close

Вход

Забыли?

вход по аккаунту

?

Development of an observation method for assessing pain behaviors in children with juvenile rheumatoid arthritis.

код для вставкиСкачать
ARTHRITIS & RHEUMATISM
Vol. 38, No. 8, August 1995, pp 1142-1151
0 1995, American College of Rheumatology
1142
DEVELOPMENT OF AN OBSERVATION METHOD FOR
ASSESSING PAIN BEHAVIORS IN CHILDREN WITH
JUVENILE RHEUMATOID ARTHRITIS
THERESA M. JAWORSKI, LAURENCE A. BRADLEY, LOUIS W. HECK,
ALEX ROCA, and GRACIELA S. ALARCON
Objective. To develop an observation method for
assessing pain behaviors in children with juvenile rheumatoid arthritis (JRA).
Methods. Thirty children with JRA performed a
standardized sequence of activities for video recording,
and correlations between the pain behaviors observed
on the videotapes and established measures of pain,
depression, and functional disability were determined.
Results. Pain behaviors were reliably observed
(kappa coefficients 0.53-0.79). Total pain behaviors
were significantly correlated with subjective reports of
pain (r = 0.50) and disability levels (r = 0.64). These
behaviors were not significantly associated with children’s depression ratings (r = 23).
Conclusion. The results indicate that the behavioral observation method provides a reliable and valid
measure of pain associated with JRA. Measurement of
pain behaviors may be especially useful in treatment
outcome studies because these behaviors are relatively
independent of depression.
Most studies of pain in children have used
self-report measures such as visual analog scales
(VAS) and category scales (1-5). Unfortunately, limitations in children’s cognitive development can make
it difficult for them to understand the concept of pain
(6-8) and to respond accurately and reliably to selfPresented in part at the annual meetings of the American
College of Rheumatology, Boston, MA, November 1991, and the
American Pain Society, New Orleans, LA, November 1991.
Supported by NIH grant P60-AR-20614.
Theresa M. Jaworski, PhD, Laurence A. Bradley, PhD,
Louis W. Heck, MD, Alex Roca, MD, Graciela S . Alarch, MD,
MPH: University of Alabama at Birmingham.
Address reprint requests to Theresa M. Jaworski, PhD,
Section of Psychology, Mayo Clinic Foundation, Rochester, MN
55905.
Submitted for publication August 5 , 1994; accepted in
revised form February 7, 1995.
report measures. These measures also may not adequately assess the subjective pain experiences of children (9). The use of measurement methods that rely on
direct observations of behavior that is indicative of
pain resolves some of the difficulties found with selfreports. These behavioral observation methods provide standardized, objective, and quantifiable measures rather than global impressions of subjective
experiences (10-13). They also may be used to provide
information regarding the ability to perform activities
of daily living and to determine the benefits of treatment (10,14).
Behavioral observation methods have been
used to assess chronic pain in adults with rheumatoid
arthritis (RA). McDaniel and colleagues have shown
that direct observations of pain behavior are significantly correlated with patients’ self-reports of subjective pain and functional disability, as well as with
disease activity measures (10). However, unlike selfreport measures of pain, pain behavior is independent
of emotional distress (15) and is sensitive to change
following a relaxation-based intervention designed to
reduce pain (14).
Behavioral observation methods have been
shown to be particularly helpful when used with
children who have pain that is acute in nature (1619).
Among children with chronic pain conditions such as
juvenile rheumatoid arthritis (JRA), however, behavioral observation methods have thus far been used
only to assess functional disability (20,21).
The first purpose of the present study, therefore, was to develop a behavioral observation method
for assessing pain behavior in children with JRA.
Subsequent objectives were to assess (a) the interrater
reliability of the method; (b) the concurrent validity of
the method, using subjective ratings of pain intensity
made by children and their parents as criterion mea-
1143
PAIN ASSESSMENT IN JRA
Table 1. Selected demographic and clinical characteristics of the study subjects with juvenile
rheumatoid arthritis (JRA)
Variable
6-1 1-year-olds
12- 17-y ear-olds
Total group
No. of boys/no. of girls
No. whiteho. black
Mean 2 SEM age, years
Mean 2 SEM months since
diagnosis
Initial diagnosis, no.
Polyarticular JRA
Systemic-onset JRA
Pauciarticular JRA
Diagnosis at evaluation, no.
Polyarticular JRA
Systemic JRA
Pauciarticular JRA
5/10
1411
8.5 & 1.8
38.4 f 7.4
3/12
8/22
2317
13.8 2 1.7
59.5 5 9.8
9
8
4
3
4
17
7
6
10
22
0
5
0
8
2
12
0
3
sures; and (c) the construct validity of the method,
using an objective measure of functional disability
and children’s self-reports of depression as criterion
measures.
The following hypotheses were proposed to
assess the reliability and validity of this method for
children with JRA: 1) With regard to interobserver
reliability, it was proposed that observers would exhibit high levels of agreement in recording each identified pain behavior (i.e., percentage agreement and
percentage effective agreement 20.70, K 2 0.60). 2)
With respect t o concurrent validity, it was proposed
that the total frequency of observed pain behaviors
would be significantly and positively correlated with
the children’s self-reports of pain intensity and the
parents’ ratings of their children’s pain intensity. 3)
Regarding construct validity, it was proposed that the
total frequency of pain behaviors would be significantly and positively correlated with the children’s
observed levels of functional disability. In addition, it
was proposed that self-reports of depression would not
be associated with the total frequency of pain behaviors displayed by children with JRA, but that selfreports of depression would be significantly correlated
with the children’s self-reports of pain and the parents’
reports of their children’s pain.
PATIENTS AND METHODS
This study was approved by the University of Alabama at Birmingham institutional review board.
Subjects. Pilot subjects included 6 children who had
JRA and were being followed up through the Children’s
Rehabilitation Services of Alabama. Two were boys and 4
were girls. Attempts were made to include younger and older
children (age range 6-17 years). The mean age of these pilot
subjects was 11.3 years (SEM 3.6 years). The group of pilot
916
11.2
2
3.2
49.0
+-
6.7
subjects was representative of the larger study population
described below in that all children had either polyarticular
JRA (at least 4 joints) or pauciarticular JRA (1 large joint
such as the hip or knee; total number ofjoints <4) according
to the criteria of the American College of Rheumatology
(ACR) (22) and had evidence of active disease, manifested
by tenderness, pain, or swelling in the involved joints.
The 30 subjects for the final study were identified
through the Children’s Rehabilitation Services of Alabama,
the Alabama chapter of the Arthritis Foundation, and the
private practices of several rheumatologists throughout the
state. The children ranged in age from 6 to 17 years. All met
the ACR criteria for the diagnosis of either poly- or pauciarticular JRA and had active joint involvement as described
above. For children with pauciarticular JRA, only those with
involvement in a large or medium-sized lower extremity joint
such as the hip, knee, or ankle were included in the present
study. This was done to help maximize our ability to observe
pain behaviors, given that individuals with large or mediumsized lower extremity joint involvement might be more likely
to demonstrate a high frequency of pain behaviors on
ambulation and transfer (e.g., from sitting to a standing
position) than would those with involvement of smaller
joints of the fingers or arm. Children had no more than 1
prior clinic visit without active joint involvement. Those
who were confined to bed or a wheelchair (Steinbrocker
functional class IV) (23) were excluded from the study
because of their inability to perform the activities of the
observation session. One parent of each child also agreed to
participate in the study. Table 1 summarizes the demographic and clinical information on the subjects.
Given the evidence that pain perception varies as a
function of age, the sample was stratified so that 15 of the
children were between 6 and 1 1 years, and 15 were between
12 and 17 years. The mean age of the 30 children was 11.2
years (SEM 3.2 years), and the average duration of disease
was 49.0 months (SEM 6.7 months), At disease onset, 17
children were diagnosed as having polyarticular JRA, 7 as
having systemic-onset JRA, and 6 as having pauciarticular
JRA. At the time of evaluation for this study, 22 children had
polyarticular JRA and 8 had pauciarticular JRA; none had
evidence of systemic manifestations.
1144
Table 2.
JAWORSKI ET AL
Operational definitions of pain behaviors for children*
Pain behavior
Guarding
Bracing
Rigidity
Passive Rubbing
Active Rubbing
Weight Shifting
Single Flexing
Multiple Flexing
Sighing
Definition
Abnormally stiff, hesitant, interrupted, jerky, or rigid movement during shifting or pacing. Can be seen during position
shifts and transfers. It includes the use of canes, crutches, or braces but cannot occur during a stationary position
(i.e., sitting, standing, reclining). The movement must be hesitant, jerky, interrupted, or stiff/rigid rather than merely
slow.
Position in which an almost fully extended limb (e.g., arm) supports and maintains an abnormal distribution of weight.
It cannot occur during movement (i.e., pacing or shifting), and must be held for at leust 3 consecutive seconds. It
most frequently is the gripping of the edge of a table, chair, cane, crutch or brace while sitting or standing. Bracing
can occur with a leg if the subject leans against a wall using no other support, but it is not simply the shifting of
weight when standing. What appears to be bracing during movement is guarding.
Excessive sfiffnessof an uflected joint or body part (with the exception of fingers or toes) that is not directly involved
in locomotion. Rigidity can be scored during movement (e.g., an arm held stiffly while pacing) or in stationary
positions if it results in an abnormal body posture (e.g., legs fully extended while sitting; a shoulder or back arched
stiffly; an elbow held rigidly away from the body). Any such stiffness or posturing is scored even if deformity may be
causing the limb to be held that way.
Touching, resting, or holding an aflected joint or body part on another body part for at least 3 consecutive seconds.
This is not scored if affected hands or fingers are simply resting independently on the lap or abdomen during sitting
or reclining (but are not being held together). However, holding affected hands together during sitting, standing, or
reclining is scored for passive rubbing. Other examples would be resting ankles against one another or resting a hand
on an affected knee.
Massaging an uflected joint or body part for at least 3 consecutive seconds. Massaging may be performed against
another body part (e.g., hand, leg, arm) or an object (e.g., side or back of a chair or table).
Unloading weight from an affected limb to an unaffected limb or from affected limb to affected limb while engaging in a
stationary activity (e.g., standing). Can be scored when only one shift or when more than one shift in weight occurs.
The weight shift can include a step to one side accompanied by a physically obvious shift of weight. It can also
include instances where the person only bends the knee which takes weight off of one limb and shifts it to the other
side of the body.
A single movement of an affectedjoint or body part that is not directly involved in a position change (e.g., flexing a
wrist; flexing one’s ankle when sitting, standing, or reclining; turning out or popping a knee when reclining, standing,
or sitting).
Repetitive movement (Le., at least twice) in a 3-second period of time of an uflected joint or body part that is not
directly involved in a position change (e.g., opening and closing of fingers, flexing ankles multiple times when
sitting).
An obvious and exaggerated exhalation of air usually accompanied by a rise and fall of the shoulders. The cheeks may
be expanded.
* Readers interested in the full details of the procedures for performing the behavioral observations and more specific details regarding the
operational definitions for the pain behaviors should contact the principal investigator (TMJ). Sample videotapes and scoring protocols will also
be available for review so that other investigators can compare their observations of pain behaviors displayed by patients on the videotapes with
the observations recorded by TMJ.
Instrument development. A pilot study was first conducted to identify children’s pain behaviors that appeared to
indicate subjective experience of pain. A protocol similar to
that used by McDaniel et al(10) was developed, which required
the 6 children to perform activities such as sitting, walking,
standing, and reclining for ten 1-minute intervals. The activities were recorded on videotape, and these videorecordings
were then reviewed to identify frequently displayed pain
behaviors. Four professionals who work in the area of
chronic pain viewed the 6 video recordings and provided the
operational definitions. Six pain behaviors associated with
the observation method used for adults with RA (lo), and 3
pain behaviors that were specific to this sample of children
and adolescents with JRA were identified. Table 2 provides
the operational definitions of pain behaviors for this behavioral observation system. Following the identification of the
pain behaviors, the reliability and validity of the observation
method were evaluated using the methods described below.
Comparison measures. The Varni/Thompson Pediatric Pain Questionnaire (PPQ) (24,25). The PPQ is a comprehensive, multidimensional instrument specifically designed
for the study of acute, recurrent, and chronic pain in children
and adolescents. The child and adolescent forms include a
VAS to assess present and worst pain intensity for the past
week. The parent form elicits information regarding children’s pain and functional activity levels. Reliability and
validity studies of this instrument with JRA patients suggest
that it is an appropriate measure of subjective pain in
developing populations (children and adolescents) (24,25).
The Juvenile Arthritis Functional Assessment Scale
(JAFAS)(21).The JAFAS was developed specifically for the
assessment of functional disability in patients with JRA. It
requires observation of a child as she/he performs 10 simple
tasks that are relevant to common daily functioning and
involve use of all joints and muscle groups. For each task, a
rating is made as to whether the child is (a) successfully able
to complete the task within a specified interval of time, (b)
able to complete the task but at a much slower pace, or (c)
unable to complete the task. Reliability and validity data
suggest that it is an appropriate measure of disability in
children with JRA.
One rheumatology trainee performed all of the JAFAS
PAIN ASSESSMENT IN JRA
ratings for the present study. He was trained by the primary
investigator (TMJ) in JAFAS administration and coding and
reached a criterion of 85% agreement with the primary
investigator.
The Children’s Depression Inventory (CDI) (26,27).
The CDI is a 27-item self-report, symptom-oriented scale
that assesses depression in children and adolescents between the ages of 7 and 17 years. Adequate levels of
reliability and validity have been established (26,27). For the
present study, difficult words were clarified during administration of the CDI to the 2 youngest children (both 6 years
old). The CDI has been used in multiple studies to assess
depression in children and adolescents with a variety of
chronic (cancer, diabetes) and recurrent (sickle cell anemia)
conditions (28,29). These studies indicate that pediatric
patients do not report higher levels of depression than do
those without chronic medical conditions. Thus, the CDI
does not yield inflated estimates of depression for pediatric
patients, and the instrument is considered to be appropriate
for use with children with chronic medical conditions.
Behavioral observation session. All children participated in a 10-minute session during which they performed,
for video recording, a series of behaviors including sitting,
walking, standing, and reclining. The standardized sequence
consisted of two 1-minute sitting periods, two 1-minute
standing periods, two 1-minute reclining periods, and four
1-minute walking periods. Although the time spent for each
activity was the same for all children, the order of the
activities was randomly varied across children. During video
recording, the investigator communicated with the children
only to instruct them to perform specific behaviors (i.e., sit,
stand, walk, recline). No detailed information regarding the
behavioral observation method or the specific behaviors of
interest was provided. It should be noted that this procedure
was identical to that used by McDaniel et a1 in the development of the pain behavior observation method for adults
with RA (10).
Procedure. Written informed consent and permission
to videotape the standard protocol of behaviors were obtained from parents and children prior to the beginning of the
study. Children first were asked to identify the joints or body
areas in which they were presently experiencing pain or
discomfort. One-half of the children then participated in the
10-minute observation session. They subsequently were
administered the PPQ, JAFAS, and CDI. The other half of
the children were first administered the PPQ, JAFAS, and
CDI and then participated in the 10-minute observation
session. The order of the questionnaire administration and
observation session was counterbalanced in this manner to
reduce potential biases. The order of administration of the
PPQ, JAFAS, and CDI also was counterbalanced to minimize potential biases. While the children underwent the
procedures described above, their parents completed the
parent form of the PPQ.
Recording of observations. Four research assistants
were trained in the behavioral observation method using a
protocol similar to that outlined by Keefe et a1 (30). The
assistants initially viewed videotaped samples of pain behavior displayed by adults and children with RA, and issues
related to defining and recording the occurrence of specific
behaviors were discussed. A manual that provided defini-
1145
tions and descriptions of the pain behaviors was then given
to each assistant. After studying the manual, assistants were
asked to respond to videotaped examples of ambiguous
situations in which it was unclear whether to record the
occurrence of a behavior. When their understanding of the
observation method seemed clear, assistants viewed selected segments from videotapes that provided a full range of
pain behaviors. As assistants watched and recorded behavior, they were given immediate feedback on their performance, and uncertainties or ambiguities were discussed.
Finally, each assistant observed and recorded pain behavior
frequencies in a series of six 10-minute videotapes of children with JRA. A criterion of 85% agreement with the
primary investigator was required before assistants were
asked to observe and record videotapes of the 30 children
whose data were to be used for the reliability and validity
studies.
Each child’s videotape was observed independently
by the primary investigator and 2 research assistants using
twenty 30-second intervals. These intervals consisted of a
20-second observation phase followed by a 10-second recording phase. To ensure consistency of timing among the
observers, each videotape was dubbed with an audio signal
to indicate the start of observation and recording intervals.
Regardless of the number of times a behavior was observed
during a 20-second interval, it was recorded only once. A
single score was computed for each pain behavior by summing the number of times it was observed for a given
individual across the 20 intervals. A total pain behavior
score also was computed by summing all of the individual
pain behavior scores.
To minimize observer drift, the research assistants
viewed a training tape after every eighth subject and recorded pain behaviors. The assistants’ recordings were
compared against a master recording made by the primary
investigator. If the reliability level between the recordings of
the assistants and the primary investigators fell below the
criterion (85% agreement), a retraining session that was
similar to the initial training program was conducted. On
only 1 occasion did an assistant’s ratings fall below the
criterion of 85% agreement with the primary investigator
(82% agreement). A retraining session with this assistant,
which included reviewing pain behavior definitions as outlined in the manual, rating sample videotapes with both
straightforward and ambiguous demonstrations of pain behaviors, and providing feedback on the assistant’s scoring,
was conducted. Following retraining, the reliability of the
assistant’s scoring never fell below the criterion level and
showed a mean agreement level of 90%.
Prior to inclusion in the final observation system,
each pain behavior was evaluated according to 2 criteria: (a)
a high level of interrater reliability even after correction for
chance agreements among independent observers (i.e., K
2 0.60), and (b) coefficient of correlation 20.10 between
pain behavior frequency and children’s reports of pain on the
PPQ. This level of correlation was selected after reviewing
associations between individual pain behaviors and selfreports of pain in previous observation studies (range of r =
0.10-0.15; [10,12,31]). Also, given that the present study
represents the initial development of this method for children with JRA, it was appropriate to include all behaviors
JAWORSKI ET AL
1146
Table 3. Mean frequencies of pain behaviors, measures of central tendency, and reliability estimate
of pain behaviors
Behavior
Mean f SEM
frequency
Minimum
frequency
Maximum
frequency
Percentage
agreement
Percentage
effective
agreement
K
~
Guarding
Bracing
Active Rubbing
Passive Rubbing
Rigidity
Weight Shifting
Single Flexing
Multiple Flexing
Sighing
5.63 2 4.24
0.66 5 1.49
1.31 5 1.71
3.78 f 4.74
2.34 ? 4.48
0.69 2 0.97
4.16 2 3.40
2.84 ? 2.30
0.63 2 1.07
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
14.0
6.0
6.0
16.0
14.0
3.0
12.0
9.0
5 .O
94
95
95
94
93
97
87
84
66
85
84
90
63
70
84
94
96
*
0.79
0.61
0.78
0.77
0.53
0.58
0.60
0.67
0.40
* Could not be calculated due to relatively low frequencies of observed behaviors.
that could be reliably assessed and that might contribute to a
reliable association between frequency of total pain behavior
and criterion measures of pain used in the validity analysis.
Statistical procedures. The relationships among demographic variables and between the demographic and pain
criterion variables were assessed using chi-square and t-test
analyses. The reliability of the behavioral observation
method was evaluated using percentage agreement, percentage effective agreement, and kappa statistic. Percentage
agreement was computed by adding the total number of
intervals in which both observers agreed on either the
occurrence or the nonoccurrence of a given behavior, dividing this sum by 20 (the total number of intervals), and
multiplying by 100. Percentage effective agreement was
calculated by dividing the number of intervals in which both
observers agreed on the occurrence of a behavior by the
number of intervals in which agreements and in which
disagreements occurred, and multiplying by 100. Kappa
coefficients for each behavior also were calculated in order
to obtain reliability estimates that were corrected for the
proportion of chance agreements between observers (32).
Pearson product-moment correlations were used to evaluate
the validity of the behavioral observation method. An intercorrelation matrix of pain behaviors revealed some correlation between the individual pain behaviors. To account for
this intercorrelation, a relatively stringent alpha level of 0.01
was set for all correlational analyses. Fisher’s Z-transformation was used to assess differences in the magnitudes of
specific correlation coefficients associated with the older and
younger children or their parents.
RESULTS
Demographic variables. Table 1 shows that
there were no significant relationships between age
groups and the variables of sex (2= 0.68 [l degree of
freedom (df), n = 301, P not significant [NS]), initial
diagnosis (2= 0.87 [2 df, n = 301, P NS), diagnosis at
time of evaluation (2 = 0.68 [l df, n = 301, P NS), or
months since initial diagnosis (t[28] -1.36, P NS). In
addition, no significant associations were found be-
tween sex and initial diagnosis (? = 0.24 [2 df, n =
301, P NS), diagnosis at time of evaluation (2 = 0.66
11 df, n = 301, P NS), or months since initial diagnosis,
(t[28] 0.39, P NS). Ethnic background was not associated with the pain criterion variables of total number
of observed pain behaviors (t[28] 0.59, P NS), children’s reports of pain (t[28] -0.60, P NS), or parents’
reports of children’s pain (t[28] 0.23, P NS).
Reliability analyses. Table 3 shows the mean
frequencies of the pain behaviors during the videotaped observation session, measures of central tendency, and reliability estimates for the pain behaviors
displayed by the children with JRA. Percentage agreement for all behaviors was high, ranging from 90% to
97% (mean f SEM 94.00 +- 2.99). Percentage effective
agreement for these behaviors was lower, although the
reliability levels for the majority of the pain behaviors
continued to be acceptable, ranging from 63% to 87%
(mean f SEM 78.00 +- 12.39%). The more conservative kappa coefficients also were acceptable for the
majority of behaviors, ranging from 0.40 to 0.79 (mean
k SEM 0.64 & 0.20).
The results of these analyses as well as the
initial validity analyses allowed us to reduce the number of pain behaviors in the observation method. First,
Passive Rubbing was deleted from the observation
method. This behavior could be reliably observed
( K = 0.77), but it was not strongly associated with
the criterion variable of children’s pain ratings
(r = -0.02). Three additional behaviors, Sighing ( K =
0.40), Rigidity ( K = 0.53), and Weight Shifting ( K =
0.58) did not meet the initial stringent criterion of 0.60.
However, given that a kappa coefficient of 0.40 is
generally considered to reflect an acceptable level of
PAIN ASSESSMENT IN JRA
1147
Table 4. Correlation matrix for the individual pain behaviors (n = 30)
Guarding
Bracing
Active Rubbing
Rigidity
Single Flexing
Multiple Flexing
Guarding
Bracing
1 .oo
0.36*
-0.05
-0.07
-0.10
-0.06
1 .oo
0.004
0.14
0.10
0.09
-
Active
Rubbing
Rigidity
Single
Flexing
Multiple
Flexing
1 .oo
-0.25
0.36*
0.40*
-
-
I .oo
-0.25
-0.20
-
-
-
-
1 .oo
0.49t
-
1 .oo
* P 5 0.005.
t P 5 0.01.
agreement between observers (33,34), we carefully
examined the validity coefficients before making the
final decision to retain or delete each of these behaviors.
Sighing was deleted from the observation
method because it showed a negative correlation with
children’s pain ratings (r = -0.22). Similarly, Weight
Shifting was deleted because its relationship with
children’s pain ratings was relatively modest (r =
0.13). Rigidity, however, was maintained in the final
system because its correlations with the criterion
variables of self-reports of pain (r = 0.22) and parents’
reports of children’s pain (r = 0.39) were higher than
those for the majority of validity coefficients assessed
in this study (ranges 0.164.34 for children and 0.030.39 for parents). The validity coefficients for Rigidity
also were higher than most of those reported for adults
with RA or other chronic pain syndromes (range
0 . 1 0 4 3 3 [3 1,351). Thus, we concluded that Rigidity
was observed with sufficient reliability to meet the
standards of other investigators (33,34), and its strong
relationship with the pain rating variables supported
its validity as an observable dimension of pain behavior. Six pain behaviors were retained for all subsequent validity analyses. These were Guarding, Bracing, Active Rubbing, Rigidity, Single Flexing, and
Multiple Flexing.
Correlation and validity analyses. Table 4 shows
the intercorrelation matrix of individual pain behaviors
included in the final observation method. The pain
behaviors generally were independent of one another.
However, significant correlations were found between
Guarding and Bracing (r = 0.36, P 5 0.005), Single
Flexing and Active Rubbing (r = 0.36, P 5 0.005),
Single Flexing and Multiple Flexing (r = 0.49, P 5
0.006), and Multiple Flexing and Active Rubbing (r =
0.40, P 5 0.005). The correlations indicate that the
amount of variance shared by all possible pairs of
these 5 pain behaviors ranged from 13% to 25%.
Concurrent validity. Table 5 presents the correlations between individual and total observed pain
behaviors and (a) subjects’ self-reports of pain and (b)
parents’ reports of their children’s pain. The mean
Table 5. Correlations between subjective reports of pain on a VAS and frequency of observed pain
behaviors*
Pain behavior
Bracing
Active
Rubbing
Rigidity
0.20
0.03
0.19
0.05
0.34
0.23
0.22
0.39
0.24
0.18
0.51
0.39
0.30
0.05
0.20
0.27
0.21
0.60
0.19
-0.06
0.01
-0.28
0.15
0.05
0.55
0.21
0.22
0.14
0.32
0.46
Guarding
Total data (n = 30)
VAS (child)
VAS (parent)
6-11-year-olds (n = 15)
VAS (child)
VAS (parent)
12-17-year-olds (n = 15)
VAS (child)
VAS (parent)
Single
Flexing
Multiple
Flexing
Total
0.16
0.32
0.50t
0.48$
-0.07
-0.07
0.34
0.71t
0.64$
0.73t
0.46
0.35
* Values are the coefficients of correlation between the behavior and the visual analog scale (VAS)
score.
tP
5 0.005.
j: P
5
0.01.
1148
JAWORSKI ET AL
Table 6. Correlations between disability levels as determined by the JAFAS and frequency of
observed pain behaviors*
Pain behavior
Total data (n = 30)
6-11-year-olds (n = 15)
12-17-year-olds (n = 15)
Guarding
Bracing
0.48t
0.10
0.711
0.59$
0.19
0.747
Active
Rubbing
Rigidity
Single
Flexing
Multiple
Flexing
Total
0.09
0.34
-0.07
0.37
0.49
0.28
0.08
0.22
0.02
0.03
-0.07
0.11
0.640
0.66t
0.708
* Values are the coefficients of correlation between the behavior and the Juvenile Arthritis Functional
Assessment Scale (JAFAS) rating (ref. 20).
t P 5 0.01.
$ P 5 0.001.
5 P 5 0.0001.
n P c 0.005.
VAS pain rating reported by children on the PPQ was
3.40 (SEM 0.57). The mean VAS pain rating reported
by parents was 3.24 (SEM 0.47). The pain ratings
given by children and their parents were significantly
and positively correlated with the total frequency of
observed pain behaviors (r = 0.50, P 5 0.005 and r =
0.48, P 5 0.007, respectively). When separate analyses were performed with the younger and older children, it was found that the younger children’s reports
of pain tended to be more strongly correlated with
total frequency of observed pain behaviors (r = 0.64,
P 5 0.01) than were those of the older children (r =
0.46, P 5 0.08). Similarly, parents’ ratings of the
younger children’s pain tended to be more strongly
correlated with total observed pain behaviors (r =
0.73, P 5 0.002) than were the pain ratings given by
the parents of the older children (r = 0.35, P = 0.20).
However, comparisons of these correlation coefficients by Fisher’s 2 transformation did not achieve
statistical significance.
Construct validity. Table 6 presents the correlations between frequencies of subjects’ pain behaviors and their levels of functional disability as measured by the JAFAS. Total pain behavior was
positively and significantly correlated with subjects’
levels of functional disability (r = 0.64, P 5 0.0001).
Correlations between functional disability and total
pain behavior were similar for younger and older
subject groups.
Table 7 shows the correlations between frequencies of subjects’ total pain behaviors and subjects’ self-reports of depression (CDI scores). Total
pain behavior was not significantly correlated with
CDI scores (r = 0.23). However, subjects’ CDI scores
tended to be associated with their self-reports of pain
(r = 0.41, P = 0.03) and were significantly correlated
with parents’ reports of children’s pain (r = 0.50, P =
0.004).
DISCUSSION
The present study examined the reliability and
validity of an observation method for assessing pain
behaviors displayed by children with JRA. Overall,
the results suggest that the behavioral observation
method provides reliable and relatively objective information regarding pain associated with JRA. Five of
the original 9 pain behaviors were recorded with very
high consistency by observers, even after controlling
for chance agreements; these behaviors also were
correlated with subjects’ self-reports of pain, at sufficiently high levels to be retained in the final observation system. Although a sixth behavior, Rigidity, did
not quite meet our initial stringent reliability criterion
of 0.60, it was retained in the observation method
because its kappa coefficient of 0.53 is deemed to
reflect acceptable interobserver agreement (33,34) and
its validity coefficients were quite robust (r = 0.22-
Table 7. Correlations of ratings of depression on the CDI with
subjective reports of pain on a VAS and with observed pain
behaviors*
Total data (n = 30)
6-11-year-olds (n = 15)
12-17-year-olds (n = 15)
Total observed
pain behaviors
VAS
(child)
VAS
(parent)
0.23
0.21
0.24
0.41
0.30
0.17t
0.501
0.50
0.62f
* Values are the coefficients of correlation between the behavior or
the visual analog scale (VAS) score and the Children’s Depression
Inventory (CDI) score (refs. 26 and 27).
t P < 0.005.
f P < 0.01.
PAIN ASSESSMENT IN JRA
0.39). We believe that inclusion of only 6 behaviors in
the observation method permits assessment of a broad
spectrum of behaviors that communicate different
dimensions of the pain experience while minimizing
measurement error and other sources of unreliability
such as observer fatigue.
Strong support was provided for the validity of
this method for use with children with active JRA.
With respect to concurrent validity, the total frequency of subjects’ pain behaviors was significantly
associated with their self-reports of pain intensity and
with parents’ ratings of subjects’ pain intensity. Total
pain behavior frequency also was significantly correlated with observed levels of functional disability.
Further evidence for the construct validity of the
behavioral observation method was produced by the
finding that the total frequency of subjects’ pain behaviors was not significantly associated with their
self-reports of depression. In contrast, subjects’ depression ratings were significantly correlated with parents’ ratings of their children’s pain intensity; subjects’ depression ratings also tended to be associated
with their self-reports of pain intensity. These results
are consistent with those found in the laboratory
setting for adults with RA (10,15,36) and suggest that
the behavioral observation method for children provides a more objective assessment of pain than do
self-report measures. The fact that the behavioral
observation method is relatively free of the influence
of affective states, in conjunction with the preliminary
evidence of its reliability and validity, strongly suggest
that the method may provide an important measure of
pain among children with JRA.
It should be noted, however, that whether the
pain behaviors displayed by children during physical
examinations also are independent of the influence of
affective states remains to be determined. One investigation performed with adults RA patients indicated
that patients’ pain behaviors during physical examinations tended to be associated with affective distress
(35). Moreover, these patients displayed different frequencies of pain behaviors during physical examination, relative to the videotaped observation session.
For example, the patients exhibited more Bracing and
Grimacing when physicians were examining their
joints, whereas they displayed more Guarding during
the videotaped observation session that required several body position changes (e.g., sitting to standing)
and walking periods. Thus, different frequencies of
pain behaviors may be observed as a function of the
demand characteristics of various situations.
1149
Although there are important differences between children and adults in the ability to verbally
communicate pain experiences (9), several of the pain
behaviors demonstrated by the children with JRA
were consistent with those previously identified
among adult RA patients (10,15,36). Pain behaviors
common to both pediatric and adult patients included
Guarding, Bracing, Rigidity, and Active Rubbing. Single Flexing and Multiple Flexing were included only
within the pediatric behavioral observation system. In
contrast, Grimacing was observed among adult patients but was not demonstrated by children with JRA.
Observed frequencies of pain behaviors were comparable for younger (6-1 I-year-old) and older (12-17year-old) children. Overall, Guarding and Single Flexing were the most frequently exhibited behaviors,
whereas Bracing and Active Rubbing were least frequently displayed by JRA patients.
These findings are similar to those of McDaniel
et a1 (lo), who found that Guarding was displayed most
frequently by adults with RA, whereas Sighing and
Grimacing were displayed least frequently. The intercorrelation matrix revealed significant correlations between Guarding and Bracing and between the flexing
behaviors and Active Rubbing. It may be that Guarding and Bracing tend to be related because both
involve movement of large joints. In contrast, the
rubbing and flexing behaviors may be associated because they typically involve smaller joints in the hands
and feet. It should be noted, however, that the amount
of variance shared by all possible pairs of these 5 pain
behaviors ranged only from 13% to 25%. Thus, the
behaviors all contribute to the communication of the
experience of pain, but they do not show sufficient
overlap (r 2 0.90) to be considered duplicates of one
another.
There are 3 limitations associated with the
current study. First, it is possible the children may
have modified their behavior because they knew they
were being videotaped, which may have influenced the
pain behaviors they exhibited. It should be noted that
the present study used procedures nearly identical to
those used in previous observational studies of adult
patients (10,12,13). Thus, the question of alterations in
behavior in the presence of a camera is an issue
common to these types of studies and should be
addressed in further research. For example, the reliability and validity of the behavioral observation
method could be assessed in everyday settings (e.g.,
home, school) without using the video camera, in
order to further address this potential influence.
1150
The relatively small sample size also could be
considered a limitation of the present study. It is
important to note, however, that virtually all of the
children with active JRA who were seen through a
statewide Children’s Rehabilitation Services’ Arthritis
Clinic were included in this study. Several children
with active disease who were followed up in private
rheumatology practices throughout Alabama also were
study participants. Thus, while the sample size was
not large, efforts were made to ensure that the sample
was representative of children and adolescents with
active JRA in a southeastern state with a total population of nearly 4 million people. It would be of interest
in future studies to compare the types and frequencies
of pain behaviors exhibited by larger samples of children with polyarticular versus pauciarticular JRA. In
addition, it would be useful to attempt to develop a
similar observation method for measuring pain behavior in children with JRA who are younger than those in
our sample and for whom verbal reports of pain are
difficult to obtain.
Finally, there is a limitation in using self-report
ratings of pain as criterion variables. This problem
could not be resolved completely since there is no
objective gold standard for measuring the multidimensional construct of pain. Nevertheless, efforts were
made to ensure that the criterion measures used in the
present study were state-of-the-art self-report scales
that had been evaluated previously with pediatric
populations and that were associated with normative
data with good reliability and validity. Also, an observational measure of functional disability was used to
help validate the behavioral observation method.
Thus, rather than relying solely on subjective reports,
we used a relatively objective measure of children’s
functional abilities as an additional criterion. Given
that total pain behavior was significantly correlated
with both the self-report and the observational criterion variables, we have confidence that our data
support the validity of this method for use in children
with JRA.
In summary, our results indicate that the behavioral observation method is a reliable, valid, and
relatively objective measure of pain in children with
JRA. At present, this measure may be most useful in
outcome studies of innovative therapies (e.g., analgesic or disease-modifying agents) for these children.
That is, the behavioral observation method might
serve as a measure of treatment efficacy, in addition to
the traditional VAS and self-report scales, in order to
control for the relationship between the self-report
JAWORSKI ET AL
measures and depression. We acknowledge that scoring of the observational method currently is too time
consuming for routine use by physicians and other
health professionals in a clinical practice setting. Nevertheless, the present study suggests that, instead of
relying upon informal and subjective evaluations of
patients’ displays of pain behaviors, health professionals may reliably identify a small number of pain
behaviors displayed by children with JRA that are
associated both with subjective pain experiences and
with displays of functional disability. It will be important to refine and simplify the pain behavior observation method so that it may be used reliably, with brief
training, by health professionals in clinical practices.
The overall goal is to arrive at a simple and practical
system which health professionals can use in a costeffective manner to better care for children with JRA.
ACKNOWLEDGMENTS
We thank Dr. William Koopman and the staff of the
University of Alabama at Birmingham Multipurpose Arthritis and Musculoskeletal Disease Center for their support of
and assistance with this project. We also thank the staffs of
the Children’s Rehabilitation Services Agencies throughout
the state of Alabama for their support. Finally, we thank
Janie Murray, Isabel Scarinci, Allyson Sorenson, and Parker
Storey for their assistance in coding data.
REFERENCES
J: The Oucher: A User’s Manual and Technical Report.
Denver, University of Colorado Health Sciences Center, 1988
2. Bieri D, Reeve RA,Champion GD, Addicoat L, Ziegler JB: The
Faces Pain Scale for the self-assessment of the seventy of pain
experienced by children: development, initial validation, and
preliminary investigation for ratio scale properties. Pain 41 :139150, 1990
3. Hester NKO: The preoperational child’s reaction to immunization. Nurs Res 28:25&255, 1979
4. Abu-Saad H: Assessing- children’s responses to pain. Pain
19:163-17 1, 1984
5 . McGrath PJ. Johnson G. Goodman JT, Schillinger J, Dunn J,
Chapman J: The Children’s Hospital of Eastem Ontario Pain
Scale (CHEOPS): a behavioral scale for rating post-operative
pain in children. In, Advances in Pain Research and Therapy:
Proceedings of the 4th World Congress on Pain, Volume 9.
Edited by H L Fields, R Dubner, F Cervero. New York, Raven
Press, 1985
6. Beales JG, Holt PJL, Keen JH, Mellor VP: Children with
juvenile chronic arthritis: their beliefs about their illness and
therapy. Ann Rheum Dis 42:481-486, 1983
7. Beales JG, Keen JH, Holt PJL: The child’s perception of the
disease and the experience of pain in juvenile chronic arthritis.
J Rheumatol 10:61-65, 1983
8, GafFney A, Dunne EA: Developmental aspects of children’s
definitions of pain. Pain 26:105-117, 1986
9. McGrath PA: Pain in Children: Nature, Assessment, and Treatment. New York, Guilford Press, 1990
1. Beyer
PAIN ASSESSMENT IN JRA
10. McDaniel LK, Anderson KO, Bradley LA, Young LD, Turner
RA, Agudelo CA, Keefe FJ: Development of an observation
method for assessing pain behavior in rheumatoid arthritis
patients. Pain 24:165-184, 1986
11. Bradley LA: Psychosocial factors and disease outcomes in
rheumatoid arthritis: old problems, new solutions, and a future
agenda (editorial). Arthritis Rheum 32:1611-1614, 1989
12. Keefe FJ, Block AR: Development of an observation method
for assessing pain behavior in chronic low back pain patients.
Behav Ther 13:363-375, 1982
13. Ahles TA, Coombs DW, Jensen L, Stukel T, Maurer LH, Keefe
FJ: Development of a behavioral observation technique for the
assessment of pain behaviors in cancer patients. Behav Ther
21:449460, 1990
14. Bradley LA, Young LD, Anderson KO, Turner RA, Agudelo
CA, McDaniel LK, Pisko EJ, Semble EL, Morgan TM: Effects
of psychological therapy on pain behavior of rheumatoid arthritis patients: treatment outcome and six-month followup. Arthritis Rheum 30:1105-1 114, 1987
15. Anderson KO, Bradley LA, McDaniel LK, Young LD, Turner
RA, Agudelo CA, Gaby NS, Keefe FJ, Pisko EJ, Snyder RM,
Semble EL: The assessment of pain in rheumatoid arthritis:
disease differentiation and temporal stability of a behavioral
observation method. J Rheumatol 14:700-704, 1987
16. Jay SM, Ozolins M, Elliott CH, Caldwell S: Assessment of
children’s distress during painful medical procedures. Health
Psychol 2:133-147, 1983
17. Katz ER, Kellerman J, Siege1 SE: Behavioral distress in children with cancer undergoing medical procedures: developmental considerations. J Consult Clin Psychol 48:356365, 1980
18. LeBaron S , Zeltzer L: Assessment of acute pain and anxiety in
children and adolescents by self-reports, observer reports, and
a behavior checklist. J Consult Clin Psychol 52:729-738, 1984
19. Jay SM: Pain in children: an overview of psychological assessment and intervention. In, Health Psychology: Treatment and
Research Issues. Edited by AR Zeiner, D Bendell, CE Walker.
New York, Plenum Press, 1985
20. Howe S, Levinson J, Shear E, Hartner S, McGirr G, Schulte M,
Lovell D: Development of a disability measurement tool for
juvenile rheumatoid arthritis: the Juvenile Arthritis Functional
Assessment Report for children and their parents. Arthritis
Rheum 34S73-880, 1991
21. Lovell DJ, Howe S, Shear E, Hartner S, McGirr G, Schulte M,
Levinson J: Development of a disability measurement tool for
juvenile rheumatoid arthritis: the Juvenile Arthritis Functional
Assessment Scale. Arthritis Rheum 32:1390-1395, 1989
1151
22. Brewer EJ Jr, Bass J, Baum J, Cassidy JT, Fink C, Jacobs J,
Hanson V, Levinson JE, Schaller J, Stillman JS: Current
proposed revision of JRA criteria. Arthritis Rheum 20 (suppl
2):195-199, 1977
23. Steinbrocker 0, Traeger CH, Batterman RC: Therapeutic criteria in rheumatoid arthritis. JAMA 140:659462, 1949
24. Varni JW, Thompson KL, Hanson V: The Varni-Thompson
Pediatric Pain Questionnaire. I. Chronic musculoskeletal pain in
juvenile rheumatoid arthritis. Pain 28:27-38, 1987
25. Varni JW: Behavioral management of pain in children. In,
Advances in Pain Research Therapy, Volume 15. Edited by DC
Tyler, EJ Krane. New York, Raven Press, 1990
26. Kovacs M: The Children’s Depression Inventory. Psychopharmacol Bull 21:995-998, 1985
27. Saylor CF, Finch AJ, Spirit0 A, Bennett B: The Children’s
Depression Inventory: a systematic evaluation of psychometric
properties. J Consult Clin Psychol 52:955-967, 1984
28. Lemanek KL, Moore, SL, Gresham FM, Williamson DA,
Kelley ML: Psychological adjustment of children with sickle
cell anemia. J Pediatr Psychol 11:397-410, 1986
29. Worchel FF, Rae WA, Olson TK, Crowley SL: Selective
responsiveness of chronically ill children to assessments of
depression. J Personality Assessment 59:605415, 1992
30. Keefe FJ, Bradley LA, Crisson JE: Behavioral assessment of
low back pain: identification of pain behavior subgroups. Pain
40:153-160, 1990
31. Jensen IB, Bradley LA, Linton SJ: Validation of an observation
method of pain assessment in non-chronic back pain. Pain
39~267-274, 1989
32. Hartmann DP: Considerations in the choice of interobserver
reliability estimates. J Appl Behav Anal 10:103-116, 1977
33. Kramer MS, Feinstein AH: Clinical biostatistics: LIV. The
biostatistics of concordance. Clin Pharmacol Ther 29: 1 1 1-123,
1981
34. Landis RJ, Koelt GG: The measurement of observer agreement
for categorical data. Biornetrics 33: 159-174, 1977
35. Anderson KO, Bradley LA, Turner RA, Agudelo CA, Pisko EJ,
Salley AN Jr, Fletcher KE: Observation of pain behavior in
rheumatoid arthritis patients during physical examination. Arthritis Care Res 5:49-56, 1992
36. Anderson KO, Bradley LA, McDaniel LK, Young LD, Turner
RA, Agudelo CA, Keefe FJ, Pisko EJ, Snyder RM, Semble EL:
The assessment of pain in rheumatoid arthritis: validity of a
behavioral observation method. Arthritis Rheum 30:3643, 1987
Документ
Категория
Без категории
Просмотров
2
Размер файла
1 018 Кб
Теги
development, pain, behavior, observations, method, arthritis, assessing, juvenile, children, rheumatoid
1/--страниц
Пожаловаться на содержимое документа