Idea Transcript
Journal of Pediatric Psychology, Vol. 20, No. 4, 1995, pp. 527-544
Mark R. Dadds,2 Ruth E. K. Stein,3 and Ellen Johnson Silver Albert Einstein College of Medicine IMontefiore Medical Center Preventive Intervention Research Center for Child Health Received December 6, 1993; accepted December 6, 1994
Studied the ways that mothers interpret illness behavior in their children to assess whether maternal psychological adjustment predicts maternal perceptions of children's behavioral limitations and attribution of these behaviors to chronic illness. Functional status ratings and attributions to illness by 365 mothers of 5to 8-year-old children with chronic illnesses were associated with children's overall adjustment but not with mothers' own psychological distress. Illness attributions also were related to the child's medical visits and hospitalizations. Thus, mother's illness attributions are related to her perceptions of the child's health and more general behavioral adjustment, but not to her own mental health. Results support the validity of the FS H(R) as a measure of functioning related to children's health status that is not influenced by maternal psychological adjustment. KEY WORDS: childhood chronic illness; functional status; child adjustment; maternal mental health.
•This work was supported in part by research grant P50-MH 38280 from the Branch for Prevention Research of the National Institute of Mental Health. We gratefully acknowledge the cooperation of the New York City Health and Hospitals Corporation. Thanks to Laurie Bauman, Dorothy Jessop, Jennifer Lauby, and Lauren Westbrook for their helpful comments, to Henry Ireys for his collaboration on the STEP project, and to all staff and participating families for their assistance. Currently affiliated with the Behaviour Research and Therapy Centre, The University of Queensland, Brisbane, Australia. 3 A1I correspondence should be sent to Ruth E. K. Stein, Preventive Intervention Research Center for Child Health, Department of Pediatrics, Albert Einstein College of Medicine, Jack & Pearl Resnick Campus, 1300 Morris Park Avenue, NR 7 South 15, Bronx, New York 10461. 527 0I46-8693/93/080O-O527SO7.3O/0 O 1995 Plenum Publishing Corporation
Downloaded from http://jpepsy.oxfordjournals.org/ at Pennsylvania State University on April 8, 2016
The Role of Maternal Psychological Adjustment in the Measurement of Children's Functional Status1
528
Dadds, Stein, and Silver
Downloaded from http://jpepsy.oxfordjournals.org/ at Pennsylvania State University on April 8, 2016
Because more children who have ongoing physical health conditions are surviving (Gortmaker & Sappenfield, 1984; Newacheck & Taylor, 1992), there is interest in the psychological consequences of chronic childhood illness and in the reliable assessment of the functioning and behavior of such children. One major technique for assessing these domains involves measures that depend on parents' ratings (Stein, 1991). An implicit component of these ratings is that the parent is required to make judgments and inferences about the nature and causes of the child's functioning or behavior. There are a number of reasons for conducting a thorough investigation into the nature of this decision-making process. First, research shows that mothers' reports of their childrens' behavior may be influenced by factors other than the child, particularly the mother's own personal and social adjustment (e.g., Breslau, Davis, & Prabucki, 1988; Griest, Wells, & Forehand, 1979; Webster-Stratton, 1988). This has been demonstrated both for well children and for children with behavioral and emotional disorders. Second, measures of the effects of chronic illness, such as those employed in the Rand Corporation's Health Insurance Study (Eisen, Ware, Donald, & Brook, 1979) and the National Health Interview Survey (National Center for Health Statistics, 1989), typically ask parents to make combined evaluations about the occurrence and causes of their children's behavior. That is, questions often are phrased in ways that ask the parent whether the child's activities (e.g., mobility, communication, play, going to school) are limited in kind or amount by the presence of a health condition. Thus, the rater is required to evaluate a range of behaviors that are thought to reflect current functioning and simultaneously to judge the extent to which behavioral variations are due to illness. If a parent answers "never" or "not at all," it cannot be determined if impaired behavior never occurs or if it occurs but is thought not to be due to the child's health condition. Research and theory developed over the last few decades has indicated that human causal attributions are influenced by a complex interplay between characteristics of the behavior being attributed, characteristics of the attributor, and the setting in which it occurs (Bugental, 1987). Further, the type of attributions made about another person's behavior influences the response that is made. Attribution theory has been employed in a number of applied areas. For example, in the area of child abuse, the parent's perception that the child can control his or her behavior, and thus that misbehavior is deliberate, may be influenced by the personal and social competencies of the parent, his or her belief systems about child behavior, and the child's general behavioral adjustment (Bugental, Blue, & Cruzcosa, 1989). One measure of the effects of chronic illness on children's functioning is the Functional Status II(R) (FS II(R); Stein & Jessop, 1990a, 1991). The FS II(R) is a behavioral rating scale that was designed to measure the extent to which chronic illness interferes with a child's ability to perform roles and tasks of daily
Maternal Adjustment to Children's Functional Status
529
METHOD Sample Recruitment occurred at two urban medical centers, both serving a predominantly minority, inner-city, low-income population. The sample included 365 families who were enrolled in a longitudinal investigation of a support intervention designed to strengthen childrens' and families' adaptation to chronic illness in the child (Ireys, Silver, Stein, Bencivenga, & Koeber, 1991; Silver et al.,
Downloaded from http://jpepsy.oxfordjournals.org/ at Pennsylvania State University on April 8, 2016
living in domains such as sleeping, eating, communication, energy, and mood. Poor functional status reflects limitations in the performance of these ageappropriate behaviors due to a health condition. In contrast, good functional status is demonstrated by a child who exhibits the full range of age-appropriate physical, cognitive, emotional, and social behaviors, or whose limitations in behavior are not attributed to the presence of a health condition. Unlike other measures, items in the FS II(R) are divided into two sections. First, the parent rates the extent to which the child exhibited each listed behavior during the last 2 weeks (never or rarely, sometimes, or almost always). In Part 2, the respondent is asked to rate the extent to which each identified behavior is due to the health condition (fully, partly, or not at all). In keeping with the intention of the measure to assess dysfunction due to illness, the item reduces the child's functional status score only if the parent's rating indicates that the behavior is related to the illness (Stein & Jessop, 1990a, 1991). As noted earlier, most types of parent-report measures of health status implicitly ask the rater to make a two-step judgment, with only the final response recorded. However, because the format of the FS II(R) makes explicit the twostep decision process of behavioral rating followed by attribution to illness, it provides a unique opportunity to study the nature of this decision-making process and to examine variables related to mothers' responses. We conducted this study to explore a number of elements that might contribute to mothers' reports of children's functioning in the presence of a serious ongoing health condition. Specifically, we asked if maternal ratings of children's behavior and their attributions of poor functioning to an ongoing illness were related to any of the following variables: sociodemographic characteristics of the mother or child, illness factors such as health care utilization, mother's perceptions of behavior disturbances and general psychosocial adjustment of the child, and her own mental health. We aimed to determine whether scores on the FS II(R) were relatively independent of maternal psychological adjustment, since this finding would support its validity as a measure of illness-related functioning in children.
530
Dadds, Stein, and Silver
Downloaded from http://jpepsy.oxfordjournals.org/ at Pennsylvania State University on April 8, 2016
1994). Only the baseline data, collected before randomization to the experimental or control groups, were used in the present analyses. Mothers were eligible if they had a child between 5 and 8 years of age who had a chronic illness, lived in the Bronx or lower Westchester County, and spoke at least conversational English. Families were excluded if the child was moderately or severely mentally retarded or had a life expectancy less than 18 months. Chronic illness was defined as an ongoing health condition that either had lasted or was expected to last for at least 3 months or had required hospitalization for 30 days or more in the previous year (Pless & Douglas, 1971). Questionable diagnoses were reviewed by a physician on the study team. Life expectancy for children with potentially fatal illnesses (e.g., some forms of cancer) was verified by the clinic that provided their medical care. The most common illnesses found in children in the sample included asthma, sickle cell disease, congenital heart disease, epilepsy and other neurological disorders, cleft lip or palate, and cancer. We used a noncategorical approach to chronic illness (Hobbs, Perrin, & Ireys, 1985; Pless & Pinkerton, 1975; Stein & Jessop, 1982, 1989b) and combined all children with illnesses into one group for analysis. An initial pool of 1,195 potential participants was obtained by identifying eligible children from hospital impatient and outpatient records covering the previous 2 years. We were able to contact 752 families by telephone to seek their enrollment in the study. Of those contacted, 240 had no mother in the home or were excluded on the basis of failing to meet other eligibility requirements. In addition, 86 mothers refused to participate, and 61 who expressed interest were not able to complete a interview during the enrollment period. This left 365 mothers (71% of eligible mothers contacted) who were recruited into the study. Participants did not differ from families whom we did not contact or those who chose not to participate with respect to child's age, sex, or type of illness. Among participants, means and standard deviations for mother and child ages were 34.4 (SD = 7.0) and 7.1 (SD = 1.2) years, respectively. Forty-six percent of the children were female. Average years in school for the mothers was 12.4 (SD = 2.2), and 48% of the families were receiving welfare. The distribution of mothers' ethnicity was: Hispanic (45%), black (37%), white (14%), and other (3%).
Procedure Data were obtained during face-to-face interviews with the mothers that included questions on standard demographic information, as well as measures of the child's and mother's psychological adjustment and the child's functional and medical status. Mothers were given the option of being interviewed in English or
Maternal Adjustment to Children's Functional Status
531
Spanish, and were paid $20 for completing the interview. Informed consent was obtained from all participants.
Measures Functional Status FS II(R) Downloaded from http://jpepsy.oxfordjournals.org/ at Pennsylvania State University on April 8, 2016
This measure assessed illness-related behavioral disturbance, defined as the extent to which an illness limits and interferes with a child's performance of normal roles and tasks (Stein & Jessop, 1990a, 1991). We used the short 14-item version of the FS II(R), applicable to children across the age span of 0 to 16 years. This version has good internal consistency (a ^ .83 across age groups), and correlates in the expected direction with other indices of child health status (Stein & Jessop, 1990a, 1991). Further descriptions of the development and evaluation of this measure may be found in Stein and Jessop (1982, 1984, 1990a, 1991). Completing the FS II(R) is a two-stage process. The mother first rates the frequency with which her child exhibited each of the 14 behaviors during the previous 2 weeks. After all items are rated, each response indicating poor functioning is further probed, and the mother is asked to rate the extent to which the impaired behavior was due to illness. To score the FS II(R), Part 1 behavioral frequency items are recoded according to the mother's attributions. The item reduces the total FS II(R) score only if the mother indicates that the behavior was partially or fully due to the child's illness. For example, if a mother indicates that a child slept well "never or rarely," but then rated this behavior as "not at all" due to illness, the original item is recoded and scored as if the mother said "almost always." Using this Recommended Score, the overall FS U(R) score can range from 0 (indicating total impairment due to illness) to 100 (for no impairment due to illness). In our sample, FS II(R) scores had an internal consistency of .81, confirming previous research that this scale measures a unitary dimension (Stein & Jessop, 1990a, 1991). For this study, we created two additional scores from the functional status measure. The first assessed the unadjusted Behavioral Frequency of the 14 listed behaviors, scored on a scale of 0 to 100 similar to the recommended FS II(R) score, but without considering the mother's perceptions of whether these behaviors were due to illness. That is, we ignored the recommended procedure of using mothers' attributions to adjust the scores, and utilized the measure as if it were a behavioral rating form only. Second, an Illness Attribution score was calculated by counting up the number of items that the mother attributed either fully or partly to illness, and dividing this sum by the number of behaviors rated. This
532
Dadds, Stein, and Silver
score represents the proportion of impaired behaviors attributed to illness. Of necessity, the Illness Attribution score was available only for those mothers who rated at least one maladaptive behavior as present or one adaptive behavior as absent; this was true for 356 of 365 respondents (98%). Mother's Adjustment
Downloaded from http://jpepsy.oxfordjournals.org/ at Pennsylvania State University on April 8, 2016
To measure mothers' mental health, we used the Psychiatric Symptom Index (PSI; Ilfeld, 1976), a 29-item checklist of psychiatric symptoms comprising four dimensions of psychological distress: depression, anxiety, anger, and cognitive disturbance. The PSI measures intensity of 29 common symptoms, and is not intended to define psychiatric diagnoses or "caseness." However, PSI Total scores range from 0 to 100, and scores of 20 or greater are considered to indicate high symptoms. The PSI was developed on a community sample of 2,299 men and women. Its concurrent validity with other criteria indicating emotional distress has been well-established, and the internal reliability of the total scale and the subscales are high (Ilfeld, 1976). For our sample, the alpha coefficient was .92. The PSI has been used successfully with urban, minority populations (Stein & Jessop, 1989a). Child's Adjustment The child's mental health was operationalized in terms of overall psychosocial adjustment and by a separate measure of conduct problems. To measure overall adjustment, we used the 28-item version of the Personal Adjustment and Role Skills Scale III (PARS III; Ellsworth & Ellsworth, 1982; Stein & Jessop, 1984, 1990b; Walker, Stein, Perrin, & Jessop, 1990). The PARS III is appropriate for minority populations and is particularly suitable for those who have a chronic illness because it does not include somatic items that might inflate an ill child's maladjustment score (Walker et al., 1990). The PARS III contains six subscales (Dependency, Hostility, Withdrawal, Anxiety-Depression, Productivity, and Peer Relations) as well as producing a Total score, which was the measure used in this investigation. Higher PARS III scores indicate better adjustment. Internal consistency coefficients are reported to be very high for the PARS III Total score (>.88); for our sample, the alpha was .87. Conduct problems were measured with the Eyberg Child Behavior Inventory (ECBI; Eyberg & Ross, 1978; Robinson, Eyberg, & Ross, 1980), a parent checklist of children's behavior problems, including aggression, noncompliance, rule breaking, and lying. It has good reliability and validity and yields a total Intensity score, produced by summing the frequency of occurrence of all 36 problem behaviors [items scored from never (1) to always (7); scale score,
Maternal Adjustment to Children's Functional Status
533
possible range = 36-252] and a Problem score, which is the tally of the number of behaviors that the parent rates as a problem. In this study, only the Intensity score was used. The alpha coefficient for our sample was .93. Child's Illness Status
Data Analysis We initially examined the associations of predictor variables with the three functional status scores [Recommended FS II(R), unadjusted Behavioral Frequency, and Illness Attribution] using correlation coefficients. Following this, separate regression analyses were performed in which the three different functional status scores were used as dependent variables and groups of predictors were added in hierarchical fashion in the following four blocks: (a) demographics, including child's age, child's sex (male = 1, female = 2), mother's age, mother's education (last grade completed), welfare status (1 = yes, 0 = no), and black, Hispanic, or mixed/other ethnicity (each dummy coded as 1 = yes, 0 = no, with white as the residual category); (b) illness status, including number of days in hospital and number of visits to a medical practitioner in the last 12 months; and (c) mother's mental health, measured by her PSI Total score; and (d) child's psychological adjustment, including the ECBI Intensity score and PARS III Total score. Sociodemographic characteristics were entered in the first block to determine which of these variables might independently relate to the three functional status scores and to remove their potentially confounding influence when examining the relationships of illness factors and adjustment to functional status in later steps. By entering mother's PSI total score on the third step, we addressed our main research question and examined the relationship between FS II(R) and maternal mental health while controlling for variables in the model that were external to the child's own adjustment. The final step in each of these analyses
Downloaded from http://jpepsy.oxfordjournals.org/ at Pennsylvania State University on April 8, 2016
To examine the relationship of functional status to other indices of the child's health that were not based on the mother's judgments and attributions, we used number of days in hospital and number of visits to medical providers in the last 12 months as assessed by the maternal interview. Although these measures still rely on maternal reports and may be influenced by recall distortion and mothers' illness behaviors, they incorporate substantially less subjective judgment than the FS II(R). Recent research also shows that respondents' recall of health-related events such as disability days, medical visits, and treatments corresponds very well with patient records (Brown & Adams, 1992; Revicki, Irwin, Reblano, & Simon, 1994).
534
Dadds, Stein, and Silver
examined independent relationships of the three functional status scores to the predictors with all other variables controlled, and allowed us to determine if adding child's adjustment attenuated any association between mother's mental health and child's functional status.
RESULTS Downloaded from http://jpepsy.oxfordjournals.org/ at Pennsylvania State University on April 8, 2016
Means and standard deviations for the three functional status ratings were: Recommended FS II(R), M = 91.2, SD = 11.3; unadjusted Behavioral Frequency score, M = 81.6, SD = 13.2; and Illness Attribution score, M = 0.4, SD = 0.4. A large proportion of children were rated at the maximum of 100 using the Recommended FS II(R) score (n = 137; 40%), but only 9 children achieved the maximum when using the unadjusted Behavioral Frequency rating. For numbers of days hospitalized and visits to medical providers, the means and standard deviations were 5.2 (SD = 18.8) and 11.5 (SD = 13.9), respectively. Finally, means and standard deviations for the psychological adjustment measures were mother's PS1 Total, M = 21.9, SD = 14.8; child's PARS III Total, M = 88.1, SD = 12.0; and ECBI Intensity, M = 106.5, SD = 35.1. Table I shows the bivariate relationships of the sociodemographic and illness variables, and the maternal child adjustment measures to the three functional status scores. The variables having significant correlations with the Recommended FS (II)R score were mother's education, the number of visits the child made to a doctor or other medical practitioner in the last 12 months, and all three psychological adjustment variables, PARS III Total, ECBI Intensity, and PSI Total scores. Fewer medical visits, more maternal education and lower symptom scores, and better child psychological adjustment were related to better functional status. The unadjusted Behavioral Frequency scores also were correlated with mother's educational level and child's PARS III Total scores, such that more maternal education and better overall adjustment in the child were related to better behavioral functioning. In addition, receiving welfare, Hispanic ethnicity, the child's ECBI Intensity score, and the mother's psychiatric symptom scores all were correlated negatively with the Behavioral Frequency score. Illness Attribution scores were correlated negatively (less attribution) with mother's educational level and child's PARS III scores. They were correlated positively with receiving welfare, number of days in the hospital, and number of medical visits in the previous year. Table I also gives the intercorrelations among the independent variables used in the analyses as a check on multicollinearity among predictors. Generally, correlations were not substantial, although, not surprisingly, the two child adjustment measures, PARS III and ECBI Intensity scores, were correlated significantly (p < .001). We also determined that mothers' symptom scores had moder-
FUNCSTAT FSFREQ FSATTRIB CHILD AGE CHILD SEX MOM AGE MOM ED BLACK HISPANIC WHITE MIXOTH WELFARE HOSP MDVISIT PARSTOT INTENS PSITOT .73 -.69 .02 -.02 -.03 .18 .05 -.09 .05 .05 -.07 -.07 -.15 .44 -.26 -.12
1
-.33 -.01 -.03 .05 .22 .10 -.12 .03 .04 -.14 -.06 -.06 .67 -.54 -.35
—
2
-.00 -.02 .07 -.15 -.02 .09 -.09 -.06 .13 .16 .21 -.14 .01 .10
3
.05 .09 .04 .05 -.06 .04 -.04 -.06 -.03 -.04 .05 -.12 -.02
4
— -.03 -.08 -.08 .11 -.08 .03 -.02 -.02 -.02 -.07 .03 -.05
5
— -.02 .10 -.14 .07 -.04 -.23 -.08 .14 .07 -.07 .03
6
.10 -.10 .01 .02 -.19 -.13 -.01 .19 -.12 -.12
7
-.67 -.30 -.17 -.10 -.04 .01 .11 -.11 -.08
8
— -.35 -.20 .15 .02 -.07 -.13 .13 .02
9
—
-.09 -.12 .00 .13 .04 -.03 .06
to
— .03 .06 -.06 .03 -.06 -.00
11
.16
.05
— .08 .05 -.14
12
— .21 -.02 -.01 .06
13
-.07 .01 .09
14
-.73 -.39
15
— .42
16
—
17
"FUNCSTAT = Recommended FS (II)R score; FSFREQ = Behavioral Frequency score; FSATTRIB = Illness Attribution score; MOM ED = mother's education; MIXOTH = mixed/other ethnicity (mother); HOSP = days hospitalized in last 12 months; MDVISIT = number of visits to medical practitioner in last 12 months; PARSTOT = PARS III Total score; INTENS = ECBI Intensity score; PSITOT = PSI Total score. For BLACK, HISPAN, WHITE, MIXOTH, WELFARE, 1 = yes, 0 = no; For CHILD SEX, 1 = female, 0 = male. Minimum pairwise number of cases = 329. Associated significance levels: r = . 11 or greater, p S .05; r — .14 or greater, p S .01; r = . 17 or greater, p £ .001.
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17.
Variable
Table I. First-Order Correlations Among Functional Status, Demographic, Illness, and Psychological Variables"
Downloaded from http://jpepsy.oxfordjournals.org/ at Pennsylvania State University on April 8, 2016
536
Dadds, Stein, and Silver
Downloaded from http://jpepsy.oxfordjournals.org/ at Pennsylvania State University on April 8, 2016
ate correlations with poorer child mental health on both the PARS III Total and ECBI Intensity scales (ps < .001). Table II shows the results of the three regression analyses. Degrees of freedom vary slightly according to the analysis being performed because of small differences in missing data. In the following description, higher scores on the original FS II(R) measure and the unadjusted Behavioral Frequency scale, as well as the PARS III Total indicate that more healthy, and fewer disturbed, behaviors were present. In contrast, higher scores on the ECBI Intensity and PSI Total scores indicate poorer adjustment. Prediction of FS (II)R Recommended Score The first regression analysis employed the recommended FS II(R) score as the dependent variable, with the demographic variables, illness status, and psychological predictors added hierarchically in four steps (Table II). Demographic variables alone were not significantly associated with FS (U)R scores. The addition of the illness predictors on the second step produced a significant regression equation; however, the amount of explained variance was small (Adjusted R2 = .03). Significant univariate predictors were mother's education and number of visits to medical providers in the previous year. Entering mother's PSI score added little to the prediction. However, addition of the set of children's psychological predictors resulted in a substantial increase in variance explained and in a significant regression equation. With all independent variables in the equation, two variables contributed significantly to the prediction, child's PARS III Total score and number of visits to medical providers. The poorer the child's overall psychosocial adjustment on the PARS III and the greater the number of medical visits, the lower the child's functional status. With all predictors entered, none of the sociodemographic variables was related to functional status. Notably, mother's adjustment, as rated by the PSI Total score, was not related to child's FS (II)R score in this analysis, although the bivariate relationship had been significant. Prediction of Unadjusted Behavioral Frequency Score The second regression analysis used the unadjusted Behavioral Frequency score computed from the functional status data as the dependent variable (see Table II). The predictors were the sets of demographic, illness, and adjustment variables described above. The higher final R2 shows that we were able to account for a greater proportion of the variance in this equation than in the previous analysis, most likely because the Behavioral Frequency score had a higher standard deviation and was less vulnerable to ceiling effects.
R2 = .17 F(l 1,320) = 6.00 MOM ED, P = .16 PSITOT, p = - . 3 2 R2 = .11 F ( l l , 314) = 3.37 MDVISIT, p = .16
R2 = .04 F(10, 321) = 2.51 MOM ED, p = .19
R2 = .10 F(10, 315) = 3.61 HOSP, p = .18 MDVISIT, p = . 11
R2 = .05 F(8, 323) = 2.95 MOM ED, p = .19 R2 = .05 F(8, 317) = 2.20 MOM ED, p = - . 1 1 WELFARE, p = .13
FS II(R) behavioral frequency
FS II(R) illness attribution
R2= .13 F(13, 312) = 3.54 HOSP, P = .11 MDVISIT, p = .16 INTENS, P = - . 1 9 PARSTOT, P = - . 2 3
R 2 = .46 F(13, 318) = 22.26 MOM ED, p = .09 PARSTOT, P = .54
R2 = .23 F(13, 315) = 7.37 PARSTOT, P = .51 MDVISIT, p = - . 1 1
Demographics + illness + mother adjustment + child adjustment
"NS = no significant variables; MOM ED = mothers' education; MDVISIT = number of visits to medical practitioner in last 12 months; HOSP = Days hospitalized in last 12 months; PSITOT = PSI Total score; PARSTOT = PARS HI Total score; INTENS = ECBI Intensity score. All R 2 reported above are Adjusted R2; p values are £ .05.
R2 = .07 f ( l l , 317) = 2.17 MOM ED, p = .15 MDVISIT, p = - . 1 4
R2 = .03 F(IO, 318) = 2.14 MOM ED, p = .16 MDVISIT, p = - . 1 4
NS
Demographic + illness
FS II(R) recommended scoring
Demographics
Demographics + illness + mother adjustment
Predictor group
Table II. Results of Three Hierarchical Regression Analyses Predicting Maternal Ratings on the FS II(R) from Four Predictor Groups"
Downloaded from http://jpepsy.oxfordjournals.org/ at Pennsylvania State University on April 8, 2016
538
Dadds, Stein, and Silver
Downloaded from http://jpepsy.oxfordjournals.org/ at Pennsylvania State University on April 8, 2016
The only demographic variable associated with mothers' ratings of behavior frequency in the first step was maternal education; less educated mothers reported more behavior problems. In the second step, the child's illness status, in terms of medical visits and hospitalizations during the previous year, was not significantly related to the unadjusted Behavioral Frequency score. Although mother's adjustment added significantly to the prediction when entered on Step 3, once again it did not enter as a predictor when other variables were controlled. Conduct problems also did not relate to the unadjusted Behavioral Frequency score. By contrast, child's overall psychological adjustment entered as a significant predictor, accounting for a substantial proportion of explained variance in this analysis. The better the child's adjustment as judged by the PARS III, the less disturbed behavior was reported using the unadjusted Behavior Frequency score. Mother's education continued to be related to their ratings of behavior on this measure when other variables were controlled.
Prediction of Illness Attribution Score The third regression analysis used the computed Illness Attribution score as the dependent measure (Table II). When demographic variables were entered alone, both mother's educational level and receiving welfare predicted Illness Attribution scores. However, when additional variables were entered in the analysis, demographic variables were unrelated to mothers' attributions. Both illness status variables, days in hospital and number of medical visits, entered as significant predictors of maternal attribution scores on the second step. The greater the number of visits the child made to medical providers and the greater the number of days hospitalized in the previous year, the more likely the mother was to attribute disturbed behaviors to illness. However, the amount of variance in the Illness Attribution measure explainable from the demographic and illness status measures together was small. In the final step, the child's psychological adjustment variables added significantly to the prediction of mothers' illness attributions, but the total variance explained remained fairly small (R2 = .13). With all variables entered, significant predictors included both measures of child adjustment, PARS III Total, and ECBI Intensity scores, as well as the two illness measures, hospital days, and number of visits to medical providers. The poorer the child's overall adjustment on the PARS III, the more likely the mother was to attribute the FS II(R) behaviors to the illness. However, the negative beta weight of the ECBI Intensity score (on which high scores are worse) suggested that mothers were less likely to attribute poor functioning to illness when their children had more conduct problems. The mother's score on the PSI was not a significant predictor of her illness attribution.
Maternal Adjustment to Children's Functional Status
539
DISCUSSION
Downloaded from http://jpepsy.oxfordjournals.org/ at Pennsylvania State University on April 8, 2016
The aim of this study was to examine sociodemographic, illness, and psychological adjustment factors that may predict the ways that mothers make judgments about the effects of chronic illness on their children's functioning. Our particular objective was to examine the role of mothers' psychological adjustment in the measurement of functional status on the FS H(R). Notably, mothers' mental health predicted the unadjusted behavior frequency ratings when added into the regression analysis after demographic and illness variables, but it was not associated with either the recommended FS U(R) scoring or with maternal attributions to illness on this scale. In addition, maternal adjustment was not related to any of the three functional status scores in the full regression that included the child adjustment measures. In an earlier study, Jessop, Riessman, and Stein (1988) found a moderate relationship (r = .35) between functional status of the child and mental health status of the mother. However, they also reported that the bivariate relationship disappeared when they conducted a regression analysis using mothers' psychiatric symptom scores as the dependent variable and included functional status in a set of demographic, social, and medical predictors. This result is analogous to our own findings, in which the correlations between mothers' mental health and the functional status scores are small to moderate and significant, but disappear when other variables were controlled using multiple regression. These data suggest that if mothers' mental health has a slight relationship to functional status, it is through its shared association with the child's overall adjustment. The finding that suggests a direct relationship between mothers' psychological distress and the unadjusted behavior scores computed from the FS II(R) is not unexpected. It confirms reports of a number of previous studies that have shown an association between maternal mental health and ratings of behavioral adjustment of children with ongoing physical health conditions (e.g., Daltroy et al., 1992; Daniels, Moos, Billings, & Miller, 1987; Mulhern, Fairclough, Smith, & Douglas, 1992; Steinhausen, Schindler, & Stephan, 1983; Thompson, Gil, Burbach, Keith, & Kinney, 1993; Thompson, Gustafson, Hamlett, & Spock, 1992; Walker, Ortiz-Valdes, & Newbrough, 1989). However, other investigators have not found relationships between mothers' functioning and measures of adjustment for children with various physical conditions (e.g., Barakat & Linney, 1992; Perrin, Ayoub, & Willet, 1993; Vami & Setoguchi, 1993). These discrepant findings are inconclusive in relating maternal mental health to child adjustment in children with chronic illness. In addition, because this previous work has been correlational, any conclusions about causal pathways must be speculative at best. Although poor adjustment in mothers may produce distress in their children, it is just as possible that poor psychological functioning in mothers is a response to their children's emotional or behavioral problems or that both the
540
Dadds, Stein, and SUver
Downloaded from http://jpepsy.oxfordjournals.org/ at Pennsylvania State University on April 8, 2016
mother and child are reacting to a common stressor such as the child's illness. In any case, the relationship of maternal and child adjustment was not the main focus of our study. More important, we conclude from our results that mothers' attributions of dysfunctional behavior to physical problems in children with chronic illness are not influenced significantly by their own psychological adjustment. In addition, the lack of a relationship between maternal distress and children's FS II(R) scores when using the recommended scoring procedures indicates that, despite the potential effects on appraisals of children's more general adjustment, maternal adjustment per se does not appear to confound ratings of the behavioral manifestations of children's health status. Both findings strongly support the use of the FS II(R) and its currently recommended scoring as a valid measure for assessing child functioning related to health status that is independent of maternal mental health. It is clear from our results, however, that both the mother's behavior ratings and attributions to illness are associated consistently with her perceptions of the child's general psychosocial adjustment, especially as measured by the PARS III Total score. The relationship of the PARS III Total score to functional status when it is scored only as a behavioral frequency measure is not surprising, given that both instruments measure similar aspects of the child's mood and behavior. Nevertheless, correlations of illness-related functioning with the PARS III Total and ECBI Intensity scores were much weaker when the recommended FS II(R) scoring was employed. This supports the judgment that recoding of items based on the probes is necessary in assessing behavioral manifestations of illnessrelated functioning. In terms of attributions, the more behavioral limitations shown, the more opportunity the mother has to ascribe them to the illness. However, the negative relationship of the ECBI Intensity score to mothers' attribution scores also suggests that the more conduct problems the mother reported, the less likely she was to attribute limited functioning to illness. The fact that children who are perceived to have a high degree of conduct problems are likely to have variance in their behavior independent of the presence of a chronic illness makes sense in a purely statistical way. It also suggests that failure to recognize that some behavior problems may be independent of the child's health condition might lead to overestimate of illness-related behavior if only a behavior rating scale or the unadjusted form of the functional status measure were used. Again, this supports the authors' recommendation to score the FS II(R) using the attribution probes. On the other hand, the findings regarding the PARS III also indicate that a mother's likelihood of attributing her child's functioning to illness is related to her perceptions regarding her child's overall adjustment. Previous research has alerted us to the possible interplay of psychological and physical health in children (e.g., Cadman, Boyle, Szatmari, & Offord, 1987; Drotar & Bush, 1985;
Maternal Adjustment to Children's Functional Status
541
Downloaded from http://jpepsy.oxfordjournals.org/ at Pennsylvania State University on April 8, 2016
LaVigne & Faier-Routman, 1992; Stein & Jessop, 1984; Wallander, Vami, Babani, Banis, & Wilcox, 1988). These data extend this research by invoking attributional theory, which specifies that a parent's attributions regarding a child's specific behaviors will be influenced by the parent's perceptions of more general characteristics of the child, particularly his or her overall adjustment. At extremes, a child who has a chronic illness and substantial psychosocial or emotional problems may have his or her illness-related behavior ignored or punished because the parent responds as if it is simply more misbehavior. We found that scores on the three forms of the FS II(R) did not vary significantly according to many sociodemographic variables. Although lesseducated mothers were more likely to report maladaptive behavior in their child, when the full set of predictor variables was considered, this relationship was found only for the unadjusted Behavioral Frequency score. This is consistent with Wallander et al. (1988), who also found a significant negative relationship between mothers' educational level and behavior problems in a large sample of children with a variety of chronic illnesses. The two measures of child illness status, number of days in hospital and number of visits to a medical provider over the last year, were not associated with mothers' ratings of the presence of problem behaviors in their children as indicated by the unadjusted Behavioral Frequency score. However, these measures were associated with mothers' attributions of behavior to illness on the FS II(R). The more the child was in hospital and visited a medical provider, the more likely the mother was to attribute maladaptive behavior to the child's illness. It may be that mothers who perceive their children as sicker are more willing or able to see illness as the reason for their children's behavior. On the other hand, children seen as having more behavior problems due to illness may be taken to the doctor more often. In any case, the size of the association with mothers' attributions was small. Medical visits also were associated with the recommended FS II(R) score. Nevertheless, it appears that mothers' ratings and attributions on the FS II(R) are related only marginally to her child's utilization of the health care system. Several limitations to the study must be considered. Because the sample consisted principally of urban, minority families of lower socioeconomic status, investigations with more demographically diverse samples are needed to establish the generalizability of our findings. Yet, despite this relative homogeneity in some demographic characteristics, our sample embodied a wide range of health conditions and had adequate distributions of both health care utilization variables and maternal and child adjustment measures for the analyses we conducted. Second, the study population also may have influenced our participation rate. Enrollment was high among families we were able to contact. However, we did not restrict eligibility to active patients or those with long-term ties to the hospital. That we were unable to locate more than a third of the potential sample was not unexpected, as our patient population moves frequently and has a pattern of
542
Dadds, Stein, and Silver
Downloaded from http://jpepsy.oxfordjournals.org/ at Pennsylvania State University on April 8, 2016
losing and reestablishing telephone service. The Bronx also has one of the highest proportions of households with unlisted numbers. In addition, we do not know if mail was received or forwarded properly to families who moved, as records provide only the child's name and many children and parents in this community have different surnames. This also made attempts to locate new telephone numbers quite challenging. Although we are unable to establish whether psychosocial characteristics of these missing families might have differed in ways that could influence our pattern of results, we do not expect the findings would be changed substantially. Finally, all data were obtained through maternal report. Mothers' ratings of their children's functional status and behavior were essential to addressing the study questions. Using mothers' own estimates of their mental health also contributed to our understanding of how this variable relates to their perceptions of their children. Nevertheless, adding data from other sources regarding maternal adjustment would strengthen future studies. Research has established that poor functioning related to chronic illness can serve as a risk factor for adjustment problems in children (e.g., Stein & Jessop, 1984). More research is needed that examines the specific mechanisms through which physical and psychological maladjustment covary. This study contributes to the process by showing (a) that the FS II(R) measures impact of chronic illness in children without undue contamination from reporting bias due to psychological distress in the mother, and (b) that attributions of the role of illness in the child's adjustment may need to be considered in the context of the child's general psychosocial adjustment.
REFERENCES Barakat, L. P., & Linney, J. A. (1992). Children with physical handicaps and their mothers: The interrelation of social support, maternal adjustment, and child adjustment. Journal ofPediatric Psychology, 17, 725-739. Breslau, N., Davis, G. C , & Prabucki, K. (1988). Depressed mothers as informants in family history research—Are they accurate? Psychiatry Research, 24, 345-359. Brown, J. B., & Adams, M. E. (1992). Patients as reliable reporters of medical care process. Medical Care, 30, 400-411. Bugental, D. B. (1987). Attributions as moderator variables within social-interactional systems. Journal of Social and Clinical Psychology, 5, 469-484. Bugental, D. B., Blue, J., & Cruzcosa, M. (1988). Perceived control over caregiving outcomes: Implications for child abuse. Developmental Psychology, 25, 532-539. Cadman, D., Boyle, M., Szatmari, P., & Offord, D. R. (1987). Chronic illness disability and mental and social well-being: Findings of the Ontario child health study. Pediatrics, 79, 805-813. 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, 277-289. Daniels, D., Moos, R. H., Billings, A. G., & Miller, J. J. (1987). Psychosocial risk and resistance factors among children with chronic illness, health siblings, and healthy controls. Journal of Abnormal Child Psychology, 15, 295-308.
Maternal Adjustment to Children's Functional Status
543
Downloaded from http://jpepsy.oxfordjournals.org/ at Pennsylvania State University on April 8, 2016
Drotar, D., & Bush, M. (1985). Mental health issues and services. In N. Hobbs & J. M. Penin (Eds.), Issues in the care of children with chronic illness (pp. 514-550). San Francisco: Jossey-Bass. Eisen, M., Ware, J. E., Donald, C. A., & Brook, R. H. (1979). Measuring components of children's health status. Medical Care, 17, 902-921. Ellsworth, R., & Ellsworth, S. (1982). CAAP Scale: The measurement of child and adolescent adjustment. Palo Alto, CA: Consulting Psychologists Press. Eyberg, S. M., & Ross, A. W. (1978). Assessment of child behavior problems: Validation of a new inventory. Journal of Clinical Child Psychology, 7, 113-116. Gortmaker, S. L., & Sappenfield, W. (1984). Chronic childhood disorders: Prevalence and impact. Pediatric Clinics of North America, 31, 3-18. Griest, D. L., Wells, K. C , & Forehand, R. (1979). Examination of predictors of maternal perceptions of maladjustment in clinic referred children. Journal of Abnormal Psychology, 88, 277-281. Hobbs, N., Perrin, J. M., & Ireys, H. T. (1985). Chronically ill children and their families. San Francisco: Jossey-Bass. Ilfeld, F. W. (1976). Further validation of a psychiatric symptom index in a normal population. Psychological Reports, 39. 1215-1228. Ireys, H. T., Silver, E. J., Stein, R. E. K., Bencivenga, K., & Koeber, C. (1991, April). Evaluating a parent support program for mothers of chronically ill children. Paper presented at the annual meeting of the Ambulatory Pediatric Association, New Orleans, LA. Jessop, D. J., Riessman, C. K., & Stein, R. E. K. (1988). Chronic childhood illness and maternal mental health. Journal of Developmental and Behavioral Pediatrics, 9, 147-156. LaVigne, J. V., & Faier-Routman, J. (1992). Psychological adjustment to pediatric physical disorders: A meta-analytic review. Journal of Pediatric Psychology, 17, 133-157. Mulhern, R. K., Fairclough, D. L., Smith, B., & Douglas, S. M. (1992). Maternal depression, assessment methods, and physical symptoms affect estimates of depressive symptomatology among children with cancer. Journal of Pediatric Psychology, 17, 313-326. National Center for Health Statistics. (1989). Current estimates from the Health Interview Survey: United States, 1988. (DHEW Publication No. PHS 89-1501). Hyattsville, MD: Author. Newacheck, P. W., & Taylor, W. R. (1992). Chronic childhood illness: Prevalence, severity, and impact. American Journal of Public Health, 82, 364-371. Perrin, E. C , Ayoub, C. C , & Willett, J. B. (1993). In the eyes of the beholder: Family and maternal influences on perceptions of adjustment of children with a chronic illness. Journal of Developmental and Behavioral Pediatrics, 14, 94-105. Pless, I. B., & Douglas, J. W. B. (1971). Chronic illness in childhood. I. Epidemiological and clinical characteristics. Pediatrics, 47, 405-414. Pless, I. B , & Pinkerton, P. (1975). Chronic childhood disorder: Promoting patterns of adjustment. Chicago: Yearbook Medical Publishers. Revicki, D. A., Irwin, D., Reblano, J., & Simon, G. E. (1994). The accuracy of self-reported disability days. Medical Care, 32, 401-404. Robinson, E. A., Eyberg, S. M., & Ross, A. W. (1980). The standardization of an inventory of child conduct problem behaviors. Journal of Clinical Child Psychology, 9, 22-28. Silver, E. J., Ireys, H. T., Bauman, L. J., Savarese, K. B., Koeber, C. K., Ault, S. A., & Stein, R. E. K. (1994, May). Mental health effects of STEP—A support intervention for mothers of children with serious ongoing health conditions. Paper presented at the annual meeting of the Ambulatory Pediatric Association, Seattle, WA. Stein, R. E. K. (1991). Measuring health status in children: Children with chronic health conditions. In H. Hibbard, P. A. Nuttig, & M. L. Grady (Eds.), Primary care research: Theory and methods (AHCPR Conference Proceedings, pp. 25-29). Washington, DC: U.S. Department of Health and Human Services, Public Health Service, Agency for Health Care Policy and Research. Stein, R. E. K., & Jessop, D. J. (1982). A non-categorical approach to chronic childhood illness. Public Health Reports, 97, 354-362. Stein, R. E. K., & Jessop, D. J. (1984). Relationship between health status and psychological adjustment among children with chronic conditions. Pediatrics. 73, 169-174. Stein, R. E. K., & Jessop, D. J. (1989a). Measuring health variables among hispanic and nonhispanic children with chronic conditions. Public Health Reports, 104, 377-384.
544
Dadds, Stein, and SUver
Downloaded from http://jpepsy.oxfordjournals.org/ at Pennsylvania State University on April 8, 2016
Stein, R. E. K., & Jessop, D. J. (1989b). What diagnosis does not tell: The case for a non-categorical approach to chronic illness in childhood. Social Science and Medicine, 29, 769-778. Stein, R. E. K., & Jessop, D. J. (1990a). Functional status II(R): A measure of child health status. Medical Care, 28. 1041-1055. Stein, R. E. K., & Jessop, D. J. (1990b). Manual for the Personal Adjustment and Role Skills Scale 111 (PARS 111). Bronx, NY: Albert Einstein College of Medicine. Stein, R. E. K., & Jessop, D. J. (1991). Manualfor the Functional Statusll(R)Measure. Bronx, NY: Albert Einstein College of Medicine. Steinhausen, H., Schindler, H., & Stephan, H. (1983). Correlates of psychopathology in sick children: An empirical model. Journal of the American Academy of Child Psychiatry, 22, 559-564. Thompson, R. J., Gustafson, K. E., Hamlett, K. W., & Spock, A. (1992). Psychological adjustment of children with cystic fibrosis: The role of child cognitive processes and maternal adjustment. Journal ofPediatric Psychology, 17, 741-755. Thompson, R. J., Gil, K. M., Burbach, D. J., Keith, B. R., & Kinney, T. R. (1993). Role of child and maternal processes in the psychological adjustment of children with sickle cell disease. Journal of Consulting and Clinical Psychology, 61, 468-474. Vami, J. W., & Setoguchi, Y. (1993). Effects of parental adjustment on the adaptation of children with congenital or acquired limb deficiencies. Journal of Developmental and Behavioral Pediatrics, 14, 13-20. Walker, L. S., Ortiz-Valdez, J. A., & Newbrough, J. R. (1989). The role of maternal employment and depression in the psychological adjustment of chronically ill, mentally retarded, and well children. Journal of Pediatric Psychology, 14, 357-379. Walker, D. K., Stein, R. E. K., Perrin, E. C , & Jessop, D. J. (1990). Assessing psychosocial adjustment of children with chronic illnesses: A review of the technical properties of PARS III. Journal of Developmental and Behavioral Pediatrics, 11, 116-121. Wallander, J. L., Varni, J. W., Babani, L., Banis, C. B., & Wilcox, K. T. (1988). Children with chronic physical disorders: Maternalreportsof their psychological adjustment. Journal of Pediatric Psychology, 13, 197-212. Webster-Stratton, C. (1988). Mothers' and fathers' perceptions of child deviance: Roles of parent and child behaviors and parent adjustment. Journal of Consulting and Clinical Psychology, 56, 909-915.