Agency for Health Care Policy and Research, Public Health Service, U.S. Department of Health and Human Services, 1993. [Content Link]
2. Maikler VE. Pharmacologic pain management in children: a review of intervention research. J Pediatr Nurs 1998;13(1):3-14. [Content Link]
3. Savedra M, Eland JM, Tesler M. Pain management. In: Craft MJ, Denehy JA, eds. Nursing interventions for infants and children. Philadelphia: WB Saunders, 1990, 304-325. [Content Link]
4. Foster R, Hester N. The relationship between pain ratings and pharmacologic interventions for children in pain. Adv Pain Res Ther 1990;15:31-36. [Content Link]
5. Sutters KA, Miaskowski C. Inadequate pain management and associated morbidity in children at home after tonsillectomy. J Pediatr Nurs 1997;12(3):178-185. [Content Link]
6. Beyer J, Aradine C. Content validity of an instrument to measure younger children's perception of the intensity of their pain. J Pediatr Nurs 1986;1(6):386-395. [Content Link]
7. Burokas L. Factors affecting nurse's decisions to medicate pediatric patients after surgery. Heart Lung 1985;14(4):373-379. [Content Link]
8. Eland J, Anderson J. The experience of pain in children. In: Jacox A, ed. Pain: a sourcebook for nurses and other health professionals. Boston: Little & Brown, 1977:453-476. [Content Link]
9. Schechter N, Allen D, Hanson K. Status of pediatric pain control: a comparison of hospital analgesic usage in children and adults. Pediatrics 1986;77(1):11-15. [Content Link]
10. Schechter N. The undertreatment of pain in children. Pediatr Clin North Am 1989;36:781-794. [Content Link]
11. Tesler MD, Holzemer WL, Savedra MC. Pain behaviors: postsurgical responses of children and adolescents. J Pediatr Nurs 1998;13(1):41-47. [Content Link]
12. Jaros J. The concept of pain. Crit Care Nurs Clin North Am 1991;3:1-10. [Content Link]
13. Sutters KA, Levine JD, Dibble SA, Savedra M, Miaskowski C. Analgesic efficacy and safety of single-dose intramuscular ketorolac for postoperative pain management in children following tonsillectomy. Pain 1995;61:145-153. [Content Link]
14. Eland J. Pain in children. Nurs Clin North Am 1990;25(4):871-883. [Content Link]
15. Margolius FR, Hudson KA, Michel Y. Beliefs and perceptions about children in pain: a survey. Pediatr Nurs 1995;21(2):111-115. [Content Link]
16. Mooney N. Pain management in the orthopedic patient. Nurs Clin North Am 1991;26(1):73-85. [Content Link]
17. O'Brien S, Konsler G. Alleviating children's postoperative pain. Matern Child Nurs J 1988;13:183-186. [Content Link]
18. Page G, Halvorson M. Pediatric nurses: the assessment and control of pain in preverbal infants. J Pediatr Nurs 1991;6(2):99-106. [Content Link]
19. Stamler LL, Thomas B. Patient stories: a way to enhance continuing education. J Contin Educ Nurs 1997;28(2):64-68. [Content Link]
20. McCaffery M, Ferrell BR. Influence of professional vs. personal role on pain assessment and use of opioids. J Cont Educ Nurs 1997;28:(2)69-77. [Content Link]
21. Beyer B, Clegg J, Foster R, Hester N. Clinical judgment in managing the crisis of children's pain. In Mosby Yearbook (Ed.) Development Diff Pain Manag 1992:295-297. [Content Link]
