New research led by Institute member Dr Jialing Lin provides a structured guide for addressing missing data in administrative health datasets, strengthening the reliability of real-world health research.
Administrative health data – such as hospital admissions, insurance claims, and billing records – are increasingly used to inform health services research and policy. While these datasets offer large-scale, population-level insights, they are often not designed for research purposes. As a result, missing or incomplete data, particularly relating to social determinants of health, can introduce bias and impact the validity of findings.
Dr Lin and colleagues outline a step-by-step approach to identifying and managing missing data, drawing on the Treatment And Reporting of Missing data in Observational Studies (TARMOS) framework.
The paper explains key types of missingness – including data missing completely at random, missing at random, and missing not at random – and highlights appropriate analytical strategies such as multiple imputation, inverse probability weighting, and sensitivity analyses. The authors also emphasise the role of causal diagrams to clarify assumptions about why data are missing and to guide robust analytical decisions.
“Administrative health data are powerful tools for advancing healthcare research and policy,” says lead author Dr Jialing Lin. “Without careful planning, transparent reporting, and appropriate methods to address missingness, we risk drawing biased or incomplete conclusions. Our framework provides practical guidance to enhance the rigour and policy relevance of real-world data studies.”
“Better data linkage and integrated digital health systems are critical to reducing fragmentation in care and essential for generating reliable evidence to inform health system decision-making.”