Journal of Pediatric Psychology Advance Access published online on March 28, 2008
Journal of Pediatric Psychology, doi:10.1093/jpepsy/jsn030
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
The Impact of Adult Behaviors and Vocalizations on Infant Distress during Immunizations
1University of Georgia, 2Duke Children's Primary Care, 3Children's Hospital Boston, and 4University of North Carolina Chapel Hill
All correspondence concerning this article should be addressed to Ronald L. Blount, PhD, Department of Psychology, University of Georgia, Athens, GA 30602-3013, USA. E-mail: rlblount{at}uga.edu
 |
Abstract |
|---|
 Top Abstract MethodResultsDiscussionAcknowledgmentsReferences |
|---|
Objective The Child–Adult Medical Procedure Interaction Scale-Infant Version (CAMPIS-IV) was used to examine the influence of adult and infant behaviors on infant distress following injections. Methods In this naturalistic observation study, videotaped interactions of 49 infants, parents, and nurses were coded using the CAMPIS-IV. A series of three lag sequential analyses were used to examine the immediate and delayed effects of each of the CAMPIS-IV criterion behaviors, as well as the effects of the onset of each behavior, on infant distress. Results Strong support was found for beneficial effects of the infants playing with an object and sucking, and for adults’ belly-to-belly contact and nonprocedural talk to infant. Some benefit was found for bouncing, patting, and rocking the infant. Apologizing, empathizing, and reassuring the infant received no support, with some indication of detrimental effects. Conclusions The CAMPIS-IV was useful for identifying modifiable risk and protective factors for infants undergoing injections.
Key words: adult–infant interactions; immunizations; infant distress; procedural pain.
The recommended immunization schedule for healthy infants from the Centers for Disease Control (CDC) includes 24 vaccination doses administered through multiple injections, in addition to two flu injections, all by 15 months of age (CDC, 2007
). Although common and valued as critical components to preventative care, immunization injections are painful procedures. It is imperative to identify factors that impact distress during immunizations and develop strategies to effectively manage infant discomfort.
Strategies to manage infant pain have focused on five areas (for review, see Anand, 2001), including: environmental (Franck & Lawhon, 2000; Stevens, Gibbins, & Franck, 2000); behavioral positioning/skin-to-skin contact (Johnston et al., 2003); sensory stimulation (Felt et al., 2000); oral stimulation (Stevens, Taddio, Ohlsson, & Einarson, 1997); and parent or nurse training (Cohen, 2002). Environmental interventions include reducing noxious environmental stimuli, such as bright lights, loud noises, frequent handling, and repeated painful procedures (Franck & Lawhon, 2000; Stevens et al., 2000). Positioning infants and using body-to-body contact to reduce distress have shown that skin-to-skin contact or "kangaroo care" and swaddling are effective strategies for reducing infant distress (Johnston et al., 2003). Sensory stimulation research has focused on using nonpainful stimulation, such as massaging or rocking the infant, to reduce distress (Felt et al., 2000). The effectiveness of oral stimulation, particularly sucking sweet substances such as sucrose and nonnutritive sucking using a pacifier, has been supported in a number of investigations (Carbajal, Chauvet, Couderc, & Olivier-Martin, 1999; Stevens et al., 1997).
Parent and/or nurse training have primarily been done with school-age children, with some researchers adapting strategies found to be effective with older children for use with infants. For example, Cohen (2002) trained nurses to provide distraction in the form of prompting infants’ attention toward a movie or toy, resulting in less distress relative to infants who did not receive the distraction intervention. Another study evaluated the effects of training nurses and parents in distraction techniques versus a typical care condition, and results showed that infants receiving the intervention displayed fewer distress behaviors (Cohen et al., 2006). These studies support the effectiveness of distraction with infants.
However, other investigators evaluated training parents using distraction versus a supportive intervention (in which parents were reminded to utilize coping strategies that they have found helpful in the past) versus a typical care condition, and found results inconsistent with prior distraction research (Cramer-Berness & Friedman, 2005). These findings indicated that infants in the supportive condition demonstrated significantly lower distress during the postinjection recovery phase compared to infants in the typical care condition, but no significant differences were found between the distraction condition and the supportive or typical care conditions. Interpretation of these results is limited because the distraction condition included distraction strategies as well as other strategies, such as oral and sensory stimulation, making it impossible to examine the effects of distraction per se. Additionally, the study focused only on parent behavior rather than also including nurse behavior. In spite of these limitations, this investigation did suggest that some untrained and, thus far, unspecified parent behaviors in the supportive care condition may be beneficial for reducing infant distress.
Parent behaviors that reduce distress have been identified for preschool and school-age children using correlational and experimental research (Blount et al., 1989; Blount, Powers, Cotter, Swan, & Free, 1994; Cohen, Blount, & Panopoulos, 1997; Manimala, Blount, & Cohen, 2000). These studies were conducted using the Child–Adult Medical Procedure Interaction Scale (CAMPIS; Blount et al., 1989) and the revised version of the scale (CAMPIS-R; Blount, Sturges, & Powers, 1990; Blount et al., 1997). The CAMPIS/CAMPIS-R uniquely examines children's coping and distress, as well as the behaviors of parents and medical staff that impact distress. Several adult behaviors were found to promote children's coping and reduce distress. These coping promoting behaviors include distraction, humor, nonprocedural talk, and coaching children to use coping strategies (Blount et al., 1989, 1990). On the other hand, several adult behaviors were associated with child distress. These distress promoting behaviors, included reassuring comments, ("you’ll be okay baby"), empathic comments, ("I know this is hard"), apologies, ("I’m sorry you have to go through this"), giving control to the child, ("tell me when you’re ready [to receive the injection]"), and criticizing (Blount et al., 1989).
A few studies have utilized the CAMPIS-R to examine how adult behaviors impact infants’ distress during immunizations (Bustos, Jaaniste, Salmon, & Champion, in press; Piira, Champion, Bustos, Donnelly, & Lui, 2007; Sweet & McGrath, 1998). Sweet and McGrath (1998) found that mothers’ distress promoting behaviors predicted higher child pain, while staffs’ distracting or coping promoting behaviors predicted lower child pain. Piira et al. (2007) and Bustos et al. (in press) showed that adults’ coping-promoting statements can reduce infant distress. These investigations indicate that parents and nursing staff are critical players who influence infant coping and distress during medical procedures. However, one limitation to this research is that the CAMPIS-R was designed for use with school-age children and does not include behaviors specific to infants, such as sucking a pacifier or being positioned in particular ways.
Two observational scales have been used to examine the effects of mother–infant interactions following immunization injections. These investigations included infants in the range of 2–6 months of age. Lewis and Ramsay (1999) investigated the effects of 23 maternal behaviors presumed to be soothing to the infant during a 30 s period following the injection. They found no significant correlations between the maternal behaviors and infant distress, as assessed by cortisol levels or level of crying. Jahromi, Putnam, and Stifter (2004) evaluated the influence of eight maternal behaviors on infant crying during a 4 min period immediately following the immunization. Only one verbal code was used, thus not distinguishing the influence of the type of maternal vocalization. They found that the combination of maternal holding/rocking and vocalization, as well as feeding/pacifying, were associated with decreased infant crying. Though informative, these investigations were conducted only with young infants (2–6 months) and did not include the nurses in postprocedural interactions.
Only one instrument was found that was designed to examine behaviors exhibited by nurses, parents, and infants during infants’ medical procedures—the Measure of Adult and Infant Soothing and Distress (MAISD; Cohen, Bernard, McClellan, & MacLaren, 2005). The MAISD is a promising observational scale, with fair to excellent inter-rater reliability for discrete behaviors. However, this scale was not comprehensive in identifying adult behaviors that influence infant distress. For example, this scale measures verbal reassurance, but does not specifically examine the types of remarks (such as empathy vs. reassurance vs. talk about things unrelated to the procedure), which have been identified as important in previous research with infants and preschoolers (Manimala et al., 2000; Piira et al., 2007). Additionally, the MAISD did not examine specific body positions, such as holding the infant belly-to-belly or in a kangaroo care position, which are associated with lower distress (Johnston et al., 2003).
The primary goal of this study was to use the CAMPIS-IV (Infant Version) to examine the effects of discrete codes that, based on prior research and clinical observations, reflect an influential set of parent and medical staff behaviors that are specific to infants undergoing medical procedures. The CAMPIS-IV code categories are a downward extension of the CAMPIS (Blount et al., 1989), a valid and reliable behavioral observation scale that has identified specific coping promoting and distress promoting behaviors for school-aged children undergoing medical procedures. This study investigated the temporally contiguous relationships among specific adult and infant behaviors during immunizations using the CAMPIS-IV, with the goals of identifying behaviors that are associated with higher or lower levels of infant distress during immunizations.
The postinjection recovery phase was selected for study because the parent and nurse are more able to engage in various verbal and nonverbal behaviors that may be soothing to the infant without the restricting demands of positioning the infant in particular ways to conduct the injections or communicating primarily about the procedure. We expected that specific behaviors that were consistent with the coping promoting and coping codes of the CAMPIS, as well as strategies found to be helpful in the infant intervention literature, would be associated with less infant distress. Similarly, we expected that codes that were consistent with the distress promoting codes of the CAMPIS would be associated with greater child distress.
|
Method |
|---|
|
TopAbstractMethodResultsDiscussionAcknowledgmentsReferences |
|---|
Subjects and Setting
This study was conducted at two county health departments located in two suburban cities. Data collection spanned a 6-month period of time, with a total of seven nurses conducting the injections. Inclusion criteria were that the participants spoke English, the infants were healthy and between 2 and 20 months of age, and that they were brought to the health department for routine immunizations. Four potential participants declined due to schedule demands. No nurses declined to participate. Participants included 49 infants, 26 boys, and 23 girls, who ranged in age from 1.6 to 20.6 months (M = 8.7 months; SD = 6.1 months). Of the 49 infants, 26 (53%) were African American, 19 (39%) were Caucasian, 1 (2%) was Hispanic, and 3 (6%) did not endorse an ethnicity. Infants were accompanied by their mothers (n = 47) or mothers and fathers (n = 2). The mean age of the parents was 25.8 years (SD = 5.1) for mothers and 27.0 years (SD = 12.7) for fathers. The mean years of education was 12.2 (SD = 2.4) for mothers and 16.0 years (SD = 5.7) for fathers.
Measures
Intake Form
Parents completed a brief inventory to obtain demographic information for the family, including years of education for each parent, prenatal care, and child medical history.
CAMPIS-IV
This scale is a downward extension of the CAMPIS/CAMPIS-R (Blount et al., 1989, 1990) to include codes reflective of infant rather than child behavior, as well as adults’ behaviors when interacting with them. Given the similarity of behaviors displayed by parents and nurses in this investigation, as well as in prior research (Blount et al., 1989), nurse and parent behaviors were coded as adult behaviors rather than examined separately. A 5 s interval recording system was used during which all behavior codes were rated as occurring or not occurring. The coding system includes five adult motoric behaviors or position codes, seven adult vocal behaviors, and two infant behaviors (Table I). The adult motoric behaviors include Bounce, Rock, Pat, Stroke, and Belly-to-Belly, each of which describe typical behaviors that adults may perform when holding infants, or, in the case of Belly-to-Belly, a position in which infants may be held. These motoric codes were considered to be soothing and distracting, serving as coping promoting behaviors. Three vocal behaviors, apologizing, empathizing, and reassuring, are downward extensions of the higher-order distress-promoting behaviors from the CAMPIS-R that have been shown in correlational and experimental research to be associated with higher child distress (Blount et al., 1989; Manimala et al., 2000). Procedural Talk to Adult, Procedural Talk to Child, and Non-procedural Talk to Adult are downward extensions of the Adult Neutral category that was found in research with the CAMPIS to have minimal effect on child distress. Nonprocedural Talk to Infant is a downward extension of a similar behavioral code from the higher-order coping promoting category on the CAMPIS-R.
|
The two infant codes, Object and Sucking, were considered to be coping behaviors. Object refers to the infant playing with or exploring an object, indicative of attentional redirection or distraction. Sucking indicates that the infant was sucking a bottle or pacifier. Sucking may be distracting from pain and discomfort, as well as comforting and soothing.
Cry
A variation of the cry code from the Modified Behavioral Pain Scale (MBPS; Taddio, Nulman, Koren, Stevens, & Koren, 1995) in infants was used. Cry was scored on a 0–3 scale, with 0 = not crying, 1 = moan/whine/fuss/whimper, 2= full crying, and 3 = screaming or vigorously crying. For the analyses, crying was collapsed into a high (cry code 2 and 3) and low (cry code 0 and 1) cry categorization. Collapsing the cry codes was done for purposes of data reduction and to increase statistical power. Further, the full crying and vigorous crying codes seem to clearly be at the high intensity end of crying, whereas moan/fuss/whimper seems at the very low end. We were much more interested in differentiations between high and low distress rather than finer gradations between similar levels of cry on the MBPS ordinal scale.
Transcription, Coding, and Inter-observer Reliability
Videotaped recordings of the immunization sessions were transcribed to provide a record of the speaker and the content of the vocalization. The transcription of each tape was independently checked by a minimum of three transcribers. Since this study focused on infants’ recovery following injection(s), coding of behaviors began with the removal of the injection needle and continued until the parent and infant left the treatment room or 2 min passed, whichever came first. An interval coding system was used, with trained observers recording the occurrence or nonoccurrence of each of the infant and adult (i.e., parent or nurse) CAMPIS-IV codes and the level of cry during consecutive 5 s intervals of the postimmunization phase. Observers used the videotape, replaying as needed, and the transcript when coding. A second observer independently coded each videotape for each participant.
Kappa reliability coefficients for each of the CAMPIS-IV codes ranged from.77 to 1.0 (M =.92; SD =.06; Table I). The 
reliability coefficient was.90 for four levels of cry, and was.94 for the two level (i.e., high or low) classification of cry. Each of these values represents excellent inter-observer agreement according to guidelines proposed by Fleiss (1981).
Procedures and Data Analyses
All research procedures were approved by a university IRB. After the participant completed the registration for a well-child immunization and was identified as meeting the inclusion criteria, medical staff notified the researchers. Parents were approached in the waiting room prior to their meeting with the nurse, the study was described, and informed consent was obtained. Parents completed the demographic information form in the waiting room. In the treatment room, a stationary videocamera with a remote omnidirectional microphone recorded the interactions among the nurse, parent, and infant. Date and time in minutes and seconds were recorded onto a corner of the videotape. The parent was present with the infant throughout all examinations and immunizations. Infants typically were weighted and examined prior to moving to a chair for the immunization. Infants were usually held by the parent, facing out, and positioned as needed for the immunization injections. After the last injection, the nurse placed an adhesive bandage, the mother and nurse interacted with the infant, the nurse and mother talked about reactions to watch for post-immunization, parent's questions were answered, and the family left the clinic.
The analyses followed the logic of sequential analysis (Bakeman & Gottman, 1997), examining the effects of each criterion behavior (i.e., the CAMPIS-IV behavioral code categories) on the probability or proportion of infants’ level of cry in a series of three analyses. In all analyses, the main dependent variable of interest was infant cry, measured as either low or high cry within each interval. Proportions are equivalent to probability in these analyses. For example, suppose that nonprocedural talk occurs in three intervals for a particular infant. In two of those intervals the infant had a low level of cry and in one there was a high level of cry. The conditional probability of low cry for that infant, given the criterion behavior of nonprocedural talk occurring in the same 5 s interval, would be.67.
From zero to multiple instances of any particular CAMPIS-IV behavior code might occur for each participant. This reflects the flow of interaction for the untrained participants in this naturalistic observation research. The proportion of intervals in which each code occurred is presented in Table I. The number of participants with one or more instances of a given CAMPIS-IV behavior is indicated in Tables II–IV in the column labeled N.
|
|
|
The effect of each CAMPIS-IV behavioral category on infant cry was examined in the 5 s interval in which the criterion behavior occurred (Lag 0—immediate effects) and the 5 s interval following the interval in which the criterion behavior occurred (Lag 1—delayed effects). In the third set of analyses, the probability of high levels of cry in the 5 s interval prior to the onset of each criterion behavior was compared to the probability of high cry in the 5 s interval following the onset of the behavior (Lag –1 and Lag 1—pre- to post-effects of onset). For example, if a particular CAMPIS-IV behavioral code occurred in several consecutive intervals, only the first interval during which it occurred was considered. However, a participant may have two or more onsets of a behavioral code, provided there were intervals during which the criterion behavior did not occur separating intervals of occurrence. This third set of analyses allowed for an examination of the context preceding the initiation of particular CAMPIS-IV behaviors; that is, whether the CAMPIS-IV code was used most often when the infant exhibited low or high distress, as well as their effect on infant distress from before to after their occurrence. For each set of analyses, if two or more CAMPIS-IV behaviors occurred in the same interval, the associations between those behaviors and infant cry were examined separately for each CAMPIS-IV code.
In the first two sets of analyses, nonparametric Wilcoxon signed-rank tests (indicated by the statistic Z) were used because there was dependency built into the data that were being compared (low cry probability = 1 minus high cry probability, for each infant). Paired samples t-tests were used for the third set of analyses, for which dependency was not an issue. For both statistical procedures two-tailed tests of significance were used. Effect sizes were indicated and may range from 0 (no effect) to ±1.0 (a perfect effect), with generally accepted guidelines indicating that.1 is indicative of a small effect.3 a medium effect, and.5 a large effect (Cohen, 1988, 1992). Only participants for whom there were instances of a particular CAMPIS-IV code were included in analyses to examine the effects of that code.
|
Results |
|---|
|
TopAbstractMethodResultsDiscussionAcknowledgmentsReferences |
|---|
Preliminary Analyses of CAMPIS-IV Codes and Cry by Age of Infant
Based on the bimodal distribution of infant age in our sample, we divided the sample into two age groups: younger (<10 months; n = 28) and older (>10 months; n = 21). Mann–Whitney tests were conducted to examine differences in the mean proportion of time each CAMPIS-IV code occurred depending on the age of the infant. Results indicated significant differences between the two age groups for Non-Procedural Talk to Infant (Younger M =.06, SD =.09, Median =.00; Older M =.35, SD =.24, Median =.30, U = 76.50, p <.001, r = –.65), Object (Younger M =.03, SD =.12, Median =.00; Older M =.33, SD =.25, Median =.40, U = 75.50, p <.001, r = –.72), and Pat (Younger M =.25, SD =.27, Median =.14; Older M =.12, SD =.18, Median =.00, U = 197.50, p <.05, r =.29). Thus, nonprocedural talk to infant and engaging with an object were more frequently used with older infants, while patting was more frequently used with younger infants. There were no significant differences between age groups for the remaining 11 CAMPIS-IV codes. In addition, there were no significant differences in cry between the two age groups. Given these results, the remaining analyses will use the entire sample rather than age groups to maximize power to examine the relationship between each CAMPIS-IV code and infant distress.
Immediate Effects Associated with Criterion Behaviors: Lag 0
The results for these analyses are displayed in Table II. The CAMPIS-IV behaviors of Sucking, Object, Belly-to-Belly, Nonprocedural Talk to Adult, Nonprocedural Talk to Infant, Procedural Talk to Adult, and Procedural Talk to Infant were each associated with a significantly higher probability of low rather than high levels of cry during the intervals in which they occurred. The probabilities of low cry were.98.92.65.75.87.72, and.71, respectively, for each behavior. The behaviors of Sucking, Object, Nonprocedural Talk to Adult, Nonprocedural Talk to Infant, and Procedural Talk to Infant were associated with a.50 or greater difference between the mean probability of low cry rather than high cry occurring in conjunction with those behaviors. There was a trend at p 
.10 for more low than high levels of cry during intervals in which Stroke occurred (probability of low cry =.69). There was also a trend at p .10 for more high than low levels of cry during intervals in which reassurance occurred (probability of low cry =.42). Apologizing, Empathizing, Bounce, Rock, and Pat were not significantly associated with the level of cry. Effect sizes for significant tests were medium to large.
Delayed Effects Associated with Criterion Behaviors: Lag 1
The results for these analyses are displayed in Table III. The CAMPIS-IV behaviors of Sucking, Object, Belly-to-Belly, Stroke, Pat, Bounce, Nonprocedural Talk to Adult, Nonprocedural Talk to Infant, and Procedural Talk to Adult each had a significantly higher probability of low cry versus high cry in the 5 s interval immediately following their occurrences. These probabilities of low cry ranged from.70 to.95, with Sucking and Object showing the highest probabilities (.92 and.95, respectively). The behaviors of Sucking, Object, Nonprocedural Talk to Adult, and Nonprocedural Talk to Infant were associated with greater than.50 differences in the likelihood of low rather than high cry in the interval following their occurrence. Differences approached significance for Rock being associated with a higher probability of low than high cry (p =.055). Apologizing, Empathizing, Reassuring, and Procedural Talk to Infant were not associated with a significant difference in the probability of high versus low cry during the intervals immediately following their occurrence. Effect sizes for significant tests were medium to large.
Pre- to Post-effects of Onset of each Criterion Behavior on Low Cry: Lag –1 to Lag 1
For this set of analyses, the probability of high cry in the interval immediately prior to the onset of the criterion behavior was compared to the probability of high cry in the interval immediately following the criterion. As seen in Table IV, Sucking, Object, Belly-to-Belly, Pat, Rock, Bounce, and Nonprocedural Talk to Infant were found to have significantly higher mean probabilities of high cry at lag –1, preceding the onset of these behaviors, when compared to lag 1, after their occurrence. The behaviors of Sucking and Belly-to-Belly were the only CAMPIS-IV behavioral codes associated with a.50 or greater decrease in the levels of high cry from lag –1 to lag 1. There was a change in the probability of high cry from.77 prior to the initiation of Sucking to.10 after its initiation. For Belly-to-Belly, the probability of high cry changed from.88 to.35 following the initiation of Belly-to-Belly contact. Apologizing, Empathizing, Reassuring, Procedural Talk to Infant or to Adult, and Stroke were not associated with a significant reduction in cry from lag –1 to lag 1. Nonprocedural Talk to Adult was associated with an increase in cry from lag –1 to lag 1. Note that Procedural Talk to Adult and Nonprocedural Talk to Adult were initiated most often following an interval of low crying, as indicated by conditional probabilities for high cry at Lag –1 of.21 and.17, respectively. Effect sizes for significant tests were medium to large. A summary of the results from the three series of analyses are displayed in Table V.
|
|
Discussion |
|---|
|
TopAbstractMethodResultsDiscussionAcknowledgmentsReferences |
|---|
In this naturalistic observation study, the effects of parent, nurse, and infant behaviors which typically occur during infants’ recovery from painful injections were assessed. None of the parents or nurses had been trained in how to assist the infants. A series of three analyses was used to assess immediate, delayed, and pre- to post-onset effects for each criterion behavior. The combination of these analyses should be seen as providing corroborative evidence for the effects of the CAMPIS-IV behavioral codes on distress. Results which were consistent across the three sets of analyses should be viewed as providing the greatest support for the likely function of those CAMPIS-IV behaviors. Findings revealed that several behaviors were consistently associated with low rather than high levels of infant crying, while some behaviors might have beneficial effects, and other behaviors were not helpful. The behaviors that were associated with clear benefit included Sucking, Belly-to-Belly, Object, and Nonprocedural Talk to the Infant. Of these four beneficial behaviors, only Nonprocedural Talk to the Infant is exclusively verbal. The other three involve use of a pacifier or bottle, play with an object, or positioning and physical comfort. These behaviors may or may not occur in conjunction with talking with the infants.
Sucking showed strong support across all three analyses, with differences between the probabilities of low versus high cry, as well as decreases in high cry from before to after the onset of sucking, being greater than.50. Sucking a pacifier or bottle is appealing to infants, soothing, and, if performed in full measure, is incompatible with crying. Sucking was used during times of high distress and produced rapid changes to low levels of cry, as indicated by the onset analysis. These results coincide with the literature that suggests sucking has an analgesic effect for infants during minor invasive procedures (Carbajal et al., 1999). One potential cautionary note is that sucking was also a low frequency behavior, occurring for only 10 of the 49 infants. Low frequency may contribute to a degree of instability for these results.
Belly-to-Belly, also an untrained behavior, is similar to kangaroo care, which has been shown in experimental investigation to be helpful for reducing infant distress (Johnston et al., 2003). Whereas kangaroo care maximizes parent–infant skin-to-skin contact, and was originally used in neonatal intensive care units, Belly-to-Belly contact describes a clothed parent holding a clothed or partially clothed infant close to their body with the infant facing toward the parent. Belly-to-Belly often occurred following intervals of high cry. It was consistently associated with lower levels of cry and evidenced rapid reduction in cry following its onset. This suggests that belly-to-belly contact, along with sucking a bottle or pacifier, may be two easily applied behavioral interventions of choice for rapidly reducing high levels of infant distress.
The behaviors of Nonprocedural Talk to the Infant and infants’ playing with Object are examples of distraction or attentional redirection. Both of these behaviors were associated with a.50 or greater difference in the mean probability of low rather than high cry in the same interval they occurred and in the interval immediately after their occurrence. Nonprocedural Talk to Infant and Object were used more often with older infants, although they could be initiated more often with younger infants as well. These beneficial behaviors may be facilitated by both parents and nurses, who may have attractive objects available during routine painful medical treatments for infants. Playing with an object and nonprocedural talk to the infant were initiated most often following intervals that were, on the average, moderate in the level of cry (mean probability of high cry =.47 to.50). They resulted in reduction in distress (Table IV). In combination with the findings for the behaviors of Suck and Belly-to-Belly, these findings suggest the possibility of a two tier intervention program, with Suck and Belly-to-Belly being most applicable for high infant distress situations, and the distraction-based interventions of Object and Nonprocedural Talk being most appropriate for times of less intense infant distress.
There were additional behaviors that were associated with a greater likelihood of low than high levels of cry, or a decrease in high levels of cry, in one or two of the three analyses. Those behaviors for which beneficial effects were expected included Bounce, Rock, Pat, and Stroke. These are behaviors that parents primarily perform when infants become distressed. Each of these behaviors provides physical stimulation and proprioceptive feedback to infants and may serve to distract the child. The findings from this investigation suggest that these behaviors may be thought of as probably beneficial, but not as likely to result in benefit as Suck, Object, Belly-to-Belly, or Nonprocedural Talk. Of these four behaviors, Stroke has the least support. It was supported only by a trend in the first set of analyses and occurred more often following intervals of low cry rather than as an effort to reduce high levels of infant distress. In contrast, the behaviors of Bounce, Rock, and Pat appeared to be attempts to soothe the infant. These three behaviors most often followed intervals of high cry.
There were also three behaviors that on first impression appeared to be beneficial though they were not predicted to be so. These included Nonprocedural Talk to Adult, Procedural Talk to Adult, and Procedural Talk to Infant. It is interesting to note that none of these three behaviors were found to have a beneficial effect in the third set of analyses that examined the effect of the onset of these behaviors, and Nonprocedural Talk to Adult was associated with an increase in high cry following its onset. Further, Nonprocedural and Procedural Talk to Adult occurred almost exclusively following intervals of low rather than high infant cry. Therefore, the apparently beneficial results for these two codes from the first two sets of analyses seem to be more influenced by the context in which they occurred, rather than the immediate or delayed consequences of their occurrence. The increase in cry in the third set of analyses for Nonprocedural Talk to Adult may have been due to a floor effect, in which cry was more likely to increase than decrease, or a lack of attending to the infant at that time, which could lead to a decrease in soothing behaviors. Procedural Talk to the Infant also showed apparently beneficial effects in the first analysis, but unlike Procedural or Nonprocedural Talk to Adults, the onset of this behavior occurred following intervals that had an equal likelihood of high versus low cry. It is possible that the nurses or parents talked to the child in animated and playful ways that may have had some beneficial effect, for example, by offering colorful bandages for the injection site.
In contrast to the clearly helpful behaviors, three adult verbal behaviors showed no indications that were helpful. Apologizing, Empathizing, and Reassuring appeared to do little to relieve infants’ distress in the interval in which they occurred, following their occurrence, or from before to after their onset. In the first set of analyses, there was a trend indicating that reassuring the infant was associated with more high than low distress during the interval during which it occurred. These findings suggest that, at best, these behaviors are not helpful. Reassuring and empathizing occurred for most of the participants in this investigation, with reassurance occurring for 46 and empathizing for 33 of the 49 participants. These findings are very similar to those from the original CAMPIS study by Blount et al. (1989) that found Reassurance was the most frequent behavior directed toward children, and that Reassurance, Empathy, and Apology were associated with a greater probability of child distress. Similarly, Sweet and McGrath (1998) found that the adult codes that were together considered to be distress promoting on the CAMPIS-R (reassuring comments, empathic comments, apologies, giving control to the child, and criticism) were positively correlated with infants’ distress during immunization procedures. The present investigation expands those findings with infants to examining individual behaviors (e.g., Reassurance vs. the composite of Distress promoting) and their associations with levels of infant distress in close temporal proximity to their occurrence.
In sum, the CAMPIS-IV indicated that there were a number of adult behaviors that seemed to lead to reductions in infant's crying following immunizations, as well as some that seemed to be ineffective. Unlike prior investigations of infant procedural pain using the CAMPIS-R (Piira et al., 2007; Sweet & McGrath, 1998), CAMPIS-IV codes were designed to reflect infant behaviors and adult–infant interactions. Also, the CAMPIS-IV provides several unique codes relative to other measures, such as the MAISD, which proved to be valuable for identifying specific adult behaviors and vocalizations that influence distress during immunizations. Also, the CAMPIS-IV was applicable across a wider range of infants’ ages than has been demonstrated in some other investigations using other measures (Jahromi et al., 2004; Lewis & Ramsay, 1999)
In addition to the positive findings of this investigation, there are also limitations. First, the measure of distress, high versus low levels of cry, is not as comprehensive an index as is included in most investigations of infant pain. To capture typical nurse and parent behaviors without being overly intrusive or restrictive in set-up, the camera was placed in a convenient angle that did not always allow views of the children's faces for assessing grimaces or other subtle nonverbal indicators of distress, limiting indications of distress to cry levels. We also note that in this naturalistic investigation, no causal statements can be made regarding whether the CAMPIS-IV criterion behaviors actually cause the outcome of higher or lower distress. Instead, the probabilities of high or low levels of distress are provided for each of the criterion behaviors investigated. Evaluations of causality await experimental investigations. Also, although not a focus of this investigation, it is obvious that infants influence adults, and adults influence each other during the infants immunizations. For example, attempts to soothe or reassure the infant are probably more likely following high rather than low levels of infant distress. Also, it is possible that changes in infant behaviors, as well as adult behaviors, differ as a function of the age of the infant, and whether the adult is the nurse or parent, though measures of this were beyond the scope of the present study. Our assessment of age effects did indicate that Nonprocedural Talk to the Infant and the infant playing with an Object were more often used with older than younger infants. In contrast, Patting was more often used with younger infants. No other significant age differences were found.
An additional issue, given the naturalistic observation nature of this study, is that there were variable numbers of participants engaging in the different CAMPIS-IV behaviors (n = 10–46). Behaviors that were displayed by fewer participants would result in lower power to detect differences. In spite of this, some large effect sizes were found even for some of the less frequently used behaviors. On a related point, there were no Bonferroni corrections for multiple statistical tests. Therefore, some of the results could have occurred by chance. We chose to not use a correction because this is a new approach to this area of research, and because we wished to thoroughly examine the effects of the CAMPIS-IV behaviors in the three series of analyses. We should also note that although the findings from this investigation were consistent with the hypotheses, the generalizability of these findings to other populations should be assessed. This sample was undergoing one type of painful medical treatment and over 50% of the participants were from one ethnic minority group. It is unclear whether similar results would be found with other populations and medical procedures. Another issue relates to the effects of the CAMPIS-IV codes that occur either alone in an interval, in the same interval as other behaviors, or even in behavioral sequences that may traverse intervals. Singular occurrences, co-occurrences, or even chain occurrences of adult behaviors may be associated with different effects on infant distress. This issue reflects the complexity of unstructured dyadic or triatic interactions. To reduce complexity to manageable levels, we intentionally focused on the associations between single CAMPIS-IV codes and infants’ level of cry. Although not comprehensive, this analytic strategy has yielded clinically meaningful findings in prior research with the CAMPIS (Blount et al., 1989), with the validity of the assessment findings being confirmed in later experimental treatment investigations (Blount et al., 1994; Cohen et al., 1997). Finally, although not a limitation of this investigation, we should mention that while this investigation assessed the association between untrained parent and nurse behaviors and infants’ level of crying, it is likely that a greater therapeutic effect could be obtained from training adults and encouraging a greater reliance on the behaviors that seem to alleviate distress.
Future research in this area should continue to assess for parent and nurse behaviors that may be associated with increases and decreases in infants’ distress following medical stressors. The CAMPIS-IV proved to be a reliably scored instrument that may assist in that process. The psychometric properties of the instrument warrant further research, giving the promising findings thus far. In addition, research should be conducted using the CAMPIS-IV during medical procedures infants undergo in order to assess adults’ influence on each other, as well as on the infants. The CAMPIS-IV also might be useful in evaluating developmental trends in strategies to reduce distress. The combination of the CAMPIS-IV and the CAMPIS-R allows for comparisons between adult behaviors found to be effective or ineffective with infants relative to those found to be effective or ineffective with school-aged children. Also, the CAMPIS-IV can be used to provide a manipulation check of adults’ use of interventions for infant pain following their training in coping skills programs. Monitoring adult behaviors following coping skills training is necessary to determine if the desired behaviors were performed during the medical treatment.
Finally, the results of this study suggests that some easily trained behavioral interventions, such as belly-to-belly and sucking a bottle or pacifier, may be most appropriate for assisting infants when they are highly distressed, and other distraction-based interventions (nonprocedural talk to the child, playing with objects) might be most effective for moderate levels of infant distress. The utility of this level of distress-intervention matching approach should be assessed in experimental investigations.
|
Acknowledgments |
|---|
|
TopAbstractMethodResultsDiscussionAcknowledgmentsReferences |
|---|
We wish to thank Drs Roger Bakeman, Richard Marsh, and Tiina Jaaniste for their kind assistance in the planning and data analysis of this investigation, and for their critique of an earlier version of this article. We would also like to thank Barbie Bushey, R. N., Northeast Georgia Immunization Coordinator, and Dr Claude A. Burnett, Director of the Northeast Georgia Health District, as well as members of the Barrow and Clarke County Health Departments, for their facilitation and participation during of this research.
Conflicts of interest: None declared.
Received November 1, 2007; revision received February 29, 2008; accepted March 4, 2008
|
References |
|---|
|
TopAbstractMethodResultsDiscussionAcknowledgmentsReferences |
|---|
Anand K, & International Evidence-Based Group for Neonatal Pain. Consensus statement for the prevention and management of pain in the newborn. Archives of Pediatric and Adolescent Medicine (2001) 155:173–180.
Bakeman R, Gottman J. Observing interaction: An introduction to sequential analysis (1997) 2nd. New York: Cambridge University Press.
Blount RL, Cohen LL, Frank NC, Bachanas PJ, Smith AJ, Manimala MR, et al. The child-adult medical procedure interaction scale-revised: An assessment of validity. Journal of Pediatric Psychology (1997) 22:73–88.
Blount RL, Corbin SM, Sturges JW, Wolfe VV, Prater JM, James LD. The relationship between adults’ behavior and child coping and distress during BMA/LP procedures: A sequential analysis. Behavior Therapy (1989) 20:585–601.[CrossRef][Web of Science]
Blount RL, Powers SW, Cotter MW, Swan SC, Free K. Making the system work: Training pediatric oncology patients to cope and their parents to coach them during/LP procedures. Behavior Modification (1994) 18:6–31.
Blount RL, Sturges JW, Powers SW. Analysis of child and adult behavioral variations by phase of medical procedure. Behavior Therapy (1990) 21:33–48.[Medline]
Bustos T, Jaaniste T, Salmon K, & Champion G. Evaluation of a brief parent intervention teaching coping-promoting behavior for the infant immunization context: a randomized controlled trial. In: Behavior Modification. (in press).
Carbajal R, Chauvet X, Couderc S, Olivier-Martin M. Randomised trial of analgesic effects of sucrose, glucose, and pacifiers in term neonates. British Medical Journal (1999) 319:1393–1397.
Centers for Disease Control and Prevention. Recommended immunization schedules for persons aged 0–18 years—United States. MMWR (2007) 55(51 and 52):Q1–Q4. 2006.
Cohen J. Quantitative methods in psychology: A power primer. Psychological Bulletin (1992) 112:155–159.[CrossRef][Medline]
Cohen J. Statistical power analysis for the behavioral sciences (1988) 2nd. New York: Academic Press.
Cohen L. Reducing infant immunization distress through distraction. Health Psychology (2002) 21:207–211.[CrossRef][Web of Science][Medline]
Cohen LL, Bernard RS, McClellan CB, MacLaren JE. Assessing medical room behavior during infants’ painful procedures: The measure of adult and infant soothing and distress (MAISD). Children's Health Care (2005) 34:81–94.[CrossRef][Web of Science]
Cohen LL, Blount RL, Panopoulos G. Nurse coaching and cartoon distraction: An effective and practical intervention to reduce child, parent, and nurse distress during immunizations. Journal of Pediatric Psychology (1997) 22:355–370.
Cohen LL, MacLaren JE, Fortson BL, Friedman A, DeMore M, Lim CS, et al. Randomized clinical trial of distraction for infant immunization pain. Pain (2006) 125:165–171.[CrossRef][Web of Science][Medline]
Cramer-Berness LJ, Friedman AG. Behavioral interventions for infant immunizations. Children's Health Care (2005) 34:95–111.[CrossRef][Web of Science]
Felt B, Mollen E, Diaz S, Renaud E, Zeglis M, Wheatcroft G, et al. Behavioral interventions to reduce infant distress at immunization. Archives of Pediatric and Adolescent Medicine (2000) 154:719–724.
Fleiss JL. Statistical methods for rates and proportions (1981) New York: Wiley.
Franck L. Environmental and behavioral strategies to prevent and manage neonatal pain. Pain Research and Clinical Management—Anand K J S, Stevens B J, McGrath P J, eds. (2000) 10:203–216. Pain in neonates, 2nd revised and enlarged edition.
Jahromi LB, Putnam SP, Stifter CA. Maternal regulation of infant reactivity from 2 to 6 months. Developmental Psychology (2004) 40:477–487.[CrossRef][Web of Science][Medline]
Johnston C, Stevens B, Pinelli J, Gibbins S, Filion F, Jack A, et al. Kangaroo care is effective in diminishing pain response in preterm neonates. Archives of Pediatric and Adolescent Medicine (2003) 157:1084–1088.
Lewis M, Ramsay DS. Effect of maternal soothing on infant stress response. Child Development (1999) 70:11–20.[CrossRef][Web of Science][Medline]
Manimala M, Blount RL, Cohen LL. The influence of parental reassurance and distraction on children's reactions to an aversive medical procedure. Children's Health Care (2000) 29:161–177.[CrossRef][Web of Science]
Piira T, Champion GD, Bustos T, Donnelly N, Lui K. Factors associated with infant pain response following an immunization injection. Early Human Development (2007) 83:319–326.[CrossRef][Web of Science][Medline]
Stevens B, Gibbins S, Franck L. Treatment of pain in the neonatal intensive care unit. The Pediatric Clinics of North America, Acute Pain in Children (2000) 47:633–650.
Stevens B, Taddio A, Ohlsson A, Einarson T. The efficacy of sucrose for relieving procedural pain in neonates – a systematic review and meta-analysis. Acta Paediatrica (1997) 86:837–842.[Web of Science][Medline]
Sweet S, McGrath P. Relative importance of mothers’ versus medical staffs’ behavior in the prediction of infant immunization pain behavior. Journal of Pediatric Psychology (1998) 23:249–256.
Taddio A, Katz J, Ilersich AL, Koren G. Effect of neonatal circumcision on pain response during subsequent routine vaccination. Lancet (1997) 349:599–604.[CrossRef][Web of Science][Medline]
Taddio A, Nulman I, Koren BS, Stevens B, Koren G. A revised measure of acute pain in infants. Journal of Pain and Symptom Management (1995) 10:456–463.[CrossRef][Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||