February 2026
We’re excited to share our latest publication in the International Journal of Population Data Science:
Cleaning and Validating Longitudinal Postcode Histories for Environmental Epidemiology
As part of our Air Pollution, housing and respiratory tract Infections in Children: National Birth Cohort (PCNIC) study, we developed a rigorous approach to cleaning and validating residential address histories for mothers and children using national administrative health records in England.
At first glance, cleaning postcode history data might not sound like ground-breaking environmental health research. However, it is an important contribution to our work because it strengthens the accuracy of our exposure assessment, which can directly affect the quality of environmental health evidence. We often assess exposure to environmental factors (like pollution, temperature, green space or neighbourhood deprivation) based on where people live. These postcode histories track where families live across pregnancy and infancy. If residential histories are incomplete, inaccurate, or not cleaned systematically, exposure estimates can be misclassified, which weakens our results and potentially biases findings.
Similar work on address histories has been done Scotland and Wales, and in some recruited cohorts in England, but there is limited published guidance on how to do this at a national scale in England. By cleaning and validating these postcode histories across pregnancy and early life, we demonstrated a practical way to build high-quality residential histories for environmental health research from routinely collected NHS records.
In the paper, we highlight key practical challenges in working with longitudinal postcode datasets, including missing or implausible dates, duplicate records, invalid postcodes, and inconsistencies across sources. After cleaning and validating these data, we found that around half of the children in the PICNIC Study moved during infancy, and around one third of the mothers moved during pregnancy. This shows that changes in address during pregnancy and early life are common and should be taken into account when estimating environmental exposures. We also outline clear lessons learned and recommendations for researchers using similar administrative data.
We hope these methods and lessons learned will help other researchers using postcode history data to improve exposure assessment in environmental health research across the UK.
For the full research article, see IJPDS.
