Development of a clinical rating scale to assess mother-infant interaction in a psychiatric mother and baby unit.

R Kumar , A E Hipwell

Abstract

BACKGROUND Psychiatrists are sometimes asked to express an opinion about the competence and motivation of a mentally ill mother to be a consistent, adequate and safe parent of her recently born infant. There are no reliable methods for assessing mother-infant interaction and relationship in a psychiatric context.

METHOD The Bethlem Mother-infant interaction Scale (BMIS) was developed by the staff of a specialised psychiatric mother and baby unit. Of the seven subscales, four measured different aspects of the mother's contribution to the dialogue with her baby, one measured her capacity to organise and maintain routine care, one attempted to rate the perception by staff of risk to the child, and the remaining subscale rated the baby's contribution to their interaction. In normal clinical practice the ratings were made consensually by nurses and they encompassed observations that had been made by them during the previous week.

RESULTS Despite the simple and global nature of the ratings of the BMIS, moderate to high coefficients of inter-rater reliability were obtained for weekly ratings by nurses as well as for intra-rater reliability of ratings of videotaped interactions between mothers and their infants. Ratings of mothers and babies in day-to-day activity in the unit that were based on one hour's observation only were much less reliable. The internal consistency of the BMIS was high and comparisons with other methods used in non-clinical settings showed good criterion-related validity.

CONCLUSIONS Using the BMIS facilitated attempts to reliably assess disturbances of mother-infant interaction and the scale was acceptable and clinically useful. The nurses' ratings of their perceptions of 'risk' to the infant, which are in some ways the most important, were the least reliable. Repeated weekly ratings gave a valuable impression of change or lack of it over time. The predictive validity of the scale has yet to be determined.