Women followed up (1990-5)

It was possible to continue to study 154 of women from the first programme of study, an average of three years later.

Aim

The aim was to see how the women changed over the intervening years in terms of their vulnerability, adverse experiences and disorder. Also to examine recovery from disorder and positive experience.

The sample

The follow-up selected women in three groups representing around half the original series:

Adult vulnerability characteristics (n=57) consisting of negative interaction with partner or child,

Childhood adversity (n=55) involving neglect or abuse before age 17,

Comparison group (n=42), a consecutive group of women. Good cooperation (80%) was achieved for the women approached for re-interview. There were no differences found between those followed-up and those not in terms of demographic or risk characteristics which might suggest the follow-up sample was skewed in some way.

Demographic characteristics, attachment style, support and self-esteem assessments were undertaken at first interview together with measures of depression, and these repeated at follow-up interview. Assessments reflected change between time 1 and time 2. At first interview, around a third of the women (35%) were working-class, over two thirds (67%) were in employment, just over half were married or cohabiting (55%) and most (76%) were mothers, with just under a third of the series (26%) were single mothers at first interview. These rates were no different from those in the larger series of 303 women. All women were interviewed in their own homes at both waves of the study.

Measures used

• Continuation of the Adult Life Phase Interview for new phases and adversities
• Attachment Style Interview and self-esteem repeated
• Parenting
• SCID for disorder since last interview

Findings

Analyses focused particularly on attachment style. This involved testing whether insecure attachment at first interview related to new onsets of depression and anxiety at follow-up. This was confirmed, with highly Fearful or Angry-dismissive styles in particular acting as mediating factors with childhood experience of neglect and abuse and new onset of disorder.

Changes in attachment style were also examined. There was high evidence of stability of secure attachment style with 73% of women remaining within broadly secure or insecure categories at both time points. Such stability was higher in the comparison women (83%) than in vulnerable women (60%). Although change occurred equally in a more positive (more secure) or negative (less secure) directions (14% versus 13%), negative change only occurred in comparison women (17%) and positive change was twice as common in the vulnerable women (20% versus 11%, p<.008). Highest stability was shown for Fearful style (correlation at both times 0.44) and those Clearly Secure (0.43).

A more sensitive marker of change in attachment style was also devised using the same measure. In this scheme 53% of women showed small degrees of change, again equally in a negative or positive direction. This showed that positive changes in support predicted such positive attachment style change, together with vulnerable status at first contact. Negative attachment style change was associated with an index of loss of employment status, negative change in non-partner support or negative change in self-esteem. Changes in depression had no impact on changes in attachment style.