Predicting posttraumatic stress disorder after childbirth

Abstract

Objective

around 50% of women report symptoms that indicate some aspect of their childbirth experience was ‘traumatic’, and at least 3.1% meet diagnosis for PTSD six months post partum. Here we aimed to conduct a prospective longitudinal study and examine predictors of birth-related trauma – predictors that included a range of pre-event factors – as a first step in the creation of a screening questionnaire.

Method

of the 933 women who completed an assessment in their third trimester, 866 were followed-up at four to six week post partum. Two canonical discriminant function analyses were conducted to ascertain factors associated with experiencing birth as traumatic and, of the women who found the birth traumatic, which factors were associated with those who developed PTSD.

Findings

a mix of 16 pre-birth predictor variables and event-specific predictor variables distinguished women who reported symptoms consistent with trauma from those who did not. Fourteen predictor variables distinguished women who went on to develop PTSD from those who did not.

Conclusions

anxiety sensitivity to possible birthing problems, breached birthing expectations, and severity of any actual birth problem, predicted those who found the birth traumatic. Prior trauma was the single most important predictive factor of PTSD. Evaluating the utility of brief, cost-effective, and accurate screening for women at risk of developing birth-related PTSD is suggested.

Introduction

The arrival of a new baby is typically an event that is associated with much anticipatory and experienced joy. However, despite low mortality rates in developed countries (World Health Organization, 2004) the birthing process can, for a few, also be accompanied by feelings of terror, fear for the mother׳s or infant׳s life and a sense of helplessness or lack of control (Soderquist et al., 2002, Geller, 2004). While experiencing some anticipatory anxiety may almost be viewed as normative, there are some women whose actual birthing experience results in them meet diagnostic criteria for either partial or full PTSD. In a large Australian sample of pregnant women (Alcorn et al., 2010), 3.6% met full PTSD criteria (diagnosed using the Posttraumatic Diagnostic Scale; Foa et al., 1997) four to six weeks post partum. By six months, this figure had risen to 5.8%. Controlling for antenatal psychopathology (pre-existing trauma and clinically significant depression and anxiety) only reduced these rates to 1.2% and 3.1% respectively. These rates are comparable to other reports where the rates of PTSD following childbirth have ranged from 1% to 6% (Creedy et al., 2000, Ayers and Pickering, 2001). These numbers are also similar to the 12-month prevalence rate of PTSD after experiencing a potentially traumatic event. Creamer et al. (2001) found the proportion of women who met criteria for PTSD following any trauma to be 2.9%.

These prevalence rates are of concern as PTSD following childbirth is associated with significant problems in mother-infant attachment (Allen, 1998), partner relationships (Beck, 2004) and increased use of the health care system (Switzer et al., 1999). It is reasonable to propose that further investigation of factors that predict PTSD will inform and improve clinical practice. Several studies have made a significant contribution to the field already. Soderquist and colleagues (Soderquist et al., 2006, Soderquist et al., 2009) in one of the most comprehensive studies to date, found that depression in early pregnancy, stress, coping capacity and severe fear of childbirth in late pregnancy to be significant predictors of PTSD, although pre-birth state anxiety was not. Other researchers found anxiety sensitivity (Keogh et al., 2002, Fairbrother and Woody, 2007), depression (Soderquist et al., 1996, van Son et al., 2005, Maggioni et al., 2006) and dissociation (van Son et al., 2005) to be key predictors. Obstetric intervention (Creedy et al., 2002; Soet et al., 2003, Fairbrother and Woody, 2007) also has been identified as important, as has a negative relationship with hospital staff (Creedy et al., 2000, Soet et al., 2003). Pain, a history of sexual trauma and feeling powerless during the birth have also been found to predict PTSD-type symptomatology (Soet et al., 2003). Testing a cognitive model, Ford et al. (2010) found a direct effect of social support at three months with some additional variance accounted for by cognitive variables. It is also possible that extreme pre-existing stressors such as child sexual abuse that result in post traumatic stress prior to childbirth also increase the likelihood of PTSD, a finding obtained in other areas of trauma research (Yehuda and McFarlane, 1995, McNally, 2003).

A major problem with identifying risk factors is that typically there is only access to individuals after a traumatic event. A sample of childbearing women, on the other hand, provides an opportunity to screen prospectively for predisposing factors. The current literature provides a wide range of possible predictors of birth-related trauma; however, no study has conducted a prospective, longitudinal, comprehensive examination of individual risk factors and pre-event factors and birth factors and their association with the development of PTSD, as defined in the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV-TR; American Psychiatric Association, 2000). Thus, taking this body of research as a starting point we surveyed childbearing women in their third trimester regarding their history, birthing expectations (based upon previous research, as exemplified above), and then followed-up these women again one month post partum. We then assessed women for (a) finding the birth traumatic and (b) meeting the criteria for PTSD. Predictors (both pre-birth predictors and event related predictors) for both of these states were then computed.

In taking a two-stage approach, it is possible to investigate the process of the development of PTSD in a more refined and systematic manner using current diagnostic nomenclature rather than symptom severity as the key outcome variable.