ارزیابی افسردگی پس از زایمان: شواهد برای نیاز به روش های متعدد

Different methods and instruments are frequently used to measure postpartum depression (PPD) in research, e.g. PPD-specific scales, DSM-based diagnostic interviews and rating scales assessing general depression. However, it is unsure whether these measures would lead to the same results, e.g. in the identification of “depressed” women or in their relations to third variables. Objective(s) We compared different measures of PPD and their relations with a third variable, namely the mothers’ marital satisfaction. Method We recruited 65 mothers to take part in a study about the impact of PPD on the development of early family relations. Maternal PPD was assessed with multiple methods (i) a PPD-specific scale, (ii) a DSM-based diagnostic interviews and (iii) a rating scale designed to assess the severity of depressive symptoms. We assessed mothers’ marital satisfaction with the Marital Adjustment Test (MAT). Results Results showed weak overlap between PPD-specific scale and DSM-based diagnosis of PPD, and modest correlations between the PPD-specific scale and the general depression rating scale. Only the score on the PPD-specific scale could predict marital satisfaction. Conclusion As we found discrepancies between different measures of PPD, we suggest being cautious in the choice of measures and using multiple methods to measure PPD in a comprehensive way.

Maternal postpartum depression (PPD) is a common disorder, that affects 10–15% mothers in the first year following delivery (Sheeder, Kabir, & Stafford, 2009). Numerous studies were conducted to investigate the disorder, with a particular attention drawn on its negative consequences for child development and for the construction of healthy mother-child relationships in the very first months (Field, 2010 and Goodman et al., 2011). From the first studies and clinical descriptions (Marcé, 1858 and Pitt, 1968), PPD was defined as an “atypical” depression. PPD includes the symptomatic features common to any depressive disorder, such as low mood, fatigue, sleep disturbances, or reduced appetite (O’Hara, 1997). However, a set of specific symptomatic expressions have also been regularly described, such as an increased anxiety related to parenthood, feelings of being inadequate in the parental role or unable to take care of the child, or specific fears concerning the child's health (Pitt, 1968 and Ross et al., 2003). Although these atypical features have been repeatedly documented in the clinical descriptions of PPD, they are not included in the list of relevant criteria for the diagnosis of PPD in nosographic classifications. For example, the fifth edition of the Diagnostic and Statistical Manual for Mental Disorders (DSM-5) defines PPD as a subtype of Major Depressive Disorder (MDD), with an onset in the four months after delivery (American Psychiatric Association, 2013). This situation has progressively blurred the definition of the disorder, as it is still unclear whether PPD should be considered as a disorder distinct from MDD – as it includes specific symptomatic features – or as a MDD with a specific onset. This indetermination has led to heterogeneous practices in the studies about PPD, more particularly in the methods of measurement and in the instruments that are used to assess the disorder. The most widely used instrument to assess PPD symptoms is the Edinburgh Postnatal Depression Scale (EPDS; Cox, Holden, & Sagovsky, 1987). The EPDS is now considered by many as the gold standard in the field. It was specifically created to screen for PPD, since scales designed for screening or diagnosing general depressive disorders did not consider the atypicality of the disorder (Cox et al., 1987). To fill this gap, the authors of the EPDS proposed to take this atypicality into account, notably including the anxious symptoms that are particularly relevant in clinical descriptions of PPD (Ross et al., 2003). The scale indeed consists in the rating of common depressive symptoms, as well as symptoms specific to PPD, that are combined in a total score. A cutoff score was established and women scoring above this cutoff can be considered at risk for PPD. Less frequently, diagnostic interviews, such as the Structured Clinical Interview for DSM-III-R (SCID-I; Spitzer, Williams, Gibbon, & First, 1992) or the Diagnostic Interview for Genetic Studies (DIGS; Nurnberg et al., 1994 and Preisig et al., 1999) are used to elicit diagnostic criteria for PPD. They are based on the DSM definition of PPD as a subtype of MDD, excluding the atypical features of PPD. Finally, screening or severity scales designed to measure general depression, such as the Montgomery-Åsberg Depression Rating Scale (MADRS; Montgomery & Åsberg, 1979) or the Beck Depression Inventory (BDI; Beck, Ward, Mendelson, Mock, & Erbaugh, 1961) and also based on the definition of PPD as a MDD, are sometimes used to measure PPD. These scales generally consist in the rating of a general set of depressive symptoms along Likert scales according to their presence or intensity. In addition to the fact that these methods are based on various definitions of PPD, each of them also has methodological implications, specificities, costs, and benefits. For example, using self-report questionnaires (EPDS, BDI) are convenient and easy-to-use methods, but they will not be appropriate to establish a diagnosis of PPD, as they were not designed to do so. However, in many comparative studies, the definition of clinical and control groups of mothers will only be based on self-reported evaluations of depression and will rarely be confirmed afterwards with a deeper diagnostic investigation, for example, including a diagnostic interview. On the other hand, the use of diagnostic interviews requires greater resources than self-report measures such as the involvement of trained clinicians in the data collection process. Moreover, being DSM-based, most of the existing semi-structured interview suitable for research will not take the specific clinical features of PPD into account. A major problem is that, in the literature, the characteristics of each method of assessment are rarely taken into account or even considered. The results of studies using different measures of PPD are frequently compared, although the choice of the measure may have a major influence on these results. Ignoring the characteristics of the measures may explain why the conclusions of many studies are heterogeneous or even contradictory. The primary cause for this heterogeneity may be due to the fact that it is frequently taken for granted to consider a priori that the different definitions of PPD, as well as the methods of measurement that result from these definitions, are equivalent. These evidences raise many concerns. Indeed, one could wonder whether different instruments measuring PPD applied in a single sample would lead to different results in terms of rates of “depressed” cases and whether the cases identified as “depressed” would be the same across the different measures. Most studies reported that the EPDS provided fairly similar results than diagnostic interviews and other assessment scales of depression (Condon and Corkindale, 1997, Evins et al., 2000, Gibson et al., 2009 and Hanusa et al., 2008). However, others also reported a certain amount of unexplained variance in the links between the instruments (Eberhard-Gran, Eskild, Tambs, Opjordsmoen, & Ove Samuelsen, 2001). As most of these studies were validation studies, discussions on the results were primarily oriented toward the similarity between the outcomes of different methods of assessment of PPD, however only a few studies specifically addressed the discrepancies between PPD assessment strategies (Beck, Kurz, & Gable, 2012). In a recent study investigating the impact of maternal depression on the construction of family relationships in the postpartum period, we decided to deal with the indetermination in the definition and measurement of PPD in using three different measures of PPD, in order to be able to compare their results as a secondary objective of our study. The depressive state of mothers was assessed with the EPDS (the PPD-specific scale), the DIGS (DSM-based diagnostic interview), and the MADRS (general depression scale). This paper aims at presenting the results of these different scales, as well as a comparison and a discussion of these results. Moreover, as part of a study about the consequences of PPD on the early family functioning, several relational outcomes were measured, such as marital satisfaction. Marital satisfaction was frequently shown to be highly related with depressive disorders, both inside and outside the postpartum period (Beck, 2001, Choi and Marks, 2008, Gotlib et al., 1991, Mead, 2002 and O’Hara and Swain, 1996). In the postpartum, low marital satisfaction is associated with increased depression, both variables being a potential antecedent or a sequela of one another (Barnett & Gotlib, 1988). In the present report, we used mother's marital satisfaction as a third variable, in order to investigate the criterion-related validity of different PPD assessments.

We present in this section the descriptive statistics for the instruments used in the study. First, the results of the EPDS showed that mean and median scores of the sample were globally low (M = 6.77, Mdn = 6.0, SD = 4.64) and ranged from 0 to 24 points. The results of the EPDS indicated that n = 5 mothers (7.5%) had a score above the originally validated cutoff of 12/13 points and were thus potentially depressed. With a lower cutoff point of 8/9, the rate of depressed mothers jumped to 31%. Scores of the mothers on the MADRS were also relatively low (M = 4.9, Mdn = 4.0, SD = 5.2). These scores remained lower than the first cutoff of mild depression set to 8/9 points and n = 9 mothers (14.8%) scored above this cutoff. The mothers were mostly in the “mild depression” range (n = 6), whereas the others (n = 3) obtained scores in the “moderate depression” range. There was no mother scoring in the “severe depression” range in this study. During the diagnostic interview (DIGS), n = 4 women (6%) reported having suffered from at least five DSM-IV depression symptoms during at least two weeks in the previous month and therefore met the criteria for a major depressive episode. Half of them just reached the minimal number of symptoms (5). Concerning marital satisfaction, the scores of mothers on the MAT ranged from 51 to 156 points. Most mothers (83%) obtained scores above the cutoff of 100 points, whereas 17% of the mothers reported a global dissatisfaction regarding their couple relationship with scores below this cutoff. Mean and median scores were above the cutoff (M = 120.8, Mdn = 125.5, SD = 22.3) and the sample could thus be globally considered as satisfied.