Evaluating health visitor assessments of mother-infant interactions: a mixed methods study.

Background: Given the significance of reliably detecting cases where mother-infant relationships are not developing successfully, it is important that initial assessment processes are as sensitive and specific as possible.

Objectives: This study sought to examine the processes by which health visitors identify problems in mother-infant relationships in the post-natal period.

Design: Mixed methods.

Settings: Two universities and two primary care trusts.

Participant: In Phase One 17 first-time mothers and their 6- to 16-week-old infants were recruited. In Phase Two, a sample of 12 health visitors participated.

Methods: The study incorporated two data collection phases. In Phase One, each mother's interaction with her baby was video-recorded for 20min in an observation laboratory. The video-recordings were coded and resultant data were analysed to derive a number of quantitative measures of interaction quality, including mothers' responsiveness and sensitivity to their infants as rated by the Global Ratings Scales (GRS) of Mother-Infant Interaction. In Phase Two, 12 health visitors rated and assessed 9 clips of the video-recorded mother-infant interactions. The rationales for their ratings were then explored through in-depth interviews. Health visitor ratings of the video clips were compared to the GRS ratings. The relationship between the main focus of each health visitor rating, as reported in the interview, and the consistency of ratings with the GRS ratings were then investigated.

Results: Correlations between individual health visitors' ratings and the GRS ratings ranged from .17 to .83 and were statistically significant in only four cases. There was a weak relationship with health visitors' years of experience (r(s)=.47, NS). When explaining their judgements, health visitors tended to comment on the mother's behaviours or the relationship between the mother and baby and often ignored the behaviour of the baby. There was a highly significant relationship between the consistency of health visitor/GRS ratings and the number of references to the baby in the health visitors' explanations (r(s)=.75, p=.005).

Conclusion: This study contributes to the understanding of how health visitors make assessments of mother-infant interactions. The frequent lack of attention and reference to the baby's behaviour suggests an area for further training.