Assessing for psychosocial morbidity in pregnant women

In recent years an increasing number of investigators have supported routine antenatal and postnatal screening for psychological distress or depression.1^,2 Performing antenatal screening is reasonable since we know that postnatal mood disorders often begin during or before pregnancy. We also know that, in the absence of systematic screening, most postnatal mood disorders are not detected.3 Furthermore, we know that women with antenatal psychosocial risk factors are more likely to have a postnatal mood disorder.4 There is also some evidence that antenatal interventions can be beneficial for women.5

Efforts have been made, therefore, to develop suitable screening measures during pregnancy. Cooper and associates6 created a 17-item prediction scale for postpartum depression consisting of questions from 6 domains: emotional experience of the pregnancy, physical experience of the pregnancy, psychiatric history, maternal bereavement before age 11, quality of close relationships and social factors. Honey and associates7 recently reported on work in which they expanded this predictive index to include items such as the woman's appraisal of anticipated child care stress after the birth, her coping style and her current mood. Beck8 has developed the Postnatal Depression Prediction Inventory, a clinician checklist of 13 risk areas for postpartum mood difficulties. The Edinburgh Postnatal Depression Scale,9with a validated cut-off score for antenatal use,10 has also been suggested. Campagne2 has put forward the Pregnancy Mood Profile, which consists of just 2 questions assessing mood using criteria for depression from the Diagnostic and Statistical Manual of Mental Disorders (fourth edition),11 and Austin12 has developed a self-report antenatal risk factor questionnaire consisting of 9 questions in the domains of psychiatric history, life stresses, personality, support, childhood experiences and relationship with the woman's mother. Matthey and associates13^,14 recently reported on a set of 12 questions from the domains of support, life stress, personality, psychiatric history, family violence and childhood experience administered to women along with the Edinburgh Postnatal Depression Scale by midwives during the antenatal period.

Reports on the tools developed by Campagne and Austin are at this stage descriptive in nature — that is, no data have yet been produced against which judgments can be made as to their clinical usefulness. Beck provides an excellent overview of how to use her Postnatal Depression Prediction Inventory (and a postpartum measure) in a clinical setting,15 and Hanna and associates have reported on the frequency of the risk factors detected using the inventory with a sample of 107 women, with around half the women found to be feeling anxious, depressed or stressed.16 The measure by Cooper and associates underwent a rigorous development and assessment process involving about 6000 women. The authors concluded that it had limited clinical utility with respect to predicting postnatal depression: only about one-quarter of women who antenatally scored above the cut-off point experienced depression postnatally. Honey and colleagues7 improved the index, but its clinical utility remained modest. The assessment questionnaire developed by Matthey and colleagues was used to collect empirical information on the frequency of psychosocial risk factors in over 2000 pregnant women in Sydney, Australia, as well as the impact on clinical caseloads.13 The most frequent risk factors were personality and life stresses (about 24%), followed by psychiatric history (20%), current mood (12%), childhood abuse (9%), family violence (5%: interestingly, more women reported being violent toward their partner than their partner being violent toward them); the least frequent risk factor was lack of support (3%). In addition, this team has gathered empirical evidence for the acceptability of routine antenatal psychological assessments to both the staff and women (from both English-speaking and non– English-speaking backgrounds).14 However, the assessment questionnaire, as with most of the others, lacks peer-reviewed empirical data about its clinical usefulness, although work on this is currently in progress. To my knowledge, this questionnaire is the only measure now embedded in routine clinical practice across a variety of hospitals and health services.

In this issue (see page 253) Carroll and associates17 report the results of a randomized controlled trial that examined the effectiveness of the Antenatal Psychosocial Health Assessment (ALPHA) form in detecting psychosocial risk factors in pregnant women. The form presents a series of suggested questions on a range of psychosocial issues (www.cmaj.ca/cgi/content/full/173/3/253/DC1). The authors found that the group of physicians, obstetricians and midwives who were randomly assigned to use the form were almost twice as likely to detect psychological concerns than the group who provided usual care (odds ratio 1.8, 95% confidence interval 1.1–3.0, p = 0.02). The majority of women with health care providers in the ALPHA group were favourable to discussing issues on the form (72.7%). Of the two-thirds of health care providers in the ALPHA group who responded to a feedback questionnaire, 64% found the tool easy to use and 86% would use it if it were recommended as standard practice.

However, beyond the stated limitations regarding losses to follow-up and, given that most of the providers were family physicians, generalizability, there are 2 important additional limitations to the study. First, 65% of eligible clinicians declined to take part, even though payment was included. This high rate of refusal indicates that it is premature to believe that the ALPHA form would be widely incorporated into antenatal care. Further work needs to be undertaken to investigate ways of encouraging its use (the authors do mention that this may be overcome with a self-report form they have developed).

Second, the fact that the health care providers were paid for providing information on each woman they recruited, and that this payment was greater for those using the ALPHA form ($50) than for those providing usual care ($20), may have significantly affected the findings. As with the “cognitive dissonance” phenomenon, payment is more likely to lead to positive views of an instrument. Payment may also lead to greater motivation to detect psychosocial concerns than if no payment is provided. Further work where financial incentive is not an issue will be helpful.

It is interesting that, although use of the ALPHA form leads to more concerns being detected, it is only in the area of abuse being witnessed or experienced during childhood that its use is significant. The authors did not find any differences between the 2 groups in the detection of family factors, maternal factors or issues regarding substance abuse. This may be because clinicians providing usual care, having expressed interest in participating, are those who are more diligent in enquiring about psychosocial concerns. We may only know if this is the case if the ALPHA form can be introduced as part of routine care in one region and the results compared with those of a region where it is not used.

This study builds on other work that shows that exploring antenatal psychosocial mental health issues is acceptable to women and health professionals.14 It also provides some tentative information (see caveats above) for the increased detection of antenatal concerns, in a similar vein to the improved detection of postnatal mood disorders when systematic screening has been implemented.

It is encouraging that not only are different centres around the world calling for the implementation of routine perinatal mental health screening but teams are attempting to evaluate different ways of doing this. Evaluation methodologies will continue to improve with the variety of studies being undertaken in this field and clinical services provided to women and their families. One of the most difficult — and most important — aspects to evaluate is that of effectiveness. Today's ethical standards would rightly not allow a study in which women assessed as being at risk of poor postnatal outcomes would not be offered an appropriate intervention in order that the efficacy of those interventions be evaluated. The challenge, therefore, is for investigators to consider how best to undertake studies that can provide some answers to this question. In addition, it is likely that routine assessments will become more sophisticated in the weighting of different risk factors.

Although some reviews18 have questioned the validity of undertaking routine psychosocial assessments, it is important to remember that routine assessment does not occur only through a formalized, published measure. It also occurs when the patient is asked how she is feeling, and whether she has any worries or concerns either now or for after her baby is born. Any reply that indicates possible stress should be taken seriously and appropriate support or intervention considered. Thus clinicians should not wait upon the results of empirical trials before deciding whether or not routine antenatal screening is something they or their service should implement — rather, they should consider which questions seem reasonable. By asking such questions, clinicians show their concern not just for the patient's physical well-being but also for her whole well-being and for that of her family.

𝛃 See related article page 253