Conference Agenda

Background: The use of observational measures of parent–infant interaction is established in research contexts, but less so in clinical services due to problems with feasibility and unknown acceptability. A programme of research has examined these issues to inform the implementation of observational measures in PMHS.

Study 1: This study assessed the construct validity of the global sensitivity scale and a sensitivity composite from the NICHD Parent–Infant Interaction Scales. Parent–infant interaction videos from parents recruited from specialist PMHS in England (N=275) were analysed and inter‑rater reliability was tested in a subsample of dyads with infants under 3 months. Both the global sensitivity scale and sensitivity composite showed evidence of construct validity and good inter‑rater reliability, although not all NICHD scales were reliable for very young infants.

Study 2: A qualitative study explored parents’ views of an observational measure being used as part of care. Semi‑structured interviews were conducted with parents from PMHS (N=29). Interviews are being analysed using the Framework Method. Analysis is ongoing; early themes relate to parents’ assumptions, their relationship with the service, and perceptions of usefulness; full results will be available at the conference.

Study 3: This study will examine the psychometric properties of a brief practitioner‑version of parental sensitivity derived from the NICHD scales when applied to a UK community sample of mothers and infants aged 6 months (N=250). Inter‑rater agreement with the original NICHD system and predictive validity to later child outcomes will be tested. Findings will inform whether the measure is ready for implementation and evaluation in PMHS.

Conclusion: Together, these studies will add to an evidence base on feasibility, acceptability, and measurement validity to guide the implementation of observational parent–infant interaction measures in PMHS so they can be used to better tailor support for families.

Attachment assessment in older children and adolescents has largely relied on representational measures, reflecting the move from attachment behavioral patterns to a focus on internal working models. While informative, these approaches provide limited access to how attachment is expressed within specific parent–child relationships. In our work we have found that observing attachment behavior in older children and adolescents continues to offer important clinical and research value.

As part of an ongoing clinical-research program, 60 families with children aged 8–18 participated in attachment evaluations that included observation of parent-child/adolescent interactions. Preliminary findings from a subsample of 39 foster and adopted school-aged children indicate significant associations between children’s observed attachment patterns and caregiving patterns, as well as caregivers’ Parental Reflective Functioning (PRF). In a smaller subsample, we are finding that the same appears to also be true for adolescents. In this presentation we will describe the development and use of a separation–reunion procedure adapted for older children and adolescents and provide some video examples of such observations.

In this same clinical group, observational results have been used to inform intervention planning, identify relational risk, and guide work with caregivers in clinical contexts. Together, these early findings demonstrate that attachment patterns can be observed beyond early childhood and that observational results provide useful and actionable information that directly supports intervention for families with older children and adolescents.

Various narrative-based measures assess attachment quality and representation in adolescence. These measures mostly originate from childhood research, such as the Child Attachment Interview (CAI; Target et al., 2007) and Friends and Family Interview (FFI; Steele & Steele, 2005), or from adult research, like the Adult Attachment Interview (AAI; Main et al., 1985; Hesse, 2016). However, few assessments focus on developmental tasks specific to adolescence. The Adolescent Separation Anxiety Interview (ASAI; Richard et al., 1998; Brown & Wright, 2003) addresses this gap by eliciting narratives about age-typical separations from parents to evaluate attachment strategies and representations.

The ASAI is a semi-structured, semi-projective interview combining visual stimuli and open-ended questions. Five age-appropriate images depicting increasingly stressful separations are used to activate the attachment system. Adolescents describe the depicted individual’s feelings, thoughts, and behaviors, as well as those of caregivers. Responses are coded categorically (secure, insecure–avoidant, insecure–preoccupied) and dimensionally across nine subscales, including emotional openness, narrative coherence, devaluation of attachment, preoccupied anger, and optimism versus anxiety. Research shows the ASAI has good reliability and validity, correlating meaningfully with adolescent clinical outcomes.

Findings from a 1.5-year longitudinal study including clinical and community adolescents show that ASAI classifications, coherence scales, and separation resolution predict internalising and externalising symptoms, as well as personality functioning measured by the LoPF Q 12–18 (Kerr et al., 2023) and OPD-CA SQ (Schrobildgen et al., 2019). The ASAI is also used in an ongoing longitudinal study following a community sample from birth, linking adolescent attachment representations with early relational experiences such as maternal sensitivity, reciprocal caregiving, and infant attachment quality in the Strange Situation Procedure.

Background: The lack of pragmatic tools to measure attachment on a large scale in preschool-middle childhood has contributed to limited data for this age range. The Attachment Screening Assessment (ASA; Glaser et al., 2013), a 45-item caregiver-report questionnaire developed at the Anna Freud Centre, was designed to identify variations in attachment behaviour in children aged 4-11 years. Unlike observational or representational measures, the ASA offers a practical, quick screening tool. This study presents tentative findings on the ASA’s factor structure.

Methods: The ASA was administered to caregivers from three samples: a community sample (N=107), children in foster and residential care from an agency (N=406), and children in foster care participating in a trial (N=57). Caregivers rated their child’s behaviour on a 5-point Likert scale (e.g. “When physically hurt does your child go to someone for help or comfort?”). Exploratory factor analyses examined theoretical subscales measuring Secure, Avoidant, Resistant, and Disorganised attachment (including “Infant”-Disorganisation, Controlling-Caregiving, and Controlling-Punitive), as well as Disinhibited social engagement.

Results: The organised attachment subscales—Secure, Avoidant, and Resistant—and the Disinhibited subscale demonstrate acceptable factor structure. The Disorganised subscales present a more complex picture: “Infant”-Disorganised and Controlling-Caregiving items function better as a combined composite, while there is no support for a Controlling-Punitive subscale. However, a 5-factor structure including all organised subscales and the one disorganised composite does not achieve robust psychometric properties.

Discussion: The ASA shows promise as a screening tool, particularly for organised attachment patterns and disinhibited behaviours. We propose possible interpretations for the complex results regarding disorganised items. Further work is needed to revise item wording, examine whether items are interpreted as intended, and collect data from new participants with revised items. Following these refinements, the ASA has the potential to be a useful tool for research and clinical purposes.