IJPDS International Journal of Population Data Science

The use of population administrative data has grown rapidly in recent years. However, because these data are collected for administrative rather than research purposes, they often lack information on key factors that influence both exposures and outcomes. This can make it difficult to fully account for confounding (a common cause of exposure and outcome distorting the effect of interest), potentially biasing research findings.

A new study published in the International Journal of Population Data Science (IJPDS) by researchers at University College London proposes a novel approach to address this challenge. The new method performed well across a range of simulated and real-world scenarios.

The researchers propose supplementing population administrative data with information from cohort studies. Cohort studies typically follow smaller samples of individuals over long periods of time and collect rich data across multiple domains. As a result, they often include important confounding variables that are missing from administrative datasets.

Administrative records for cohort members are increasingly being linked to their cohort study data to enhance research potential. In this study, the authors demonstrate how these linked data can be used alongside a statistical method known as multiple imputation – commonly applied to handle missing data – to better address confounding in analyses of population administrative data. The paper also considers situations where linked cohort and administrative data cannot be accessed together within the same secure environment, such as when datasets are held in separate Trusted Research Environments (TREs). 

The approach is illustrated using simulated data across three different scenarios, before being applied to a real-world example examining the association between pupil mobility (changing schools at non-standard times) and Key Stage 2 (age 11) attainment. The analysis used data from the UK National Pupil Database (NPD), supplemented by multiple measures of socioeconomic deprivation from linked Millennium Cohort Study (MCS) data.

The study included 509,670 pupils in the NPD, of whom 7,768 (1.5%) were MCS cohort members. Pupil mobility was associated with lower Key Stage 2 attainment, but the estimated effect was reduced by almost 20% when socioeconomic information from the MCS was incorporated alongside NPD variables.

Lead author, Prof Richard Silverwood, said, “The underlying principles of this novel approach are widely applicable. Any analysis of administrative data could potentially be strengthened by linking a subset of individuals into richer cohort data”. However, he added that more research is needed to understand how best to apply these methods across different settings.

 

Click here to read the full article

Professor Richard Silverwood, Centre for Longitudinal Studies, University College London, UK

Silverwood, R., Baranyi, G., Calderwood, L., De Stavola, B., Ploubidis, G., White, I. and Harron, K. (2026) “Adjusting for confounding in population administrative data when confounders are only measured in a linked cohort”, International Journal of Population Data Science, 11(1). doi: 10.23889/ijpds.v11i1.3015.