A woman's perception of the degree to which she feels she has satisfactory emotional and social support has been found to be related to postnatal depression. Clearly there is clinical value in being able to assess social support simply and accurately. The study by Webster et al sets out to assess a 6-item scale that the authors have been using clinically in terms of its correlations with various pregnancy and postnatal measures. The scale showed modest associations with scores for the Edinburgh Postnatal Depression Scale. The authors also suggest links with poorer health and health care use during pregnancy and postnatally, although the evidence that they provide to demonstrate this is somewhat limited.
So, have they produced the kind of simple and straightforward measure of social support that many would welcome for clinical use? The answer can only be, maybe. What this study does not tell us is whether or not this particular set of questions is either a valid measure (do they indeed measure what they set out to measure?) or a reliable measure (do they give consistent results?). Simply putting together a set of questions and calling them the “Maternity Social Support Scale” and showing some associations with outcome measures does not, of course, mean that the scale is either reliable or valid. Does it have the qualities that good screening measures need—is it sensitive, specific, and a good predictor? Among other things we need to know something about the structure of the scale—how far do the separate questions correlate with each other? If the 6 questions, or at least some of them, correlate highly with one another, the scale could be simplified. Indeed, does the scale have any advantage over the kinds of direct questions that others have used; for example, “Do you feel you are getting the emotional support you would like from your husband/partner?” ( 1). Or perhaps a couple of such questions relating to partners and the wider circle of family and friends may be more appropriate.
Although there may be clinical value in having knowledge of social support for its own sake, much better measures may exist that could be used in screening for health, well-being, and postpartum depression. In the present study, the group rated as having poor social support clearly have a number of expected demographic and social characteristics. They are more likely to be under 20 years old and not in a marriage or de facto relationship, and they turn up later for antenatal care. Does the scale provide a better predictor of outcomes than would these variables? As far as scores on the Edinburgh Postnatal Depression Scale is concerned, a good predictor is the score on the Edinburgh scale during pregnancy—we should not forget that depression in pregnancy is at least as common as it is after delivery ( 2). This study of Green's, as others have shown, indicates that other factors exist which sharpen the predictive power of a pregnancy depression measure for postpartum well-being. These include physical symptoms, tiredness, social support (“Do you feel you are getting the social support you would like from your husband/partner”?), having no one to talk to postnatally, and mothers saying that they had not been pleased to find themselves pregnant when this became apparent.
There is no doubt of the need to have simple, nonintrusive ways of assessing women's well-being and psychological and social problems in pregnancy and the postnatal period that have the necessary requirements to be useful for effective screening. But I do not think we are there yet. A lot of research centers are doing parts of what may be required. What is now needed is the putting together of a package of measures that is appropriate for each particular population and health care system. We must remember that questions that seem meaningful and relevant to women in one community may seem irrelevant or even offensive to those in another. These packages will need to be carefully assessed to ensure that the measures are valid and reliable and fulfill the necessary criteria for an effective screening program. In addition to measuring these aspects of scales, it is important to assess their value to open discussion on difficult issues like domestic violence, as Webster et al report for their scale.
I think that measuring social support in pregnancy, as well as the other required factors, can be done in ways that are simple and meaningful, but more work is needed to reach that point.