Challenges of defining and measuring perinatal anxiety

Unlike many physical health problems where diagnosis can often be determined by a specific set of symptoms which are either visible or measureable in a relatively unambiguous and quantifiable way, the diagnosis of mental health problems is often more problematic. This is because both the definition and measurement of mental health problems is challenging. The challenge is greater still when considering the potential impact of a major life event such as childbirth. Most women experience anxiety in the perinatal period; it is a common emotional response, which is rational and sometimes beneficial in certain situations. However, when anxiety is extreme or persistent, it may indicate the presence of an anxiety disorder.

An additional challenge is that we do not diagnose ‘anxiety’ per se, rather anxiety symptoms are classified into different types of anxiety disorder, including generalised anxiety disorder, panic disorder, and phobia. These disorders can have different known or unknown triggers and different manifestations but are all characterised by core symptoms of anxiety, such as cognitive distortions, physiological arousal and behavioural avoidance. Obsessive-compulsive disorder and post-traumatic stress disorder are also characterised by symptoms of anxiety and, until the most recent updates to the DSM and ICD classifications of mental health problems, were also categorised as anxiety disorders (American Psychiatric Association [APA], 2013; World Health Organization [WHO], 2018).

Current research suggests that levels of anxiety symptoms are high at between 10-15% in the perinatal period (Dennis et al., 2017) and that pre-existing anxiety disorders can be exacerbated during this time (Forray, Focseneanu, Pittman, McDougle, & Epperson, 2010). These prevalence rates are extracted from systematic reviews which have included heterogeneous studies with different methods of assessment. Inconsistency in assessment methods can lead to large variability in prevalence rates and possible misrepresentation of the true extent of anxiety.

The gold standard for diagnosing anxiety disorders is via a structured clinical interview with a suitably trained health professional. However, in certain research, such as population-based, cross-sectional studies or studies where anonymity is required, it isn’t feasible to do this. Also, in maternity care it is important to identify women who are experiencing problems with anxiety through simple screening questions. It is neither desirable nor possible to do a structured clinical interview with all women. Therefore, self-report questionnaires are relied upon to provide an assessment of anxiety symptoms. On a practical level, self-report questionnaires are efficient, requiring little time to administer and score. Cut-off points validated against the gold standard are used to determine when symptoms are at a level that may indicate a disorder and require further investigation. However, if further investigation is not possible, self-report questionnaires may be used in isolation.

This brings us to what is probably the biggest challenge: how to interpret self-report questionnaires when used in isolation. Firstly, they overlook the importance of clinical observation to gain additional information about symptomatology and clinical judgement to tailor the assessments. Secondly, they do not necessarily consider all components that would constitute a structured clinical interview and would be required to reach a diagnosis, such as the onset, course and duration of symptoms or the extent to which symptoms impact on daily life. Thirdly, they provide only a snapshot of an individual’s mental health. Anxiety symptoms can be transient and variable and, unless repeated administrations are used, a snapshot may capture an atypical time-point. Therefore, self-report questionnaires may provide a useful assessment aid, but used in isolation and at one time-point, they do not provide a full clinical picture of an individual’s mental health.

The psychometric properties of self-report questionnaires provide an indication of how confidently we can rely on the information they provide, for example, a self-report questionnaire with good specificity and sensitivity should return a positive or negative score for individuals with or without clinical levels of anxiety respectively. A questionnaire with poor specificity could inflate prevalence rates and overwhelm clinical services whereas a questionnaire with poor sensitivity could underestimate the burden of the problem and fail to identify individuals who are in need of support. Even with robust psychometric properties, self-report questionnaires do not provide a substitute for the gold standard method of assessment. If we are to ensure that we are reporting accurate prevalence rates for anxiety during the perinatal period, that we are correctly identifying those women who need support, and that we are monitoring their progress, we need to reflect carefully on how self-report questionnaires are used in research and practice.