Improving practice: impact of a social services data quality investigation

Since 2018, ADRC Northern Ireland has been collaborating with the Northern Irish Department of Health (DoH) on a programme of work surrounding mental ill-health and children known to social services. During this time issues surrounding data quality and completeness have been raised, reinforcing the message within the department of the importance of high-quality data recording on the ground for high quality research further downstream. This collaboration is resulting in an important impact on practice. 

Led by Dr Aideen Maguire, this programme of work makes use of data from Northern Ireland’s unique integrated health and social care system which dates back to 1985. Social services data has been linked to primary care, prescription medication, emergency department presentations and mortality data. This has created the UK’s largest population-wide longitudinal study of children known to social services to date. The linked, de-identified data is held within the secure data environment of the Business Service’s Organisations’ Honest Broker Service.   

This programme of research is the first to utilise the social services data for research purposes on this large a scale, demonstrating the value and potential of these datasets. Not only has this led to publications, citations in government strategies and many additional projects from other researchers, it has also highlighted the importance of high-quality data to the DoH, and identified areas where they can address improvements to data collection and quality.

Linkage of these datasets are dependent on accurate recording of the individuals’ Health and Care Number (HCN). The project team found that over the 30-year study period, HCN coverage (the existence of HCN in the records) was high in all datasets except the social services data, with over 10% of all “Children in Need” records missing a complete HCN. In some years, this was as high as 20% for “Children in Care” records. It is not known from the data whether these gaps are down to mis-recording of HCN or individuals lacking a HCN. This was fed back to the DoH with a full Data Quality report. Our partners are looking into what improvements and mitigations can be made to improve HCN coverage in both historic and current data given these issues which were uncovered by allowing researchers data access. 

Administrative data research is only possible if we have high quality data to link and interrogate. As well as improving access to data, improving data quality will ensure that linked administrative data can be used effectively to its fullest potential. To that end, we have appointed a Steering Group to discuss project progression, co-produce and prioritise research questions and report on data completeness and quality. In addition, working closely with the DoH has allowed us to share the importance of accurate data keeping and the uses and impact that this data can have beyond direct service provision through to research and evidence informed policy development.