2 Jun 2026
The goal is to bring about lasting improvements in the quality, validity, and transparency of studies and thus strengthen the foundations of reliable, evidence-based medicine.
An international research consortium has published a pioneering guideline for the use of routinely collected data (RCD) in healthcare.
The paper appeared in the prestigious journal The BMJ and constitutes the first comprehensive guideline of this kind. It seeks to bring about lasting improvements in the quality, validity, and transparency of studies based on such data and thus strengthen the foundations of reliable, evidence-based medicine.
The guideline is the result of a unique interdisciplinary collaboration, with leading doctors from the fields of cardiology and heart surgery working closely with top international researchers from the domains of statistics, methodology, and artificial intelligence.
This close cooperation between clinical and methodological expertise enabled the team to properly address the complex challenges of routinely collected data and develop robust scientific solutions.
Sabine Hoffmann, Leiterin des Statistischen Beratungslabors am Institut für Statistik der LMU | © privat
Routinely collected data – from sources such as electronic health records, registers, and invoicing data – offer huge opportunities for medical research, as they present large patient groups under real care conditions. At the same time, they are associated with considerable methodological challenges.
The newly published guideline systematically analyzes these data and describes key problem areas, including lack of representativity, insufficient data quality, lack of coordination on timing of measurements and interventions, non-randomized treatment decisions, and a multitude of possible analytical pathways.
A particular focus is on the risk of distorted results and on problems in dealing with missing and erroneous data. Furthermore, the role of modern analytical methods – especially the methods of artificial intelligence – is critically evaluated. The guideline shows that although they possess great potential, they can also lead to misleading results without due methodological rigor.
As a central innovation, the paper presents a structured roadmap together with concrete recommendations for action which enable researchers to improve the quality of their analyses. These include strategies for ensuring data quality, for correctly defining time points, and for the transparent and reproducible reporting of studies.
“Through these comprehensive recommendations, the guideline makes a decisive contribution to improving research quality,” says Dr. Sabine Hoffmann, Head of the Statistical Consulting Unit at LMU’s Department of Statistics and lead author of the guideline. “It helps researchers avoid faulty interpretations, enhance the reproducibility of studies, and strengthen confidence in findings based on routinely collected data.”
The paper arose in the context of growing initiatives to promote high-quality evidence, including through institutions such as the Institute for Quality and Efficiency in Health Care (IQWiG), which emphasize the importance of methodologically robust studies for health policy decisions.
“Routinely collected data open up huge opportunities for investigating medical questions faster and on a broader basis. At the same time, we must be aware of the methodological challenges if we are to obtain valid and reliable results,” explains Hoffmann.
With its publication in The BMJ, the guideline sets a new international reference standard for the analysis of routinely collected data. It offers key guidance to researchers, clinicians, and decision-makers and makes an important contribution to the further development of evidence-based medicine in the digital era.
Sabine Hoffmann et al.: Using routinely collected data for research purposes: Challenges and mitigation strategies. The BMJ 2026