Can and should we? Use of maternal administrative data to identify children at risk of poor outcomes

Abstract

Introduction
Millions of children in the UK experience early adversity, with many requiring children’s social care (CSC) support. National longitudinal data on these children is essential to understand lifelong impacts on health and development, yet they are often excluded from the UK’s leading child development studies. This study explored the feasibility of using linked administrative records for sampling, supported by discussions with data owners.

Methods
We analysed populationlevel administrative data from the SAIL Databank (Wales) and the ECHILD database (England) to identify maternal psychosocial risk factors (such as mental health conditions, substance misuse and domestic violence) and CSC involvement before age six. A cohort was created of mothers with live births between April 2009 and March 2016, linked to secondary healthcare and CSC records. In parallel, we engaged with data providers across the UK to assess practical and governance considerations.

Results
In Wales, 15% of mothers whose child was involved in CSC had psychosocial risk factors. By contrast, only 4% of mothers without these risk factors had a child with CSC involvement. Further analyses are ongoing. Data providers confirmed that while administrative records have precedent for recruitment, ethical and sensitivity concerns pose barriers.

Conclusion
Findings highlight links between maternal vulnerability and CSC involvement, alongside rising psychosocial adversities. Administrative data offers valuable insights but using it to identify children at risk remains complex. Differences in legislation, governance, and data access across UK nations hinder a unified sampling frame, especially given the sensitivity of maternal and social care records.