Predictors of Stress in Pregnant Women: Findings from the Maternity Experiences survey

Dear Eric - Thank you for your thoughtful questions and comments. "You said: «Backward stepwise multivariate logistic regression was used to develop separate models of predictors for life event stress and perceived stress, and included all variables that were significantly related to the outcomes in bivariate analysis » .
a. Was the “outcomes”, the presence of one or more of the 13 stressful life events sugested by Newton & Hunt, 1984?
b. If not, what outcomes were they (pre-term birth, low-birth weight, post-partum depression, etc.)?"
Question a)
For this analysis, we were really looking at predictors of stress, and the 2 outcomes were: (i) perceived life stress, and (ii) stressful life events. Measurement of stress based on "stressful life events" remains one of the most prominent ways of measuring stress, but has some limitations in that the nature of the events themselves may not cause negative stress for people ( e.g., moving to a new house). Most often, it is the number of events that are used to class an individual as having "high" or "low" stress. In our study, women who had greater than or equal to 3 stressful life events were considered to be high stress. There were 13 events assessed, and the low "cut-off" point for determining low vs higher stress is a reflection of the "rarity" of many of the events in the scale. Again, this is a limitation of the life event measures. Given their popularity in the stress literature, it was deemed important to include this validated scale in the survey along with the perceived stress measure.

You have also made some interesting comments, Eric, about perception of stress. Perception of stress (regardless of the source of stress, which the life event scales attempt to quantify) is a more direct measure in some ways in that it allows an individual to describe the level of stress that they personally experience. There are a few validated tools available (e.g., Perceived STress Scale) which are quite widely used, and you have made salient points about the need for valid/reliable measures of perception. A few studies have linked perception of stress to physiological responses. For the most part, there has been little distinction in the stress-perinatal outcome literature regarding the "type" of stress e.g., chronic versus acute. This is a more recent focus.

The ongoing debate regarding how to measure stress, this very difficult construct, motivated us to look at predictors of stress using both measures. Although I have described some differents in the Results, one additional striking finding is that more modifiable predictors were captured with both stressful life events and perceived stress instruments.
Thanks, again, Eric for the discussion.
DAwn

Hi, Jo -
Thank you for your comments. Your comment regarding the impact of stress and its effect on outcomes is an important consideration when considering the "bioloigcal plausibility" of the relationship between stress and some poor birth outcomes. I do think, too, that the impact of stress on the poor outcome will vary depending on the outcome e.g., preterm birth - is the impact on elevating "stress" hormones, which cascades to initiate labour versus low birth weight - is its impact more on risky behaviours e.g, smoking? These are questions yet to be addressed.

Jo - thank you for your comments on sleep quality. The Perceived Stress Scale (not used in this study) does use "symptoms" of stress e.g., "in the last month, how often have you felt that you were on top of things", and "in the last month, how often have you been angered because of things that were outside your control". There is not a specific item related to sleep. I could certainly see sleep being a mediator of stress, or even a moderator in some cases. I wonder if this would make it difficult to use sleep dep as a proxy?
Thoughts?
Thanks, Jo
Dawn

Dear Heather -
Your comments about linking subjective stress measures to physiol measures is interesting, and there is only one or two papers (by the same group of researchers) that have explored this and found a linkage. I agree that this is really needed, because otherwise we are just hypothesizing about the effects of stress.

Your note about delaying childbearing is also an important point, and you have "hit" upon what some have referred to as a "weathering" response and introduces the idea of not only chronic, but CUMULATIVE stress. This is also of particular importance to me, and one that I am looking at in my doctoral work. Bruce McEwan has used the terms "allostasis" to describe the wear and tear that cumulative stress has on the body, and a few perinatal researchers have adopted this framework. IN particular, looking at the disparities in preterm birth rates in black vs white women, some have suggested that the cumulative effects of racism may partially explain this difference.

Dawn

Dear Alex -
Thank you for your comments. Surprisingly, the correlation between perceived stress and stressful life events was not high (.23), but was statistically significant. Of interest, cross-tabs showed that among women who had high stress based on stressful life events (>3), 28.7% of them also had high perceived stress.

Interesting comment, Alex, regarding daily hassles versus major life events and the cumulative nature of daily hassles, and the use of SEM to disentangle some of the pathways to stress. I am exploring cumulative stress using SEM for my thesis work. One of the challenges with cumulative / chronic stress evaluation is certainly distinguishing (if possible) how current circumstances colour or influence recollection of previous circumstances.

Re: the content of the brochures - it does address the linkage between stress and social support, and recommends that women seek sources of support.
Thanks, Alex,
Dawn

Hello -
Thank you for your comments and questions

In this analysis, social support is conceptualized as a moderator (effect modifier) of the relationship between stressor and perceived stress. In this regard, we have based the analysis of perceived stress on Lazarus' framework. I do think that it would be interesting to follow-up on this analysis of predictors of stress with a structural equation model that would provide the opportunity to fully model the mediators and moderators of stress. There has been little work done in this area, and my sense is that it would have been difficult to construct a model without a bit of preliminary work. There are other variables that could potentially act as modifiers as well e.g., depression (e.g.,the perception of stress is likely to be different among women who are depressed, not depressed)

You are correct - we did not have info on women's employment status. I cannot comment on whether this would have acted as a mediator specifically between education level and stress; however, I have seen one study recently that tied lack of control in women's working environment to stress and poor birth outcomes. Given some of the Whitehall study work, this is not surprising.

There have been trials which have attempted to reduce stress in pregnant women through "social support". However, most were not effective, and their lack of success may have been related to the nature of the support (not targeted at area of stress; provided by "strangers" without prior relationship). I tend to think that we need to understand sources of stress in order to effectively intervene.
Based on these findings, women experience stress when they have medical conditions that they perceive may impact their pregnancy outcome, or have had a previous poor birth outcome (stillbirth,miscarriage, fertility procedures) that, again, may affect their current pregnancy. Health care providers can assess women's level of stress related to these factors and make them points of discussion and support during visits, ensuring that women's understanding about the impact of these factors on pregnancy is accurate. We do not know why women were unhappy about their pregnancy, or its timing , but exploring women's feelings about their pregnancy during visits may also raise opportunities for support. Women who experience depression prenatally may experience stress because they are concerned about the development of post-partum depression. Assessment of "antenatal" depression can provide opportunity for education, support, and encouraging women to prepare and educate their support network re: signs of postpartum depression.

Based on other findings of this survey, women's main sources of information are physician/nurse, books, other mothers, and previous experience. Trials of peer support (mother-to-mother) on post-partum depression have been effective in reducing depressive symptommatology. Peer support (mother to pregnant mother; pregnant woman to pregnant woman)using trained peer "counsellors" may also be a promising intervention for reducing stress, and one which is cost-sparing for the health care system. Again, my comments on this are based on the realization that other mothers are a key source of support and information for pregnant women and new mothers.

From an assessment perspective, formally identifying these predictors of stress as "risk factors" could lead to their inclusion in risk assessments that are performed at prenatal care visits. There has been criticism that some of our risk assessment approaches for poor birth outcomes are based solely on physiological assessments to the exclusion of psychosocial factors.

Finally, just a brief final thought - the most ideal situation I believe would be to assist women to identify and reduce their stress in the context pre-conception well-woman visits. In other words, assisting women with stress-management is something that belongs in a regular health check-up (nurse practitioner, family physician), and not only in pregnancy.

Thank you for the discussion.

Dawn