Psychosomatics 48:476-481, November-December 2007
doi: 10.1176/appi.psy.48.6.476
© 2007 Academy of Psychosomatic Medicine
Parental Physical Illness as a Risk for Psychosocial Maladjustment in Children and Adolescents: Epidemiological Findings From a National Survey in Germany
Claus Barkmann, Ph.D., M.P.H.,
Georg Romer, M.D.,
Maggie Watson, Ph.D., and
Michael Schulte-Markwort, M.D.
Received June 15, 2006; revised June 23, 2006; accepted July 7, 2006. From University Hospital, Hamburg-Eppendorf, Germany; Royal Marsden Hospital; and the Institute of Cancer Research, Sutton, U.K. Send correspondence and reprint requests to Claus Barkmann, Ph.D., M.P.H., Child Psychosomatics (W29), University Hospital Hamburg-Eppendorf, Martinistrasse 52, 20146 Hamburg, Germany. e-mail: barkmann{at}uke.uni-hamburg.de
© 2007 The Academy of Psychosomatic Medicine
|
ABSTRACT
|
The authors assessed the risk for psychosocial maladjustment in a population-based sample of dependent children of parents with serious physical illness. In the context of The Hamburg Health Survey, a wide range of data on current life situation and health status was collected from a representative sample of families with children and adolescents between 4 and 18 years old (N=1,950). For 4- to 18-year-old children and adolescents in Germany, the prevalence of a serious physical illness in a parent was 4.1%. The adjusted risk of psychosocial maladjustment in this target group, depending on the case definition, is elevated, with internalizing problems being more prevalent than externalizing problems. Although problems for affected boys decline with puberty, they increase for girls. In this age-group, girls and boys appear to be under particular strain when the respective same-sex parent has a serious physical illness. Exposure to serious parental physical illness is an epidemiologically relevant risk factor for psychosocial maladjustment in children and adolescents that needs to be better recognized by medical professionals.
|
INTRODUCTION
|
Despite long-standing evidence that the serious physical illness of parents is a risk factor for psychological problems in their children,1 prevalence data are still rare. Extrapolating from epidemiological records of the U.S. National Center for Health Statistics, it was concluded that, in Western societies, 5%–15% of children and adolescents may live with a seriously ill parent.2 As compared with existing data on children of mentally ill parents, which is an equally established risk group for mental health problems, the risk for behavioral problems in dependent children exposed to serious parental physical illness appears to be slightly lower.3 This difference may be partly due to the accumulation of genetic and psychosocial risk determinants in the offspring of mentally ill parents. Alongside a nonspecific increase of psychosocial risk factors in families with an ill parent of any kind, it has been reported that, for example, in schizophrenic parents, the risk of this mental illness in their children is eight times higher than in the general population, whereas the risk of children of depression patients becoming ill from a depressive disease has been found to be twice as high as in the general population.4,5
Most research in the field of offspring of physically ill parents has been done with cancer patients, and, within this disease group, mothers diagnosed with breast cancer and their adolescent offspring are overrepresented in the samples studied.6 Children of physically ill parents showed higher scores on symptom scales than control subjects, with a tendency toward internalizing symptoms. In the only study to-date using a community-based sample, 50% of school-age children (6–18 years old; N=116) with a parent who had cancer showed more symptoms in the clinical range, especially within the internalizing spectrum—problems such as anxiety, depression, or social withdrawal, when assessed by parent and self-report screening measures.7 Most empirical studies generate hypotheses on associated risk and protective factors, such as age, gender, and parental disease characteristics. Adolescent daughters of ill mothers were reported to be at higher risk for mental health problems than other combinations of age and gender;8–10 however, it remains unclear whether this has been due to sampling biases, because data on younger children and ill fathers have been lacking.
Because of these previous findings, the aim of this study was to determine the prevalence of serious parental physical illness and its associated variables, and to test the following hypotheses in a nationally representative sample: 1) A serious physical illness in a parent is a risk for psychosocial maladjustment in their dependent children, especially for internalizing problems. 2) Adolescent girls show more emotional and behavioral problems than any other age/sex combination in affected children. 3) The latter effect is increased for girls with seriously ill mothers.
|
METHOD
|
Design
The subjects of this study were represented in a population-based data-set aiming to define the prevalence of psychosocial maladjustment among children and adolescents between the age of 4 and 18 years in Germany in 2001 (The Hamburg Health Survey).11 Originally designed as a cross-sectional epidemiological survey using a national representative sample, analyses used a case–control design. Data were obtained by a questionnaire because of its advantages in anonymity and economy. Parents or their surrogates were used as a source of information.
Serious parental physical illness was defined as a currently existing serious physical disease in one or both parents, which can be classified as life-threatening and/or having a severe impact on a patient’s quality of life. Children’s psychosocial maladjustment were assessed with a reliable and well-validated measure (see below). Further variables were selected according to the literature on empirically verified risk factors for the development of emotional and behavioral problems or disorders in children and adolescents.12–14 In all, 58 variables were included: sociodemographic data (age, gender, nationality, siblings, etc.), parents (age, nationality, education, profession, relational status, etc.), living conditions (number of persons, living space, location, etc.), kindergarten, school, and jobs (description, problems during attendance, etc.), stressful life events (in the last 6 months or earlier), mental health (emotional or behavioral problems/disorders), physical health (weight and length at birth, pre- and postpartum complications, development, somatic illnesses, accidents, etc.) and utilization of healthcare institutions (type, age at first use, medication, etc.).
Instruments
The questionnaire included study-specific items, as well as a standardized questionnaire. Information on parental illness was obtained using specific questions asking about illnesses in the family: Mental and physical illnesses, as well as mental or physical handicaps, could be reported separately for mother, father, index child, siblings, or any other family member. A more detailed description of the disease was ascertained with an open-ended question. The answers were rated by two independent physicians ( =0.91; for distribution of ICD-10 categories, see Results, below).
Psychosocial maladjustment in children and adolescents was assessed by one or both parents completing the Child Behavior Checklist (CBCL;15 German translation by Doepfner et al.16). The CBCL comprises eight syndrome scales (see Table 1), two broad-band scales (Internalizing and Externalizing behavior), and a Total score. Satisfactory-to-good psychometric properties for the German version are described by Doepfner et al.16 The instrument can be seen as an international standard in the screening of psychopathology in children and adolescents.17
View this table:
[in this window]
[in a new window]
|
TABLE 1. Frequencies of Criteria for Psychosocial Maladjustment, According to the Case-Definitions of the Child Behavior Checklist (CBCL) Total Score
|
In order to determine the prevalence of clinically relevant maladjustment (defined as a need for clinical assessment, counseling, and/or treatment), we used the two different norm-based cutoff values developed by Achenbach15 and a criterion-based odds-ratio cutoff (OR-crit) by Remschmidt and Walter (R/W).18 Sensitivity and specificity for the three case definitions are 69.7% and 92.2% for values within the clinical range ( 90th percentile), 83.6% and 83.9% for values within the borderline and clinical range (82nd percentile), and 81.9% and 67.1% for cases with at least two critical symptoms.19
Sample
Children and adolescents age 4–18 years who are German-speaking residents and had at least one German (surrogate) parent were included. The sample size (N=1,950) was considered adequate for case ascertainment, given a previously reported 18% prevalence rate of clinical cases in the population.16 Data were collected by a professional marketing research institute (PSYDATA; Frankfurt, Germany). On the basis of population data, a three-step sampling procedure was chosen: 1) selection of professional interviewers, distributed representatively according to communities; 2) selection of families by the interviewers according to children’s age and gender quota; 3) randomized selection of one index child to be evaluated in families with more than one child (the child with his or her birthday closest to the examination date). Data were collected between February and April 2001. Participation was voluntary, and written informed consent was obtained. The mean response rate was 73.1%. The representativeness of the sample was counterchecked by data from the Federal Department of Statistics (detailed sample description in Barkmann11).
Statistical Analysis
Single-item indicators were combined to scales by factor analyses and reliability analyses whenever possible (criterion: KR20=0.70). For explorative bivariate analyses, Pearson correlation, point-biserial correlation, or the phi coefficient were used because these are statistically equivalent and therefore comparable. Sometimes, Cohen’s d was calculated for standardized differences in means. The risk for psychosocial maladjustment in children and adolescents with physically ill parents was estimated by the odds ratio (OR) adjusted for confounders by use of multiple logistic-regression analysis. Furthermore, the percent of estimated attributable risk (ARP) was used to derive a measure of impact of the study factor on exposed children in the population. For parameter estimations, local 95% confidence intervals (CI) were computed. Differences between the risks for the three different case definitions of the CBCL were not tested since they are not simultaneously stated, but given as alternatives. Hypothesis 2 and 3 were analyzed by a univariate three-way ANOVA ( =0.05; 21%). Interpretation of effect sizes followed the recommendations of Cohen.20
|
RESULTS
|
If not otherwise stated, variables refer to the child.
Prevalence of Serious Parental Physical Illness Among German Families
The prevalence of parental physical illness among German families including 4- to 18-year-old offspring was 4.1% (N=79; 95% CI: 3.2–5.0). The present sample includes mainly parents with cancer (28.8%, neoplastic illnesses; 13.7%, cardiovascular system; 12.3%, intestinal tract; 12.3%, muscle, bone, connective tissue; 6.8%, endocrine and metabolic system; 6.8%, nervous system; 5.5%, respiratory tract; 5.5%, trauma, intoxication; and 8.2%, other diseases). In 55.7% of affected families, the mother, in 40.5% the father, and in 3.8% both parents, were ill. In comparison, the prevalence of mental illnesses in parents was 2.3%; the prevalence of a serious physical illness in the index child was 1.2%. The prevalence of parental physical illness in the population of child and adolescent psychiatric patients was 8.3% (clinical sample: N=1,310 in- and outpatients of the Department of Child and Adolescent Psychiatry at the University Hospital Hamburg-Eppendorf, continuously assessed by clinical routine documentation between 1997 and 2003; 95% CI: 6.8–9.8), which is twice as high as in the field population.
Correlates of Serious Parental Physical Illness
Only 3 of 58 indicators revealed bivariate correlations of at least small size: The probability of a child being exposed to parental physical illness was slightly increased if both parents were alive ( =0.11; 95% CI: 0.02–0.20), if the father did not work full-time (=0.10; 95% CI: 0.03–0.17), and/or a parent suffers from mental illness (=0.18; 95% CI: 0.08–0.27). No other variables, including sociodemographic characteristics, living conditions, kindergarten/school attendance, distressing life-events, health characteristics, nor the utilization of healthcare institutions showed systematic covariation with a parental physical illness. Also, age, gender, and socioeconomic status (SES; education, occupation, and income: KR20=0.71) were unrelated to rates of parental physical illness (r[pbis]=0.04; =–0.01; r[pbis]<=0.01).
Estimated Risk of Psychosocial Maladjustment
Separate logistic-regression analyses for the three different case definitions of the CBCL, adjusted for the three confounders mentioned in the section above, yielded the following values (Table 1): 10% OR: 1.21, 95% CI: 0.60–2.45; 18% OR: 2.08, 95% CI: 1.25–3.46; and ORR/W: 2.04; 95% CI: 1.15–3.62. The estimated proportion of cases in the exposed group presumably attributed to the exposure for 10% ARP is 17.4%; for 18% ARP: 51.9%, or ARPR/W: 51.0%, respectively. To compare: the adjusted risk linked to the factor "parental mental illness" is 10% OR: 3.77 (95% CI: 1.84–7.69); 18% OR: 2.65 (95% CI: 1.39–5.04); and ORR/W: 2.90 (95% CI: 1.15–3.62); The adjusted risk linked to the factor "severe physical illness of the index child" is 10% OR: 2.34 (95% CI: 0.83–6.59); 18% OR: 4.08 (95% CI: 1.77–9.41), and ORR/W: 4.53 (95% CI: 1.92–10.71).
The increase of risk from OR: 1.2 to OR: 2.1 when borderline cases are included indicates that parental physical illness correlates with subclinical levels of psychosocial maladjustment. However, stratification of the proportion of affected families by decile classes of the CBCL Total score reveals that the largest differences exist in the 9th and 10th decile (+13.3 percentage points and +7.3 percentage points; for the 1st to the 8th decile: –5.2 to –0.8 percentage points).
Although the parental CBCL score is of proven validity, ratings of children’s psychopathology by the ill parent (55.7% of the cases) could be systematically biased. However, the standardized differences in means of the total score between ratings of ill and healthy parents within the group of affected families showed just a slightly higher judgment of the ill parent ( =0.11; 95% CI: –0.33–0.55; differences between results when one or both parents completed the scale could not be calculated because of small cell size.)
Specificity of the Estimated Risk
According to the two CBCL broad-band scales instead of the total score for separate logistic-regression analyses under adjustment for the three confounders, children and adolescents exposed to serious parental physical illness seem to have a higher risk for internalizing than externalizing problems (10% OR: 1.81, 95% CI: 0.96–3.40; 18% OR: 1.70, 95% CI: 1.01–2.88 versus 10% OR: 1.51, 95% CI: 0.76–3.01; 18% OR: 1.48, 95% CI: 0.86–2.54).
Table 2 presents the eight syndrome-specific, adjusted risks (case definition according to CBCL syndrome scales from the 95th percentile by Achenbach15) and confirms this pattern: The highest risks exist for anxiety/depressiveness, social problems, and somatic complaints; the lowest risks exist for attention problems and antisocial (delinquent) behavior. In contrast, children of affected parents show a doubly decreased chance of being aggressive.
View this table:
[in this window]
[in a new window]
|
TABLE 2. Syndrome-Specific, Adjusted Risks for Psychosocial Maladjustment in Children With Physically Ill Parents
|
Age and Sex Effects
A 2 x 2 x 2 ANOVA (4–10 versus 11–18-year-olds; male versus female; sex match of child and ill parent [Yes/No]), yielded, with respect to the interaction between age and sex, the expected effect stated in Hypothesis 2 (F[1]=5.612; p=0.021; 7.3% explained variance). Whereas the extent of mental health problems decreased for boys from the children’s group to the adolescents’ group (from a mean of 31.7 to 22.0; d=0.54; N=32), it increased for girls (from 16.1 to 27.9; d=0.79; N=37). Such an effect is not found in the children’s and adolescents’ groups with physically healthy parents (F[1]=3.022; p=0.082; 0.2% explained variance), and it furthermore remains constant when control variables are considered (age, sex, SES, etc.).
In contrast, the analysis of a third-order interaction between age, sex, and same/other-sex of ill parent did not yield the hypothesized effect (F[1]=0.596; p=0.443; 0.8% explained variance). Descriptive analysis showed no difference between same-sex and other-sex child–parent constellations in the group of 4–10-year-olds (mean: 24.5 versus 25.9; d=0.08; N=32); indeed, however, there was an effect in the group of 11–18-year-olds (30.0 versus 25.4; d=0.34; N=47), although this effect was different than expected: when the ill parent was of the same sex, boys demonstrated even greater problems than girls (26.1 versus 18.9; d=0.42 for 23 boys; 30.0 versus 25.4 for 24 girls; d=0.23).
|
DISCUSSION
|
With a prevalence of 4.1% for serious parental physical illness, this is an issue for a substantial minority of German children and adolescents. The occurrence of physical illness is twice as frequent as parental mental illness and three times as frequent as serious physical illness in children. Compared with serious parental physical illness, both parental mental illness and serious physical illness in children have been extensively studied and are widely acknowledged risk factors for psychological maladjustment during childhood and adolescence.12–14 However, assessing the prevalence of parental illness by a questionnaire may lead to an underestimation due to nondisclosure of illnesses. Yet the more precisely defined term of "chronic illness" could not be used in this context because this would have included disorders that are neither life-threatening nor severely reduce quality of life (e.g., mild allergies). The fact that prevalence of parental physical illness is twice as high in populations utilizing child and adolescent psychiatric services underlines its significance as a risk factor for mental health problems. In any case, it has to be taken into account that only a small number of children and adolescents with emotional and behavioral problems or disorders will reach existing healthcare services.21
The level of risk for psychosocial maladjustment—defined as a need for clinical assessment, counseling, and/or treatment according to the cut-offs of the CBCL, was found to depend on the chosen case definition. Whereas the 10% cutoff results in a statistically nonsignificant effect, the 18% cutoff of Achenbach as well as the criterion-based case definition of Remschmidt and Walter showed a statistically significant risk, which meant a doubled chance of being a clinical case when a parent was physically ill. As hypothesized, the symptomatology of those children who were reported to have clinically relevant scores was predominantly an internalizing one; these were depression, anxiety, or somatic complaints.
Although there are no controlled studies on children’s outcome based on representative samples with physically ill versus mentally ill parents, the reviewed literature supports the assumption that physical illness in a parent represents a lower risk for children’s psychosocial maladjustment than parental psychiatric illness.3 In sum, our data support the assumption that the risk of psychosocial maladjustment in children with a physically ill parent is lower than for children having their own chronic illness and also lower than for children with a mentally ill parent. In any case, a deterministic, bivariate dependency between parental physical illness and childhood psychopathology was not assumed; otherwise every exposed child would appear to have mental health problems. Thus, the child’s psychosocial well-being is assumed to be affected by secondary, illness-related stressors on family life, such as fears for the future, financial burdens, role changes, changes in parental personality traits, or parents’ self-esteem.7,10,22 Mutual regulatory processes, including effects of children’s mental health problems on a parent’s health status, have to be considered as well.
We must take into account the fact that information on children’s psychosocial maladjustment in this study was based solely on parents’ reports and that there might be a bias in ratings by ill parents. However, any tendency of parents in general to either over- or underreport their children’s emotional and behavioral problems is compensated for by using representative norms within the same reporting perspective. Still, one could argue that physical illness in a parent makes for an additional systematic reporting bias in either direction. Studies in which physically ill parents’ reports on their children’s behavioral problems were compared with children’s self-ratings, using the CBCL and its self-report version, the YSR, showed that physically ill parents tended to underreport their children’s problems, especially internalizing ones.7,10,23 If one presumes this tendency as a way of defensive responding in physically ill individuals, results here may be even more relevant.
As to the hypotheses on age and gender effects, the results reveal a mixed picture. Whereas affected adolescent girls showed the expected higher symptom scores than younger girls, this effect was reversed in adolescent boys. This result did not occur in the group of children with two healthy parents, and it thus appears to be characteristic of the exposed group. In families with an ill parent, adolescent girls may often take over a motherly caring role for the ill parent as well as for other family members, such as younger siblings.8–9 This may interfere with the developmental tasks of individuation or autonomous identity formation. However, the hypothesis that adolescent girls show even more emotional and behavioral problems when it is the mother who is ill could not be confirmed. Whereas, in younger children, there was no effect of having the same- versus the opposite-sex parent who is ill, in the adolescent group, sons of ill fathers showed more emotional and behavioral problems than did daughters of ill mothers. Because this represents an unexpected effect, and the sample sizes are small, we can only conclude that the specific vulnerability of adolescent daughters having an ill mother (which has been reported in several previous studies) requires further investigation in larger samples.
This study refers to the general features of a parental physical illness as a family stressor, such as a threat to life, side effects of treatment, diminished quality of life, etc. Comparative studies with various disease groups are necessary to examine which effects on family and child adjustment they have in common and which are specific for respective diseases. Furthermore, the developmental outcome of an individual child exposed to a specific stressor may be psychopathologically dramatic; however, it may also take a progressive course by stimulating the child’s full utilization of his or her potential in coping and mastery. More research is needed to determine what kinds of differential circumstances predict either outcome and to clarify factors associated with psychosocial risk. This study goes some way toward achieving this by providing data on prevalence, and it suggests the need for further large-scale studies.
|
ACKNOWLEDGMENTS
|
The research questions posed here were further investigated in the Trans-national Collaborative Multisite Research Project "Mental Health Prevention in a Target Group at Risk: Children of Somatically Ill Parents (COSIP)," which was funded by the EU in the 5th Framework Program "Quality of Life" (Project #QLG4-CT-2001–02378).
|
REFERENCES
|
- Rutter M: Children of Sick Parents: An Environmental and Psychiatric Study. London, UK, Oxford University Press, 1966
- Worsham NL, Compas BE, Ey S: Children’s coping with parental illness, in Handbook of Children’s Coping: Linking Theory and Intervention. Edited by Sandler JN, Wolchik SA. New York, Plenum, 1997, pp 195-213
- Roy R: Consequences of parental illness on children: a review. Soc Work Soc Sci Rev 1990; 2:109–121
- Weissman MM, Leckman JF, Meringas KR, et al: Depression and anxiety disorders in parents and children: results from The Yale Family Study. Arch Gen Psychiatry 1984; 41:845–852[Abstract/Free Full Text]
- Remschmidt H, Mattejat F: Kinder Psychotischer Eltern (Children of Psychotic Parents). Göttingen, Germany, Hogrefe, 1994
- Romer G, Schulte-Markwort M, Riedesser P: Kinder körperlich kranker eltern am beispiel kinder krebskranker Mütter. Geburtshilfe und Frauenheilkunde 2002; 62:537–542[CrossRef]
- Birenbaum LK, Yancey DZ, Phillips DS, et al: School-age children’s and adolescents’ adjustment when a parent has cancer. Oncol Nurs Forum 1999; 26:1639–1645[Medline]
- Compas BE, Worsham NL, Epping-Jordan J, et al: When Mom or Dad has cancer: markers of psychological distress in cancer patients, spouses, and children. Health Psychol 1994; 13:507–515[CrossRef][Medline]
- Grant KE, Compas BE: Stress and anxious-depressed symptoms among adolescents: searching for mechanisms of risk. J Consult Clin Psychol 1995; 63:1015–1021[CrossRef][Medline]
- Welch AS, Wadsworth ME, Compas BE: Adjustment of children and adolescents to parental cancer: parents and children’s perspectives. Cancer 1996; 77:1409–1418[CrossRef][Medline]
- Barkmann C: Emotional and Behavioral Problems of Children and Adolescents in Germany: An Epidemiological Screening. Hamburg, Germany, Dr. Kovac, 2004
- Rutter M: Psychosocial adversity and child psychopathology. Br J Psychiatry 1999; 174:480–493[Abstract/Free Full Text]
- Fortin L, Bigras M: Risk factors exposing young children to behaviour problems. Emot Behav Difficulties 1997; 2:3–14
- Kazdin AE, Kraemer HC, Kessler RC, et al: Contributions of risk-factor research to developmental psychopathology. Clin Psychol Rev 1997; 1:375–406
- Achenbach TM: Manual for the Child Behavior Checklist/4-18 and 1991 profile. Burlington, VT, University of Vermont, 1991
- Doepfner M, Plueck J, Boelte S, et al: Elternfragebogen Über das Verhalten von Kindern und Jugendlichen. Deutsche Bearbeitung der Child Behavior Checklist (CBCL/4-18). Koeln, KJFD, Arbeitsgruppe Kinder-, Jugend-, und Familiendiagnostik, 1998
- Verhulst FC: The cross-cultural generalizability of the CBCL, in European Approaches to Hyperkinetic Disorder. Edited by Sergeant J. Zürich, Germany, EUNETHYDIS, Fotorotar, 1995, pp 19-35
- Remschmidt H, Walter R: Psychische Auffaelligkeiten bei Schulkindern. Goettingen, Germany, Hogrefe, 1990
- Schmeck K, Poustka F, Doepfner M, et al: Discriminant validity of the Child Behavior Checklist CBCL/4-18 in German samples. Eur Child Adolesc Psychiatry 2001; 10:240–247[CrossRef][Medline]
- Cohen J: Statistical Power Analysis for the Behavioral Sciences. Hillsdale, NJ, Lawrence Erlbaum Associates, 1988
- Angold A, Costello EJ: Developmental epidemiology. Epidemiol Rev 1995; 17:74–82[Free Full Text]
- Christ GH: Impact of development on children’s mourning. Cancer Pract 2000; 8:72–81[CrossRef][Medline]
- Lewandowski LA: A parent has cancer: needs and responses of children. Pediatr Nurs 1996; 22:518–521[Medline]
Get information about faster international access.
Privacy Policy
Copyright © 2007
Academy of Psychosomatic Medicine.
All rights reserved.
Home
| Search
| Current Issue
| Past Issues
| Subscribe
| All APPI Journals
| Help
| Contact Us
|
TOP
ABSTRACT
INTRODUCTION
