- Email: [email protected]
The impact of postnatal and concurrent maternal depression on child behaviour during the early school years
Journal of Affective Disorders 119 (2009) 116–123
Contents lists available at ScienceDirect
Journal of Affective Disorders j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j a d
Research report
The impact of postnatal and concurrent maternal depression on child behaviour during the early school years Irene Fihrer ⁎, Catherine A. McMahon, Alan J. Taylor Psychology Department, Macquarie University, Sydney, Australia
a r t i c l e
i n f o
Article history: Received 22 November 2008 Received in revised form 2 March 2009 Accepted 2 March 2009 Available online 1 April 2009 Keywords: Maternal depression Postnatal depression Child behaviour Internalising problems Externalising problems Middle childhood
a b s t r a c t Background: This prospective study explores the ongoing impact of early and subsequent maternal depression on offspring behaviour in the early school years. Methods: Seventy five mothers recruited into a longitudinal study were assessed for symptoms of depression when their children were 4, 12 and 15 months, 4 years and later when the children were 6–8 years of age. Mothers, fathers, and school teachers were asked to report on children's internalising and externalising behaviour problems. Results: Exposure to maternal depression during the first postpartum year was related to mother reports of child internalising and externalising problems in the early school years. Although depression in the first year predicted later internalising problems, effects for externalising behaviour problems were mediated by concurrent depression. Relations between concurrent maternal depression and externalising problems were confirmed by teacher ratings. Interestingly, the severity of symptoms at four months was significantly correlated with behaviour problems seven years later. Limitations: Attrition over successive study contacts and therefore limited statistical power is acknowledged. Findings may be a conservative estimate of associations between maternal depression and later child behaviour problems. Also, the high prevalence of depressive symptomatology in the population from which the sample was drawn may limit the generalisability of results. Conclusions: Findings confirm the importance of early identification and treatment for mothers with postnatal depression, given the likelihood of ongoing depression and relations with later child behaviour problems. From a practical point, severity of early symptoms may be a reliable index of those mothers and children at greatest risk. © 2009 Elsevier B.V. All rights reserved.
1. Introduction Depression is a highly prevalent disorder among women of childbearing age. Approximately 13% of women will experience a postpartum episode (O'Hara and Swain, 1996) and while the majority resolve spontaneously within the first few months after birth, a postpartum episode increases the likelihood of recurrent episodes of depression (Cooper and Murray, 1995). In this study, we report on a cohort of mothers and their children first recruited at four months postpartum from the residential unit of a parentcraft centre, in suburban Sydney, Australia, for ⁎ Corresponding author. E-mail address: ifi[email protected] (I. Fihrer). 0165-0327/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.jad.2009.03.001
support with infant sleeping, settling or feeding. Sixty percent of the 127 mothers recruited met diagnostic criteria for postnatal depression (McMahon et al., 2001). Subsequently, 30% experienced persistent episodes of depression into the second postnatal year (McMahon et al., 2005), and 56% during the preschool years (McMahon et al., 2008). There is substantial evidence of adverse consequences of exposure to maternal depression for the socio-emotional wellbeing of offspring. Not only is maternal depression a predictor of depression and other internalising disorders in young children (for a review see Goodman and Gotlib, 1999), but the offspring of depressed mothers also display higher rates of externalising disorders relative to control groups (e.g. Beardslee et al., 1998; Lyons-Ruth et al., 2000).
I. Fihrer et al. / Journal of Affective Disorders 119 (2009) 116–123
117
Depression, however, is heterogeneous in presentation, course and severity and can range from a mild episode with little impairment, to a major depressive episode involving psychosis, immobilization and risk of suicide. Moreover, a child may be exposed to maternal depression immediately after birth (i.e., postnatal depression) or later in the child's life, or both. The present study examined the impact of depression on child behaviour in the early school years, focusing on three characteristics of depression, namely: a) timing, b) chronicity or recurrence, and c) severity.
Chronicity, severity, and concurrently elevated symptoms at the time of assessment are typically confounded (Brennan et al., 2000; Campbell et al., 1995). A second objective of this study was to explore the interrelationships among symptom severity and depression recurrence over time, and how these various characteristics of depression were related to child behaviour. In particular we were interested in exploring whether any effects of early depression on later child behaviour might be mediated by the mothers' depression symptoms at the time of the assessment.
1.1. Timing of maternal depression
2. Method
Many developmental theorists argue that exposure to maternal depression in the first 12 months of a child's life may be particularly detrimental, since during this “sensitive period” the infant is maximally dependent on the primary caretaker for stimulation and optimization of learning opportunities, which in turn, impact brain development, as well as perceptual and memory abilities (Hay et al., 2001; Sohr-Preston and Scaramella, 2006). During this time the foundations of attachment relationships and emotion regulation are also established, which have been shown to play a mediating role in later socio-emotional problems (Essex et al., 2001). Findings from two British prospective studies suggest that exposure to depression during this stage of development may have lasting adverse effects on the child's cognitive and behavioural functioning (see Hay et al., 2003, 2001, 2008; Murray, 1992; Murray et al., 1999). Other researchers argue, however, that it is not depression in the postnatal period, but later (generally concurrent) depression that impacts on the mother's interactions with the child and leads to adverse behavioural outcomes (e.g. Brennan et al., 2000). Consistent with this view, earlier research with the current sample showed that parent reports of behaviour problems at 15 months and 4 years were apparent only in cases where mothers had ongoing depression (Cornish et al., 2006; Trapolini et al., 2007). Given these divergent findings, a major objective of this study was to explore whether depression confined to the first postnatal year, the so called “sensitive” period of development, had a long-term impact on child behaviour, or whether any such effects were mediated by ongoing exposure to maternal depression.
2.1. Participants
1.2. Chronicity and severity of maternal depression Still others argue that the effects of depression on the child depend upon the duration and/or severity of maternal depression that the child is exposed to (Lyons-Ruth et al., 2000). Chronicity of depression (generally conceptualised by number of study contacts where mothers report elevated symptoms or meet diagnostic criteria) has been linked to lower maternal sensitivity (NICHD, 1999) and more child behaviour difficulties (Beardslee et al., 1998; Sohr-Preston and Scaramella, 2006). Surprisingly few studies have explored the impact of severity. Some argue that severity is one of the strongest determinants of socio-emotional problems in youths (Hammen and Brennan, 2003), whereas others have reported only modest associations between severity and later child pathology e.g., rank order correlations of − .21 to − .28 (Radke-Yarrow and Klimes-Dougan, 2002).
One hundred and twenty-seven mothers and their first born infants attending a residential parentcraft centre for assistance with sleeping and settling issues were recruited into the study when the child was 4 months. Sixty percent met criteria for an episode of major depression since childbirth (see McMahon et al., 2001 for more details). Mothers and children were subsequently seen at 12 and 15 months, 4 years and between 6 and 8 years (Mean = 82.17 months, SD = 6.91, range 71–102 months). See Table 1 for numbers assessed at each contact and characteristics of the sample. At child age 6–8 years, 75 mothers, 60% of the original sample and 82% of those seen at 4 years, participated. Of the 17 lost to follow-up after the 4 year contact, six (8%) could not be contacted, four (5%) had moved overseas, and seven (9%) declined participation due to time constraints. Sixty-three percent (n = 47) of partners, all biological fathers, participated at 6–8 years; (Mean age = 39.41 years, SD = 4.26, range 32–53). Fathers who were separated or divorced were less likely to participate (10 of the 14 did not complete questionnaires). Fathers completed depression symptom measures on two occasions — at child age four years and 6– 8 years. Of the 75 children's primary school class teachers invited to participate, 67 (89%) reported on the child's behaviour at school mid-way through the school year to ensure they knew the children adequately. When compared on demographic indices (child gender, marital status, education, language background, initial depression status), there were no significant differences between those retained in the study at 6–8 years and those in the original sample (all ps N .10). Moreover the reasons for drop out did not appear to be related to the variables of interest. 2.2. Measures 2.2.1. Depression At each of the five study points current symptomatology was measured using the Centre for Epidemiological Studies Depression Scale (CES-D: Radloff, 1977), a 20-item self-report measure, with total scores of 16 or more indicating levels of symptomatology likely to be of clinical significance. Fathers were also asked to complete this symptom checklist at the 4 and 6–8 year assessments, as it was considered possible that paternal depression may further contribute to child problems and/or influence father reporting of child behaviour. The depression module of the Composite International Diagnostic Interview (CIDI: World Health Organization, 1997)
118
I. Fihrer et al. / Journal of Affective Disorders 119 (2009) 116–123
Table 1 Sample characteristics.
Na Gender Male:female % Tertiary education % Non English speaking back-ground % Diagnosed with PND at 4 months Mean CES-D score: Mother Mean CES-D score: Father a
4 months
12 months
15 months
4 years
6–8 years
127 67:60
114 58:56
114 58:56
92 46:46
75 39:36
69% (n = 87) 21% (n = 26) 60% (n = 76) 11.71 –
74% (n = 84) 20% (n = 23) 62% (n = 73) 12.12 –
74% (n = 84) 20% (n = 23) 62% (n = 73) 12.12 –
83% (n = 76) 16% (n = 15) 73% (n = 67) 8.99 7.71
76% (n = 57) 5% (n = 4) 76% (n = 57) 7.74 7.04
Number of participants with complete data at each time point.
was used to diagnose depressive episodes retrospective to the previous assessment, according to criteria specified in the DSM IV. Since the 12 and 15 month assessments occurred so close together, the CIDI was only administered at four of the five time points (i.e. 4 and 12 months; 4 and 6–8 years). Trained interviewers achieved agreement on diagnosis of depression from 30 tape recorded interviews of the sample, given as a weighted coefficient (k), ranged from .87 to .91 at the different study contacts. 2.2.2. Depression variables Mothers were considered to have experienced depression in the first postnatal year if they met criteria for major depression on the CIDI at 4 months and/or between 5 and 12 months. To test overall exposure, we calculated the total number of time periods (0 to 4) during which the child was exposed to at least one episode meeting diagnostic criteria, a similar approach to that employed in the NICHD study (NICHD, 1999). An overall severity score was created by summing CES-D scores from each of the study contacts and taking an average which allowed for less impact of missing data, at any single time point. 2.2.3. Child behaviour To ensure the validity of the data, and to assess the level of functional disruption related to the child's behaviour problems, we used a structured clinical interview with mothers (ADIS-P) in addition to questionnaires completed by mothers, fathers and teachers (CBCL). Both these measures were completed during the child's first two years of school, aged 6–8 years.
ety, sleep terrors, etc) and externalising disorders (e.g. ADHD, oppositional defiant disorder etc). Mothers answered several screener questions. If answered affirmatively, the interviewer further explored symptoms to obtain frequency, intensity, and interference ratings (0–8 scale). Interference ratings of four or more indicated that the child met DSM IV diagnostic criteria for a childhood disorder. Trained postgraduate-student clinicians, blind to the depression status of the mother, conducted the interviews which were tape recorded. Interrater reliability was calculated from 30 with agreement across the various diagnoses, given as a weighted coefficient (k), ranging from .70 to .77. Examination of the ADIS-P (for DSM-IV) has yielded acceptable to excellent 7–14-day test–retest reliability (Silverman et al., 2001). Following protocol from previous studies (e.g. Grills, Ollendick and Thomas, 2003), disorders were grouped into internalising and externalising disorders. Children were then classified according to the presence (one or more problems receiving a rating ≥4) or absence of both. 2.3. Procedure All relevant Ethics Committees approved the project. Questionnaires were mailed prior to the interview in the family home. Interviewers were trained clinical psychologists blind to previous depression status. With maternal consent, teachers were asked to complete a questionnaire regarding the children's behaviour at school. 3. Results 3.1. Preliminary analyses
2.2.4. Questionnaires The Child Behaviour Checklist (CBCL) and The Teacher Report Form (TRF) are widely used in both clinical and research settings and have demonstrated reliability, validity, and ease of administration and scoring (Dutra and Campbell, 2004). Validity is confirmed as scores discriminate between clinicreferred and non-referred children (Achenbach and Rescorla, 2001). 2.2.5. Interview The Anxiety Disorders Interview Schedule — Parent version (ADIS-P: (Silverman and Albano, 1996)) is a semi-structured diagnostic interview that assesses the clinical significance of reported behaviour problems, in particular the interference in the child's and family's life. The interview explores childhood problems, including both internalising (e.g. separation anxi-
3.1.1. Demographics Results revealed a trend for mothers from a non-English speaking background to report higher levels of externalising behaviours, (p b .10). Maternal education was related to fathers' reports of externalising behaviours (p b 05) with higher scores reported for children whose mothers had higher education levels. A conservative approach was taken whereby maternal education and non-English speaking background were included in all analyses as covariates. 3.1.2. Paternal depression symptoms Three fathers (6%) reported depressive symptoms of clinical significance (≥16) at child age 6–8 years. Bivariate correlations between mothers' and fathers' CES-D scores revealed no significant relationship, but a positive trend (r = .26, p b .10).
I. Fihrer et al. / Journal of Affective Disorders 119 (2009) 116–123
119
Table 2 Correlations between mother, father and teacher ratings of behavioural problems. CBCL scores
Mother Father Teacher
Mother
Internalising Externalising Internalising Externalising Internalising
Father
Teacher
Internalising
Externalising
Internalising
Externalising
Internalising
Externalising
–
.62⁎⁎ –
.46⁎⁎ .43⁎ –
.36⁎⁎ .49⁎⁎ .35⁎⁎ –
.07 .05 .05 .24 –
.19 .32⁎⁎ .06 − .20 .32⁎⁎
⁎p b .05. ⁎⁎p b .01.
3.1.3. Child behaviour Mothers' and fathers' ratings were significantly correlated for both internalising and externalising behaviours, r = .46 and r = .43, respectively, (both ps b .01). Mothers' ratings were significantly correlated with teachers' ratings for externalising behaviours, r = .32, p b .01, but not for internalising behaviours (See Table 2). Mothers' scores were consistently higher (i.e. more problematic) than fathers' and teachers'. In the ADIS-P interview 35% (n = 26) of mothers rated their children as having a clinically significant internalising problem, and 27% (n = 20) a clinically significant externalising problem. Mothers' ratings on the CBCL subscales were significantly related to diagnostic data from the ADIS-P for internalising and externalising behaviours, t (70) = 4.02, p b .01 and t (70) = 3.42, p b .01, respectively: Mean for those reporting clinical internalising problems= 58.60 (SD = 11.39); Mean for nonclinical internalising = 48.37 (SD = 7.99); Mean for those reporting clinical externalising problems= 62.63 (SD = 6.21); Mean non-clinical externalising= 50.58 (SD = 9.69). 3.1.4. Relations among depression variables over time Meeting diagnostic criteria at each study contact predicted recurrence prior to the next contact: from 4–12 months, χ2 (1, N = 74) = 19.06, p b 01; from 12 months–4 years, χ2 (1, N = 75) = 15.24, p b 01; from 4 years–6–8 years, χ2 (1, N = 75) = 19.22, p b 01. Further, mothers' CES-D scores at 6– 8 years were significantly related to a diagnosis of depression in the first year postpartum: Mean CES-D score at 6–8 years for “Mothers depressed in first year” 14.40 (SD = 9.24); Mean CES-D score 6–8 years for “Mothers not depressed in first year” 5.17 (SD = 4.65), t (70) = 5.62 p b .001. Overall there were moderate to high correlations among CES-D scores over time (see Table 3). To explore the inter-relationships among symptom severity and recurrence over time, a number of correlations were tested. The total number of depressive episodes the mother reported, used as an index of chronicity, was correlated with
Table 3 Correlations between CES-D scores at 4, 12, 15 months, 4 years and 6–8 years. CES-D CES-D CES-D CES-D CES-D 4 months 12 months 15 months 4 years 6–8 years CES-D 4 months – CES-D 12 months CES-D 15 months CES-D 4 years ⁎⁎p b .01.
.536⁎⁎ –
.591⁎⁎ .800⁎⁎ –
.667⁎⁎ .653⁎⁎ .639⁎⁎ –
.610⁎⁎ .676⁎⁎ .662⁎⁎ .613⁎⁎
initial symptom severity at 4 months (r = .46; p b .01); mean symptom severity score (r = .52; p b .01), and concurrent symptoms at 6–8 years (r = .46; p b .01). These data confirm that severity, chronicity and concurrent symptoms are confounded. 3.1.5. Depression in the first postnatal year and later child behaviour problems Depression in the first postnatal year was significantly correlated with mothers' reports of later internalising (r = .36; p b .01) and externalising problems (r = .28; p b .05); with a trend for an association with teachers' ratings of externalising problems (r = .21; p b .10). Father behaviour ratings were not significantly related to maternal depression in the first year. A regression analysis including covariates (maternal education, NESB) was conducted. All variables were entered together. Results confirmed that, compared to mothers not depressed in the first postnatal year, mothers depressed in the first year reported more internalising and externalising behaviours in their children (see Table 4). No significant relationships were found among depression in the first postnatal year and fathers' or teachers' reports, all ps N .10. According to the ADIS interview, 9% (n = 5) of mothers who were not depressed in the first postnatal year rated their child as having a clinically significant internalising problem at 6–8 years, compared with 34% (n = 12) who had been depressed in the first postnatal year, χ2 (1, N = 75) = 12.61, p b .001. With regard to externalising problems, 7% (n = 4) of mothers not depressed in the first postnatal year (n = 53) rated their child as having a clinically significant externalising problem, compared with 18% (n = 4) of the postnatally depressed mothers (n = 22). This difference, while in the expected direction, was not significant. 3.1.6. Later/concurrent depression and child behaviour problems Bivariate correlations revealed that concurrent symptoms were significantly correlated with mothers' reports of internalising (r = .38; p b .01) and externalising problems (r = .41; p b .01) and teachers' ratings of externalising problems (r = .32; p b .01) at child age 6–8 years. Total number of depressive episodes was also correlated with mothers' reports of internalising (r = .36; p b .01) and externalising (r = .31; p b .01) problems, as was overall symptom severity; internalising (r = .34; p b .05), and externalising (r = .29; p b .05) problems. Concurrent maternal depression was not related, however, to fathers' reports of child internalising (r = .04; p = .80) or externalising problems (r = .11; p = .46). A regression analysis including covariates confirmed these
120
I. Fihrer et al. / Journal of Affective Disorders 119 (2009) 116–123
Table 4 Regression analysis: early maternal depression and child behaviour problems. DV
Variable
B
Standard error
Beta
t
p
Internalising
Education NESB Dep. Yr 1
− .05 4.61 7.40
2.74 4.75 2.43
− .002 .11 .35
− .02 .97 3.04
.99 .34 .003
Education NESB Dep. Yr 1
.67 7.81 5.62
2.91 5.05 2.59
.03 .18 .25
.23 1.55 2.17
Externalising
a
.82 .13 .03
Adjusted means a
Depressed 55.87 (2.06)
Not depressed 50.34 (1.35)
Depressed 56.01 (2.20)
Not depressed 50.34 (1.35)
The adjusted means are the estimated marginal means of the dependent variable, adjusted for the other variables in the model.
associations for mothers' ratings of internalising and externalising problems (see Table 5). 3.1.7. Testing the mediating role of later depression symptoms A mediation model (Baron and Kenny, 1986) (Figs. 1 and 2) was tested to explore both direct effects of early depression on current child behaviour and indirect effects mediated by ongoing maternal depression. As noted earlier, chronicity (number of episodes) severity (mean symptom score) and concurrent symptoms were all highly correlated. Since there are theoretical grounds for believing concurrent symptoms may influence both child behaviour (e.g. Brennan et al., 2000) and the mother's reporting of child behaviour (e.g. Richters, 1992) we have opted to use concurrent symptoms, a continuous variable, as our primary index of ongoing depression in the mediation analyses. The analyses in Figs. 1 and 2 have satisfied the first three pre-requisites specified by Baron and Kenny (1986) for testing mediation. First, a significant relationship between early maternal depression and child behaviour was established. Then, the relationship between early and concurrent depression was established, finally, the relationship between concurrent depression and child behaviour was established. The next steps were carried out using an SPSS macro produced by Preacher and Hayes (2008). As well as providing the conventional estimates of path b (the effect of concurrent depression on child behaviour, adjusted for early maternal depression), and the effect of path c′ (the direct effect of early depression on child behaviour, adjusted for concurrent depression), this program provides a bootstrapped estimate of the significance of the indirect effect (path a ⁎ b). Bootstrapping methods have been found to be among the most powerful and accurate when producing confidence intervals
for indirect effects (Fritz and Mackinnon, 2007; Mackinnon et al., 2004). The results (Figs. 1 and 2) suggest that in the case of externalising problems, as reported by mothers, the effects of early maternal depression are fully mediated by concurrent symptoms. That is, early maternal depression affects behaviour through its effect on concurrent depression, which itself has a unique effect on externalising behaviour. In the case of internalising problems, however, there is no evidence that concurrent depression attenuated the effect of the mother's early depression in explaining the children's current behaviour. The mediation model was not tested for teacher ratings as there was no significant relationship with depression in the first postnatal year. Finally, a second mediation model was tested to determine whether recurrent maternal depression (indexed by total number of episodes) as opposed to concurrent depression might mediate the effects of early depression. There was no attenuation of the effect of early
Table 5 Regression analysis: later/concurrent depression and child behaviour problems. DV Internalising
Variable
Education NESB Concurrent depression Externalising Education NESB Concurrent depression
B
Standard Beta error
− .21 2.72 3.59 4.76 .47 .15 .28 2.77 6.04 4.86 .53 .15
t
p
− .009 − .07 .94 .09 .75 .45 .36 3.18 .002 .01 .14 .39
.10 .92 1.24 .22 3.49 .001
Fig. 1. a = effect of early maternal depression on concurrent maternal depression, b = effect of concurrent depression on child behaviour, adjusted for early maternal depression, c = total effect of early depression on child behaviour, c′ = direct effect of early depression on child behaviour, adjusted for concurrent depression.
I. Fihrer et al. / Journal of Affective Disorders 119 (2009) 116–123
Fig. 2. a = effect of early maternal depression on concurrent maternal depression, b = effect of concurrent depression on child behaviour, adjusted for early maternal depression, c = total effect of early depression on child behaviour, c′ = direct effect of early depression on child behaviour, adjusted for concurrent depression.
depression when recurrent depression (rather than concurrent symptoms) was included in the model. 3.1.8. Post hoc analysis — relationship of initial symptom severity to later behaviour problems Finally, given the strong correlation between the index episode and later episodes of depression, including scores at 6–8 years (see Table 3), a post hoc analysis was also conducted to determine the relationship between the severity of the index CES-D score at four months and behaviour problems at 6–8 years. There were significant correlations with both internalising and externalising behaviours, r = .32, p = .01, and r = .25, p = .01, respectively, and also with teachers' ratings for externalising behaviours, r = .27, p = 04. 4. Discussion This prospective study aimed to clarify the impact of timing, chronicity and severity of maternal depression on offspring behaviour in the early school years. With regard to timing, both early and later depression seem to be important. Exposure to maternal depression during the first postnatal year was related to mothers' reports of child internalising and externalising problems in the early school years. Interestingly, the severity of symptoms at four months was significantly correlated with behaviour problems six to eight years later. Further, a substantial number of mothers (34%) who had been depressed during the first postnatal year perceived their child's internalising problems as causing significant interference in the child's and/or family's life six years later. The effect of early depression on later internalising problems was not attenuated by subsequent depression, but effects for externalising beha-
121
viour problems were mediated by concurrent depression. Concurrent maternal depression was related to both internalising and externalising problems as reported by the mother, and to externalising problems reported by the teacher. There are a number of possible explanations for the relationship between depression in the first postnatal year and later internalising problems in the children. Intergenerational transmission of risk whether through problematic relationship exchanges between the mother and child and/or a genetic vulnerability seems plausible. Previous research (Essex et al., 2001) suggests that the first postnatal year represents a “sensitive period” for later vulnerability to internalising symptoms as it is an important time for the development of a secure mother–child attachment which in turn impacts emotion regulation skills (Campbell et al., 1995). Earlier findings with this sample suggest compromised mother–child interaction may have contributed. Mothers with postnatal depression persisting throughout the first postnatal year were more likely to have insecure attachment relationships with their children (McMahon et al., 2006). Findings regarding the stability of elevated depression scores over time and relations between postnatal depression and child internalising behaviour problems, on the other hand, also suggest the possibility of a genetic vulnerability to internalising problems in both mother and child. It seems reasonable to conclude that in the case of externalising problems, early depression may impact the child indirectly, by increasing the mother's risk for later depressive episodes. Our findings show that early depression predicts later episodes, and there are moderate to high correlations for depression symptoms over time. It is likely that these later episodes, which occur closer to the assessment time points, may be causally related to the problematic externalising behaviours, and that the children were reacting behaviourally to their mothers' depressed mood. Since mothers reported higher levels of problem behaviour than other raters, (and given the importance of concurrent symptoms) the possibility that mothers may be projecting their own negative thoughts and feelings onto their children must also be considered. The moderate correlations with fathers' reports for both internalising and externalising problems and teachers' reports for externalising behaviours, however, make this interpretation less likely. Research into inter-rater agreement on behaviour has commonly found greater levels of correspondence for informants' ratings of child externalising problems than internalising problems (De los Reyes and Kazdin, 2005), as was the case in this study. Teachers' reports of externalising problems are also important in showing that the effect of maternal depression is evident outside the home, albeit with milder problems reported. It is possible that the stricter boundaries and greater structure inherent in the school system serve to contain the children's problematic behaviours. There are a number of possible explanations for the null findings with respect to fathers. The majority worked fulltime and their involvement in care-giving responsibilities was considerably less than mothers, so it is possible that their perceptions of their children's behaviour were less reliable. It is also important to note that relatively few fathers participated, yielding limited power for these analyses. Also, most fathers who were separated or divorced (whose children may have had more problematic behaviour) did not participate.
122
I. Fihrer et al. / Journal of Affective Disorders 119 (2009) 116–123
4.1. Study limitations While this study represents a comprehensive, long-term follow up of a socio-economically low risk sample of clinically depressed mothers, the high prevalence of depressive symptomatology in the population from which the sample was drawn may limit the generalisability of our results. Further, participants were all mothers who experienced early difficulties with infant sleeping and settling. It could be argued that this represents a more rigorous test of depression effects on later behaviour, as there is some control for infant/ child factors. It is possible, therefore, that a comparison between depressed mothers and a community sample without a history of mother–infant difficulties may have revealed greater depression effects. A major limitation is the attrition over successive study contacts with just 60% of the original sample participating at child age 6–8 years and variability in the number with full questionnaire data at each timepoint. We have confirmed, however, that ongoing participants are similar to the original sample with regard to demographic variables and early depression. We acknowledge limited statistical power (particularly for analyses involving fathers' ratings) and believe the findings may be a conservative estimate of associations between maternal depression and later child behaviour problems, but this requires further confirmation in larger samples. Future longitudinal studies might deal with attrition problems by recruiting larger sample sizes at the outset in anticipation of inevitable attrition. Researchers could also consider providing incentives for continued participation through monetary compensation, and use online questionnaire delivery to minimize missing data. A substantial number of mothers in this sample have reported moderate to severe symptoms at each of the study time points. This pattern of ongoing depression is suggestive of a “depressive personality style” (Boyce et al. 1991) or a possible Axis II personality disorder, neither of which were directly assessed in the current study. We have also previously noted that mothers with an insecure working model of attachment are much more likely to experience persistent depression, at least until the preschool years (McMahon et al., 2006, 2008). Research into co-morbid personality disturbance and attachment difficulties provides an interesting direction for future research. It is also possible, given the pattern of depression over time, that mothers in this sample may have been depressed in pregnancy. There is a growing body of evidence suggesting that maternal mood in pregnancy may be related to later child behaviour difficulties and that this effect may be due to fetal programming of the Hypothalamic Pituitary Axis (see Talge et al., 2007, for a recent review). Future prospective studies including measures of maternal mood in pregnancy are important in testing this explanatory model. Finally, causation in this study is generally inconclusive, particularly with regard to concurrent depression symptoms. It remains unclear whether maternal depression causes behaviour problems in offspring or vice versa. In infancy, when offspring of depressed mothers interact with a non-depressed adult, the nondepressed adult becomes more negative and their affect is more flat suggesting that the infants' “depressed behaviour” may elicit and then
continue to reinforce maternal depression (e.g. McCormick, 1995). Other studies have also noted that maternal stress and depressive symptoms tend to subside when their children's behaviour improves (Elgar et al., 2003), and increase as the behaviour worsens (Gartstein and Sheeber, 2004). It is possible that maternal depression and child behaviour problems may have a deleterious reciprocal influence. 5. Conclusions Findings confirm the importance of early identification and treatment for mothers with postnatal depression, given the likelihood of ongoing depression and relations with later child behaviour problems. From a practical point of view, the finding that symptom severity at four months can predict which mothers and children are at greatest risk could enable healthcare workers to target the most at risk group of mothers early on in the child's life. Role of the funding source This research was supported by a grant from Macquarie University.
Conflict of interest There are no conflicts of interest associated with this paper.
Acknowledgements
We are grateful to the families who participated in the study and to Felicity Chandler, Lisa Lia and Alia Palethorpe for help with data collection and data management. References Achenbach, T.M., Rescorla, L.A., 2001. Manual for ASEBA School-Age Forms & Profiles. Burlington, VT: University of Vermont, Research Center for Children, Youth & Families. Baron, R.M., Kenny, D.A., 1986. The moderator–mediator variable distinction in social psychological research: conceptual, strategic, and statistical consideration. Journal of Personality and Social Psychology 51, 1173–1182. Beardslee, W., Versage, E., Gladstone, T., 1998. Children of affectively ill parents: a review of the past 10 years. Journal of the American Academy of Child and Adolescent Psychiatry 37 (11), 1134–1141. Boyce, P., Parker, G., Barnett, B., Cooney, M., 1991. Personality as a vulnerability factor to depression. British Journal of Psychiatry 159, 106–114. Brennan, P.A., Hammen, C., Andersen, M., Bor, W., Najman, J., Williams, G., 2000. Chronicity, severity, and timing of maternal depressive symptoms: relationship with child outcome at age five. Developmental Psychology 36, 759–766. Campbell, S.B., Cohn, J.F., Meyers, T.A., 1995. Depression in first-time mothers: mother–infant interaction and depression chronicity. Developmental Psychology 31, 349–357. Cooper, P.J., Murray, L., 1995. Course and recurrence of postnatal depression: evidence for the specificity of the diagnostic concept. British Journal of Psychiatry 166, 191–195. Cornish, A.M., McMahon, C.A., Ungerer, J.A., Barnett, B., Kowalenko, N., Tennant, C., 2006. Maternal depression and the experience of parenting in the second postnatal year. Journal of Reproductive and infant psychology 24 (2), 121–132. De los Reyes, A., Kazdin, A.E., 2005. Informant discrepancies in the assessment of childhood psychopathology: a critical review, theoretical framework, and recommendations for future study. Psychological Bulletin 131 (4), 483–509. Dutra, L., Campbell, L., 2004. Quantifying clinical judgement in the assessment of adolescent psychopathology: reliability, validity and factor structure of the Child Behavior Checklist for clinician report. Journal of Clinical Psychology 60 (1), 65–85.
I. Fihrer et al. / Journal of Affective Disorders 119 (2009) 116–123 Elgar, F.J., Curtis, L.J., Mcgrath, P.J., Waschbusch, D.A., Stewart, S.H., 2003. Antecedent-consequence conditions in maternal mood and child adjustment: a four-year cross-lagged study. Journal of Clinical Child and Adolescent Psychology 32 (3), 362–374. Essex, M.J., Klein, M.H., Miech, R., Smider, N.A., 2001. Timing of initial exposure to maternal major depression and children's mental health symptoms in kindergarten. British Journal of Psychiatry 179, 151–156. Fritz, M.S., Mackinnon, D.P., 2007. Required sample size to detect the mediation effect. Psychological Science 18 (3), 233–239. Gartstein, M.A., Sheeber, L., 2004. Child behavior problems and maternal symptoms of depression: a mediational model. Journal of Child and Adolescent Psychiatric Nursing 17 (4), 141–150. Goodman, S.H., Gotlib, I.H., 1999. Risk for psychopathology in the children of depressed mothers: a developmental model for understanding mechanisms of transmission. Psychological Review 106 (3), 458–490. Grills, A.E., Ollendick, M.S., Thomas, H., 2003. Multiple informant agreement and the anxiety disorders interview schedule for parents and children. Journal of the American Academy of Child and Adolescent Psychiatry 42 (1), 30–40. Hammen, C., Brennan, P.A., 2003. Severity, chronicity, and timing of maternal depression and risk for adolescent offspring diagnoses in a community sample. Archive of General Psychiatry 60, 253–258. Hay, D.F., Pawlby, S., Sharp, D., Asten, P., Mills, A., Kumar, R., 2001. Intellectual problems shown by 11-year old children whose mothers had postnatal depression. Journal of Child Psychology and Psychiatry 42, 871–889. Hay, D.F., Pawlby, S., Angold, A., Harold, G.T., 2003. Pathways to violence in children of mothers who were depressed postpartum. Developmental Psychology 39 (6), 1038–1094. Hay, D.F., Pawlby, S., Waters, C., Sharp, D., 2008. Antepartum and postpartum exposure to maternal depression: different effects on different adolescent outcomes. Journal of Child Psychology and Psychiatry 49 (10), 1079–1088. Lyons-Ruth, K., Wolfe, R., Lyubchik, A., 2000. Depression and the parenting of young children: making the case for early preventative mental health services. Harvard Review of Psychiatry 8 (3), 148–153. Mackinnon, D.P., Lockwood, C.M., Williams, J., 2004. Confidence limits for the indirect effect: distribution of the product and resampling methods. Multivariate Behavioral Research 39 (1), 99–128. McCormick, L.H., 1995. Depression in mothers of children with attention deficit hyperactivity disorder. Family Medicine 27, 176–179. McMahon, C., Barnett, B., Kowalenko, N., Tennant, C., 2005. Psychological factors associated with postnatal depression: past and current relationships, defence styles and the mediating role of insecure attachment styles. Journal of Affective Disorders 84 (1), 15–24. McMahon, C., Barnett, B., Kowalenko, N., Tennant, C., 2006. Maternal attachment state of mind moderates the impact of postnatal depression on infant attachment. Journal of Child Psychology and Psychiatry 47 (7), 660–669. McMahon, C., Barnett, B., Kowalenko, N., Tennant, C., Don, N., 2001. Postnatal depression, anxiety and unsettled infant behaviour. Australian and New Zealand Journal of Psychiatry 35, 581–588.
123
McMahon, C., Trapolini, T., Barnett, B., 2008. Maternal state of mind regarding attachment predicts persistence of postnatal depression in the preschool years. Journal of Affective Disorders 107 (1–3), 199–203. Murray, L., 1992. The impact of postnatal depression on infant development. Journal of Child Psychology and Psychiatry 33, 543–561. Murray, L., Sinclair, D., Cooper, P., Ducournau, P., Turner, P., 1999. The socioemotional development of 5-year-old children of postnatally depressed mothers. Journal of Child Psychology and Psychiatry 40 (8), 1259–1271. NICHD, 1999. Chronicity of maternal depressive symptoms, maternal sensitivity, and child functioning at 36 months. Developmental Psychology 35, 1297–1310. O'Hara, M.W., Swain, A.M., 1996. Rates and risk of postpartum depression: a meta-analysis. International Review of Psychiatry 8, 37–54. Preacher, K.J., Hayes, A.F., 2008. Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior Research Methods 40, 879–891. Radke-Yarrow, M., Klimes-Dougan, B., 2002. Parental depression and offspring disorders: a developmental perspective. In: Goodman, S.H., Gotlib, I.H. (Eds.), Children of Depressed Parents: Mechanisms of Risk and Implications for Treatment Washington. American Psychological Association DC. Radloff, J.S., 1977. The CES-D Scale: a self-report depression scale for research in the general population. Applied Psychological Measurement 1, 385–401. Richters, J.E., 1992. Depressed mothers as informants about their children: a critical review of the evidence for distortion. Psychological Bulletin 112 (3), 485–499. Silverman, W.K., Albano, A.M., 1996. The Anxiety Disorders Interview Schedule for Children for DSM:IV: Child and Parent Versions. Psychological Corporation, San Antonio. Silverman, W.K., Saavedra, L.M., Pina, A.A., 2001. Test-retest reliability of anxiety symptoms and diagnoses with the anxiety disorders interview schedule for DSM-IV: Child and parent versions. American Academy of Child and Adolescent Psychiatry 40 (8), 937–944. Sohr-Preston, S.L., Scaramella, L.V., 2006. Implications of timing of maternal depressive symptoms for early cognitive and language development. Clinical Child and Family Psychology Review 9 (1), 65–83. Talge, N., Charles, N., Vivette, G., 2007. Antenatal maternal stress and longterm effects on child neurodevelopment: how and why? Journal of Child Psychology and Psychiatry 48 (3–4), 245–261. Trapolini, T., McMahon, C., Ungerer, J.A., 2007. The effect of maternal depression and marital adjustment on young children's internalizing and externalizing behaviour problems. Child, Care, Health and Development 33 (6), 794–804. World Health Organization, 1997. Composite International Diagnostic Interview (CIDI, Version 2.1). World Health Organization, Geneva, Switzerland.
Contents lists available at ScienceDirect
Journal of Affective Disorders j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j a d
Research report
The impact of postnatal and concurrent maternal depression on child behaviour during the early school years Irene Fihrer ⁎, Catherine A. McMahon, Alan J. Taylor Psychology Department, Macquarie University, Sydney, Australia
a r t i c l e
i n f o
Article history: Received 22 November 2008 Received in revised form 2 March 2009 Accepted 2 March 2009 Available online 1 April 2009 Keywords: Maternal depression Postnatal depression Child behaviour Internalising problems Externalising problems Middle childhood
a b s t r a c t Background: This prospective study explores the ongoing impact of early and subsequent maternal depression on offspring behaviour in the early school years. Methods: Seventy five mothers recruited into a longitudinal study were assessed for symptoms of depression when their children were 4, 12 and 15 months, 4 years and later when the children were 6–8 years of age. Mothers, fathers, and school teachers were asked to report on children's internalising and externalising behaviour problems. Results: Exposure to maternal depression during the first postpartum year was related to mother reports of child internalising and externalising problems in the early school years. Although depression in the first year predicted later internalising problems, effects for externalising behaviour problems were mediated by concurrent depression. Relations between concurrent maternal depression and externalising problems were confirmed by teacher ratings. Interestingly, the severity of symptoms at four months was significantly correlated with behaviour problems seven years later. Limitations: Attrition over successive study contacts and therefore limited statistical power is acknowledged. Findings may be a conservative estimate of associations between maternal depression and later child behaviour problems. Also, the high prevalence of depressive symptomatology in the population from which the sample was drawn may limit the generalisability of results. Conclusions: Findings confirm the importance of early identification and treatment for mothers with postnatal depression, given the likelihood of ongoing depression and relations with later child behaviour problems. From a practical point, severity of early symptoms may be a reliable index of those mothers and children at greatest risk. © 2009 Elsevier B.V. All rights reserved.
1. Introduction Depression is a highly prevalent disorder among women of childbearing age. Approximately 13% of women will experience a postpartum episode (O'Hara and Swain, 1996) and while the majority resolve spontaneously within the first few months after birth, a postpartum episode increases the likelihood of recurrent episodes of depression (Cooper and Murray, 1995). In this study, we report on a cohort of mothers and their children first recruited at four months postpartum from the residential unit of a parentcraft centre, in suburban Sydney, Australia, for ⁎ Corresponding author. E-mail address: ifi[email protected] (I. Fihrer). 0165-0327/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.jad.2009.03.001
support with infant sleeping, settling or feeding. Sixty percent of the 127 mothers recruited met diagnostic criteria for postnatal depression (McMahon et al., 2001). Subsequently, 30% experienced persistent episodes of depression into the second postnatal year (McMahon et al., 2005), and 56% during the preschool years (McMahon et al., 2008). There is substantial evidence of adverse consequences of exposure to maternal depression for the socio-emotional wellbeing of offspring. Not only is maternal depression a predictor of depression and other internalising disorders in young children (for a review see Goodman and Gotlib, 1999), but the offspring of depressed mothers also display higher rates of externalising disorders relative to control groups (e.g. Beardslee et al., 1998; Lyons-Ruth et al., 2000).
I. Fihrer et al. / Journal of Affective Disorders 119 (2009) 116–123
117
Depression, however, is heterogeneous in presentation, course and severity and can range from a mild episode with little impairment, to a major depressive episode involving psychosis, immobilization and risk of suicide. Moreover, a child may be exposed to maternal depression immediately after birth (i.e., postnatal depression) or later in the child's life, or both. The present study examined the impact of depression on child behaviour in the early school years, focusing on three characteristics of depression, namely: a) timing, b) chronicity or recurrence, and c) severity.
Chronicity, severity, and concurrently elevated symptoms at the time of assessment are typically confounded (Brennan et al., 2000; Campbell et al., 1995). A second objective of this study was to explore the interrelationships among symptom severity and depression recurrence over time, and how these various characteristics of depression were related to child behaviour. In particular we were interested in exploring whether any effects of early depression on later child behaviour might be mediated by the mothers' depression symptoms at the time of the assessment.
1.1. Timing of maternal depression
2. Method
Many developmental theorists argue that exposure to maternal depression in the first 12 months of a child's life may be particularly detrimental, since during this “sensitive period” the infant is maximally dependent on the primary caretaker for stimulation and optimization of learning opportunities, which in turn, impact brain development, as well as perceptual and memory abilities (Hay et al., 2001; Sohr-Preston and Scaramella, 2006). During this time the foundations of attachment relationships and emotion regulation are also established, which have been shown to play a mediating role in later socio-emotional problems (Essex et al., 2001). Findings from two British prospective studies suggest that exposure to depression during this stage of development may have lasting adverse effects on the child's cognitive and behavioural functioning (see Hay et al., 2003, 2001, 2008; Murray, 1992; Murray et al., 1999). Other researchers argue, however, that it is not depression in the postnatal period, but later (generally concurrent) depression that impacts on the mother's interactions with the child and leads to adverse behavioural outcomes (e.g. Brennan et al., 2000). Consistent with this view, earlier research with the current sample showed that parent reports of behaviour problems at 15 months and 4 years were apparent only in cases where mothers had ongoing depression (Cornish et al., 2006; Trapolini et al., 2007). Given these divergent findings, a major objective of this study was to explore whether depression confined to the first postnatal year, the so called “sensitive” period of development, had a long-term impact on child behaviour, or whether any such effects were mediated by ongoing exposure to maternal depression.
2.1. Participants
1.2. Chronicity and severity of maternal depression Still others argue that the effects of depression on the child depend upon the duration and/or severity of maternal depression that the child is exposed to (Lyons-Ruth et al., 2000). Chronicity of depression (generally conceptualised by number of study contacts where mothers report elevated symptoms or meet diagnostic criteria) has been linked to lower maternal sensitivity (NICHD, 1999) and more child behaviour difficulties (Beardslee et al., 1998; Sohr-Preston and Scaramella, 2006). Surprisingly few studies have explored the impact of severity. Some argue that severity is one of the strongest determinants of socio-emotional problems in youths (Hammen and Brennan, 2003), whereas others have reported only modest associations between severity and later child pathology e.g., rank order correlations of − .21 to − .28 (Radke-Yarrow and Klimes-Dougan, 2002).
One hundred and twenty-seven mothers and their first born infants attending a residential parentcraft centre for assistance with sleeping and settling issues were recruited into the study when the child was 4 months. Sixty percent met criteria for an episode of major depression since childbirth (see McMahon et al., 2001 for more details). Mothers and children were subsequently seen at 12 and 15 months, 4 years and between 6 and 8 years (Mean = 82.17 months, SD = 6.91, range 71–102 months). See Table 1 for numbers assessed at each contact and characteristics of the sample. At child age 6–8 years, 75 mothers, 60% of the original sample and 82% of those seen at 4 years, participated. Of the 17 lost to follow-up after the 4 year contact, six (8%) could not be contacted, four (5%) had moved overseas, and seven (9%) declined participation due to time constraints. Sixty-three percent (n = 47) of partners, all biological fathers, participated at 6–8 years; (Mean age = 39.41 years, SD = 4.26, range 32–53). Fathers who were separated or divorced were less likely to participate (10 of the 14 did not complete questionnaires). Fathers completed depression symptom measures on two occasions — at child age four years and 6– 8 years. Of the 75 children's primary school class teachers invited to participate, 67 (89%) reported on the child's behaviour at school mid-way through the school year to ensure they knew the children adequately. When compared on demographic indices (child gender, marital status, education, language background, initial depression status), there were no significant differences between those retained in the study at 6–8 years and those in the original sample (all ps N .10). Moreover the reasons for drop out did not appear to be related to the variables of interest. 2.2. Measures 2.2.1. Depression At each of the five study points current symptomatology was measured using the Centre for Epidemiological Studies Depression Scale (CES-D: Radloff, 1977), a 20-item self-report measure, with total scores of 16 or more indicating levels of symptomatology likely to be of clinical significance. Fathers were also asked to complete this symptom checklist at the 4 and 6–8 year assessments, as it was considered possible that paternal depression may further contribute to child problems and/or influence father reporting of child behaviour. The depression module of the Composite International Diagnostic Interview (CIDI: World Health Organization, 1997)
118
I. Fihrer et al. / Journal of Affective Disorders 119 (2009) 116–123
Table 1 Sample characteristics.
Na Gender Male:female % Tertiary education % Non English speaking back-ground % Diagnosed with PND at 4 months Mean CES-D score: Mother Mean CES-D score: Father a
4 months
12 months
15 months
4 years
6–8 years
127 67:60
114 58:56
114 58:56
92 46:46
75 39:36
69% (n = 87) 21% (n = 26) 60% (n = 76) 11.71 –
74% (n = 84) 20% (n = 23) 62% (n = 73) 12.12 –
74% (n = 84) 20% (n = 23) 62% (n = 73) 12.12 –
83% (n = 76) 16% (n = 15) 73% (n = 67) 8.99 7.71
76% (n = 57) 5% (n = 4) 76% (n = 57) 7.74 7.04
Number of participants with complete data at each time point.
was used to diagnose depressive episodes retrospective to the previous assessment, according to criteria specified in the DSM IV. Since the 12 and 15 month assessments occurred so close together, the CIDI was only administered at four of the five time points (i.e. 4 and 12 months; 4 and 6–8 years). Trained interviewers achieved agreement on diagnosis of depression from 30 tape recorded interviews of the sample, given as a weighted coefficient (k), ranged from .87 to .91 at the different study contacts. 2.2.2. Depression variables Mothers were considered to have experienced depression in the first postnatal year if they met criteria for major depression on the CIDI at 4 months and/or between 5 and 12 months. To test overall exposure, we calculated the total number of time periods (0 to 4) during which the child was exposed to at least one episode meeting diagnostic criteria, a similar approach to that employed in the NICHD study (NICHD, 1999). An overall severity score was created by summing CES-D scores from each of the study contacts and taking an average which allowed for less impact of missing data, at any single time point. 2.2.3. Child behaviour To ensure the validity of the data, and to assess the level of functional disruption related to the child's behaviour problems, we used a structured clinical interview with mothers (ADIS-P) in addition to questionnaires completed by mothers, fathers and teachers (CBCL). Both these measures were completed during the child's first two years of school, aged 6–8 years.
ety, sleep terrors, etc) and externalising disorders (e.g. ADHD, oppositional defiant disorder etc). Mothers answered several screener questions. If answered affirmatively, the interviewer further explored symptoms to obtain frequency, intensity, and interference ratings (0–8 scale). Interference ratings of four or more indicated that the child met DSM IV diagnostic criteria for a childhood disorder. Trained postgraduate-student clinicians, blind to the depression status of the mother, conducted the interviews which were tape recorded. Interrater reliability was calculated from 30 with agreement across the various diagnoses, given as a weighted coefficient (k), ranging from .70 to .77. Examination of the ADIS-P (for DSM-IV) has yielded acceptable to excellent 7–14-day test–retest reliability (Silverman et al., 2001). Following protocol from previous studies (e.g. Grills, Ollendick and Thomas, 2003), disorders were grouped into internalising and externalising disorders. Children were then classified according to the presence (one or more problems receiving a rating ≥4) or absence of both. 2.3. Procedure All relevant Ethics Committees approved the project. Questionnaires were mailed prior to the interview in the family home. Interviewers were trained clinical psychologists blind to previous depression status. With maternal consent, teachers were asked to complete a questionnaire regarding the children's behaviour at school. 3. Results 3.1. Preliminary analyses
2.2.4. Questionnaires The Child Behaviour Checklist (CBCL) and The Teacher Report Form (TRF) are widely used in both clinical and research settings and have demonstrated reliability, validity, and ease of administration and scoring (Dutra and Campbell, 2004). Validity is confirmed as scores discriminate between clinicreferred and non-referred children (Achenbach and Rescorla, 2001). 2.2.5. Interview The Anxiety Disorders Interview Schedule — Parent version (ADIS-P: (Silverman and Albano, 1996)) is a semi-structured diagnostic interview that assesses the clinical significance of reported behaviour problems, in particular the interference in the child's and family's life. The interview explores childhood problems, including both internalising (e.g. separation anxi-
3.1.1. Demographics Results revealed a trend for mothers from a non-English speaking background to report higher levels of externalising behaviours, (p b .10). Maternal education was related to fathers' reports of externalising behaviours (p b 05) with higher scores reported for children whose mothers had higher education levels. A conservative approach was taken whereby maternal education and non-English speaking background were included in all analyses as covariates. 3.1.2. Paternal depression symptoms Three fathers (6%) reported depressive symptoms of clinical significance (≥16) at child age 6–8 years. Bivariate correlations between mothers' and fathers' CES-D scores revealed no significant relationship, but a positive trend (r = .26, p b .10).
I. Fihrer et al. / Journal of Affective Disorders 119 (2009) 116–123
119
Table 2 Correlations between mother, father and teacher ratings of behavioural problems. CBCL scores
Mother Father Teacher
Mother
Internalising Externalising Internalising Externalising Internalising
Father
Teacher
Internalising
Externalising
Internalising
Externalising
Internalising
Externalising
–
.62⁎⁎ –
.46⁎⁎ .43⁎ –
.36⁎⁎ .49⁎⁎ .35⁎⁎ –
.07 .05 .05 .24 –
.19 .32⁎⁎ .06 − .20 .32⁎⁎
⁎p b .05. ⁎⁎p b .01.
3.1.3. Child behaviour Mothers' and fathers' ratings were significantly correlated for both internalising and externalising behaviours, r = .46 and r = .43, respectively, (both ps b .01). Mothers' ratings were significantly correlated with teachers' ratings for externalising behaviours, r = .32, p b .01, but not for internalising behaviours (See Table 2). Mothers' scores were consistently higher (i.e. more problematic) than fathers' and teachers'. In the ADIS-P interview 35% (n = 26) of mothers rated their children as having a clinically significant internalising problem, and 27% (n = 20) a clinically significant externalising problem. Mothers' ratings on the CBCL subscales were significantly related to diagnostic data from the ADIS-P for internalising and externalising behaviours, t (70) = 4.02, p b .01 and t (70) = 3.42, p b .01, respectively: Mean for those reporting clinical internalising problems= 58.60 (SD = 11.39); Mean for nonclinical internalising = 48.37 (SD = 7.99); Mean for those reporting clinical externalising problems= 62.63 (SD = 6.21); Mean non-clinical externalising= 50.58 (SD = 9.69). 3.1.4. Relations among depression variables over time Meeting diagnostic criteria at each study contact predicted recurrence prior to the next contact: from 4–12 months, χ2 (1, N = 74) = 19.06, p b 01; from 12 months–4 years, χ2 (1, N = 75) = 15.24, p b 01; from 4 years–6–8 years, χ2 (1, N = 75) = 19.22, p b 01. Further, mothers' CES-D scores at 6– 8 years were significantly related to a diagnosis of depression in the first year postpartum: Mean CES-D score at 6–8 years for “Mothers depressed in first year” 14.40 (SD = 9.24); Mean CES-D score 6–8 years for “Mothers not depressed in first year” 5.17 (SD = 4.65), t (70) = 5.62 p b .001. Overall there were moderate to high correlations among CES-D scores over time (see Table 3). To explore the inter-relationships among symptom severity and recurrence over time, a number of correlations were tested. The total number of depressive episodes the mother reported, used as an index of chronicity, was correlated with
Table 3 Correlations between CES-D scores at 4, 12, 15 months, 4 years and 6–8 years. CES-D CES-D CES-D CES-D CES-D 4 months 12 months 15 months 4 years 6–8 years CES-D 4 months – CES-D 12 months CES-D 15 months CES-D 4 years ⁎⁎p b .01.
.536⁎⁎ –
.591⁎⁎ .800⁎⁎ –
.667⁎⁎ .653⁎⁎ .639⁎⁎ –
.610⁎⁎ .676⁎⁎ .662⁎⁎ .613⁎⁎
initial symptom severity at 4 months (r = .46; p b .01); mean symptom severity score (r = .52; p b .01), and concurrent symptoms at 6–8 years (r = .46; p b .01). These data confirm that severity, chronicity and concurrent symptoms are confounded. 3.1.5. Depression in the first postnatal year and later child behaviour problems Depression in the first postnatal year was significantly correlated with mothers' reports of later internalising (r = .36; p b .01) and externalising problems (r = .28; p b .05); with a trend for an association with teachers' ratings of externalising problems (r = .21; p b .10). Father behaviour ratings were not significantly related to maternal depression in the first year. A regression analysis including covariates (maternal education, NESB) was conducted. All variables were entered together. Results confirmed that, compared to mothers not depressed in the first postnatal year, mothers depressed in the first year reported more internalising and externalising behaviours in their children (see Table 4). No significant relationships were found among depression in the first postnatal year and fathers' or teachers' reports, all ps N .10. According to the ADIS interview, 9% (n = 5) of mothers who were not depressed in the first postnatal year rated their child as having a clinically significant internalising problem at 6–8 years, compared with 34% (n = 12) who had been depressed in the first postnatal year, χ2 (1, N = 75) = 12.61, p b .001. With regard to externalising problems, 7% (n = 4) of mothers not depressed in the first postnatal year (n = 53) rated their child as having a clinically significant externalising problem, compared with 18% (n = 4) of the postnatally depressed mothers (n = 22). This difference, while in the expected direction, was not significant. 3.1.6. Later/concurrent depression and child behaviour problems Bivariate correlations revealed that concurrent symptoms were significantly correlated with mothers' reports of internalising (r = .38; p b .01) and externalising problems (r = .41; p b .01) and teachers' ratings of externalising problems (r = .32; p b .01) at child age 6–8 years. Total number of depressive episodes was also correlated with mothers' reports of internalising (r = .36; p b .01) and externalising (r = .31; p b .01) problems, as was overall symptom severity; internalising (r = .34; p b .05), and externalising (r = .29; p b .05) problems. Concurrent maternal depression was not related, however, to fathers' reports of child internalising (r = .04; p = .80) or externalising problems (r = .11; p = .46). A regression analysis including covariates confirmed these
120
I. Fihrer et al. / Journal of Affective Disorders 119 (2009) 116–123
Table 4 Regression analysis: early maternal depression and child behaviour problems. DV
Variable
B
Standard error
Beta
t
p
Internalising
Education NESB Dep. Yr 1
− .05 4.61 7.40
2.74 4.75 2.43
− .002 .11 .35
− .02 .97 3.04
.99 .34 .003
Education NESB Dep. Yr 1
.67 7.81 5.62
2.91 5.05 2.59
.03 .18 .25
.23 1.55 2.17
Externalising
a
.82 .13 .03
Adjusted means a
Depressed 55.87 (2.06)
Not depressed 50.34 (1.35)
Depressed 56.01 (2.20)
Not depressed 50.34 (1.35)
The adjusted means are the estimated marginal means of the dependent variable, adjusted for the other variables in the model.
associations for mothers' ratings of internalising and externalising problems (see Table 5). 3.1.7. Testing the mediating role of later depression symptoms A mediation model (Baron and Kenny, 1986) (Figs. 1 and 2) was tested to explore both direct effects of early depression on current child behaviour and indirect effects mediated by ongoing maternal depression. As noted earlier, chronicity (number of episodes) severity (mean symptom score) and concurrent symptoms were all highly correlated. Since there are theoretical grounds for believing concurrent symptoms may influence both child behaviour (e.g. Brennan et al., 2000) and the mother's reporting of child behaviour (e.g. Richters, 1992) we have opted to use concurrent symptoms, a continuous variable, as our primary index of ongoing depression in the mediation analyses. The analyses in Figs. 1 and 2 have satisfied the first three pre-requisites specified by Baron and Kenny (1986) for testing mediation. First, a significant relationship between early maternal depression and child behaviour was established. Then, the relationship between early and concurrent depression was established, finally, the relationship between concurrent depression and child behaviour was established. The next steps were carried out using an SPSS macro produced by Preacher and Hayes (2008). As well as providing the conventional estimates of path b (the effect of concurrent depression on child behaviour, adjusted for early maternal depression), and the effect of path c′ (the direct effect of early depression on child behaviour, adjusted for concurrent depression), this program provides a bootstrapped estimate of the significance of the indirect effect (path a ⁎ b). Bootstrapping methods have been found to be among the most powerful and accurate when producing confidence intervals
for indirect effects (Fritz and Mackinnon, 2007; Mackinnon et al., 2004). The results (Figs. 1 and 2) suggest that in the case of externalising problems, as reported by mothers, the effects of early maternal depression are fully mediated by concurrent symptoms. That is, early maternal depression affects behaviour through its effect on concurrent depression, which itself has a unique effect on externalising behaviour. In the case of internalising problems, however, there is no evidence that concurrent depression attenuated the effect of the mother's early depression in explaining the children's current behaviour. The mediation model was not tested for teacher ratings as there was no significant relationship with depression in the first postnatal year. Finally, a second mediation model was tested to determine whether recurrent maternal depression (indexed by total number of episodes) as opposed to concurrent depression might mediate the effects of early depression. There was no attenuation of the effect of early
Table 5 Regression analysis: later/concurrent depression and child behaviour problems. DV Internalising
Variable
Education NESB Concurrent depression Externalising Education NESB Concurrent depression
B
Standard Beta error
− .21 2.72 3.59 4.76 .47 .15 .28 2.77 6.04 4.86 .53 .15
t
p
− .009 − .07 .94 .09 .75 .45 .36 3.18 .002 .01 .14 .39
.10 .92 1.24 .22 3.49 .001
Fig. 1. a = effect of early maternal depression on concurrent maternal depression, b = effect of concurrent depression on child behaviour, adjusted for early maternal depression, c = total effect of early depression on child behaviour, c′ = direct effect of early depression on child behaviour, adjusted for concurrent depression.
I. Fihrer et al. / Journal of Affective Disorders 119 (2009) 116–123
Fig. 2. a = effect of early maternal depression on concurrent maternal depression, b = effect of concurrent depression on child behaviour, adjusted for early maternal depression, c = total effect of early depression on child behaviour, c′ = direct effect of early depression on child behaviour, adjusted for concurrent depression.
depression when recurrent depression (rather than concurrent symptoms) was included in the model. 3.1.8. Post hoc analysis — relationship of initial symptom severity to later behaviour problems Finally, given the strong correlation between the index episode and later episodes of depression, including scores at 6–8 years (see Table 3), a post hoc analysis was also conducted to determine the relationship between the severity of the index CES-D score at four months and behaviour problems at 6–8 years. There were significant correlations with both internalising and externalising behaviours, r = .32, p = .01, and r = .25, p = .01, respectively, and also with teachers' ratings for externalising behaviours, r = .27, p = 04. 4. Discussion This prospective study aimed to clarify the impact of timing, chronicity and severity of maternal depression on offspring behaviour in the early school years. With regard to timing, both early and later depression seem to be important. Exposure to maternal depression during the first postnatal year was related to mothers' reports of child internalising and externalising problems in the early school years. Interestingly, the severity of symptoms at four months was significantly correlated with behaviour problems six to eight years later. Further, a substantial number of mothers (34%) who had been depressed during the first postnatal year perceived their child's internalising problems as causing significant interference in the child's and/or family's life six years later. The effect of early depression on later internalising problems was not attenuated by subsequent depression, but effects for externalising beha-
121
viour problems were mediated by concurrent depression. Concurrent maternal depression was related to both internalising and externalising problems as reported by the mother, and to externalising problems reported by the teacher. There are a number of possible explanations for the relationship between depression in the first postnatal year and later internalising problems in the children. Intergenerational transmission of risk whether through problematic relationship exchanges between the mother and child and/or a genetic vulnerability seems plausible. Previous research (Essex et al., 2001) suggests that the first postnatal year represents a “sensitive period” for later vulnerability to internalising symptoms as it is an important time for the development of a secure mother–child attachment which in turn impacts emotion regulation skills (Campbell et al., 1995). Earlier findings with this sample suggest compromised mother–child interaction may have contributed. Mothers with postnatal depression persisting throughout the first postnatal year were more likely to have insecure attachment relationships with their children (McMahon et al., 2006). Findings regarding the stability of elevated depression scores over time and relations between postnatal depression and child internalising behaviour problems, on the other hand, also suggest the possibility of a genetic vulnerability to internalising problems in both mother and child. It seems reasonable to conclude that in the case of externalising problems, early depression may impact the child indirectly, by increasing the mother's risk for later depressive episodes. Our findings show that early depression predicts later episodes, and there are moderate to high correlations for depression symptoms over time. It is likely that these later episodes, which occur closer to the assessment time points, may be causally related to the problematic externalising behaviours, and that the children were reacting behaviourally to their mothers' depressed mood. Since mothers reported higher levels of problem behaviour than other raters, (and given the importance of concurrent symptoms) the possibility that mothers may be projecting their own negative thoughts and feelings onto their children must also be considered. The moderate correlations with fathers' reports for both internalising and externalising problems and teachers' reports for externalising behaviours, however, make this interpretation less likely. Research into inter-rater agreement on behaviour has commonly found greater levels of correspondence for informants' ratings of child externalising problems than internalising problems (De los Reyes and Kazdin, 2005), as was the case in this study. Teachers' reports of externalising problems are also important in showing that the effect of maternal depression is evident outside the home, albeit with milder problems reported. It is possible that the stricter boundaries and greater structure inherent in the school system serve to contain the children's problematic behaviours. There are a number of possible explanations for the null findings with respect to fathers. The majority worked fulltime and their involvement in care-giving responsibilities was considerably less than mothers, so it is possible that their perceptions of their children's behaviour were less reliable. It is also important to note that relatively few fathers participated, yielding limited power for these analyses. Also, most fathers who were separated or divorced (whose children may have had more problematic behaviour) did not participate.
122
I. Fihrer et al. / Journal of Affective Disorders 119 (2009) 116–123
4.1. Study limitations While this study represents a comprehensive, long-term follow up of a socio-economically low risk sample of clinically depressed mothers, the high prevalence of depressive symptomatology in the population from which the sample was drawn may limit the generalisability of our results. Further, participants were all mothers who experienced early difficulties with infant sleeping and settling. It could be argued that this represents a more rigorous test of depression effects on later behaviour, as there is some control for infant/ child factors. It is possible, therefore, that a comparison between depressed mothers and a community sample without a history of mother–infant difficulties may have revealed greater depression effects. A major limitation is the attrition over successive study contacts with just 60% of the original sample participating at child age 6–8 years and variability in the number with full questionnaire data at each timepoint. We have confirmed, however, that ongoing participants are similar to the original sample with regard to demographic variables and early depression. We acknowledge limited statistical power (particularly for analyses involving fathers' ratings) and believe the findings may be a conservative estimate of associations between maternal depression and later child behaviour problems, but this requires further confirmation in larger samples. Future longitudinal studies might deal with attrition problems by recruiting larger sample sizes at the outset in anticipation of inevitable attrition. Researchers could also consider providing incentives for continued participation through monetary compensation, and use online questionnaire delivery to minimize missing data. A substantial number of mothers in this sample have reported moderate to severe symptoms at each of the study time points. This pattern of ongoing depression is suggestive of a “depressive personality style” (Boyce et al. 1991) or a possible Axis II personality disorder, neither of which were directly assessed in the current study. We have also previously noted that mothers with an insecure working model of attachment are much more likely to experience persistent depression, at least until the preschool years (McMahon et al., 2006, 2008). Research into co-morbid personality disturbance and attachment difficulties provides an interesting direction for future research. It is also possible, given the pattern of depression over time, that mothers in this sample may have been depressed in pregnancy. There is a growing body of evidence suggesting that maternal mood in pregnancy may be related to later child behaviour difficulties and that this effect may be due to fetal programming of the Hypothalamic Pituitary Axis (see Talge et al., 2007, for a recent review). Future prospective studies including measures of maternal mood in pregnancy are important in testing this explanatory model. Finally, causation in this study is generally inconclusive, particularly with regard to concurrent depression symptoms. It remains unclear whether maternal depression causes behaviour problems in offspring or vice versa. In infancy, when offspring of depressed mothers interact with a non-depressed adult, the nondepressed adult becomes more negative and their affect is more flat suggesting that the infants' “depressed behaviour” may elicit and then
continue to reinforce maternal depression (e.g. McCormick, 1995). Other studies have also noted that maternal stress and depressive symptoms tend to subside when their children's behaviour improves (Elgar et al., 2003), and increase as the behaviour worsens (Gartstein and Sheeber, 2004). It is possible that maternal depression and child behaviour problems may have a deleterious reciprocal influence. 5. Conclusions Findings confirm the importance of early identification and treatment for mothers with postnatal depression, given the likelihood of ongoing depression and relations with later child behaviour problems. From a practical point of view, the finding that symptom severity at four months can predict which mothers and children are at greatest risk could enable healthcare workers to target the most at risk group of mothers early on in the child's life. Role of the funding source This research was supported by a grant from Macquarie University.
Conflict of interest There are no conflicts of interest associated with this paper.
Acknowledgements
We are grateful to the families who participated in the study and to Felicity Chandler, Lisa Lia and Alia Palethorpe for help with data collection and data management. References Achenbach, T.M., Rescorla, L.A., 2001. Manual for ASEBA School-Age Forms & Profiles. Burlington, VT: University of Vermont, Research Center for Children, Youth & Families. Baron, R.M., Kenny, D.A., 1986. The moderator–mediator variable distinction in social psychological research: conceptual, strategic, and statistical consideration. Journal of Personality and Social Psychology 51, 1173–1182. Beardslee, W., Versage, E., Gladstone, T., 1998. Children of affectively ill parents: a review of the past 10 years. Journal of the American Academy of Child and Adolescent Psychiatry 37 (11), 1134–1141. Boyce, P., Parker, G., Barnett, B., Cooney, M., 1991. Personality as a vulnerability factor to depression. British Journal of Psychiatry 159, 106–114. Brennan, P.A., Hammen, C., Andersen, M., Bor, W., Najman, J., Williams, G., 2000. Chronicity, severity, and timing of maternal depressive symptoms: relationship with child outcome at age five. Developmental Psychology 36, 759–766. Campbell, S.B., Cohn, J.F., Meyers, T.A., 1995. Depression in first-time mothers: mother–infant interaction and depression chronicity. Developmental Psychology 31, 349–357. Cooper, P.J., Murray, L., 1995. Course and recurrence of postnatal depression: evidence for the specificity of the diagnostic concept. British Journal of Psychiatry 166, 191–195. Cornish, A.M., McMahon, C.A., Ungerer, J.A., Barnett, B., Kowalenko, N., Tennant, C., 2006. Maternal depression and the experience of parenting in the second postnatal year. Journal of Reproductive and infant psychology 24 (2), 121–132. De los Reyes, A., Kazdin, A.E., 2005. Informant discrepancies in the assessment of childhood psychopathology: a critical review, theoretical framework, and recommendations for future study. Psychological Bulletin 131 (4), 483–509. Dutra, L., Campbell, L., 2004. Quantifying clinical judgement in the assessment of adolescent psychopathology: reliability, validity and factor structure of the Child Behavior Checklist for clinician report. Journal of Clinical Psychology 60 (1), 65–85.
I. Fihrer et al. / Journal of Affective Disorders 119 (2009) 116–123 Elgar, F.J., Curtis, L.J., Mcgrath, P.J., Waschbusch, D.A., Stewart, S.H., 2003. Antecedent-consequence conditions in maternal mood and child adjustment: a four-year cross-lagged study. Journal of Clinical Child and Adolescent Psychology 32 (3), 362–374. Essex, M.J., Klein, M.H., Miech, R., Smider, N.A., 2001. Timing of initial exposure to maternal major depression and children's mental health symptoms in kindergarten. British Journal of Psychiatry 179, 151–156. Fritz, M.S., Mackinnon, D.P., 2007. Required sample size to detect the mediation effect. Psychological Science 18 (3), 233–239. Gartstein, M.A., Sheeber, L., 2004. Child behavior problems and maternal symptoms of depression: a mediational model. Journal of Child and Adolescent Psychiatric Nursing 17 (4), 141–150. Goodman, S.H., Gotlib, I.H., 1999. Risk for psychopathology in the children of depressed mothers: a developmental model for understanding mechanisms of transmission. Psychological Review 106 (3), 458–490. Grills, A.E., Ollendick, M.S., Thomas, H., 2003. Multiple informant agreement and the anxiety disorders interview schedule for parents and children. Journal of the American Academy of Child and Adolescent Psychiatry 42 (1), 30–40. Hammen, C., Brennan, P.A., 2003. Severity, chronicity, and timing of maternal depression and risk for adolescent offspring diagnoses in a community sample. Archive of General Psychiatry 60, 253–258. Hay, D.F., Pawlby, S., Sharp, D., Asten, P., Mills, A., Kumar, R., 2001. Intellectual problems shown by 11-year old children whose mothers had postnatal depression. Journal of Child Psychology and Psychiatry 42, 871–889. Hay, D.F., Pawlby, S., Angold, A., Harold, G.T., 2003. Pathways to violence in children of mothers who were depressed postpartum. Developmental Psychology 39 (6), 1038–1094. Hay, D.F., Pawlby, S., Waters, C., Sharp, D., 2008. Antepartum and postpartum exposure to maternal depression: different effects on different adolescent outcomes. Journal of Child Psychology and Psychiatry 49 (10), 1079–1088. Lyons-Ruth, K., Wolfe, R., Lyubchik, A., 2000. Depression and the parenting of young children: making the case for early preventative mental health services. Harvard Review of Psychiatry 8 (3), 148–153. Mackinnon, D.P., Lockwood, C.M., Williams, J., 2004. Confidence limits for the indirect effect: distribution of the product and resampling methods. Multivariate Behavioral Research 39 (1), 99–128. McCormick, L.H., 1995. Depression in mothers of children with attention deficit hyperactivity disorder. Family Medicine 27, 176–179. McMahon, C., Barnett, B., Kowalenko, N., Tennant, C., 2005. Psychological factors associated with postnatal depression: past and current relationships, defence styles and the mediating role of insecure attachment styles. Journal of Affective Disorders 84 (1), 15–24. McMahon, C., Barnett, B., Kowalenko, N., Tennant, C., 2006. Maternal attachment state of mind moderates the impact of postnatal depression on infant attachment. Journal of Child Psychology and Psychiatry 47 (7), 660–669. McMahon, C., Barnett, B., Kowalenko, N., Tennant, C., Don, N., 2001. Postnatal depression, anxiety and unsettled infant behaviour. Australian and New Zealand Journal of Psychiatry 35, 581–588.
123
McMahon, C., Trapolini, T., Barnett, B., 2008. Maternal state of mind regarding attachment predicts persistence of postnatal depression in the preschool years. Journal of Affective Disorders 107 (1–3), 199–203. Murray, L., 1992. The impact of postnatal depression on infant development. Journal of Child Psychology and Psychiatry 33, 543–561. Murray, L., Sinclair, D., Cooper, P., Ducournau, P., Turner, P., 1999. The socioemotional development of 5-year-old children of postnatally depressed mothers. Journal of Child Psychology and Psychiatry 40 (8), 1259–1271. NICHD, 1999. Chronicity of maternal depressive symptoms, maternal sensitivity, and child functioning at 36 months. Developmental Psychology 35, 1297–1310. O'Hara, M.W., Swain, A.M., 1996. Rates and risk of postpartum depression: a meta-analysis. International Review of Psychiatry 8, 37–54. Preacher, K.J., Hayes, A.F., 2008. Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior Research Methods 40, 879–891. Radke-Yarrow, M., Klimes-Dougan, B., 2002. Parental depression and offspring disorders: a developmental perspective. In: Goodman, S.H., Gotlib, I.H. (Eds.), Children of Depressed Parents: Mechanisms of Risk and Implications for Treatment Washington. American Psychological Association DC. Radloff, J.S., 1977. The CES-D Scale: a self-report depression scale for research in the general population. Applied Psychological Measurement 1, 385–401. Richters, J.E., 1992. Depressed mothers as informants about their children: a critical review of the evidence for distortion. Psychological Bulletin 112 (3), 485–499. Silverman, W.K., Albano, A.M., 1996. The Anxiety Disorders Interview Schedule for Children for DSM:IV: Child and Parent Versions. Psychological Corporation, San Antonio. Silverman, W.K., Saavedra, L.M., Pina, A.A., 2001. Test-retest reliability of anxiety symptoms and diagnoses with the anxiety disorders interview schedule for DSM-IV: Child and parent versions. American Academy of Child and Adolescent Psychiatry 40 (8), 937–944. Sohr-Preston, S.L., Scaramella, L.V., 2006. Implications of timing of maternal depressive symptoms for early cognitive and language development. Clinical Child and Family Psychology Review 9 (1), 65–83. Talge, N., Charles, N., Vivette, G., 2007. Antenatal maternal stress and longterm effects on child neurodevelopment: how and why? Journal of Child Psychology and Psychiatry 48 (3–4), 245–261. Trapolini, T., McMahon, C., Ungerer, J.A., 2007. The effect of maternal depression and marital adjustment on young children's internalizing and externalizing behaviour problems. Child, Care, Health and Development 33 (6), 794–804. World Health Organization, 1997. Composite International Diagnostic Interview (CIDI, Version 2.1). World Health Organization, Geneva, Switzerland.