In setting out to enhance caregiver sensitivity, the target for change is the mother's or primary caregiver's sensitivity. However, since the sensitivity needs to be tailored to that particular child, the intervention needs to include observation of the mother with the child. The issue is not the mother–child interaction but the mother's interaction with the child. Some of this work therefore requires the presence of the child, while, for instance, feedback to the mother using video will be with the mother alone. Within the parameters of attachment theory, the child's behaviour or emotional state are not a focus for intervention although these aspects are central to the evaluation of the efficacy of the intervention.
The intervention studies included in the following meta-analyses by Bakermans-Kranenburg, van IJzendoorn and Juffer1 (2003) were not restricted to a particular population. The authors point out that some samples were middle-class families with healthy infants, but studies with clinical and at-risk populations were included as well. Moreover, the meta-analyses include studies where children were living with their birth families and some studies where children had moved from their birth families to live with alternative caregivers. From the descriptions of the samples presented (Table 1, p.198), it seems that only a very small number of studies of adopted children are included in the meta-analyses, and that the majority of the samples are of at-risk or clinical populations.