Child Psychiatry Hum Dev DOI 10.1007/s10578-013-0423-5

ORIGINAL ARTICLE

Prospective Associations Between the Family Environment, Family Cohesion, and Psychiatric Symptoms Among Adolescent Girls James White • Katherine H. Shelton Frank J. Elgar



Ó Springer Science+Business Media New York 2013

Abstract The present study used a longitudinal design to investigate whether parental values, interests, and the perceived level of expressiveness, cohesiveness and control in the family were associated with changes in psychiatric symptoms during early adolescence. We used data gathered at ages 12, 14 and 17 from 1,600 adolescent girls in the National Heart, Lung and Blood Institute Growth and Health Study. Using a novel approach we found significant informant effects of family cohesion on psychiatric symptoms, with stronger associations for adolescent ratings than parental ratings. Multivariate modelling adjusting for family cohesion (from parents and adolescents perspective), and parents’ depressive and eating disorder symptoms indicated parental ratings on the intellectual and cultural orientation of the family predicted small but significant increases in eating disorder symptoms, anxiety and a reduction in self-worth 2-years later. Parental control also predicted a reduction in anxiety and family expressiveness was related to increases in self-worth.

Electronic supplementary material The online version of this article (doi:10.1007/s10578-013-0423-5) contains supplementary material, which is available to authorized users. J. White (&) Centre for the Development and Evaluation of Complex Public Health Interventions (DECIPHer), School of Medicine, Cardiff University, Heath Park, Cardiff CF14 4YS, UK e-mail: [email protected] K. H. Shelton School of Psychology, Cardiff University, Tower Building, Park Place, Cardiff CF10 3AT, UK F. J. Elgar Institute for Health and Social Policy and Douglas Institute, McGill University, Montreal, Canada

Keywords Family cohesion  Family functioning  Eating disorders  Depression  Anxiety adolescence

Introduction Stable and supportive family relationships are a wellestablished determinant of health in childhood [1] and adolescence [2]. Longitudinal studies in this area have found that open, rational, and issue-orientated communication within the family reduces the risk of poor mental health and risk taking in early to mid-adolescence [3]. This type of communication between parents and children may protect against mental health problems by helping to foster social competencies, modeling healthy coping strategies, and providing opportunities for children to express concerns and feel valued [1]. However, adolescent mental health may also depend on experiences of the family unit, over and above the role of dyadic parent–child interactions [4, 5]. Indeed, the whole family, as a unit of analysis, serves important developmental functions including security, socialization and the internalization of values [4, 6, 7]. A number of cohorts studies have demonstrated positive associations between adolescent mental health and family cohesion, communication, expressiveness and warmth [8, 9], and low levels of hostility and conflict [5, 10–12]. The mechanisms underlying these influences are complex. From infancy to adolescence, positive family influences shape the development of neuroregulatory systems in the brain that govern emotion, attention and social interactions and thus set the stage for healthy socio-emotional development [1]. Unfortunately, the current evidence on the family dynamics which contribute most to adolescent mental health lacks precision. Studies have used assessments of parenting styles [13, 14], the perceived

123

Child Psychiatry Hum Dev

cohesiveness of the family from the childs’ perspective [8, 9], and also conflict [15, 16], but the way that parental values might shape adolescent mental health has been largely neglected. Parental values are likely to manifest in parent-to-parent and parent–child conversations and also the kind of activities, sanctions and support that is provided to children. Family characteristics such as level of expressiveness may also have some bearing on the number of opportunities, and willingness of children, to communicate with parents. These characteristics may have implications, in turn, for adolescent psychological health. For example, two crosssectional studies, one in Hong Kong [17] and another with Australian children with parents recently diagnosed with cancer [18] found that adolescents from families with open and expressive communication had fewer depressive symptoms. Adolescents with depression have also been found to have families characterized by lower levels of support and higher levels of conflict compared to nondepressed controls [19]. Other studies have found conflict, cohesion and expressiveness in the family relate to internalizing symptoms in adolescents [20], but few studies have examined other potentially important aspects of family life such as the engagement of the family in religion, intellectual pursuits, or the emphasis placed by parents on achievement. It is conceivable that these factors foster psychological resilience, or exacerbate levels of stress and diminish adolescents’ perceptions of levels of support. Clarifying the link between these facets of family functioning and adolescent mental health may be of use to clinicians to identify specific types of interactions within families which might be associated with an increased risk of poor mental health in adolescents. Another gap in the current research is our understanding of the role of informant effects on parent and adolescent reports of family cohesion. Some recent cross-sectional evidence has showed mother- and father-rated family cohesion was more strongly associated with boys’ depression while mother-rated family expressiveness and conflict were more strongly associated with girls’ depression [20]. Other research has indicated that mothers and their adolescent offspring differ in their perceptions of family life; with adolescents perceiving their families as significantly less intimate and more conflicted than mothers [21]. One possible explanation for these differing views is that adolescents in the process of individuation take a more negative view compared to parents seeking validation of their investment in their families [21]. On the one hand, there is a strong theoretical rationale for including adolescent ratings of family life when investigating potential risk and protective factors for poor adolescent mental health. By appreciating that a child is

123

engaged in an ongoing process of interpreting parents’ actions, their subjective understanding of such behaviors could influence their interactions with parents, and have implications for their psychological health [22]. Previous research supports this rationale, finding that adolescent’s perceptions of family life mediate the effects of such interactions on their psychological functioning and wellbeing [23]. On the other hand, Powers et al. [22] cogently argued that an adolescent’s appraisals of family life will not capture all the ways in which family behaviors and their impact on its members can be understood. Aspects of the family environment may impinge on adolescent mental health in ways that the adolescent is unaware or unable to express. It therefore remains unclear whether parents’ and adolescents’ appraisals of the family have similar consequences for adolescent mental health [8, 9, 24]. This lack of clarity has been compounded by collapsing sources of information (e.g. combining parent and child reports) about the family or else assuming that disparities reflect measurement error [24]. The present study used a longitudinal design to examine the association between different aspects of family functioning (i.e., levels of cohesion, expressiveness, conflict, independence, orientation towards achievement, orientation to intellectual and cultural pursuits, emphasis of morals and religion, level of control) and changes in levels of anxiety, self-worth and disordered eating between the ages of 12–14 years, and depression at 17 years of age. Adolescence is a developmental period when internalizing symptoms of psychological distress as well as disordered eating behavior increase, particularly among girls [24, 25]. We therefore focused on these dimensions of mental health in a sample of female adolescents and their families. We had two objectives: (1) to explore the role of different dimensions of family environment on changes in adolescent mental health between 12 and 14 years and depression at 17 years of age, after adjustment for depressive symptoms and disordered eating reported by parents, as well as other known confounding factors (e.g. pubertal status), (2) to examine whether parent and adolescent perceptions of the level of family cohesion were independently associated with changes in adolescent mental health. As there were so few studies’ examining these dimensions in family functioning (moral–religious emphasis, expressiveness, independence) we a priori conducted an exploratory analysis of these factors with changes in girls’ mental health. We hypothesized that family cohesion, from the perspective of parents and girls would be associated with decreases in anxiety and disordered eating, improvements in self-worth between 12 and 14 years and fewer depressive symptoms at 17 years of age.

Child Psychiatry Hum Dev Table 1 Continuous measures for parents and girls on the family environment and psychological symptoms at 12 and 14 years of age (n = 1,600)

12 years Mean (SD)

14 years Mean (SD)

p value for change between 12 and 14 years

Parent report Family Environment Scale

Mann–Whitney tests were used to examine differences in skewed variables; t tests were used for normally distributed variables FACES III Family Adaptability and Cohesion Evaluation Scale version III; CES-D Centre for Epidemiologic Studies Depression Scale; RCMAS-2 Revised Children’s Manifest Anxiety Scale; EDI Eating Disorders Inventory a Parents depressive symptoms assessed when girls were 10 years of age b

Parents Eating Disorder Inventory Scores does not include body dissatisfaction

c

Medians (IQRs) are given for skewed variables

Cohesion

6.14 (1.29)

Expressiveness

5.86 (1.20)

Conflict

5.52 (1.18)

Independence

5.26 (1.14)

Achievement orientation Intellectual–cultural orientation

5.78 (1.05) 6.15 (1.19)

Moral–religious emphasis

6.57 (1.17)

Control

6.35 (0.99)

Psychological symptoms Depressive symptomsa

21.09 (3.21)

Disordered eating behaviours and symptoms (EDI)b

15.08 (9.34)

Adolescent report Family environment Perceived family cohesion (FACES III)

2.81 (0.59)

2.59 (0.66)

\0.001

Psychological symptoms Anxiety (RCMAS-2)

10.66 (6.40)

10.53 (6.26)

0.27

Global self-worth (Harter’s)

3.20 (0.61)

3.09 (0.66)

\0.001

Disordered eating behaviours and symptoms (EDI)c

4.92 (2.71, 6.57)

4.00 (2.71, 6.00)

Depressive symptoms (CES-D at 17 years)c Covariates

0.01

13.00 (8.00, 21.00)

Age

12.02 (0.57)

14.01 (0.57)

\0.001

Body mass index

20.55 (4.64)

22.59 (5.13)

\0.001

Methods Participants and Recruitment We conducted a secondary analysis of the National Heart, Lung and Blood Institute Growth and Health Study (NGHS). A complete description of NGHS procedures and measures has previously been reported [26]. Briefly, the NGHS was a 10-year multicenter longitudinal study of 2,379 girls who were followed up annually from ages nine or 10 years (1985 = year 1) to 18 or 19 years (1995 = year 10). The NGHS was designed to investigate the development of obesity and cardiovascular disease risk factors in black and white girls and as such neither males nor other racial groups were recruited. Eligibility criteria of the sample were that: (1) girls and parents declared they were either white or black (concordance of parent and child race was required), (2) girls were within 2 weeks of their 9th or 10th birthday, (3) girls gave

assent, and (4) parents provided consent and completed a household demographic assessment. The participants were recruited from three study sites: University of California at Berkeley, University of Cincinnati/Cincinnati Children’s Hospital Medical Center and Westat, Inc./Group Health Association in Rockville, Maryland. Institutional Ethical Review Boards of all participating sites approved the study protocol. We restricted our analysis to the data gathered at 12 and 14 years of age, as these were the study years when assessments of family functioning (at 12 years), family cohesion and mental health outcome measures (at 12 and 14 years) were concurrently collected. Mental health was not assessed at 13 years of age. We also used data from a measure of depressive symptoms that was only collected at 17 years. The average age of the girls at wave 3 (the first assessment) was 12.02 (SD = 0.57) years of age. The characteristics of the sample are summarized in Table 1.

123

Child Psychiatry Hum Dev

Measures and Procedures A demographic questionnaire was used to capture information from parents and girls on age, family structure and parental education. Participants’ age was recorded as age at last birthday. The highest level of education achieved by either parent and the number of parents in the household was collected at study entry from parents (or guardians). Based on previous analyses into the categorization of education in the NGHS, [27] maximum parental education was categorized into: Bhigh school, college (some post high school education), and college (C4 years). Parental education was chosen instead of household income because NGHS data were collected in three regions with different price indices, such that the cost of living may have differed between sites. Detailed descriptions of the procedure and the measurement of depression [28], eating disorders symptoms [29], self esteem [30], obesity [31] and pubertal development [32] in the NGHS have previously been reported.

scales are scored separately. One parent was randomly selected where two-parents completed the FES [35]. Psychometric data indicates the FES has adequate levels of internal consistency (0.78–0.76) and an 8-week test–retest reliability of between 0.73 and 0.86 [34]. Family Cohesion The 10-item cohesion subscale from the Family Adaptability and Cohesion Evaluation Scale (FACES III, [36]) was used to assess girls’ perceptions of family cohesion. Cohesion includes aspects of emotional bonding, support, boundaries, shared recreation, spending time together and approval of friends, e.g. ‘Family members ask each other for help’ [37]. Responses are made on a five point Likert scale. The author of this scale has presented reasonable internal reliability scores (Cronbach alpha) of 0.77 for the cohesion sub-scale and a test–retest reliability of 0.93 at a 5 weeks interval [38]. The Cronbach’s alpha was 0.82 at 12 years of age in the NGHS sample.

Body Mass Index and Pubertal Development Depressive Symptoms Examiners were centrally trained and certified annually. Height was measured to the nearest 0.1 cm with the girls wearing socks, using custom-made stadiometers. Weight was measured to the nearest 0.1 kg with calibrated Healtho-meter electronic scales (Sunbeam Products, Inc, Maitland, FL) with the participant wearing a standard size T-shirt. BMI (kg/m2) was derived from measurements of height and weight. Stage of maturational development was assigned using the Tanner stages for pubic hair [33] and Garn’s system for classifying areolar development [32]. Appropriate adaptations were made as these classification systems had not been validated in a black sample. Prepubertal girls had limited signs of development, pubertal girls had signs of maturation but had not yet experienced menarche, post menorrheal had experienced menarche for less than 2 years, and another category was used for girls for whom menarche was more than 2 years ago. Family Environment We used the eight subscales of The Family Environment Scale (FES; [34] to assess parent perceptions of family environment). These include a series of true–false statements drawn from three domains: (1) relationships: cohesion, expressiveness and conflict, (2) personal-growth: levels of independence, achievement orientation (reflecting an orientation towards achievement and competition), intellectual-cultural orientation (reflecting an orientation towards cultural, political and social issues) and moral– religious emphasis on ethical and religious values and activities, and (3) system-maintenance: control. The sub-

123

Depressive symptoms were assessed in girls’ using the Center for Epidemiological Studies—Depression Scale (CES-D) which is a well-established 20-item self-report scale assessing depressive symptoms experienced in the past week [39]. In the NGHS the CES-D was administered at 17 years of age. Parents’ depressive symptoms were assessed when girls’ were nine to 10 years old using a 10-item version of the Zung SelfRating Depression Scale [40]. Items chosen were those that related to frequency of depressed mood, crying episodes, sleep disturbances, lethargy and anxiety. Trait Anxiety Trait anxiety experienced by girls was assessed using the 27-item total anxiety subscale of the Revised Children’s Manifest Anxiety Scale (RCMAS II, [41]). The RCMAS-2 measures for the presence of academic stress, test anxiety, and anxiety around peer and family conflict. Psychometric data on this scale indicates that it has adequate to high internal consistency [42], moderate test–retest reliability [41], and the original RCMAS showed reasonable convergent validity (r = 0.85) with the State-Trait Anxiety Inventory for Children, Trait Anxiety subscale, but little correlation with the State Anxiety subscale [43]. In the NGHS the RCMAS II was administered at 12 and 14 years of age. Eating Disorder Symptoms Psychological traits associated with eating disorders and disordered eating behaviors were assessed using the

Child Psychiatry Hum Dev

64-item the Eating Disorders Inventory (EDI-2, [44]) scale, which is organized into eight sub-scales. Numerous studies support the reliability and validity of this measure [45, 46]; including previous studies with the NGHS [47]. The total of the EDI-2 was used to capture a breadth of eating disorder behaviors and symptoms in girls. For parents, the EDI-2 was assessed when girls were 12–13 years old. The body dissatisfaction sub-scale of the EDI was not assessed in the original study as other measures of body dissatisfaction were included (e.g. Figure Rating Scales, [48], body part dissatisfaction). We examined the impact of exclusion of this sub-scale in preliminary analysis by running the analysis with and without the inclusion of the different dissatisfaction sub-scales. As there was no material change in estimates, a total of the seven EDI sub-scales were used to assess parents’ disordered eating behaviors (data not tabulated). Self-Worth Self-worth was assessed using Harter’s Self-Perception Profile for Children [49], measure of self-perceived competency and self-esteem. The six items ask children to rate the extent to which they agree with statements such as, ‘‘Some kids like the kind of person they are’’, and ‘‘Some kids don’t like the way they are leading their life’’. The global self-worth sub-scale has a high level of internal consistency (Cronbach’s alpha 0.80 [50], test–retest reliability (r = 0.86) and convergent validity with the Trait Anxiety Scale of the State-Trait Anxiety Inventory for Children (r = -0.56; [43]) and the internalizing (r = -0.22) and externalizing (r = -0.30) sub-scales of the Child Behavior Checklist [51]. Statistical Analysis We used t tests and v2 tests to examine differences between continuous and categorical values for girls who were and were not included in the final analytical samples. Paired t tests examined the change between 12 and 14 years of age in levels of family cohesion and mental health outcomes. For skewed continuous variables we used the Wilcoxon signed rank test. Multivariable linear regression models were tested to determine whether parents’ ratings on different aspects of family environment, levels of family cohesion (12 years) and change (D) in cohesion scores completed by girls (12–14 years) were associated with changes in trait anxiety, eating disorder symptoms and global self-worth. All change scores were calculated as: score at 14 minus score at 12 years of age. Models were also produced for depression scores at 17 years of age. Separate models were built for each mental health outcome measure. In a

minimally adjusted model, the eight sub-scales of the FES were entered together to estimate the independent contributions of parents’ perception of different aspects of the family environment to girls’ mental health. In the fully adjusted model, we included girls’ family cohesion at 12 and change in cohesion between 12 and 14 years of age, and also controlled for other confounds linked to changes in mental health in early adolescence: age, race, parents depressive symptoms, level of disordered eating behaviors and symptoms by parents, parental education, number of parents resident in the household, and maturational stage [52, 53]. Sensitivity Analysis There is currently no consensus on the best method for modelling change in outcomes, with some recommending change scores [54], others advocating the use of time two values as the outcome with time one as a covariate (due to concerns of regression to the mean with low or high values [55]), and some suggesting both methods are equivalent [56]. We therefore used change scores then re-ran our analysis using outcomes at age 14 adjusting for prior levels at 12 years of age, and examined the difference in estimates. Missing Data As with all cohort studies, the NGHS experienced a degree of attrition and non-response which raises a concern about the potential effects of selection and non-response bias [57]. Out of the 2,379 participants recruited at age 9, 2,228 (93.7 %) attended at 12, 2,056 (86.4 %) at 14, and 1,978 (83.1 %) at 17 years of age. Of these girls, 1,600 (67.3 %) returned questionnaires with complete information on all mental health outcome measures, and 1,482 (62.2 %) parents completed all sub-scales of the FES. To increase efficiency and minimize selection bias we used multivariate multiple imputation to impute missing values on all predictor and confounding variables. We chose not to impute outcomes as some have suggested they are more likely to have data missing not at random which may bias estimates [58]. We used the MICE method, as described by Royston, [59] to generate 20 imputed datasets. We included all predictor and outcome variables and covariates in the imputation model, as well as variables associated with missingness such as household income. The main analysis results were obtained by averaging across the estimates from each of these datasets, taking into account uncertainty so that the standard errors of estimates are appropriately sized. We also repeated all analysis including only girls with complete data on all variables (N = 824).

123

Child Psychiatry Hum Dev Table 2 Categorical measures for parents and girls on the family environment and psychological symptoms at 12 and 14 years of age (n = 1,600) Categorical variables

% (n)

Race White

51.69 (844)

Black Highest level of parental education

48.31 (789)

CHigh school

23.31 (373)

College (post high school)

39.50 (632)

College (C4 years)

37.19 (595)

Maturational stage Pre-pubertal

10.56 (169)

Pubertal

57.00 (912)

Post menorrheal

28.75 (460)

C2 years post menorrheal

adjustment for girls’ family cohesion scores, parents’ psychological symptoms, with race (i.e. significant effects in white but not black girls) and parental college education yielding the greatest attenuation. Trait Anxiety In the fully adjusted model, an increase in anxiety scores between 12 and 14 years was positively associated with parental ratings of the intellectual and cultural orientation of the family (b = 0.45; 95 % CI 0.06, 0.84), whereas a reduction in anxiety scores was associated with high levels of parental control (b = -0.50; 95 % CI -0.92, -0.08; see Table 3). Eating Disorder Symptoms

3.69 (59)

Parents/guardians in the household One

29.57 (473)

Two

70.43 (1,127)

Results A comparison between girls who did and did not provide complete data for all mental health outcomes showed those with complete data were more likely to be white (39.9 vs. 50.5 %; p \ 0.001), were slightly younger (M = 12.02 vs. 12.06; p = 0.04), had parents with a college level education (55.3 vs. 44.7 %; p \ 0.001), were less likely to be resident with only one parent (32.7 vs. 50.9 %; p \ 0.001) and have experienced menarche (48.1 vs. 51.4 %; p \ 0.001). They also had parents who provided higher expressiveness ratings (M = 5.93 vs. 5.77; p = 0.03). Changes in Mental Health, Family Cohesion and Confounding Factors Tables 1 and 2 show the sample characteristics. Table 1 shows that between 12 and 14 years of age, scores on the EDI improved, Harter’s global self-worth decreased and little change occurred in RCMAS total anxiety scores. Scores on the family cohesion sub-scale of the FACES III reduced significantly between 12 and 14 years of age and body mass index increased.

Table 4 shows that in the minimally adjusted model, the intellectual and cultural orientation of the family was positively associated with increases in eating disorder symptoms. In contrast, a moral–religious emphasis was associated with a reduction in eating disorder symptoms. Adjusting for girls’ cohesion scores and other confounds significantly attenuated the relation between moral religious emphasis and EDI-2 scores (b = -0.60; 95 % CI -1.71, 0.52); whilst the association with intellectual-cultural orientation remained significant at conventional levels (b = 1.51; 95 % CI 0.48, 2.55). Increases in girls’ family cohesion scores were also associated with a significant reduction in eating disorder symptoms (see Table 4). Global Self Worth Minimally adjusted estimates showed a small association between improvements in global self-worth and parents’ ratings of expressiveness of the family and a negative association with intellectual cultural orientation. In the fully adjusted model there was no material change in estimates with expressiveness (b = 0.07; 95 % CI 0.03, 0.11) and intellectual cultural orientation (b = -0.06; 95 % CI -0.10, -0.02) remaining associated with changes in global self-worth, and increases in girls’ family cohesion scores associated with improvements in global self-worth (see Table 4). Sensitivity Analysis

Depression Table 3 shows that depressive symptoms at 17 years of age were negatively related to parent-rated family cohesion (b = -0.68; 95 % CI -1.21, -0.15) and an emphasis on morals and religion (b = -0.68; 95 % CI -1.24, -0.13). These associations were significantly attenuated following

123

There was very little difference in the estimates when outcomes were modeled as changes scores, or with values at age 14 as outcomes adjusting for prior values at 12 years of age, with 95 % CIs overlapping for all estimates across the two methods. These results are available from the first author.

Child Psychiatry Hum Dev Table 3 Regression coefficients (95 % CI) for depression (age 17) and changes in anxiety between ages 12 and 14 years according to parent and adolescent reports on the family environment at 12 years of age (n = 1,600) Depressionc

Anxietyd

Minimala

Full adjustmentb

Minimala

Full adjustmentb

b

95 % CI

b

b

b

-1.21, -0.15*

95 % CI

95 % CI

95 % CI

Parent reports Family Environment Scale Cohesion

-0.45

-1.00, 0.10

0.24

-0.12, 0.59

0.22

-0.14, 0.58

0.30

-0.22, 0.92

0.18

-0.36, 0.73

-0.07

-0.42, 0.28

-0.14

-0.50, 0.22

Conflict

-0.12

-0.68, 0.28

-0.32

-0.91, 0.28

0.26

-0.08, 0.59

0.19

-0.16, 0.53

Independence

-0.11

-0.67, 0.53

-0.10

-0.66, 0.46

0.06

-0.26, 0.38

0.07

-0.25, 0.39

Achievement orientation

0.50

-0.08, 1.23

0.55

-0.02, 1.26

-0.08

-0.44, 0.28

0.02

-0.34, 0.39

Intellectual-cultural orientation

-0.37

-0.88, 0.26

-0.08

-0.63, 0.47

0.40

Moral-religious emphasis

-0.68

-1.24, -0.13*

-0.37

-0.97, 0.22

-0.29

-0.61, 0.03

-0.12

-0.46, 0.21

0.24

-0.44, 0.92

-0.52

-0.94, -0.11*

-0.50

-0.92, -0.08*

Expressiveness

Control

-0.68

0.35

-0.32, 1.02

0.02, 0.77*

0.45

0.06, 0.84*

Psychological symptoms Depressive symptomsd

0.14

-0.04, 0.32

0.03

-0.07, 0.14

Disordered eating behaviours and symptomse

0.02

-0.04, 0.08

0.01

-0.04, 0.04

Adolescent reports Perceived family cohesion at 12 years

-3.00

-3.97, -2.02***

-0.12

-0.71, 0.47

Change in perceived family cohesion between 12 and 14 years of age

-1.86

-2.70, -1.02***

-0.34

-0.85, 0.18

FACES III Family Adaptability and Cohesion Evaluation Scale version III; CES-D Centre for Epidemiologic Studies Depression Scale; RCMAS2 Revised Children’s Manifest Anxiety Scale; CI confidence interval * p \ 0.05; ** p \ 0.01; *** p \ 0.001 a

Minimal adjustment: FES subscales completed by parents

b

Minimal plus: parental depressive symptoms, parental disordered eating behaviours and symptoms, FACES III cohesion subscale, change in FACES cohesion sub-scale, age, race, maturational stage, parental education, and number of parents in the household

c

CES-D assessed at 17 years of age

d

Parents depressive symptoms were assessed when girls were 10 years of age Parents Eating Disorder Inventory Scores does not include body dissatisfaction

e

Comparisons to the Complete Case Model There was very little difference in the strength of associations between the results from the imputed and complete case data set, with 95 % CIs overlapping for all estimates (see Tables 1 and 2, web supplement).

Discussion This study explored the impact of family functioning on changes in mental health in girls during early adolescence.

Aside from replicating associations between family cohesion with depression [8, 9], a unique contribution of the study was evidence that other aspects of family functioning predicted changes in disordered eating, anxiety and levels of self-worth in early adolescents. Girls with parents who focused on intellectual and cultural affairs showed small but significant increases in anxiety and disordered eating, and decreases in global self-worth between 12 and 14 years. In contrast, high parental control related to reductions in anxiety and expressiveness with improvements in self-worth. These associations were not explained by the established protective effect of family cohesion, or

123

Child Psychiatry Hum Dev Table 4 Regression coefficients (95 % CI) for a change in eating disorder symptoms and self-worth between 12 and 14 years according to parent and adolescent reports on the family environment at 12 years of age (n = 1,600) Disordered eating symptoms and behaviours

Global self-worthc

Minimala

Adjustedb

Minimala

Adjustedb

b

95 % CI

b

95 % CI

b

95 % CI

b

-0.07, 0.01

-0.02

95 % CI

Parent reports Family Environment Scale Cohesion Expressiveness

0.12

-0.81, 1.05

-0.05

-1.00, 0.91

-0.03

-0.08

-1.09, 0.92

-0.41

-1.44, 0.61

0.06

0.73

-0.37, 1.84

0.35

-0.75, 1.45

-0.02

-0.06, 0.02

-0.01

-0.04, 0.04

0.30

-0.74, 1.34

0.32

-0.71, 1.36

-0.01

-0.05, 0.03

-0.01

-0.05, 0.03

-0.95

-2.04, 0.14

-0.43

-1.56, 0.70

0.01

-0.03, 0.05

-0.01

-0.05, 0.03

-0.09, -0.01**

-0.06

Conflict Independence Achievement orientation Intellectual–cultural orientation

1.47

0.50, 2.44**

1.51

0.48, 2.55**

-0.05

0.02, 0.09**

0.07

-0.06, 0.02 0.03, 0.11***

-0.10, -0.02**

Moral–religious emphasis

-1.23

-2.32, -0.15**

-0.60

-1.71, 0.52

0.04

-0.01, 0.09

0.02

-0.03, 0.07

Control

-0.75

-2.00, 0.50

-0.67

-1.93, 0.58

0.01

-0.04, 0.06

0.01

-0.03, 0.06

Depressive symptomsd

0.05

-0.28, 0.38

-0.01

-0.02, 0.01

Disordered eating behaviours and symptomse

0.05

-0.06, 0.16

-0.01

-0.01, 0.01

-0.05, 0.08

Psychological symptoms

Adolescent reports Perceived family cohesion at 12 years



-0.72

-2.48, 1.05

-

0.01

Change in perceived family cohesion between 12 and 14 years of age

-

-1.82

-3.33, -0.30**

-

0.15

0.08, 0.21***

CI confidence interval * p \ 0.05; ** p \ 0.01; *** p \ 0.001 a

Minimal adjustment: FES subscales completed by parents

b

Minimal plus: parental depressive symptoms, parental disordered eating behaviours and symptoms, FACES III cohesion subscale, change in FACES cohesion sub-scale, age, race, maturational stage, parental education, number of parents in the household and body mass index at 12 years of age

c

Global self-worth scale from Harter’s Self-Perception Profile

d

Parents depressive symptoms were assessed when girls were 10 years of age

e

Parents Eating Disorder Inventory Scores does not include body dissatisfaction

changes in cohesion, on adolescent mental health [47]. Our findings replicate previous research that has found a null effect of family cohesion and expressiveness on adolescent anxiety (e.g. Queen et al. [20], after controlling for depression), and extend existing research on the family and adolescent mental health to show how day-to-day interactions within families on social, political and intellectual pursuits may effect adolescents’ psychological well-being. This study is unique in that it assessed the independent contributions of parent-and adolescent rated family cohesion on adolescent mental health using a longitudinal research design [60]. Our results indicate that adolescent reports of changes in family cohesion were associated with decreases in disordered eating and increased self-worth, as well as fewer depressive symptoms at 17 years old. In contrast, associations between parents’ reports of family

123

cohesion and adolescent mental health were much smaller and for all scales hovered around the null. This suggests that adolescent reports of the level of family cohesion may be more important than parent reports in predicting changes in adolescent mental health. These findings are consistent with other published findings in this cohort [47] and others [61] that have shown that adolescent ratings of family cohesion reduce the risk of disordered eating and increase overall self-worth. The mechanisms that underlie these associations may include the social support and other benefits derived from spending time with parents, such as a sense of belonging and emotional security [4, 62]. Support for this mechanism comes from a previous analysis of the NGHS cohort that found that girls’ ratings of family cohesion mediated the relation between the frequency of family meals and changes in

Child Psychiatry Hum Dev

adolescent mental health [47], and others have found adolescent family cohesion scores mediate the association between parent perceptions of family rituals and adolescent well-being [60]. A novel finding in this study was that even after adjustment for parent and girls’ ratings on levels family cohesion, an interest in political, social and cultural affairs was linked with increases in levels of anxiety, disordered eating and lower self-worth 2 years later. A potential explanation is that engaging girls in intellectual and cultural discussions may displace more mundane but useful time spent sharing of problems, or the offering of help and the modelling of positive coping strategies by parents which might occur if a focus on intellectual pursuits did not exist [1]. Collectively, this suggests that claims that ‘quality time’ spent together as a family is uniformly adaptive for adolescent health should be tempered by findings that the intellectual and cultural orientation of the family may be relevant to understanding variation in symptoms of distress in this age group. Strengths and Limitations Our study has several strengths. Data were collected on a large sample representing three distinct regions of the United States. To our knowledge, we are the first to investigate the association between parental perception of the family environment and adolescent mental health, after adjusting for family cohesion and parental mental health, established risk and protective factors for adolescent mental health [8, 9, 63]. The use of validated measures of mental health and detailed information on potential confounding factors, along with a large sample afforded the detection of small but meaningful effects of the family environment on mental health. Limitations are also acknowledged. First, only 62.2 % of the participants in the NGHS study had information available on both parental perceptions of the family environment and adolescent mental health at ages 12, 14 and 17. Although a degree of attrition is inevitable in cohort studies and rates are comparable to other adolescent cohorts investigating psychopathology [64], we found non-response was more common in girls who were black, resident with one parent, less likely to have experienced menarche and whose parents did not attend college. However, non-response was more likely to have biased estimates towards the null because mental health problems are typically more common in these groups. Second, we were not able to explore the internal consistency of some of the study variables, as only subscale totals were provided. Third, previous research has indicated that the relation between family environment and adolescent mental health may vary depending on genetic risk for psychopathology [65]. Future research may be in a stronger position to explain heterogeneity in adolescent

psychological health by assessing links with the family environment using a genetically sensitive design. Finally, these findings are based on a sample of female adolescents and we do not presume they generalize to males.

Summary In early adolescence, a number cohort studies have demonstrated that high levels of family cohesion, communication, expressiveness and warmth are associated with a reduced symptoms of depression and disordered eating. The way that parental values might shape the family environment and adolescent mental health, and whether parent or child ratings of family cohesion are more important predictors has, however, largely been neglected. The current study investigated specific dimensions of family functioning, namely parents ratings on the level of expressiveness, control, conflict, independence, orientation towards achievement, intellectual and cultural pursuits, moral–religious emphasis of their family. We used these ratings by parents to predict changes in symptoms of anxiety, disordered eating and self-worth of adolescent girls between 12 and 14 years and depression at 17 years of age. Following multivariable adjustment for parental mental health problems, and parent and child ratings of cohesion, a parental interest in intellectual and cultural pursuits was associated with an increased risk of girls reporting increases in symptoms of anxiety, disordered eating and low self-worth. High parent ratings of control were associated with lower anxiety and expressiveness with higher self-worth. Furthermore, we found that associations between family cohesion and psychiatric symptoms were stronger when adolescent than parent ratings were used. These are new insights into the influence of the family during a period of development when psychiatric illnesses commonly first present [66] and advance our understanding of familial risk factors for mental illness in early adolescence. These findings indicate that parents should be mindful that the expression of specific views and interests in early adolescence, as these views might unintentionally impact on girls’ risk of mental health problems in later years. Acknowledgments The NGHS was supported by a Grant from the National Heart, Lung and Blood Institute (NHLBI) (HL/DK71122). Participating NGHS Centers included Children’s Medical Center, Cincinnati, OH (Stephen R Daniels, MD, Principal Investigator, John A Morrison, PhD, Co-Investigator); Westat, Inc., Rockville, Maryland (George B Schreiber, ScD, Principal Investigator, Ruth StriegelMoore, PhD, Co-Investigator) and University of California, Berkeley, California (Zak I Sabry, PhD, Principal Investigator, Patricia B Crawford, Dr PH, RD, Co-Investigator); Maryland Medical Research Institute, Baltimore, Maryland (Bruce A Barton, PhD, Principal Investigator) served as the data coordinating center. Program Office: NHLBI (Eva Obarzanek, PhD, RD, Project Officer 1992-present, Gerald H Payne, MD, Project Officer 1985–1991).

123

Child Psychiatry Hum Dev The work was undertaken at The Centre for the Development and Evaluation of Complex Interventions for Public Health Improvement, a UKCRC Public Health Research: Centre of Excellence. Funding from the British Heart Foundation, Cancer Research UK, Economic and Social Research Council (RES-590-28-0005), Medical Research Council, the Welsh Assembly Government and the Wellcome Trust (WT087640MA), under the auspices of the UK Clinical Research Collaboration, is gratefully acknowledged (by Dr White).

References 1. Repetti RL, Taylor SE, Seeman TE (2002) Risky families: family social environments and the mental and physical health of offspring. Psychol Bull 128:330–366 2. Viner RM, Ozer EM, Denny S et al (2012) Adolescence and the social determinants of health. Lancet 379:1641–1652 3. Baumrind D (1991) The Influence of Parenting Style on Adolescent Competence and Substance Use. J Early Adolesc 11:56–95. doi:10.1177/0272431691111004 4. Cummings EM, Davies PT, Campbell SB (2002) Developmental psychopathology and family process: theory, research and clinical implications. Guilford, New York 5. McHale JP, Rasmussen JL (1998) Coparental and family grouplevel dynamics during infancy: early family precursors of child and family functioning during preschool. Dev Psychopathol 10:39–59 6. Fiese BH, Parke RD (2002) Introduction to the special section on family routines and rituals. J Fam Psychol 16:379 7. Sroufe LA, Rutter M (1984) The domain of developmental psychopathology. Child Dev 55:17–29 8. Cumsille PE, Epstein N (1994) Family cohesion, family adaptability, social support, and adolescent depressive symptoms in outpatient clinic families. J Fam Psychol 8:202 9. Harris PhD TL, Molock PhD SD (2000) Cultural orientation, family cohesion, and family support in suicide ideation and depression among African American college students. Suicide Life-threat Behav 30:341–353 10. Elgar FJ, Waschbusch DA, Dadds MR, Sigvaldason N (2007) Development and validation of a short form of the Alabama Parenting Questionnaire. J Child Fam Stud 16:243–259 11. Holahan CJ, Moos RH (1987) Personal and contextual determinants of coping strategies. J Pers Soc Psychol 52:946 12. Lucia VC, Breslau N (2006) Family cohesion and children’s behavior problems: a longitudinal investigation. Psychiatry Res 141:141–149 13. Milevsky A, Schlechter M, Netter S, Keehn D (2007) Maternal and paternal parenting styles in adolescents: associations with self-esteem, depression and life-satisfaction. J Child Fam Stud 16:39–47 14. Wolfradt U, Hempel S, Miles JNV (2003) Perceived parenting styles, depersonalisation, anxiety and coping behaviour in adolescents. Pers Indiv Diff 34:521–532 15. Cole DA, McPherson AE (1993) Relation of family subsystems to adolescent depression: implementing a new family assessment strategy. J Fam Psychol 7:119 16. Sheeber L, Hops H, Alpert A et al (1997) Family support and conflict: prospective relations to adolescent depression. J Abnorm Child Psychol 25:333–344 17. Lau S, Kwok L-K (2000) Relationship of family environment to adolescents depression and self concept. Soc Behav Pers Int J 28:41–50. doi:10.2224/sbp.2000.28.1.41 18. Edwards B, Clarke V (2004) The psychological impact of a cancer diagnosis on families: the influence of family functioning

123

19.

20.

21.

22.

23.

24.

25. 26.

27.

28.

29.

30.

31.

32. 33. 34.

35.

36. 37.

38.

and patients’ illness characteristics on depression and anxiety. Psychooncology 13:562–576. doi:10.1002/pon.773 Sheeber LB, Davis B, Leve C et al (2007) Adolescents’ relationships with their mothers and fathers: associations with depressive disorder and subdiagnostic symptomatology. J Abnorm Psychol 116:144–154. doi:10.1037/0021-843X.116.1.144 Queen AH, Stewart LM, Ehrenreich-May J, Pincus DB (2013) Mothers’ and fathers’ ratings of family relationship quality: associations with preadolescent and adolescent anxiety and depressive symptoms in a clinical sample. Child Psychiatry Hum Dev 44:351–360 Noller P, Seth-Smith M, Bouma R, Schweitzer R (1992) Parent and adolescent perceptions of family functioning: a comparison of clinic and non-clinic families. J Adolesc 15:101–114 Powers SI, Welsh DP, Wright V (1994) Adolescents’ affective experience of family behaviors: the role of subjective understanding. J Res Adolesc 4:585–600 Neiderhiser JM, Pike A, Hetherington EM, Reiss D (1998) Adolescent perceptions as mediators of parenting: genetic and environmental contributions. Dev Psychol 34:1459 Paikoff RL, Carlton-Ford S, Brooks-Gunn J (1993) Motherdaughter dyads view the family: associations between divergent perceptions and daughter well-being. J Youth Adolesc 22:473–492 Nolen-Hoeksema S (2001) Gender differences in depression. Curr Dir Psychol Sci 10:173–176 National Heart Lung Blood Institute: Growth and Health Study Group L (1992) Obesity and cardiovascular disease risk factors in black and white girls: the NHLBI Growth and Health Study. Am J Public Health 82:1613–1621 Kimm SYS, Obarzanek E, Barton BA et al (1996) Race, socioeconomic status, and obesity in 9- to 10-year-old girls: the NHLBI growth and health study. Ann Epidemiol 6:266–275. doi:10.1016/S1047-2797(96)00056-7 Franko DL, Striegel-Moore RH, Thompson D et al (2005) Does adolescent depression predict obesity in black and white young adult women? Psychol Med 35:1505–1513. doi:10.1017/ S0033291705005386 Striegel-Moore RH, McMahon RP, Biro FM et al (2001) Exploring the relationship between timing of menarche and eating disorder symptoms in black and white adolescent girls. Int J Eat Disord 30:421–433. doi:10.1002/eat.1103 Brown KM, McMahon RP, Biro FM et al (1998) Changes in selfesteem in black and white girls between the ages of 9 and 14 years: the NHLBI growth and health study. J Adolesc Health 23:7–19. doi:10.1016/S1054-139X(97)00238-3 Kimm S, Glynn NW, Obarzanek E et al (2005) Relation between the changes in physical activity and body-mass index during adolescence: a multicentre longitudinal study. Lancet 366: 301–307 Biro FM, Falkner F, Khoury P et al (1992) Areolar and breast staging in adolescent girls. Adolesc Pediatr Gynecol 5:271–272 Tanner JM (1962) Growth at adolescence, 2nd edn. Blackwell, Oxford Moos RH (1990) Conceptual and empirical approaches to developing family-based assessment procedures: resolving the case of the Family Environment Scale. Fam Process 29:199–208 Roosa MW, Beals J (1990) Measurement issues in family assessment: the case of the Family Environment Scale. Fam Process 29:191–198 Olson DHL (1985) FACES III. Family Social Science. University of Minnesota, Minnesota Maynard PE, Olson DH (1987) Circumplex model of family systems: a treatment tool in family counseling. J Couns Dev 65:502–504 Olson DH (1986) Circumplex model VII: validation studies and FACES III. Fam Process 25:337–351

Child Psychiatry Hum Dev 39. Aebi M, Winkler Metzke C, Steinhausen HC (2009) Prediction of major affective disorders in adolescents by self-report measures. J Affect Disord 115:140–149 40. Zung WW, Richards CB, Short MJ (1965) Self-Rating Depression Scale in an outpatient clinic: further validation of the SDS. Arch Gen Psychiatr 13:508 41. Reynolds CR, Richmond BO (1979) Factor structure and construct validity of ‘What I Think and Feel’: the Revised Children’s Manifest Anxiety Scale. J Pers Assess 43:281–283 42. Ang RP, Lowe PA, Yusof N (2011) An examination of the RCMAS-2 scores across gender, ethnic background, and age in a large Asian school sample. Psychol Assess 23:899–910. doi:10. 1037/a0023891 43. Spielberger C (1973) Manual for state-trait anxiety interview for children. Consulting Psychologists Press, Palo Alto, CA 44. Garner DM, Olmsted MP (1984) Manual for eating disorder inventory (EDI). Psychological Assessment Resources, Incorporated 45. Franko DL, Striegel-Moore RH, Barton BA et al (2004) Measuring eating concerns in Black and White adolescent girls. Int J Eat Disord 35:179–189 46. Wear RW, Pratz O (1987) Test–retest reliability for the eating disorder inventory. Int J Eat Disord 6:767–769 47. Franko DL, Thompson D, Affenito SG et al (2008) What mediates the relationship between family meals and adolescent health issues. Health Psychol 27:S109 48. Stunkard AJ, Sørensen TI, Hanis C et al (1986) An adoption study of human obesity. N Engl J Med 314:193–198 49. Harter S (1985) Manual for the self-perception profile for children. University of Denver, Denver 50. Muris P, Meesters C, Fijen P (2003) The self-perception profile for children: further evidence for its factor structure, reliability, and validity. Pers Individ Diff 35:1791–1802. doi:10.1016/ S0191-8869(03)00004-7 51. Achenbach TM, Edelbrock CS (1983) Manual for the child behavior checklist and revised profile. University of Vermont, Department of Psychiatry, Burlington, VT 52. Goodman E, Slap GB, Huang B (2003) The public health impact of socioeconomic status on adolescent depression and obesity. Am J Public Health 93:1844–1850 53. Patton GC, Hibbert ME, Carlin J et al (1996) Menarche and the onset of depression and anxiety in Victoria, Australia. J Epidemiol Comm Health 50:661–666 54. Allison PD (1990) Change scores as dependent variables in regression analysis. Sociol Methodol 20:93–114

55. Bland JM, Altman DG (1994) Statistics notes: some examples of regression towards the mean. BMJ 309:780. doi:10.1136/bmj. 309.6957.780 56. Laird RD, Weems CF (2011) The equivalence of regression models using difference scores and models using separate scores for each informant: implications for the study of informant discrepancies. Psych Assess 23:388–397. doi:10.1037/a0021926 57. Wood AM, White IR, Thompson SG (2004) Are missing outcome data adequately handled? A review of published randomized controlled trials in major medical journals. Clin Trials 1:368–376 58. Sterne JAC, White IR, Carlin JB et al (2009) Multiple imputation for missing data in epidemiological and clinical research: potential and pitfalls. BMJ 338:b2393 59. Carlin JB, Galati JC, Royston P (2008) A new framework for managing and analyzing multiply imputed data in Stata. Stata J 8:49–67 60. Crespo C, Kielpikowski M, Pryor J, Jose PE (2011) Family rituals in New Zealand families: links to family cohesion and adolescents’ well-being. J Fam Psychol 25:184 61. Crespo C, Kielpikowski M, Jose PE, Pryor J (2010) Relationships between family connectedness and body satisfaction: a longitudinal study of adolescent girls and boys. J Youth Adolesc 39:1392–1401 62. White J, Halliwell E (2010) Alcohol and tobacco use during adolescence: the importance of the family mealtime environment. J Health Psychol 15:526–532. doi:10.1177/1359105309355337 63. Brent DA, Perper JA, Moritz G et al (1994) Familial risk factors for adolescent suicide: a case-control study. Acta Psychiatr Scand 89:52–58. doi:10.1111/j.1600-0447.1994.tb01485.x 64. Pe´rez RG, Ezpeleta L, Domenech JM (2007) Features associated with the non-participation and drop out by socially-at-risk children and adolescents in mental-health epidemiological studies. Soc Psychiatry Psychiatr Epidemiol 42:251–258. doi:10.1007/ s00127-006-0155-y 65. Rice F, Harold GT, Shelton KH, Thapar A (2006) Family conflict interacts with genetic liability in predicting childhood and adolescent depression. J Am Acad Child Adolesc Psychiatry 45:841–848 66. Kessler RC, Berglund P, Demler O et al (2005) Lifetime prevalence and age-of-onset distributions of DSM-IV disorders in the National Comorbidity Survey Replication. Arch Gen Psychiatry 62:593

123

Prospective associations between the family environment, family cohesion, and psychiatric symptoms among adolescent girls.

The present study used a longitudinal design to investigate whether parental values, interests, and the perceived level of expressiveness, cohesivenes...
257KB Sizes 0 Downloads 0 Views