Maternal Depression and Child Growth: Definitional Issues, Longitudinal Trajectories, and Analytic Considerations
Article Outline
EPDS, Edinburgh Postnatal Depression Scale
See related article, p 401
However, findings from this study must be cautiously interpreted given the limitations of the Edinburgh Postnatal Depression Scale (EPDS) as a measure of depression and the overall strategy used for capturing the expression and course of depressive illness. The EPDS is a brief self-report measure designed to screen for possible Major Depressive Disorder.2 Several cutoffs have been suggested, and the investigators use the widely used level of ≥13 to delineate those with and without “depression” at each time point. However, the EPDS is not without its shortcomings. A recent review by Gibson et al3 concluded that the EPDS has yielded inconsistent findings regarding the validity of the measure. They particularly note the wide variability in psychometric properties reported across different language and cultural groups. Sensitivity and specificity is variable across studies, and it is notable that these indices are only modestly robust in the study by Santos et al. The authors administered the measure as part of an interview, a strategy that is infrequently used and one that is distinctly different than the more typical approach of having mothers read and independently record their responses. It is possible that mothers may have been more reluctant to verbally admit symptoms to an interviewer in contrast to the more anonymous and removed process of completing the measure independently. Comparing findings across studies using the EPDS is difficult as a function of these collective psychometric inconsistencies, and the mixed results found in studies of child growth and depression in particularly and the effects of maternal depression generally arise in part from this problem.
In addition, the EPDS has distinct limitations when used to determine the manifestation and course of depression. The measure covers the week before completion, a short interval that misses expressions of clinically significant depressive symptoms occurring at other times. Future studies should consider using measures that document and track the course of depression through combinations of prospective and retrospective reporting, such as the Longitudinal Interview Follow-Up Evaluation.4 Another concern is the use of a continuous measure to create groups of depressed and nondepressed persons. As noted, the EPDS is a screen, and cutoffs were designed to indicate the potential for having Major Depressive Disorder, which in turn should be determined by a subsequent in-depth psychiatric interview. Using cutoffs to form groups makes the assumption that categories are distinct and inherently different from one another. In fact, such a strategy leads to assignment of mothers to depressed and nondepressed categories even when they may be similar. For example, when using the cutoff of ≥13, mothers with scores of 12 and 13 would be assigned to different groups even though they have obtained overall scores that are almost identical. In the Santos et al study, the authors contrast extreme groups to partially address this issue, but such a strategy drops most of the sample, reduces statistical power, and does not directly resolve the problem of using screens to determine caseness in the primary analyses.
Another approach to determining course of functioning in longitudinal data such as that described by Santo et al is to use recently developed analytic strategies that describe trajectories over time. Latent class growth analysis, for example, provides an empiric approach to deriving groups of individuals based on similarities in patterns of functioning over time.5 This analytic strategy determines individual growth curves and groups individuals together, based on shared trajectories. Distinct groups are determined empirically, and these provide the basis for subsequent comparisons on meaningful outcomes. In contrast to forming crude groups based on cutoffs, such statistical approaches allow for consideration of both frequency and severity of symptoms across time. Campbell et al6, 7 used these strategies to describe trajectories of depressive symptoms over several years using multiple administrations of the CES-D in mothers. These studies found 6 and 5 groups, respectively, reflecting different trajectories of depressive symptoms. In addition to the expected non-depressed and chronic patterns of depressive symptoms over time, other trajectories also were revealed, including such patterns as increasing or decreasing symptoms over time, high variability with low and high levels reported at different intervals, and consistent subthreshold but clinically elevated symptoms at each assessment. These types of analytic strategies of data have the potential to yield a deeper and more illuminating understanding of the expression and course of depression in mothers, and they should be strongly considered in the examination of longitudinal data sets.
Another strength in the analytic approach by Santos et al is the rigorous approach to confounding. What was unclear is the question of whether variables were so correlated that effects may not be fully appreciated. For example, income, maternal schooling, and smoking may all contribute in a causal way to maternal depression and subsequently to aberrant child growth. Disaggregating the relative contribution of these characteristics toward maternal depression, in a theoretically driven way, could elucidate the complicated interrelationships among them. This is a challenging task. Future studies should be designed to accomplish this aim.
A final consideration is the placement of the Santos et al findings in the context of the larger literature on maternal depression and its effects on child health and development. Although the current study focused on child growth and weight, an important area warranting continued empiric examination, a large body of data has accrued documenting additional potentially devastating impact of depression on mothers and their children. These are particularly evident in the social, emotional, and cognitive development of children raised by depressed mothers. A recent report by the Institute of Medicine confirms the strength and breadth of these associations.8 Moreover, a key recommendation of the report is to screen for depression in multiple care settings, including pediatric offices, and to develop innovative approaches to addressing the needs of this population.
References
- . Long-lasting maternal depression and child growth at 4 years of age: a Cohort Study. J Pediatr. 2010;157:401–406
- . Detection of postnatal depression: development of the 10-item Edinburgh Postnatal Depression Scale. Br J Psychiatry. 1987;150:782–786
- . A systematic review of studies validating the Edinburgh Postnatal Depression Scale in antepartum and postpartum women. Acta Psychiatrica Scand. 2009;119:350–364
- . The Longitudinal Interval Follow-up Evaluation: a comprehensive method for assessing outcome in prospective longitudinal studies. Arch Gen Psychiatry. 1987;44:540–548
- . Group-based modeling of development. Cambridge, MA: Harvard University Press; 2005;
- . Trajectories of maternal depressive symptoms, maternal sensitivity, and children's functioning at school entry. Dev Psychol. 2007;43:1202–1215
- . A latent class analysis of maternal depressive symptoms over 12 years and offspring adjustment in adolescence. J Abnorm Psychol. 2009;118:479–493
- . Depression in parents, parenting, and children: Opportunities to improve identification, treatment, and prevention. Washington, DC: Committee on Depression, Parenting Practices, and the Healthy Development of Children, Board on Children, youth, and Families, Division on Behavioral and Social Sciences and Education; 2009;
PII: S0022-3476(10)00380-X
doi:10.1016/j.jpeds.2010.04.061
Published by Elsevier Inc.
Refers to article:
- Long-Lasting Maternal Depression and Child Growth at 4 Years of Age: A Cohort Study , 19 April 2010
