The validity of DSM symptoms for depression and anxiety disorders during pregnancy

Abstract

Background

The applicability of the symptom criteria for diagnostic mood and anxiety disorders has in recent times been questioned for women in the perinatal period, due to the overlap of diagnostic symptoms with normal symptoms due to the physical changes of pregnancy or postpartum, (e.g., sleep difficulties).

Method

118 women in their second or early third trimester of pregnancy participated in a telephone interview that included the depression and anxiety modules of the MINI diagnostic interview, and an attributional probe question asking the woman whether endorsement of a symptom was due to the physical changes of her pregnancy or due to her mood or worries.

Results

66% of pregnant women who met criteria for major depression reported that a sufficient number of their symptoms were due to the normal physical changes of their pregnancy, such that they would no longer meet criteria for a diagnosis of major depression. Attributional probing resulted in the rate of major depression falling from 6.8% to 1.7%. The impact on anxiety disorders appears to be less.

Limitations

An important issue is whether women's attributional perceptions are valid, though the face validity of some of their responses means that credence should be given to the findings. In addition, it would be useful from a validity perspective to undertake the same attributional probing with other populations where symptom presence is most likely due to mood, such as unemployed or recently separated adults without physical complaints.

Conclusion

Rates of depressive disorders in pregnancy, using DSM symptom criteria, are significantly over-estimated due to the natural occurrence of many of the symptoms as a result of the pregnancy. Rates of anxiety disorders are also inflated, but to a lesser degree. This means that the validation of self-report mood measures, typically done against DSM diagnoses, is likely to have produced erroneous findings (e.g., optimum cut-off scores). It is probable that a similar finding would be obtained for the postpartum period. Thus future use of DSM symptom criteria for depression and anxiety in perinatal mental health work should use some form of attributional probing question to more accurately understand the applicability of symptoms to a diagnosis in this population.

Introduction

Depression in pregnancy and postpartum has been well documented over the past 20 years, with prevalence rates ranging from 5% to 41% (e.g., see reviews by Gavin et al., 2005, Halbreich and Kartun, 2006, Jesse et al., 2010; Lancaster et al., 2010, Mann et al., 2010, O'Hara, 2009) with the higher rates usually being from studies using self-report measures, such as the Edinburgh Postnatal Depression Scale (EPDS, or antenatally the EDS; Cox et al., 1987), and the lower rates obtained from diagnostic interviews, using the Diagnostic and Statistical Manual of Mental Disorders criteria for major or minor depression (DSM, 2005).

In addition, there continues to be work done in validating self-report measures against DSM diagnostic criteria (e.g., Chaudron et al., 2010, Edmondson et al., 2010), and for using diagnoses of depression in the antenatal or postnatal period as indicators of distress, or to report on program evaluation. All of this work thus inherently accepts that the DSM diagnostic criteria for depression in perinatal women are valid, and that they are the ‘gold standard’. The diagnostic symptoms are:

• Depressed mood

• Diminished interest or pleasure in nearly all activities (anhedonia)

• Significant unintentional weight loss or weight gain, or change in appetite

• Difficulty with sleep (too much or too little)

• Psychomotor agitation or retardation

• Fatigue or loss of energy

• Feelings of worthlessness or inappropriate guilt

• Diminished ability to think or concentrate, or indecisiveness

• Recurrent thoughts of death or suicidal ideation

For Major depression, a woman needs to have five or more symptoms, and for Minor depression she needs to have two to four symptoms. In both disorders at least one of the symptoms must be depressed mood or anhedonia (the core symptoms: A1 or A2). And the symptoms must have been present for most of the day, nearly every day for the past 2 weeks. In addition, the presence of these symptoms must also a) cause clinically significant distress or impairment in everyday functioning; and b) not be due to a general medical condition, and c) not be accounted for by bereavement.

It is important that clinicians and researchers realise that these symptoms, and the number required to meet diagnostic criteria, have not been determined empirically—as van Praag (1998) has stated: “The number of symptoms necessary to qualify for a particular diagnosis is determined arbitrarily” (p. 769). Summerfield, in arguing that many normal emotions are being pathologised, stated that “Psychiatric categories are manufactured constellations emerging as DSM or ICD committee decisions…so why do we treat them as if they were facts of nature identifiable ‘out there’?” (Summerfield and Veale, 2008; p.329). Liebowitz (1993) is another investigator who has espoused the view that expert opinion, rather than empirical evidence, usually guides the requirements for a diagnostic disorder.

The validity of diagnostic depression symptom criteria for women (or possibly men) in the perinatal period has been questioned by several researchers. Ross et al. (2003) pointed out that some of the symptoms assessed on the Hamilton Rating Scale for Depression (HAM-D: Hamilton, 1960), which are also symptoms for DSM Depression, are often normal physical symptoms experienced in new parents. Such symptoms included weight, energy, sleep and appetite changes. Altshuler et al. (2008) stated that “(the) identification of a major depressive disorder in pregnancy can be difficult because many symptoms of depression, including fatigue, sleep difficulties and changes to appetite and weight, commonly occur in a normal pregnancy and do not necessarily connote depression” (p.278). Indeed, Yonkers et al. (2009) have suggested that it may be wise to increase the number of symptoms required during pregnancy to meet criteria for depression at this time for this very reason. In the clinical setting, Jesse et al. (2009) stated that as a consequence of women endorsing depressive symptoms that were just due to the physical changes of pregnancy, “more women…were referred (for mental health services) than necessary” (p. 164).

This view has also been expressed for the postnatal period by Csatordai et al. (2009), who state “The clear overlap between the normal sequel (sic) of childbirth and the symptoms of major depression underscores the need to develop guidelines for early identification” (p. 57). In addition, Martini et al. (2010) have raised concerns regarding the overlap of somatic symptoms and anxiety in the antenatal period.

If this view is correct that normal symptoms of pregnancy or the postnatal period can be misconstrued as symptoms of depression, then this has significant implications for studies that report on the prevalence of depression in the perinatal period, as well as studies validating self-report measures against diagnostic criteria. In addition, if women (or men) are then incorrectly given the diagnosis of postnatal or antenatal depression because the normal symptoms of the perinatal period are being attributed to the individual's mood, then such an incorrect diagnosis can have significant detrimental effects on the individual (Krantz et al., 2008).

Given this concern that some symptoms required for a diagnosis of depression may not be appropriate during the perinatal period, Kammerer et al. (2009) recently published a study in which they diagnostically assessed 697–892 women at 6 weeks postpartum, as well as asking them to recall the presence or absence of the depression symptoms during their pregnancy. All the DSM depression symptoms were assessed, irrespective of whether the participant endorsed either of the two core symptoms. Their analyses compared the frequency of each symptom for women who did and did not then meet the symptom diagnostic criteria for major or minor depression (but excluding the impairment criterion). The only symptom that was not significantly different between the two groups of women was that of a loss of appetite. They concluded that this therefore was the only symptom not valid for use antenatally or postnatally.

All of the above arguments may also apply to the diagnosis of an anxiety disorder in the perinatal period. There is increasing awareness that high levels of anxiety should also be detected both in pregnancy and postpartum (e.g., Chaudron and Nirodi, 2010), and indeed there is some evidence that the rate of anxiety disorders is higher than the rate of depressive disorders (Matthey et al., 2003, Wenzel et al., 2003). For many of the anxiety disorders the presence of physiological symptoms, rather than cognitive or behavioural symptoms, determines if an individual meets criteria for the disorder. There are also some symptoms that overlap with those for depression, and which are, as previously mentioned, normal concomitants of being pregnant. Thus panic disorder requires at least four symptoms of 13, most of which are physiological (e.g., sweaty or clammy hands; tingling or numbness in parts of the body); PTSD includes some symptoms that are the same as for depression (e.g., difficulty sleeping; become much less interested in hobbies or social activities); and Generalised Anxiety Disorder (GAD) includes feeling tired; difficulty concentrating, and difficulty sleeping.

As part of a study investigating antenatal mood disorders, our interest was raised on the question of diagnostic validity due to the responses being given by some of the participants in the early phase of the study. We thus inspected this issue using a different approach to that by Kammerer et al. (2009), by directly asking the women who met the core symptoms for the disorders of Depression, Panic, Post Traumatic Stress Disorder (PTSD), and GAD whether they thought a symptom was due to the physical change of their pregnancy, or their mood. This client-informed approach has been used in two studies administering the HAM-D (Da Costa et al., 2009, Ross et al., 2003). These investigators discounted symptoms on the HAM-D if they were perceived to be due to the normal physical changes associated with pregnancy or the early postpartum period, but no information was reported on the frequency of such attributions in either study.