Abstract

Objective To complete an assessment of social functioning of children with juvenile rheumatoid arthritis (JRA) and nonchronically ill controls who had been evaluated 2 years earlier (Noll et al., 2000) and to examine the impact of disease severity or disease activity over time on the social functioning of children with JRA. Methods Peer-, teacher-, and self-reports of social functioning were obtained from 57 children with JRA and 63 controls. Social reputation and social acceptance were examined cross-sectionally and longitudinally. Results Cross-sectional analyses indicated no significant differences between children with JRA and controls on measures of social functioning. For children with more severe disease, like ratings declined over the 2-year period relative to children with mild disease. Children with active disease were chosen fewer times over the 2-year period as a best friend than children in remission. Conclusions Because children with severe or active JRA may be at risk for difficulties with social acceptance over time, they are appropriate targets for interventions that ameliorate or prevent these difficulties.

Juvenile rheumatoid arthritis (JRA) is a heterogeneous group of conditions primarily distinguished by inflammation of the connective tissues (e.g., the joints), JRA affects approximately 30,000-50,000 children each year in the United States (Lawrence et al., 1998). Children with JRA can experience chronic pain, stiffness in joints, limitations in mobility, delayed puberty, and growth disturbances that can lead to permanent disability and deformity. Approximately 10% of children with JRA will experience severe functional disabilities as adults (Cassidy & Petty, 1995). Additional complications can involve the cardiac, pulmonary, ocular, endocrine, and auto-immune systems. Long-term treatment of JRA, palliative rather than curative, is multifaceted, often including medication, physical and occupational therapy, orthopedic surgery, and nutritional support. Onset of JRA appears to peak at ages 1-3 years and 8-12 years. It is more common in girls than in boys and generally follows infection or physical trauma (Cassidy & Petty, 1995; Vandvik & Hoyeraal, 1993).

With few exceptions (Harris, Newcomb, & Gewanter, 1991), studies of children with JRA suggest that they are at risk for experiencing social difficulties, often attributed to the physical effects of the disease and its treatment, as well as the many disruptions to daily living that children with JRA experience. For instance, alterations in appearance (e.g., swollen or deformed joints) may cause peers to react negatively, and many forms of play and sports that stress affected joints are restricted. Side effects from treatment (e.g., nausea and vomiting produced by anti-inflammatory drugs), as well as medical procedures and frequent visits to the doctor, may lead to repeated school absences and limited time with peers (Lovell et al., 1990).

These events have the potential to negatively affect the social functioning of children with JRA. Several studies have reported that children with JRA participate in fewer activities than nonchronically ill controls or siblings (Billings, Moos, Miller, & Gottlieb, 1987; Timko, Stovel, Moos, & Miller, 1992). Children with JRA also have been described as socially isolated (Wilkinson, 1981), as considering themselves less popular and attractive than healthy siblings (Timko et al., 1992), and as exhibiting lower social competence than population norms (Daltroy et al., 1992). Two studies have reported that adolescents with JRA are especially at risk for poor social functioning (Billings et al., 1987; Daltroy et al., 1992), suggesting that deterioration in social functioning may accompany adolescence. Finally, multiple studies have suggested that children with severe forms of JRA experience more social difficulties than patients with mild or inactive forms (Billings et al., 1987; Timko et al., 1992; Ungerer, Horgan, Chaitow, & Champion, 1988).

Despite the evidence that children with JRA experience problems in social functioning, a number of methodological problems have made findings difficult to interpret. Such problems have included a failure to report recruitment rates and the use of inappropriate comparison groups. Moreover, most information about the social relationships and adjustment of children with JRA has been provided by parents, teachers, or the child, but few studies have reported data directly from the chronically ill child's peer group.

The first purpose of this longitudinal investigation was to extend an earlier study of the peer relationships of children with JRA and nonchronically ill controls (Noll et al., 2000) by repeating an assessment of social functioning 2 years later. In both the initial study and this 2-year follow-up, methodological limitations of previous work were addressed. In particular, a control group was recruited during the initial assessment by pairing each child with JRA with a classmate who was the same race, gender, and age. In addition, data were obtained from multiple sources (e.g., teacher-, peer-, and self-reports) using standardized measures. At the first assessment, Noll et al. (2000) reported no differences between the two groups on measures of social reputation (What is the child like?) or social acceptance (Is the child liked?).

Although no differences in social functioning were observed initially, living with the physical effects and complications of JRA and its treatment, as well as enduring continued daily hassles and disruptions, might result in the emergence of social difficulties over time. In addition, 51% of the children with JRA had made the transition from grade school to junior high or high school while participating in this study. This transition is often difficult for healthy children, but may be especially challenging for children coping with a chronic illness. In this investigation, we hypothesized that at the 2-year follow-up children with JRA would be identified by peer-, teacher-, and self-reports as being more socially withdrawn and isolated, as having fewer friends, and as being less popular than their controls. The effects of gender and grade level in school were examined for possible interactions with group (JRA vs. control) across all measures. Children with JRA were also compared to controls on a number of nonsocial attributes related to chronic illness. Specifically, we hypothesized that children with JRA would be perceived by peers as less attractive and athletic than controls, as well as more tired, more physically ill, and more often absent from school. Finally, the effects of disease severity (mild or moderate/severe), activity (active or full/partial remission), and duration on social functioning were examined. We expected that more severe disease and a longer duration of disease would be associated with greater difficulties in social functioning.

The second purpose of this investigation was to examine whether the role of group (JRA vs. control), disease severity, and disease activity affected the subsequent social functioning of children with JRA. We hypothesized that children with JRA, especially those with more severe or active disease, would experience increased social difficulties over time. These social difficulties included more socially withdrawn and isolated behavior, less popularity, and fewer friends.

Method

This follow-up study was part of a larger, longitudinal project examining the psychosocial impact of pediatric chronic illness on children and their families. The initial phase involved an assessment of the social functioning of children with JRA and controls in the school (Noll et al., 2000) followed by an assessment of family functioning in the home. Children with JRA and their controls were followed prospectively 2 years later with a repeated school-based assessment of social functioning.

Participants

Participants with JRA were recruited from patient rosters at a large, pediatric tertiary care facility in the Midwest. This center is the only treatment facility with board-certified pediatric rheumatologists within a 50-mile radius, and a hospital tax levy ensures treatment for disadvantaged youths. Thus, nearly every child with JRA in the area receives treatment at this institution. Initial selection criteria for this study required that children be 8-15 years of age, attending school without full-time special education, and living within 50 miles of the medical center. All children with JRA who participated in the initial study (N = 74) were potential participants for this 2-year follow-up. Children with JRA were diagnosed according to the classification standards of the American College of Rheumatology (Cassidy et al., 1986) and were 10-17 years of age (M = 12.98 years, SD = 1.53) at the time of second data collection. Mean time since diagnosis was approximately 8 years (M = 98 months, SD = 37, range = 34-173 months). Of the 74 children participating at time 1, 5 children with JRA no longer met inclusion criteria (2 were living more than 50 miles from the medical center, 2 were no longer in school, 1 was receiving full-time special education). Six children could not be located, and 2 families and 4 schools declined to participate, leaving 57 children with JRA (19 boys and 38 girls). For this follow-up sample, type of disease at onset was classified as the following: (a) pauciarticular (e.g., affecting four or fewer joints; n = 22); (b) polyarticular (e.g., affecting five or more joints; n = 27); or (c) systemic (e.g., affecting other organ systems; n = 8). Disease course was classified as (a) pauciarticular (n = 26), (b) polyarticular (n = 25), or (c) systemic (n = 6).

Controls were identified from a pool of children who participated in the initial classroom study. This was accomplished by selecting a classmate who was the same gender and race and who had the closest date of birth to the child with JRA (see Noll et al., 2000). This procedure was used so that illness status was not confounded with multiple social, economic, or cultural risk factors (Masten, 1989). Control families were screened to ensure that they did not have a child with chronic illness. If a potential control's family declined participation, then the next classmate whose birth date was closest to that of the target child was contacted, until a family agreed to participate. Of the 74 controls who participated in the initial study, 4 children were living more than 50 miles from the medical center, 4 children could not be located, and 3 schools declined to participate, leaving 63 controls (21 boys, 42 girls).

Measures

Revised Class Play (Masten, Morison, & Pellegrini, 1985). The Revised Class Play (RCP) is a descriptive matching instrument that requires students or teachers in a classroom to imagine that they are the director of a play and to cast members of their class into roles of this hypothetical play. Earlier factor analyses of peer nominations on the RCP with younger children (grades 3-7) indicated that these roles are designed to measure social reputation along three dimensions: sociability—leadership, aggressive—disruptive, and sensitive—isolated. Children are assigned scores for each item according to the number of times they were nominated for each role, and these item scores are summed for each source to create dimension scores (e.g., aggressive—disruptive). Z-score transformations of the dimension scores are completed within each classroom to adjust for unequal class size prior to conducting analyses (Masten et al., 1985). These dimension scores have adequate psychometric properties and predict later functioning (Hymel, Rubin, Rowden, & LeMare, 1990; Masten et al., 1985; Morison & Masten, 1991; Noll, Bukowski, Rogosch, LeRoy, & Kulkarni, 1990; Noll, LeRoy, Bukowski, Rogosch, & Kulkarni, 1991; Rubin, 1993).

Recently, it has been suggested that a four-factor structure for the RCP is a more meaningful, alternative scoring system for a broader age range of children (Zeller, Vannatta, Schafer, & Noll, 1999). The four-factor structure continues to have dimensions related to aggressive—disruptive and sensitive—isolated behaviors consistent with those described by Masten et al. (1985) for younger children, but the original sociability—leadership factor was split into a popular—leadership and a prosocial factor. The popular—leadership dimension consists of items that reflect popularity with peers (e.g., someone everybody likes to be with) and leadership skills (e.g., someone who is a good leader). The prosocial factor consists of items that pertain to social etiquette with peers (e.g., someone who is polite, someone who helps others). Internal consistencies for each factor were computed based on this study's sample, and in general, the majority of factors demonstrated adequate and comparable reliability (Cronbach's αs range from .77 to .90 for peers, .46 to .73 for teachers, and .65 to .78 for self-data). In addition, evidence for validity has been provided previously in the pattern of associations with measures of peer acceptance (Zeller et al., 1999).

For this study, three roles were added that were relevant to chronic illness (e.g., “someone who is sick a lot,” “someone who misses a lot of school,” and “a person who is tired a lot”), along with two roles related to physical appearance and two roles related to athletic competence (Noll et al., 1999). The ratings for each role were standardized within gender for each classroom. A single score for both physical appearance and for athletic competence was created by combining the two items assessing each domain, with higher scores indicating greater attractiveness and athletic ability (Vannatta, Zeller, Noll, & Koontz, 1998). These roles have been used in previous research, have discriminated between children with a chronic illness and healthy children, and have been associated concurrently with multiple domains of social reputation and acceptance (Graetz & Shute, 1995; Vannatta, Gartstein, Short, & Noll, 1998; Vannatta, Zeller, et al., 1998).

Three Best Friends. Children were asked to choose the three peers in their class whom they thought of as their best friends. From these data, each child was given a social preference score based on the number of times he or she was chosen as a friend by peers and a mutual friendship score indicating how many of his or her friendship selections were reciprocated. Z-score transformations of total and reciprocated best friend scores were completed separately for boys and girls within each classroom to adjust for unequal class size. This measure has been demonstrated to provide a stable and valid index of peer acceptance (Bukowski & Hoza, 1989; Gottman, Gonso, & Rasmussen, 1975).

Like Rating Scale (Asher, Singleton, Tinsley, & Hymel, 1979). This instrument assesses social preference by providing a measure of the degree to which each child in the class is liked or disliked by peers. The children are asked to rate each of their classmates and themselves on a 5-point Likert scale ranging from 1 (i.e., “someone you do not like”) to 5 (i.e., “someone you like a lot”) (Asher et al., 1979). For each child, an average social preference score was computed and then converted to Z-scores for boys and girls within each classroom. This measure has been shown to be a reliable index of a child's relative social acceptance, with test-retest correlations of .81 to .86 over a 4-week interval (Asher et al., 1979; Ladd, 1981).

Severity Ratings. At time 1 and time 2, disease severity was rated by a pediatric rheumatologist as mild or moderate/severe based on disease onset and course, the presence of vision problems, radiographic evaluation, pharmacotherapy, the number of active joints, and the presence of functional disability. At time 1, 49% of the children (n = 36) were classified as having mild disease and 51% (n = 38) as having moderate/severe disease (Noll et al., 2000). At the 2-year follow-up, 27 of the children initially classified with mild disease and 30 of the children with moderate/severe disease continued to participate. Disease severity was unrelated to whether or not a child continued to participate at time 2 (χ2 [1, N = 74] = .16, ns). Disease severity ratings completed at time 2 were available for 53 of the 57 children who participated at follow-up, as some of these children had transitioned to other clinic or care facilities. Of these children, 40% (n = 21) continued to be classified as having mild disease at time 2, while 53% (n = 28) were classified again as having severe disease. For 8% of the children, the severity of disease changed over time, from mild to moderate/severe (n = 2) or from moderate/severe to mild (n = 2). In addition, each child's disease status was rated at times 1 and 2 as active or in full/partial remission. Active disease status was defined as the presence of synovitis (swelling or painful loss of motion), while partial remission was described as a lack of synovitis but continued use of medication. At time 1, 55% (n = 41) of the children's disease status was evaluated as active, and 45% (n = 33) had disease in full/partial remission (Noll et al., 2000). At time 1, 34 of the children who had active disease and 23 who were in full/partial remission continued to participate at the 2-year follow-up. Disease activity was unrelated to whether or not a child continued to participate at time 2 (χ2 [1, N = 74] = 1.81, ns). Disease activity ratings made at time 2 were available for 51 of the 57 children who participated at follow-up. Of these children, disease status remained active in 43% (n = 22) of the children and in full/partial remission in 29% (n = 15) of the children. Twenty-seven percent of the children's disease changed in status from active to in remission (n = 9) or from in remission to active (n = 5). Because so few children experienced changes in disease severity or activity, analyses were not completed to examine change over time.

Procedures

The initial study of peer relationships and this 2-year follow-up required informed consent/assent and were approved by the local institutional review board. The procedures detailed here were used in both studies.

Data collection at time 2 occurred in 110 different classrooms and involved meeting children as a group in their class at school, usually the English class. English was chosen because it is a required class to which students are usually assigned based on ability and level of interest. Thus, there is an increased probability that each child's friends would be in that class (Ollendick, Greene, Weist, & Oswald, 1990). Each child's teacher, along with available classmates who returned a consent form (90%; N = 2,250 of 2,489), completed the RCP. During administration of the RCP, children in the class were given a roster listing all of their classmates of the same gender as the target child to eliminate sex-role stereotyped nominations and increase the probability of choosing the target child. Children were asked to nominate one peer from the roster other than himself or herself for each role and were told that they could use a name as many times as they wanted if they thought a classmate could play more than one role. The RCP was administered again, at which time the children were asked to mark the roles that they thought they would be best at playing. For the Three Best Friends measure, the children were given another list with the names of every child in the class from which they could nominate their first, second, and third best friend. Finally, the children were asked to complete the Like Rating Scale. Although peer nomination methods of social reputation and social acceptance such as these were initially developed for school-age children, this methodology has been used successfully to assess the social functioning of older children in junior high and high school (Casiglia, Lo Coco, & Zappulla, 1998; Luthar, 1995; Luthar & McMahon, 1996).

Analyses

Two-tailed independent t tests were conducted to compare families of children with JRA and controls on demographic variables. In addition, two-tailed, independent t tests were also used to compare groups of children on measures of social reputation and social acceptance at time 2. Correlational analyses were completed to examine the associations between disease duration and all measures of social functioning, as well as the degree of individual stability in RCP scores and measures of social acceptance over 2 years. Finally, the effects of several disease variables on changes in social functioning (i.e., RCP scores and measures of social acceptance) over time were examined with a series of analyses of variance (ANOVA) with repeated measures using time as a within-subjects variable and either type of participant (JRA vs. control), disease severity (mild vs. moderate/severe), or disease activity (full/partial remission vs. active) as a between-subjects variable. Holm's procedure for multiple corrections was applied to all comparisons not hypothesized a priori to control for Type I error (Holland & Copenhauer, 1988).1 Only those significant results adjusted for multiple comparisons are discussed.

1

In an attempt to avoid spurious findings (i.e., control familywise Type I error), corrections for multiple comparisons were applied for exploratory tests. In conducting tests of hypothesized effects, a less conservative approach was taken to avoid inflation of Type II error. For these tests, nondirectional, rather than directional, tests were used to limit Type I error, but corrections were not applied for the number of analyses performed. This more liberal approach was taken to increase power, given that this investigation is one of the few studies to examine social functioning of children with a chronic illness longitudinally.

Using GPOWER for power calculations (Faul & Erdfelder, 1992), the sample of 120 children produced sufficient power (.77-.96) to detect medium effects for t tests (d = .5) and for correlational analyses (d = .3). For the repeated measures analyses using type of participant as the between-subjects variable, the sample of 120 children produced adequate power to detect medium effects for the between-subjects variable (.87), the within-subjects variable (.99), and the within-between interaction term (.99). Finally, for the repeated measures analyses using disease severity or disease activity as the between-subjects variable, the sample of 57 children permitted a medium effect size to be detected with .57 power for the between-subjects variable, .96 power for the within-subjects variable, and .96 for the within-between interaction term.

Results

Attrition Analyses

Mann-Whitney non-parametric tests were conducted to compare the subsample of children who did not continue to participate at time 2 (N = 28) with those children who did (N = 120). No significant differences were identified between families of longitudinal and nonlongitudinal participants on demographic variables (i.e., family socioeconomic status [SES] and income, family size, and parental age and education). Only one significant difference was observed on the measures of social functioning. Nonlongitudinal participants selected themselves significantly more often than longitudinal participants for aggressive—disruptive roles on the RCP-self (Mnon = .16, SD = .95; Mlong = -.22, SD = .81; U = 1161.50, p < .05). However, these nonlongitudinal participants were not perceived as different from longitudinal participants on this dimension of the RCP from the perspectives of teachers or peers.

Demographic and Background Factors

Demographic information collected at time 1 was reanalyzed for those families of children with JRA and control families who continued to participate at the 2-year follow-up. Similar to results at time 1 (see Noll et al., 2000), the mean age for children with JRA (MJRA = 12.98 years, SD = 1.53) and controls (MCC = 13.12 years, SD = 1.55) was similar, as were estimated IQ scores (based on Vocabulary and Block Design subtests of WISC-R; MJRA = 108.2, SD = 16.2; MCC = 103.9, SD = 14.0). No significant differences were observed for SES and income, family size, and parental age and education. Duncan scores of SES (TSEI2; Nakao & Treas, 1992) indicated that, on average, the heads of households for both groups held positions in sales or management (MJRA = 49.4, SD = 20.3; MCC = 48.8, SD = 19.2).

Cross-Sectional Analyses of Social Functioning

The first set of analyses was designed to examine the social reputation and social acceptance of children with JRA and controls at the 2-year follow-up. In addition, the effects of concurrent ratings of disease severity, activity, and duration on social functioning were examined.

Social Reputation

We hypothesized that children with JRA would have difficulties with social functioning relative to controls. Independent t tests were completed for each dimension of the RCP for teacher-, peer-, and self-report. No significant differences were observed (Table I).

Table I.

Peer Relationships of Children with JRA (n = 54-57) and Controls (n = 60-63): Teacher, Peer, and Self-Nominations

  JRA   Control   
 M SD M SD ta 
The teacher, peer, and self-data reported are standardized scores based on M = 0 and SD = 1. The number of participants varies slightly between measures because not all students completed all items from all questionnaires.  
aAll t values nonsignificant.  
bThe number of peer raters per child was approximately 15-25 per class.  
RCP teacher nominations       
     Popular—leadership   .33   1.04   .09   1.01   1.29  
     Prosocial   .54   1.16   .13   1.00   2.04  
     Aggressive—disruptive   -.33   .30   -.11   0.65   -2.28  
     Sensitive—isolated   -.01   1.05   .20   1.14   -1.07  
RCP peer nominations       
     Popular—leadership   .14   1.12   -.09   0.97   1.16  
     Prosocial   .11   .85   .12   1.06   -.08  
     Aggressive—disruptive   -.27   .71   -.12   .86   -1.03  
     Sensitive—isolated   .33   1.22   .09   1.02   1.16  
RCP self-nominations       
     Popular—leadership   -.16   1.18   -.09   1.04   -.34  
     Prosocial   .00   .92   -.02   .93   .11  
     Aggressive—disruptive   -.09   .88   -.14   .77   .32  
     Sensitive—isolated   .03   1.06   .17   .99   -.73  
Sociometric measuresb      
     Best friends nominations   -.03   .85   .10   .91   -.78  
     Reciprocated friendships   .03   .91   .13   .99   -.55  
     Like ratings   .03   .90   -.01   1.09   .25  
  JRA   Control   
 M SD M SD ta 
The teacher, peer, and self-data reported are standardized scores based on M = 0 and SD = 1. The number of participants varies slightly between measures because not all students completed all items from all questionnaires.  
aAll t values nonsignificant.  
bThe number of peer raters per child was approximately 15-25 per class.  
RCP teacher nominations       
     Popular—leadership   .33   1.04   .09   1.01   1.29  
     Prosocial   .54   1.16   .13   1.00   2.04  
     Aggressive—disruptive   -.33   .30   -.11   0.65   -2.28  
     Sensitive—isolated   -.01   1.05   .20   1.14   -1.07  
RCP peer nominations       
     Popular—leadership   .14   1.12   -.09   0.97   1.16  
     Prosocial   .11   .85   .12   1.06   -.08  
     Aggressive—disruptive   -.27   .71   -.12   .86   -1.03  
     Sensitive—isolated   .33   1.22   .09   1.02   1.16  
RCP self-nominations       
     Popular—leadership   -.16   1.18   -.09   1.04   -.34  
     Prosocial   .00   .92   -.02   .93   .11  
     Aggressive—disruptive   -.09   .88   -.14   .77   .32  
     Sensitive—isolated   .03   1.06   .17   .99   -.73  
Sociometric measuresb      
     Best friends nominations   -.03   .85   .10   .91   -.78  
     Reciprocated friendships   .03   .91   .13   .99   -.55  
     Like ratings   .03   .90   -.01   1.09   .25  

To test for interactions between type of participant (JRA vs. control) with gender or with grade level (based on a median split at 7th grade), ANOVAs for all RCP dimensions were performed (2 × 2 design). All two-way interactions with type of participant were nonsignificant.

Roles related to chronic illness (i.e., “sick a lot,” “ misses a lot of school,” and “tired a lot”), physical appearance, and athletic competence were also examined using independent t tests. No significant differences were observed between children with JRA and controls according to peer report. In ANOVAs, two-way interactions between type of participant and gender or grade level were also nonsignificant for these roles.

Social Acceptance

Independent t tests were conducted on all measures of social acceptance. No significant differences were identified between children with JRA and controls on the number of times they were selected by peers as a best friend or received reciprocated friendship nominations. In addition, analyses of like ratings showed no significant differences between the two groups (Table I). ANOVAs indicated that these results did not vary by gender or grade level of the child.

Disease Severity, Activity, and Duration

Differences in social functioning were examined as a function of concurrent ratings of disease severity and disease activity. No significant differences were identified on any of the measures of social functioning (i.e., RCP dimensions, Three Best Friends, and Like Rating Scale) for children with moderate/severe disease relative to children with mild disease, nor for children with disease in remission relative to children with active disease.

Analyses were also conducted to investigate the role of disease duration. Correlations between the measures of social functioning and the number of months since the child with JRA was diagnosed were completed. No significant associations were identified.

Stability of RCP Dimensions and Measures of Friendship Across Time

Stability of individual differences in RCP scores and measures of friendship and social acceptance from time 1 to time 2 were examined using correlational analyses. This was done for the subsamples of children with JRA and controls, as well as for the sample as a whole. Although considerable stability was observed across time for most peer- and self-report measures, less stability was apparent for teacher nominations on the RCP (Table II). This might be expected, given that these indices reflected the views of a single rater and that the children had different teachers at time 2, and in some cases, were at different schools (i.e., they had started junior high school or high school). To examine if the magnitude of the correlations for each measure differed between children with JRA and controls, Fisher's r to Z transformations were performed. No significant differences in stability were observed, indicating that individuals in both subsamples tended to retain their relative position across measurement occasions.

Table II.

Stability (r) of RCP Dimensions, Measures of Friendship, and Like Ratings for Children With JRA (nJRA = 54-57), Controls (nc = 57-63), and the Entire Sample (N = 115-120)

Measure   Children with JRA   Controls   Entire Sample  
The number of participants varies slightly between measures because not all students completed all items from all questionnaires.  
*p < .01, all two-tailed tests.  
RCP teacher nominations     
     Popular—leader   .20   .15   .18  
     Prosocial   .26   .51*  .38* 
     Aggressive—disruptive   .29   .18   .22  
     Sensitive—isolated   .29   .16   .21  
RCP peer nominations     
     Popular—leader   .67*  .70*  .68* 
     Prosocial   .52*  .43*  .46* 
     Aggressive—disruptive   .54*  .63*  .60* 
     Sensitive—isolated   .61*  .67*  .62* 
RCP self-nominations     
     Popular—leader   .51*  .51*  .51* 
     Prosocial   .23   .15   .18  
     Aggressive—disruptive   .38*  .55*  .46* 
     Sensitive—isolated   .47*  .47*  .47* 
Sociometric nominations     
     Best friend nominations   .21   .46*  .36* 
     Reciprocated friendships   .20   .28   .25* 
     Like ratings   .61*  .64*  .63* 
Measure   Children with JRA   Controls   Entire Sample  
The number of participants varies slightly between measures because not all students completed all items from all questionnaires.  
*p < .01, all two-tailed tests.  
RCP teacher nominations     
     Popular—leader   .20   .15   .18  
     Prosocial   .26   .51*  .38* 
     Aggressive—disruptive   .29   .18   .22  
     Sensitive—isolated   .29   .16   .21  
RCP peer nominations     
     Popular—leader   .67*  .70*  .68* 
     Prosocial   .52*  .43*  .46* 
     Aggressive—disruptive   .54*  .63*  .60* 
     Sensitive—isolated   .61*  .67*  .62* 
RCP self-nominations     
     Popular—leader   .51*  .51*  .51* 
     Prosocial   .23   .15   .18  
     Aggressive—disruptive   .38*  .55*  .46* 
     Sensitive—isolated   .47*  .47*  .47* 
Sociometric nominations     
     Best friend nominations   .21   .46*  .36* 
     Reciprocated friendships   .20   .28   .25* 
     Like ratings   .61*  .64*  .63* 

Change in Social Functioning From Time 1 to Time 2

The effects of several disease variables on changes in social functioning over time were examined. A series of ANOVAs with repeated measures was completed with time as a within-subjects variable and either type of participant (JRA vs. control), disease severity (mild vs. moderate/severe), or disease activity (full/partial remission vs. active) as a between-subjects variable. Time 1 ratings of disease severity and activity were used for all longitudinal analyses, because these variables were considered fairly stable over time, with few children experiencing change in either disease classification. These analyses were conducted on all RCP scores and on the measures of friendship and social acceptance.

Type of Participant

Mixed factorial ANOVA with time (within-subject) and type of participant (between-subjects) revealed one significant main effect for type of participant. On average, over the 2-year period, children with JRA were selected by teachers more often for prosocial roles on the RCP than controls (collapsing across time, marginal means: MJRA = .43; MCC = .03), F(1, 116) = 6.34, p < .05.2

2

Cell means for children with JRA: at time 1 (M1 = 0.34, SD1 = 1.16) and at time 2 (M2 = 0.54, SD2 = 1.16). Cell means for controls: at time 1 (M1 = -.07, SD1 = .89) and at time 2 (M2 = 0.13, SD2 = 1.00).

One interaction between type of participant and time was observed for peer report on the sensitive—isolated dimension of the RCP, F (1, 117) = 5.28, p < .05. While the overall interaction was significant, post hoc analyses indicated that this interaction reflected trends rather than significant effects of time for children with JRA (M1 = .05, SD1 = .93; M2 = .33, SD2 = 1.22), F(1, 55) = 3.99, p < .10, and controls (M1 = .22, SD1 = 1.11; M2 = .09, SD2 = 1.02), F(1, 62) = 1.37, ns. No main effects for time were identified.

Disease Severity

For children with JRA, mixed factorial ANOVA with time 1 ratings of disease severity as the between-subjects variable revealed no significant main effects for group (mild vs. moderate/severe) or for time. However, several interactions between group and time were observed. One significant interaction was identified for the aggressive—disruptive dimension of the RCP, F(1, 54) = 6.37, p < .05. Self-reported aggressive—disruptive scores on the RCP increased significantly for children with mild disease (M1 = -.43, SD1 = .52; M2 = .12, SD2 = 1.01), F(1, 26) = 7.21, p < .05, while remaining relatively constant over time for children with moderate/severe disease (M1 = -.21, SD1 = .90; M2 = -.28, SD2 = .69), F(1, 28) = .07, ns.

Two additional interactions between disease severity group and time were observed for the number of reciprocated best friends, F(1, 52) = 5.02, p < .05, and like ratings, F(1, 54) = 7.51, p < .01. While the overall interaction was significant, post hoc analyses revealed that the number of reciprocated best friends did not change significantly over time for children with mild disease (M1 = -.19, SD1 = .92; M2 = .25, SD2 = .87), F(1, 25) = 4.24, p < .10, nor for children with moderate/severe disease (M1 = .07, SD1 = .92; M2 = -.18, SD2 = .91), F(1, 27) = 1.33, ns.3 However, like ratings deteriorated significantly for children with moderate/severe disease (M1 = .14, SD1 = .82; M2 = -.22, SD2 = .98), F(1, 28) = 5.67, p < .05, but did not significantly change for children with mild disease (M1 = .10, SD1 = .97; M2 = .31, SD2 = .74), F(1, 26) = 2.24, ns.

3

To aid in the interpretation of the Reciprocated Best Friends scores, descriptive statistics are provided using the raw data. Cell means for children with mild disease: at time 1 (M1 = 1.30, SD1 = .99) and at time 2 (M2 = 1.81, SD2 = .98). For children with severe disease: at time 1 (M1 = 1.53, SD1 = 1.04) and at time 2 (M2 = 1.45, SD2 = .83). Scores ranged from 0 to 3 for both groups of children.

Disease Activity

For children with JRA, mixed factorial ANOVAs with time 1 ratings of disease activity as the between-subjects variable were completed next. One significant main effect for group (active vs. full/partial remission) was observed. Children with active disease and children with disease in remission differed on the number of times they were selected by peers as a best friend, and these differences remained stable over time. On average, over the 2-year time period, children with active disease were chosen fewer times as a best friend than were children with their disease in remission (collapsing across time, marginal means: Mact = -.22; Mrem = .15), F(1, 55) = 4.47, p < .05.4

4

Cell means for children with active disease: at time 1 (M1 = -.31, SD1 = .76) and at time 2 (M2 = -.13, SD2 = .77). For children with their disease in remission: at time 1 (M1 = .18, SD1 = .93) and at time 2 (M2 = .12, SD2 = .95). To aid in the interpretation of the Best Friends scores, descriptive statistics are also provided using the raw data. For children with active disease: at time 1 (M1 = 2.06, SD1 = 1.48) and at time 2 (M2 = 2.41, SD2 = 1.37). Scores ranged from 0 to 6 nominations at time 1 and 0 to 5 at time 2. For children with disease in remission: at time 1 (M1 = 2.83, SD1 = 1.80) and at time 2 (M2 = 2.74, SD2 = 1.42). Scores ranged from 0 to 8 nominations at time 1 and 0 to 5 at time 2.

One significant main effect for time was observed for sensitive—isolated scores on the RCP. All children with JRA, irrespective of disease activity, experienced increasing sensitive—isolated scores on the RCP over time according to peer report, (collapsing across groups, marginal means: Mtime1 = .03; Mtime2 = .32), F(1, 54) = 4.60, p < .05.5

5

For children with active disease: at time 1 (M1 = .20, SD1 = .93) and at time 2 (M2 = .39, SD2 = 1.24). For children with their disease in remission: at time 1 (M1 = -.18, SD1 = .91) and at time 2 (M2 = .24, SD2 = 1.21).

No significant interactions between disease activity group and time were found.

Discussion

This study evaluated the social functioning of children with JRA 2 years after an initial study of their peer relationships and emotional well-being (Noll et al., 2000). We expected that social difficulties might emerge as children grew older. Methodological problems apparent in earlier work were addressed in both the initial study and this 2-year follow-up by having an appropriate comparison group, data from multiple sources, and reasonable recruitment and retention rates.

At the 2-year follow-up, cross-sectional analyses revealed that children with JRA as a group were not different from controls on any of the measures of social reputation or social acceptance. In addition, children with JRA were not different from controls in terms of a number of nonsocial attributes related to chronic illness, physical appearance, and athletic competence, suggesting that peers did not perceive these children as chronically ill. These findings did not vary by gender or by grade of the child, even though the median split for grade level was near the transition from grade school to junior high or high school. Finally, concurrent measures of disease characteristics such as severity, activity, and duration did not affect social functioning at time 2. These findings of minimal social difficulties for children with JRA are similar to our initial findings (Noll et al., 2000).

In contrast, longitudinal analyses provided evidence that social functioning deteriorated slightly over time for some children with JRA. For instance, a significant interaction between type of participant (JRA vs. control) and time was identified for peer-reported withdrawn behavior. While post hoc analyses revealed no significant change in these behaviors for either group of children, there was a trend for peers to perceive children with JRA as exhibiting increasing withdrawn behaviors over time (p = .051). In fact, the increase in withdrawn behavior over time was significant in a separate analysis involving disease activity. These results need to be replicated, particularly with longer-term longitudinal research, as these trends may continue and strengthen over time. Additionally, children with more severe disease at time 1 had declines in like ratings over time, whereas these scores did not change significantly for children with mild disease. With respect to disease activity at time 1, children with active disease were chosen fewer times as a best friend at both assessment points than were children with disease in remission.

These findings are consistent with a number of studies in which children with JRA, particularly adolescents and children with more severe disease, were found to be at risk for experiencing social difficulties. For instance, studies have reported that older children with severe disease participated in fewer activities with friends than controls, according to self-report (Billings et al., 1987), and exhibited lower social competence according to the Child Behavior Checklist (CBCL; Daltroy et al., 1992). In addition, Wilkinson (1981) observed that adolescents ages 12 to 19 with severe JRA suffered from social isolation as measured by the number of peer contacts they had during a 1-month period. While these studies used cross-sectional designs, Timko et al. (1992) assessed a group of children with JRA and their healthy siblings two times over 1 year. At both assessments, children with JRA participated in fewer activities with friends than healthy siblings, and those with severe disease had fewer very good friends than children with mild disease or siblings.

Despite the comparable findings of some social difficulties for children with JRA, this study was able to extend previous work in several ways. First, many researchers have described social difficulties primarily in terms of a decrease in the frequency of contacts with friends, as well as the number of activities in which the child with JRA participated. However, such a decrease may be the result of functional limitations and life disruptions (e.g., repeated visits to the doctor) that inhibit participation in activities rather than poor social functioning (Daltroy et al., 1992; Drotar, Stein, & Perrin, 1995). This study suggested greater risk for difficulties in peer relationships for some children with JRA, specifically in social withdrawal and social acceptance. Peer relationships are considered central to healthy emotional and social development (Parker & Asher, 1987), and there is evidence that an early pattern of isolated and withdrawn behavior is a risk factor for future difficulties with social acceptance by peers, negative self-perceptions, excessive anxiety, and depression (Hymel et al., 1990; Rubin & Mills, 1988). Over time, participation in fewer activities and diminished peer contact may lead to having fewer friends (Daltroy et al., 1992).

Using multiple sources of information about social difficulties was another means by which this study extended previous findings. Most studies have reported social difficulties for children with JRA from only one perspective, usually the parent. In this study, social difficulties were assessed from multiple sources. Reliance on parent report when assessing social functioning may be particularly problematic, because parents primarily observe their children at home rather than during direct interactions with peers. For instance, the activities and social subscales of the CBCL social competence measure have demonstrated poor concordance with children's self-ratings and with peer nominations (Schneider & Byrne, 1989). In this study, most of the findings indicating social difficulties were based on peer report. Peers may be the best judges of a child's social functioning, as they have the most daily interactions with the child. In addition, peer report in this study consisted of multiple raters (i.e., the child's classmates) of each child, thereby providing a more stable measure of social functioning than single rater sources (e.g., teacher).

Several findings in this investigation that were not hypothesized are noteworthy. First, teachers perceived children with JRA as being more prosocial than controls over time. While this seems to be an indication of better functioning for children with JRA, peers did not perceive them in this manner. It is conceivable that the withdrawn, and therefore nondisruptive, behavior of children with JRA is interpreted as prosocial by teachers. Second, self-report indicated that children with mild disease perceived themselves as becoming more similar over time to their classmates with respect to aggressive and disruptive behaviors, whereas children with severe disease perceived themselves as remaining the same over time, specifically that they engaged in less aggressive and disruptive activity than peers. These findings suggest that children with mild disease may view themselves as being less passive and as adjusting over time to the physical limitations of the disease. In contrast, children with more severe disease may continue to perceive themselves as being hindered by their physical limitations (e.g., chronically fatigued). Finally, a significant interaction between disease severity (remission vs. active) and time was identified for the number of reciprocated best friends. While post hoc analyses revealed no significant change over time in these behaviors for either group of children, there was a trend for children with mild disease to exhibit increasing reciprocated best friends over time (p = .050). Improvement in social acceptance for children with mild disease may be a result of the less disruptive nature of their disease.

There are several limitations to this investigation. First, this study obtained data from children at one treatment center. Replication at different centers using a multisite approach is necessary. Second, a large number of statistical tests were completed on this sample. Thus, spurious effects must be considered when interpreting these results, and significant findings need to be replicated. Finally, the cross-sectional results of this study appeared to indicate that children with JRA exhibited considerable psychological hardiness. These results support the theory that the social development of children is highly protected, with maladjustment occurring only with exposure to severe adversity (e.g., family disruption or central nervous system damage; Masten, Best, & Garmezy, 1990). For example, children with central nervous system tumors demonstrate considerable social deficits (Vannatta, Gartstein, et al., 1998), while children with other forms of pediatric cancer do not (Noll et al., 1999). However, longitudinal analyses revealed subtle decrements in social functioning over time for children with more severe or active disease. These findings support disability/stress models that suggest that chronic illness creates ongoing strain, requiring continuous adjustment, with the potential for disrupting a child's psychosocial development (Wallander, Varni, Babani, Banis, & Wilcox, 1989). Moreover, these results appear to fit a dose response model of disease effects in which children with more severe or active disease have increasing difficulties over time compared with children with milder disease.

This investigation is one of the few studies to examine social functioning of children with chronic illness longitudinally. Future work might focus on extending longitudinal studies, particularly into young adulthood. It is also important to understand whether the slight deterioration in social functioning observed in this study is associated with more clinically significant social problems over time or with the development of difficulties in emotional well-being, particularly internalizing symptoms. Thus, further research is needed to examine the social functioning and emotional well-being of these children later in life.

In summary, our findings suggest that children with more severe or active JRA are at risk for a slight increase in social difficulties over time. From a clinical perspective, effect sizes in this study were small and, therefore, may not be sufficient to warrant intervention for most children with JRA. However, there may be subgroups of children with JRA who experience increasing difficulties over time and would be appropriate targets for interventions to ameliorate or prevent such difficulties in the future. Thus, future research should focus on identifying the mechanisms that result in increasing social difficulties over time for some children with JRA. Such work is imperative for the evaluation and development of interventions to improve the quality of social functioning among these children.

This paper was reviewed and accepted during the term of the previous editor, Anne E. Kazak, PhD, ABPP.

This study was funded in part by a grant from the National Arthritis Foundation. The authors wish to acknowledge the special contribution of the families and schools participating in this project. Portions of this article have been presented at the annual meeting of the American Psychological Association, Washington, DC, August 2000.

References

Asher, S. R., Singleton, L. C., Tinsley, B. R., & Hymel, S. (
1979
). A reliable sociometric measure for pre-school children.
Developmental Psychology
 ,
15
(4),
443
-444.
Billings, A. G., Moos, R. H., Miller, J. J., & Gottlieb, J. E. (
1987
). Psychosocial adaptation in juvenile rheumatic disease: A controlled evaluation.
Health Psychology
 ,
6
(4),
343
-359.
Bukowski, W. M., & Hoza, B. (
1989
). Popularity and friendship: Issues in theory, measurement, and outcomes. In T. Berndt & G. Ladd (Eds.),
Contributions of peer relationships to children's development
  (pp.
15
-45). New York: Wiley and Sons.
Casiglia, A. C., Lo Coco, A., & Zappulla, C. (
1998
). Aspects of social reputation and peer relationships in Italian children: A cross-cultural perspective.
Developmental Psychology
 ,
34
(4),
723
-730.
Cassidy, J., Levinson, J., Bass, J., Baum, J., Brewer, E., Fink, C., Hanson, V., Jacobs, J., Masi, A., Schaller, J., Fries, J., McShane, D., & Young, D. (
1986
). A study of classification criteria for a diagnosis of juvenile rheumatoid arthritis.
Arthritis and Rheumatism
 ,
29
(2),
274
.
Cassidy, J. T., & Petty, R. E. (
1995
).
Textbook of pediatric rheumatology
  (3rd ed.). Philadelphia: Saunders.
Daltroy, L. H., Larson, M. G., Eaton, H. M., Partridge, A. J., Pless, I. B., Rogers, M. P., & Liang, M. H. (
1992
). Psychosocial adjustment in juvenile arthritis.
Journal of Pediatric Psychology
 ,
17
(3),
277
-289.
Drotar, D., Stein, R. E. K., & Perrin, E. (
1995
). Methodological issues in using the child behavior checklist and its related instruments in clinical child psychology research.
Journal of Clinical Child Psychology
 ,
24
(2),
184
-192.
Faul, F., & Erdfelder, E. (
1992
).
GPOWER: A priori, post-hoc, and compromise power analyses for MS-DOS [computer program]
 . Bonn: Bonn University, Department of Psychology.
Gottman, J., Gonso, J., & Rasmussen, B. (
1975
). Social interaction, social competence, and friendship in children.
Child Development
 ,
46
,
709
-718.
Graetz, B., & Shute, R. (
1995
). Assessment of peer relationships in children with asthma.
Journal of Pediatric Psychology
 ,
20
(2),
205
-216.
Harris, J. A., Newcomb, A. F., & Gewanter, H. L. (
1991
). Psychosocial effects of juvenile rheumatic disease: The family and peer systems as a context for coping.
Arthritis Care and Research
 ,
4
,
123
-130.
Holland, B. S., & Copenhauer, M. D. (
1988
). Improved Bonferroni-type multiple testing procedure.
Psychological Bulletin
 ,
104
,
145
-149.
Hymel, S., Rubin, K. H., Rowden, L., & Lemare, L. (
1990
). Children's peer relationships: Longitudinal prediction of internalizing and externalizing problems from middle childhood.
Child Development
 ,
61
,
2004
-2021.
Ladd, G. (
1981
). Effectiveness of a social learning method for enhancing children's social interaction and peer acceptance.
Child Development
 ,
52
,
171
-178.
Lawrence, R. C., Helmick, C. G., Arnett, F. C., Deyo, R. A., & Felson, D. T. (
1998
). Estimates of the prevalence of arthritis and selected musculoskeletal disorders in the United States.
Arthritis & Rheumatism
 ,
41
(5),
778
-799.
Lovell, D. J., Athreya, B., Emery, H. M., Cribbas, D. L., & Levinson, J. E. (
1990
). School attendance and patterns, special services, and special needs in pediatric patients with rheumatic diseases.
Arthritis Care Research
 ,
3
,
196
-203.
Luthar, S. S. (
1995
). Social competence in the school setting: Prospective cross-domain associations among inner-city teens.
Child Development
 ,
66
,
416
-429.
Luthar, S. S., & McMahon, T. J. (
1996
). Peer reputation among inner-city adolescents: Structure and correlates.
Journal of Research on Adolescence
 ,
6
(4),
581
-603.
Masten, A. S. (
1989
). Resilience in development: Implications of the study of successful adaptation for developmental psychopathology. In D. Cicchetti (Ed.),
The emergence of a discipline: Rochester Symposium on Developmental Psychopathology
  (Vol.
1
, pp.
261
-294). Hillsdale, NJ: Lawrence Erlbaum.
Masten, A. S., Best, K. M., & Garmezy, N. (
1990
). Resilience and development: Contributions from the study of children who overcome adversity.
Development and Psychopathology
 ,
2
,
425
-444.
Masten, A. S., Morison, P., & Pellegrini, D. S. (
1985
). A revised class play method of peer assessment.
Developmental Psychology
 ,
21
(3),
523
-533.
Morison, P., & Masten, A. S. (
1991
). Peer reputation in middle childhood as a predictor of adaptation in adolescence: A seven-year follow-up.
Child Development
 ,
62
,
991
-1007.
Nakao, K., & Treas, J. (
1992
).
The 1989 Socioeconomic Index of Occupations: Construction from the 1989 Occupational Prestige Scores
  (General Social Survey Methodological Report No. 74). Chicago: University of Chicago, National Opinion Research Center.
Noll, R., Bukowski, W. M., Rogosch, F. A., LeRoy, S. S., & Kulkarni, R. (
1990
). Social interactions between children with cancer and their peers: Teacher ratings.
Journal of Pediatric Psychology
 ,
15
,
43
-56.
Noll, R., Gartstein, M., Vannatta, K., Correll, J., Bukowski, W., & Davies, W. (
1999
). Social, emotional, and behavioral functioning of children with cancer.
Pediatrics
 ,
103
(1),
71
-78.
Noll, R., Kozlowski, M. A., Gerhardt, C., Vannatta, K., Taylor, J., & Passo, M. (
2000
). Social, emotional, and behavioral functioning of children with juvenile rheumatoid arthritis.
Arthritis & Rheumatism
 ,
43
,
1387
-1396.
Noll, R., LeRoy, S., Bukowski, W., Rogosch, F., & Kulkarni, R. (
1991
). Peer relationships and adjustment of children with cancer.
Journal of Pediatric Psychology
 ,
16
,
307
-326.
Ollendick, T. H., Greene, R. W., Weist, M. D., & Oswald, D. P. (
1990
). The predictive validity of teacher nominations: A five-year follow-up of at-risk youth.
Journal of Abnormal Child Psychology
 ,
18
(6),
699
-713.
Parker, J. G., & Asher, S. R. (
1987
). Peer relations and later personal adjustment: Are low-accepted children at risk?
Psychological Bulletin
 ,
102
(3),
357
-389.
Rubin, K. H. (
1993
). The Waterloo Longitudinal Project: Correlates and consequences of social withdrawal from childhood to adolescence. In K. H. Rubin & J. B. Asendorpf (Eds.),
Social withdrawal, inhibition, and shyness in childhood
  (pp.
291
-314). Hillsdale, NJ: Lawrence Erlbaum.
Rubin, K. H., & Mills, R. S. (
1988
). The many faces of social isolation in childhood.
Journal of Consulting and Clinical Psychology
 ,
56
,
916
-924.
Schneider, B. H., & Byrne, B. M. (
1989
). Parents rating children's social behavior: How focused the lens?
Journal of Clinical Child Psychology
 ,
18
,
237
-241.
Timko, C., Stovel, K. W., Moos, R. H., & Miller J. J. (
1992
). Adaptation to juvenile rheumatic disease with a one-year follow-up.
Health Psychology
 ,
11
(1),
67
-76.
Ungerer, J. A., Horgan, B., Chaitow, J., & Champion, G. D. (
1988
). Psychosocial functioning in children and young adults with juvenile arthritis.
Pediatrics
 ,
81
(2),
195
-202.
Vandvik, I. H., & Hoyeraal, H. M. (
1993
). Juvenile chronic arthritis: A biobehavioral disease. Some unsolved questions.
Clinical and Experimental Rheumatology
 ,
11
,
669
-680.
Vannatta, K., Gartstein, M., Short, A., & Noll, R. (
1998
). A controlled study of peer relationships of children surviving brain tumors: Teacher, peer, and self-ratings.
Journal of Pediatric Psychology
 ,
23
,
279
-88.
Vannatta, K., Zeller, M., Noll, R., & Koontz, K. (
1998
). Social functioning of children surviving bone marrow transplantation.
Journal of Pediatric Psychology
 ,
23
(3),
169
-178.
Wallander, J. L., Varni, J., Babani, L., Banis, H. T., & Wilcox, K. T. (
1989
). Family resources as resistance factors for psychological maladjustment in chronically ill and handicapped children.
Journal of Pediatric Psychology
 ,
14
(2),
157
-173.
Wilkinson, V. A. (
1981
). Juvenile chronic arthritis in adolescence: Facing the reality.
Internal Rehabilitation Medicine
 ,
3
(1),
11
-17.
Zeller, M., Vannatta, K., Schafer, J., & Noll, R. (
1999
).
The Revised Class Play: A cross-age perspective
 . Paper presented at the biennial meeting of the Society for Research in Child Development, Albuquerque, NM.