Psychosocial Adjustment in Children With Kidney Disease

Journal of Pediatric Psychology, Vol. 25, No. 2, 2000, pp. 93–103

Psychosocial Adjustment in Children With Kidney Disease Elizabeth Soliday,1 PhD, Elizabeth Kool,1 BS, and Marc B. Lande,2 MD 1

Washington State University Vancouver and 2Oregon Health Sciences University

Objective: To examine family environment, levels of parenting stress, and child behavior problems in children with one of three kidney diseases compared to healthy children and to examine predictors of psychological distress in the full sample. Method: Parents of children with steroid sensitive nephrotic syndrome, chronic renal insufficiency, or kidney transplant (n ⫽ 41) were compared to 34 healthy children of similar demographic characteristics. Results: Mean scores on family functioning, parenting stress, and child behavior were within normal limits. Family environment variables significantly predicted child behavior and parenting stress for parents of ill and healthy children. Qualitative responses provided insight into developmentally specific stressors and intervention needs in the illness groups. Conclusions: These data indicate that long-term survivors of kidney disease function similarly to demographically matched peers and that the family environment may buffer stress caused by illness. Specific concerns raised by parents in the kidney disease groups indicate the need to appropriately assess and intervene with this understudied population. Key words: chronic illness; children; family functioning; nephrology; psychosocial adjustment.

Since the 1970s, treatments for kidney disease in children have improved greatly. While pediatric kidney diseases are relatively rare, with annual incidence rates ranging from 11 to 69 children per million (Wyatt, Kagy, & Kritchevsky, 1996), improved dialysis techniques, surgical procedures for kidney transplantation, and new medications provide a better prognosis for affected children. However, like most chronic illnesses of childhood, chronic kidney disease seriously affects children’s lives as they negotiate the stress and responsibilities associated with disease management and the prospect of a All correspondence should be sent to Elizabeth Soliday, Psychology and Human Development, Washington State University Vancouver, 14204 Salmon Creek Ave., Vancouver, Washington 98686. E-mail: [email protected].

䉷 2000 Society of Pediatric Psychology

shortened life span (Frauman & Lansing, 1983; Hobbs & Sexson, 1993). In addition, these illnesses would understandably affect the family. Research has found that the impact of chronic kidney disease on families varies. Clinicians generally conclude that factors such as age of onset, family structure, and the disease’s clinical features influence psychosocial adjustment and outcome (Fukunishi & Kudo, 1995; Schweitzer & Hobbs, 1995; Vance & Pless, 1983), although many assumptions remain to be supported with empirical evidence. As children live longer with chronic illnesses such as kidney disease, understanding long-term psychosocial sequelae becomes increasingly important to clinicians and families. To identify variables influencing long-term psy-

94

chosocial adjustment across kidney disease subtypes, we examined the psychosocial adjustment of children with one of three kidney disease subtypes differing according to intrusiveness of treatment. The first of these, steroid sensitive nephrotic syndrome (SSNS), is characterized by massive urinary loss of protein resulting in total body edema. Although the majority of children can achieve remission with corticosteroids, most have a chronically relapsing course. About half of children with SSNS relapse frequently, requiring several 8–12 week courses of steroids throughout the year. Although 95% of children with SSNS outgrow their disease without long-term detrimental effects, the average duration is approximately 10 years (Warshaw, 1994). Second, chronic renal insufficiency (CRI) is partial loss of kidney function. CRI is invariably progressive, with many children eventually requiring dialysis or kidney transplantation. As kidney function deteriorates, the impact on health becomes more pronounced; children require numerous medications and more frequent physician visits (Wassner, 1994). And, third, end-stage renal disease (ESRD) is reached when native kidney function can no longer sustain life. Kidney transplantation is the treatment of choice for children with ESRD. Children who receive kidney transplants face the necessity of lifelong immunosuppression treatment to prevent organ rejection, and most eventually need retransplantation (Grimm & Ettenger, 1992). Available studies on psychosocial adjustment of children with SSNS and CRI primarily focus on the prevalence of clinically significant distress. Mehta, Bagga, Pande, Bajaj, and Srivastava (1995) reported significantly higher mean behavior problem scores in children with SSNS than those for healthy controls. Overall, mean scores for both groups were within normal limits, and the expected proportion of children in both groups scored above clinical cutoffs. In another comparison between children with SSNS and healthy controls, scores on behavior rating scales, teacher reports, and a children’s selfperception scale were not significantly different between groups. However, children with SSNS showed a trend toward more introvertive traits (Vance & Pless, 1983). Garralda, Jameson, Reynolds, and Postlethwaite (1988) compared psychiatric adjustment in children with CRI, hemodialysis patients, and healthy children. They reported a trend toward higher internalizing symptoms (i.e., anxiety, worrying, depressive symptoms) for children in the illness groups. McEvoy (1990) reported elevated

Soliday, Kool, and Lande

behavior problems and lower social competence in a comparison between transplanted children and healthy siblings, though scores for both groups were within normal limits. Although these studies reported behavioral functioning within normal limits, findings of significant differences between children with kidney disease and healthy controls were not consistent. These discrepancies could reflect differences in ways researchers measure children’s behavioral adjustment. Two of the above studies (McEvoy, 1990; Mehta et al., 1995) used the Child Behavior Checklist (Achenbach, 1992), which was standardized with healthy children. Researchers have suggested that instruments standardized with healthy populations are not adequately sensitive to detect subtle illness-related differences between chronically ill and healthy children (Harris, Canning, & Kelleher, 1996; Perrin, Stein, & Drotar, 1991). In addition, authors in these studies did not report whether children were assessed during disease remission or relapse periods, when medication could affect behavior. Only one of the studies reported duration of diagnosis. Garralda et al. (1988) reported duration of diagnosis, but duration was treated as a categorical rather than continuous variable. Any behavioral differences between children with kidney disease and healthy controls in the long term remain to be clarified. In addition to the need to clarify the functioning of children with kidney disease relative to healthy children, the impact of moderating factors on children’s adjustment merits further attention. Because of the tremendous impact of children’s chronic illness on the entire family, researchers have examined how family environment variables influence chronically ill children’s and other family members’ adjustment. Family systems theorists propose that a positive family environment can buffer the effects of ongoing stressors (Moos & Moos, 1994). Positive family environments have been characterized by high cohesion, which is the support family members provide one another. In addition, a high degree of encouragement to express emotions directly, called expressiveness, and low levels of conflict among family members also characterize a positive family environment. Research examining family factors in children’s adjustment to kidney disease generally supports family systems assumptions. In one of the few available studies linking family environment variables to children’s adjustment, Fukunishi and Kudo (1995)

Children With Kidney Disease

compared adjustment of children on dialysis, with kidney transplants, and healthy controls. The authors reported positive correlations between family functioning (cohesion, expressiveness) and school adjustment. In a study of children with CRI, Crittenden and Holaday (1986) assessed children who had received treatment for CRI for at least 2 years. They found a significant, negative correlation between parents’ reports of stressful life events and staff ratings of children’s behavioral adaptation. Although these studies lend limited support to family systems assumptions, virtually no studies have examined the predictive influence of family environment variables on adjustment outcomes in the kidney disease population. We therefore reviewed studies on other chronic illnesses to provide a starting point for our research. In a study of children newly diagnosed with cancer, family cohesion was significantly, negatively correlated with parents’ depressive symptoms and child behavior problems (Manne et al., 1995). Hamlett, Pellegrini, and Katz (1992) reported that higher family cohesion and conflict predicted externalizing behavior scores in a sample of children with asthma and diabetes. Neither the direction of effect nor duration of diagnosis was reported. In contrast, high inflexibility, considered a less positive family attribute, has been associated with better outcomes in children with diabetes (Klemp & LaGreca, 1987). Other diabetes researchers have found no significant relationships between family system variables and adjustment outcomes (Kovacs, Kass, Schnell, Goldston, & Marsh, 1989). Previous studies provide some insight into the functioning of children with kidney disease and other chronic illnesses but have not consistently shown increased adjustment problems in children with kidney disease compared to healthy children. We therefore compared functioning between these two groups. Few studies on the psychosocial functioning of children with CRI and SSNS have been conducted, and we included these diagnostic groups in addition to children with ESRD. We examined the predictive effects of family environment on parents’ and children’s adjustment, as this model has not been systematically applied to the kidney disease population. We also examined the effects of kidney disease subtype on the issues we have described. Although previous studies have revealed that chronic illness affects global factors such as family environment and children’s behavior similarly across disease categories, researchers

95

recommend that disease category continue to be examined as a factor in psychosocial functioning (Holden, Chmielewski, Nelson, Kager, & Foltz, 1997; Lavigne & Faier-Routman, 1992; Wallander, Varni, Babani, Banis, & Wilcox, 1988). Finally, of the studies reporting duration of illness, the majority were conducted in the early postdiagnosis period, that is, 1 year or less (e.g., Manne et al., 1995; Northam, Anderson, Adler, Werther, & Warne, 1996). To provide information on the long-term sequelae of kidney disease in children, we assessed families who had been living with their children’s disease at least 1 year. Results of our study will contribute to the existing literature by providing heuristic insight on how family system variables influence the global adjustment of children with kidney disease. Research findings may provide guidance to help clinicians appropriately target educational and supportive interventions and to inform health care providers on families’ needs. Thus, we proposed the following questions and hypotheses: (1) what is the prevalence and severity of child behavior problems and parenting stress in a sample of children with no chronic illness, or SSNS, CRI, or ESRD 1 year or more postdiagnosis? Based on previous research, we expected elevated rates of parenting stress and behavior problems in the kidney disease subgroups compared to norms, but we expected both stress and behavior problems to be within normal limits. Based on previous research finding few disease category differences on global measures (Garralda et al., 1988; Wallander et al., 1988), we predicted no differences by disease subtype. (2) How do family environment variables predict parenting stress and children’s behavioral adjustment across groups? Based on previous research, we expected that families with higher cohesion, higher expressiveness, and lower conflict would have more positive child behavior and lower parenting stress. As in question 1, we expected no effects for kidney disease subtype.

Method Participants and Recruitment Before conducting this study, we obtained approval from each participating institution’s respective institutional review board (IRB). Seventy five parents with children ranging from 2–18 years old who spoke English as their primary language agreed to

96

participate. Forty-one families had children with kidney disease: 15 (36.59%) were diagnosed with SSNS; 12 (29.27%) had CRI; 14 (34.15%) had received a kidney transplant. Children with kidney disease were recruited from the Pediatric Nephrology clinic in a Pacific Northwest teaching hospital. Fifteen additional families were contacted but declined participation; reasons given included lack of time, dissatisfaction with hospital services, change of address with no forwarding information, and no reason given. Analyses of demographics indicated no significant differences on demographic characteristics (i.e., age and gender of child, duration of diagnosis, parent status) between participants and nonparticipants. In the kidney disease group, children’s ages ranged from 1.83 to 18.09 years (SD ⫽ 5.09), M ⫽ 10.86 (SD ⫽ 4.35) years. The average age of diagnosis for children with kidney disease was 3.53 (SD ⫽ 4.18) years, and they had been diagnosed with their illness an average of 7.33 years (SD ⫽ 5.66). The broad age range represents the relative rarity and age distribution of the kidney diseases we studied. Eighty percent of children with SSNS are younger than 6 years old at presentation of illness. By contrast, only 18% of children with ESRD are younger than 6 years old at the time of renal transplantation and 39% are 13–17 years (Kohaut & Tejani, 1996; Wyatt et al., 1996). In our sample, 50 (66.77%) children were 0–12 years old, and 25 (33.33%) were older than 12 years. Twenty-seven (65.85%) were male. A comparison sample of 34 families was recruited from a regional teaching center. An attempt was made to recruit participants comparable to the kidney disease sample on geographic location, children’s age, gender, family structure, income level, education, and ethnicity. In both groups, most children were Anglo from middle-income families (Hollingshead classes II–IV, see Table I). Chi-square analyses revealed no significant differences between groups on gender of child, social class, or ethnicity. The CRI and healthy samples contained significantly more single parents than the SSNS and transplant groups, ␹2 ⫽ 9.41, p ⬍ .05 (see Table 1). Similar to the kidney disease sample, the age range of children in the comparison sample was 1.93 to 18.17 years, (M ⫽ 9.54, SD ⫽ 4.35). A one-way ANOVA with follow-up comparisons (Scheffe`) revealed that transplant group children were significantly older than the SSNS and healthy samples, F(3, 68) ⫽ 4.53, p ⬍ .01 (see Table 1). In both groups, the majority of respondents were mothers, 38 (92.68%) in the


Table I.

Sample Characteristics of Each Diagnostic Group, n (%) SSNS (n ⫽ 15)

CRI (n ⫽ 12)

TX (n ⫽ 14)

HE (n ⫽ 34)

4 (26.7) 11 (73.3)

7 (58.3)

3 (21.4)

12 (35.3)

5 (41.7)

11 (78.6)

20 (58.8)

␹2/F

Gender Female Male

4.55

Parent Single

4 (26.7)

7 (58.3)

2 (14.3)

18 (52.9)

Two

9 (60.0)

5 (41.7)

12 (85.7)

13 (38.2)

I & II

4 (26.7)

5 (41.7)

2 (14.3)

6 (17.6)

III

5 (33.3)

5 (41.7)

2 (14.3)

7 (20.6)

IV & V

6 (40.0)

1 (8.3)

10 (71.4)

16 (47.1)

11 (73.3)

11 (91.7)

12 (85.7)

28 (82.4)

3 (20.0)

1 (8.3)

1 (7.1)

3 (8.8)

9.41*

Hollingshead

10.83

Ethnicity Anglo Minority

10.73

Child’s age M

8.08

10.42

14.21

9.55

SD

3.91

6.40

4.79

4.35

4.53*

SSNS ⫽ steroid sensitive nephrotic syndrome; CRI ⫽ chronic renal insufficiency; TX ⫽ transplant; HE ⫽ healthy. Values do not always add to 100% due to occasional missing data. *p ⬍ .05.

kidney disease group, and 30 (88.24%) in the comparison sample. Measures Family Information. Family structure, parents’ education level, occupation, marital status, and age and gender of child were collected. Hollingshead’s (1975) index was used to calculate social status. Family Environment Scale (FES; Moos & Moos; 1994). This 90-item true/false scale measures family functioning. We selected the three subscales, cohesion, expressiveness, and conflict, due to their previously demonstrated relationships with parent and child adjustment outcomes. These subscales have adequate internal consistency reliabilities, with ␣ coefficients ranging from .61 to .78. Testretest reliabilities are also adequate, ranging from r ⫽ .68–.86. Child Behavior Checklist (CBCL; Achenbach, 1991, 1992). The CBCL is designed to assess behavior problems and competencies; forms are available for children 2–3 and 4–18 years old. We selected the CBCL due to its widespread use in research on children’s behavior in a variety of populations. The CBCL also provides T scores, allowing for pooling and comparison of the scores across the age range of our sample. Finally, the CBCL contains two openended questions asking what concerns the parent


most about the child and what are the best things about the child. We felt that analyses of these questions in addition to the quantitative data would help address previous researchers’ concerns related to the inadequate sensitivity of most standardized assessments for use with chronic illness populations. The quantitative portion of the CBCL consists of 118 items assessing internalizing and externalizing behavioral dimensions. The CBCL has excellent reported internal consistency reliabilities: ␣ ⫽ .89– .93, varying by age and gender of child. One week test-retest reliabilities ranged from .87 to .95, varying by age of child. Normative data are available. Parenting Stress Index-Short Form (PSI-SF; Abidin, 1995). The PSI-SF is designed to assess stress in the parent-child system. Parents responded on a Likerttype scale; higher scores indicated higher parenting stress. The 36 items form a global rating of stress in the parent-child system consisting of three subscales: parental distress, parent-child interaction, and difficult child. The PSI-SF has high reported internal consistency reliabilities: ␣ for the total score was .91. Normative data are available for children 0–12 years old. Procedure Following IRB approval, parents of children with kidney disease were recruited by phone or letter by the third author (ML), who obtained informed consent. Comparison families were recruited by the first author (ES). Parents were told we were trying to learn more about how families with children who have a variety of illnesses are doing. Packets containing demographic forms, the FES, PSI, CBCL scales, and a postage-paid return envelope were mailed to the families whose children had kidney disease. Families who returned completed questionnaire packets received $15.00 reimbursement for their time. Parents in the comparison group completed a consent form and questionnaires at their convenience and returned them to the first author.

Results Preliminary Analyses Validity Check. As a validity check, the Defensive Responding subscale of the PSI was examined. Scores below 10 indicate defensive responding (Abidin,

97

Table II. Correlations Between Measured Variables

1. INT-T 2. EXT-T 3. PSI 4. COHES 5. EXPRESS 6.CONFLICT

1

2

3

4

5

6

—

.58**

.62**

⫺.35**

⫺.30*

.32*

—

.61**

⫺.33*

⫺.22

.48**

—

⫺.50**

⫺.43**

.39*

.40*

⫺.49**

—

⫺.15 —

INT-T ⫽ CBCL internalizing T-score; EXT-T ⫽ CBCL externalizing T-score; PSI ⫽ Parenting Stress Index; COHES ⫽ FES cohesion scale; EXPRESS ⫽ FES expressiveness scale; CONFLICT ⫽ FES conflict scale. *p ⬍ .05. ** p ⬍ .005.

1995). For the entire sample, the mean score was 15.42 (SD ⫽ 4.95), indicating valid responses. A one-way ANOVA revealed no significant differences across groups. Scale Reliabilities. Internal consistency reliabilities (Cronbach’s ␣) were conducted on CBCL internalizing and externalizing scales, FES cohesion, expressiveness, and conflict subscales, and the PSI total stress score. For the CBCL 2/3-year-old internalizing scale, ␣ ⫽ .95; CBCL 2/3 externalizing ␣ ⫽ .94. For CBCL 4/18-year-old internalizing scale, ␣ ⫽ .93, and externalizing ␣ ⫽ .90. On the FES cohesion subscale, ␣ ⫽ .67. The ␣ for the expressiveness subscale was .70, and for the conflict scale, .71. On the PSI, ␣ for 0–12-year-olds was .89, and for children 12–18, .89. Scale reliabilities were consistent with published norms. Correlations Between Measured Variables. Pearson’s rs were conducted between measured variables (see Table II). Correlations between CBCL internalizing, externalizing, and PSI scores were significant ( p ⬍ .005). Several significant correlations also resulted between adjustment measures and family functioning variables. Correlations were in expected directions: for example, CBCL scores correlated positively with PSI scores, indicating that parents reporting higher stress also perceived their children to have higher rates of behavior problems. Negative correlations between family adjustment, CBCL, and PSI scores indicated that parents reporting more positive family adjustment perceived lower rates of child behavior problems and lower rates of parenting stress. Children’s age had low correlations with measured variables, ranging from ⫺.01 to.16 (CBCL internalizing; p ⫽ .19). In the kidney disease groups, correlations were conducted between duration of diagnosis and measured variables. Longer duration of diagnosis correlated with lower externalizing

98

behavior problems, r ⫽ ⫺.38, p ⬍ .05; duration of diagnosis significantly correlated with family conflict, r ⫽ ⫺.47, p ⬍ .01. Correlations with other variables were nonsignificant. Effects of Demographics on Measured Variables. To test for gender effects, t tests were conducted between females (n ⫽ 26) and males (n ⫽ 47) on CBCL subscales, PSI total stress, and FES subscales. Gender was not reported on two children. Mean scores between genders were highly similar and none of the t tests approached statistical significance. T tests were also conducted to test for differences between single- and two-parent families. The difference between parent groups approached significance on externalizing behavior (t ⫽ 1.96, p ⫽.05) and on parenting stress (t ⫽ ⫺1.88, p ⫽ .07). On both variables, single parents scored significantly higher. One-way ANOVAs were conducted to test for effects of Hollingshead rating on internalizing and externalizing scores, PSI, and FES scale scores. Due to small cell sizes in classes I and V, they were combined with classes II and IV, respectively. No significant differences resulted on measured variables by social class.


Table III. Descriptive Statistics on Measured Variables Variable

Healthy (n ⫽ 34)

Typical Scoresa

CBCL Int-T M

52.66

51.23

50.0

SD

13.10

12.28

9.8

Range

30–87

33–87

23–100

M

51.07

50.74

50.0

SD

11.52

9.76

9.9

Range

32–72

35–77

23–100

M

76.50

78.16

69

SD

16.63

16.58

—

Range

45–116

52–110

39–112

M

7.41

7.29

6.80

SD

1.86

1.79

2.02

Range

1–9

2–10

0–10

M

5.92

6.06

5.68

SD

1.94

1.79

1.78

Range

1–9

3–9

0–10

M

3.54

3.42

3.76

SD

2.15

1.82

2.32

Range

1–8

1–8

0–10

CBCL Ext-T

PSI

FES Cohesion

FES Express

FES Conflict

Major Quantitative Analyses a

Mean Scores on Standardized Measures and Clinically Significant Distress. To evaluate the prevalence and severity of parenting stress and child behavior problems in our sample, CBCL internalizing and externalizing, PSI, and FES subscale scores were calculated for each diagnostic group, that is, SSNS, CRI, ESRD, and healthy. (Statistical tests for group differences were conducted in regression analyses using dummy coded variables for SSNS, CRI, ESRD, and healthy categories.) Descriptive statistics on measured variables are presented in Table III. Mean scores for all groups on CBCL internalizing and externalizing symptoms were within normal limits (i.e., T ⬍ 65). Parenting stress scores were also within normal limits (i.e., ⬍ 90) across groups. Family cohesion, conflict, and expressiveness were in the average to above-average range. The number of participants scoring above cutoffs for clinical distress on the CBCL scales and PSI were also examined. On the CBCL internalizing scale, an expected proportion (i. e., ⬍ 15%) of children in the SSNS, CRI, and healthy groups scored above clinical cutoffs. However, four (28.6%) of children with transplants scored above the CBCL internalizing cutoff. On the CBCL externalizing scale, less than 15% of children scored above clini-

Kidney Disease (n ⫽ 41)

Typical scores are those given for standardization samples. CBCL scores will vary slightly by age and gender of child; T scores presented here are for boys, 4–18 years old. PSI scores also vary slightly by age of child; scores presented here are for 9–12 years old, representing the mean age of our sample.

cal cutoffs in the CRI, transplant, and healthy groups. Three (20%) children in the SSNS group scored above the clinical cutoff for externalizing symptoms. On the PSI, higher than expected proportions of clinically significant distress were found in the transplant (n ⫽ 3, 21.4%) and healthy (n ⫽ 7, 19.4%) groups. Predictors of Children’s and Parents’ Distress. To examine how family environment variables predicted parenting stress and children’s behavioral adjustment, we conducted hierarchical multiple regression equations. Diagnostic group (SSNS, CRI, transplant, healthy) was entered as the first predictor. Because scores for single parents were higher on reported externalizing behavior and parenting stress scores, family structure (single- or two-parent) was entered next as a covariate. FES scales (cohesion, expressiveness, conflict) were entered as a block of predictors. CBCL internalizing and externalizing subscales and the PSI total stress score were outcome variables. Multiple R, F, F change, and ␤ coefficients are presented in Table IV.


99

Table IV. Predictors of Children’s and Parents’ Adjustment Outcome INT-T

⌬R2

F

Diag

.04

.80

Parent

.00

.01

Predictor

Overall F

df

R2

Conflict EXT-T

7, 64

.26

.22

6.04***

Diag

.02

.40

.06

4.07*

Cohes

⫺.19 .02 ⫺.22 .00

Conflict PSI

.04 .25

3.01**

Parent

Express

.01 ⫺.20

Cohes Express

␤

.48 5.41****

7, 60

.39

.31

10.09****

⫺.24

Diag

.02

.38

⫺.07

Parent

.05

3.44

⫺.21 ⫺.22

Cohes Conflict Express

.23 5.39****

7, 60

.39

.32

10.34****

⫺.31

INT-T ⫽ CBCL internalizing T-score; EXT-T ⫽ CBCL externalizing T-score; PSI ⫽ Parenting Stress Index; Diag ⫽ diagnostic group; Parent ⫽ single or two parent; Cohes ⫽ FES cohesion scale; Conflict ⫽ FES conflict scale; Express ⫽ FES expressiveness scale. ⌬R2 and F values are not printed for Cohes and Conflict because these variables were entered on one step with Express. *p ⬍ .05. **p ⬍ .01. ***p ⬍ .005. ****p ⬍ .001.

In the first equation, diagnostic group was dummy coded and entered as a block, family structure was dummy coded and entered as a block, and FES scores were entered as a block of predictors. CBCL internalizing scores were the outcome variable. The set of predictors accounted for 25.97% of the variance in CBCL internalizing scores, F(7, 64) ⫽ 3.01, p ⬍ .01. In the full model, the block of FES scores accounted for the majority of variance (22.34%); no single predictor emerged as a unique predictor of significant variance. The ␤ weights indicated that more positive FES scores predicted lower internalizing symptoms. In the second equation, diagnostic group, family structure, and FES scores were entered as predictors; CBCL externalizing scores were the outcome. The set of predictors accounted for 38.71% of the variance in CBCL externalizing scores, F(7, 60) ⫽ 5.41, p ⬍ .001. In the full model, the block of FES scores accounted for the majority of variance (30.91%). Family conflict uniquely predicted 17.06% of the variance, t ⫽ 4.09, p ⬍.001; expressiveness accounted for 4.19%, t ⫽ ⫺2.07, p ⬍ .05. Family structure accounted for 4.19% of the variance, t ⫽ ⫺2.03, p ⬍.05. The ␤ weights indicated that more positive family functioning scores and two-parent family structure predicted lower externalizing symptoms. In the third equation, diagnostic group, family structure, and FES scores were entered as predictors;

PSI total stress scores were the outcome. The set of predictors accounted for 38.71% of the variance in PSI scores, F(7, 60) ⫽ 5.39, p ⬍ .001. In the full model, the block of FES scores accounted for the majority of variance (32.07%). Family expressiveness uniquely predicted 7.45% of the variance, t ⫽ ⫺2.69, p ⬍ .01. Examination of ␤ weights indicated FES scales predicted PSI scores in expected directions. Major Qualitative Analyses Previous researchers (Harris et al., 1996; Perrin et al., 1991) have criticized use of standardized assessments with chronically ill children. Although we believe the CBCL was the best instrument to assess general behavioral adjustment, we also examined qualitative responses. Fiese and Bickham (1998) proposed that analysis of qualitative data from rarely studied populations can reveal insights into participants’ experiences and are useful for generating hypotheses and theories for further research. In addition, our sample contained children from a wide range of developmental stages, and we believed qualitative data could provide developmentally specific information. For these reasons, we analyzed responses to the two CBCL open-ended items, “What concerns you most about your child,” and “What are the best things about your child.”

100

Two trained coders independently developed coding lists from the CBCL open-ended items. A master coding list was developed and revised by consensus of the two coders to include all responses. The majority (90.24%) of parents responded to the “concerns” item, and 92.68% of parents responded to the “best things” item. For the “concerns” item, we developed seven categories, and additional categories for “no concerns” (n ⫽ 3; 17.07%) and no response. For the “best things” item, we developed five categories and one additional category for no response. Response categories and percentages of respondents are presented in Table V. The nature of parents’ reported concerns was to some degree related to the children’s developmental stage. For example, 36.59% of parents indicated illness-related concerns. Parents of school-age and older children were primarily concerned about adherence, for example, “I’m worried about him going off his diet/forgetting to take medications.” The concerns of parents of children younger than 5 years related more to long-term prognosis, exemplified by, “Will he outgrow this.” Other concerns of younger children’s (i.e., 8 years or younger) parents came primarily from three categories. Growth/development/body image concerns, reported by eight parents, were represented by statements such as “We are concerned about his growth . . . about him being really small.” As children mature physically, whether children will “catch up” in physical development may be more obvious. Externalizing behaviors such as anger and aggression and hyperactivity/inattention were also reported by parents of younger children, for example, “He acts so mean . . . [I]f he does not get his way, the whole family suffers.” Other parents mentioned excessive kicking, screaming, and concern over the impact of short attention span on future school performance. Parents of school-age and adolescent children reported school adjustment problems such as low motivation or poor attitudes toward academics. As parents of children this age were more likely to observe academic problems, they also observed good school performance and cognitive abilities and special skills/ abilities that included leadership qualities, creativity, athletic, and artistic abilities. Across ages and developmental stages, parents reported both positive personality characteristics and prosocial qualities of their children. Thirty (73.17%) parents described their children using terms such as “spirited,” loves life,” “very easy going,” and “al-


Table V. Summary of Qualitative Responses Response category

n

%

15

36.59

Concerns Illness-related Growth/development/body image

8

19.51

Anger/aggression

5

12.20

School adjustment

3

7.32

Hyperactivity/inattention

3

7.32

Miscellaneous

3

7.32

Fears

1

2.44

Positive characteristics Positive personality

30

73.17

Prosocial

19

46.34

Good school performance

8

19.51

Special skills/abilities

7

17.97

Good self-care

1

2.44

ways has a smile.” Statements about prosocial qualities included: “He is very sensitive and caring,” “[Has] compassion for people,” and “Very compassionate kind, and gentle.”

Discussion In this study, we aimed to provide insight into the prevalence of psychosocial distress and to understand the moderators thereof in three rare pediatric chronic conditions that have received little research attention. Our global assessments of family and child functioning provide information on the general status of the family environment, parenting stress, and child behavior problems in the kidney disease subgroups as compared to healthy children. Across groups of families whose children had disease or were healthy, we found that family cohesion, expressiveness, and conflict predicted child behavior problems and parenting stress. Our qualitative data provide clues for further investigation into the unique needs of the kidney disease population, including several associated with specific developmental periods within the broad age range we assessed. Among our central findings was the general similarity between the kidney disease and healthy groups on child behavior problems, parenting stress, and family environment variables. Mean scores for both groups indicated typical functioning, which is relatively consistent with previous studies. Vance and Pless (1983) found few differences between behavior scores of children with


SSNS and healthy controls. In samples of children with SSNS and other kidney diseases, child behavior was within normal limits, though scores were elevated compared to those for healthy children (Mehta et al., 1995; McEvoy, 1990). In our study, the chronically ill families and children may have had typical rates of functioning because they had been living with their illness for at least 1 year. After families and children have been managing their illness for a year or more, they may have stabilized, returning to levels of functioning more characteristic of the prediagnostic period. However, mean scores alone do not fully characterize the functioning of our study groups. For example, 28.6% of children in the transplant group had clinically significant levels of internalizing symptoms. Clinically significant levels of externalizing symptoms were present in 20% of children with SSNS. Although previous researchers have noted trends toward higher internalizing symptoms in kidney disease subgroups (Garralda et al., 1988; Vance & Pless, 1983), our finding of higher-thanexpected proportions of participants scoring above clinical cutoffs was unexpected and has not been previously reported. In examining the potential factors related to the higher rates of clinically significant internalizing and externalizing symptoms in the kidney disease groups, we considered the impact of medications. The side effects of medications used to treat kidney disorders may include both elevated internalizing and externalizing symptoms (Soliday, Grey, & Lande, 1999). However, higher levels of externalizing symptoms identified in the SSNS group cannot be attributed to medication side effects because all SSNS participants were assessed at least 2 weeks after discontinuation of medications. The immunosuppressant medications used in the transplant groups may have been associated with higher internalizing symptoms, though this suggestion must be supported with systematic research. Elevated distress was further evidenced in the approximately 20% of the transplant and healthy groups who reported clinically significant levels of parenting stress. Some parents of children with transplants would understandably experience higher rates of distress as they manage repeated hospitalizations for their children, questions over longterm prognosis, and possible struggles over immunosuppressant adherence. For the healthy sample, one might attribute elevated rates of parenting stress to the relatively high rate (52%) of single parents. However, the CRI group had a similar percent-

101

age of single parents (58%), yet rates of clinically significant distress did not exceed clinical cutoffs in this group. In the full sample, single-parent status significantly predicted higher externalizing symptoms in multiple regression equations, but it did not predict parenting stress. Because our healthy and ill child samples were otherwise demographically comparable, we can only assume that unmeasured situational variables or child characteristics underlie the elevated distress in each of the groups. No matter what the cause, we believe our finding is important for several reasons. Of primary importance is the demonstrated impact that elevated psychological symptoms, including internalizing and externalizing behavior and parenting stress, have on children’s long-term developmental outcomes (Abidin, 1995; Achenbach, 1992). Second, our results raise greater awareness of the levels of psychosocial distress and associated risk for negative outcomes in healthy and chronically ill samples drawn from a region infrequently represented in the developmental literature. In addition to identifying the prevalence and severity of child behavior problems and parenting stress, we identified several moderator variables for those outcomes. Family system variables (cohesion, conflict, expressiveness) predicted the majority, up to 32%, of the variance in parent and child outcomes. Researchers have suggested that positive family relationships can buffer stress, including that caused by childhood chronic illness (Moos & Moos, 1994). Our results indicating that higher family cohesion, higher expressiveness, and lower conflict significantly predicted more positive adjustment in children and parents provide support for such a buffering hypothesis. Further, our finding was consistent with previous kidney disease research indicating associations between family environment and children’s adjustment (Crittenden & Holaday, 1986; Fukunishi & Kudo, 1995). In contrast, previous research on diabetes patients has found either no relationship or a negative relationship between family environment variables and child adjustment (Klemp & LaGreca, 1987; Kovacs et al., 1989). We propose that the characteristics of kidney disease are at least partly responsible for the discrepant findings between chronic illness groups. Compared to diabetes, kidney disease and its associated treatments cause more visible changes in appearance. Visible symptoms include edema and stunted growth; corticosteroid treatments cause changes in facial appearance and body fat distribu-

102

tion. In addition, the prospect of a child’s potentially shortened lifespan due to need for organ transplant may be especially prominent compared to the long-term prospects associated with diabetes. Factors such as the visibility of a condition, its associated treatment, and perception of shortened lifespan may cause greater interdependence between family interaction and individual functioning in children with kidney disease. However, this speculation remains to be explored with further research comparing kidney disease and other chronic illness groups. Although our global assessments allowed us to make conclusions about general family and child functioning, several of our qualitative results indicate the need for further assessments targeted at specific age groups within the kidney disease population. For example, our qualitative analyses indicated that parents of younger children (8 years or less) were primarily concerned about their children’s acting-out behaviors and attention span. For this age group, we are in the process of conducting repeated assessments to determine whether such symptoms constitute a developmental phase or whether they tend to persist over time (Soliday, Grey, & Lande, 1999). For older children, several parents reported academic concerns, such as low motivation for or poor attitudes toward school. Because kidney disease places children at risk for educational problems (Hobbs & Sexson, 1993; Schweitzer & Hobbs, 1995), cognitive assessments conducted early in the educational process may be indicated as part of an overall early intervention effort. Our results provide clinicians additional suggestions for developing appropriate family interventions for the kidney disease population. Given that family environment variables predicted outcomes, clinicians might focus on strengthening aspects of the family environment, which in turn could lead to reduced individual distress in children and parents. Clinicians could work with families to find ways of decreasing conflict, improving family closeness, and improving expression of emotions. Previous research has found that improving emotional expression in the family environment can reduce psychological and physical symptomatology in physically healthy and chronically ill groups (Goldstein, Rosenfarb, Woo, & Nuechterlein, 1997; Minuchin, 1974; Suedfeld & Pennebaker, 1997). In addition, clinicians could promote children’s wellbeing by focusing on those strengths identified by parents in our study, including sense of humor,


spirit, athletic abilities, love of life, and prosocial behavior. In conclusion, we consider our study a preliminary, though systematic, sketch of real families dealing with serious pediatric illness. As an initial effort, we must recognize the study’s limitations. All measures were parent report, raising the issue of common method variance. Whereas the number of participating children with kidney disease reflects the relative rarity of the conditions, the sample size was modest. The sample was quite heterogeneous, ranging in economic status and family structure. The children represented a wide range of developmental stages, which again reflects the rarity of these conditions. The fact that 37% of eligible families did not participate is a potentially confounding factor. Despite its drawbacks, our study identified areas where the healthy and ill children appeared to be at psychosocial risk. Furthermore, we identified potential protective factors in the children’s family system. This information benefits researchers and clinicians in designing and implementing appropriate, effective interventions to help ensure the most optimal developmental outcome possible for children with kidney disease. Multisite, longitudinal studies would help us best achieve that goal.

Acknowledgments Portions of this study were supported by Washington State University Dean’s Office Mini Grant Funds. We thank the families participating in this study, Dennis Dyck and Evelyn Florio for reviewing this manuscript, and Shannon Grey for assistance in study design and patient contact. Received April 9, 1998; revisions received October 13, 1998; accepted December 17, 1998

References Achenbach, T. A. (1991). Manual for the Child Behavior Checklist/4–18 and 1991 Profile. Burlington: University of Vermont Department of Psychiatry. Achenbach, T. A. (1992). Manual for the Child Behavior Checklist/2–3 and 1992 Profile. Burlington: University of Vermont Department of Psychiatry. Abidin, R. R. (1995). Parenting Stress Index: Professional manual. 3rd ed. Odessa, FL: Psychological Assessment Resources.


Crittenden, M. R., & Holaday, B. (1986). Psychosocial factors associated with physical growth and behavior adaptations of children with renal dysfunction. Perceptual and Motor Skills, 62, 437–438. Fiese, B. H., & Bickham, N. L. (1998). Qualitative inquiry: An overview for pediatric psychology. Journal of Pediatric Psychology, 23, 79–87. Frauman, A. C., & Lansing, L. (1983). The child with chronic renal failure: I. Change and challenge. Issues in Comprehensive Pediatric Nursing, 6, 127–133. Fukunishi, I., & Kudo, H. (1995). Psychiatric problems of pediatric end-stage renal failure. General Hospital Psychiatry, 17, 32–36. Garralda, M. E., Jameson, R. A., Reynolds, J. M., & Postlethwaite, R. J. (1988). Psychiatric adjustment in children with chronic renal failure. Journal of Child Psychology and Psychiatry, 29 (1), 79–90. Goldstein, M. J., Rosenfarb, I., Woo, S., & Nuechterlein, K. (1997). Transactional processes which can function as risk or protective factors in the family treatment of schizophrenia. In H. D. Brenner, W. Boker, & R. Genner (Eds.), Towards a comprehensive therapy for schizophrenia (pp. 147–157). Gottingen, Germany: Hogrefe and Huber. Grimm, P. C., & Ettenger, R. (1992). Pediatric renal transplantation. Advances in Pediatrics, 39, 441–493. Hamlett, K. W., Pellegrini, D. S., & Katz, K. S. (1992). Childhood chronic illness as a family stressor. Journal of Pediatric Psychology, 17, 33–47. Harris, E. S., Canning, R. D., Kelleher, K. J. (1996). A comparison of measures of adjustment, symptoms, and impairment among children with chronic medical conditions. Journal of the American Academy of Child and Adolescent Psychiatry, 35, 1025–1032. Hobbs, S. A., & Sexson, S. B. (1993). Cognitive development and learning in the pediatric organ transplant recipient. Journal of Learning Disabilities, 26, 104–113. Holden, E. W., Chmielewski, D., Nelson, C. C., Kager, V. A., & Foltz, L. (1997). Controlling for general and disease-specific effects in child and family adjustment to chronic childhood illness. Journal of Pediatric Psychology, 22, 15–27. Hollingshead, A. B. (1975). Four factor index of social status. Unpublished manuscript, Yale University. Klemp, S. B., & LaGreca, A. M. (1987). Adolescents with IDDM: The role of family cohesion and conflict. Diabetes, 36, 18A. Kohaut, E. C., & Tejani, A. (1996). The 1994 annual report of the North American Pediatric Renal Transplant Cooperative Study. Pediatric Nephrology, 10, 422–434. Kovacs, M., Kass, R. E., Schnell, T. M., Goldston, D., & Marsh, J. (1989). Family functioning and metabolic control of school-aged children with IDDM. Diabetes Care, 12, 409–414. Lavigne, J. V., & Faier-Routman, J. (1992). Psychological adjustment to pediatric physical disorders: A meta-

103

analytic review. Journal of Pediatric Psychology, 17, 133–157. Manne, S. L., Lesanics, D., Meyers, P., Wollner, N., Steinherz, P., & Redd, W. (1995). Predictors of depressive symptomatology among parents of newly diagnosed children with cancer. Journal of Pediatric Psychology, 20, 491–510. McEvoy, R. E. (1990). A study of the relationship between family characteristics and child competence in children with chronic renal disease. Doctoral dissertation, University of Houston, 1990. Dissertation Abstracts International, 51, 455. Mehta, M., Bagga, A., Pande, P., Bajaj, G., & Srivastava, R. N. (1995). Behavior problems in nephrotic syndrome. Indian Pediatrics: Journal of the Indian Academy of Pediatrics, 32, 1281–1286. Minuchin, S. (1974). Families and family therapy. Cambridge, MA: Harvard University Press. Moos, R. H., & Moos, B. S. (1994). Family Environment Scale manual. 3rd ed. Palo Alto: Consulting Psychologists Press, Inc. Northam, E., Anderson, P., Adler, R., Werther, G., & Warne, G. (1996). Psychosocial and family functioning in children with insulin-dependent diabetes at diagnosis and one year later. Journal of Pediatric Psychology, 21, 699–717. Perrin, E. C., Stein, R. E., & Drotar, D. (1991). Cautions in using the Child Behavior Checklist: Observations based on research about children with a chronic illness. Journal of Pediatric Psychology, 16, 411–421. Schweitzer, J. B., & Hobbs, S. A. (1995). Renal and liver disease: End-stage and transplantation issues. In M. C. Roberts (Ed.), Handbook of pediatric psychology (pp. 425–445). New York: Guilford. Soliday, E., Grey, S., & Lande, M. B. (1999). Behavioral effects of corticosteroids in steroid-sensitive nephrotic syndrome. Pediatrics, 104, 1–4. Suedfeld, P., & Pennebaker J. W. (1997). Health outcomes and cognitive aspects of recalled negative life events. Psychosomatic Medicine, 59, 172–177. Vance, J. C., & Pless, I. B. (1983). The effect of chronic nephrotic syndrome on the affected child. Developmental and Behavioral Pediatrics, 4, 159–162. Wallander, J. L., Varni, J. W., Babani, L., Banis, H. T., & Wilcox, K. T. (1988). Children with chronic physical disorders: Maternal reports of their psychological adjustment. Journal of Pediatric Psychology, 13, 197–212. Warshaw, B. L. (1994). Nephrotic syndrome in children. Pediatric Annals, 23, 495–503. Wassner, S. J. (1994). Conservative management of chronic renal insufficiency. In M. A. Holliday, T. M. Barratt, & E. D. Avner (Eds.), Pediatric nephrology (pp. 1314–1338). Baltimore: Williams & Wilkins. Wyatt, R. J., Kagy, L., & Kritchevsky, S. B. (1996). Epidemiology of pediatric kidney diseases. Pediatric Annals, 25, 288–296.