New research conducted within the PHEMS project has been presented at the ISTH 2026 Congress, exploring how missing treatment information in Electronic Health Records (EHRs) can be accurately reconstructued for children with heamophilia A.
The study focuses on a common challenge in health-record data: incomplete information about treatment and medication dosing. The researchers developed a methodology to infer missing Factor VIII concentrate dosing information from routinely collected clinical data, with the aim of reconstructing more complete treatment histories.
Using data from children with haemophilia A treated at Erasmus MC Sophia Children’s Hospital, the approach was evaluated for its ability to accurately infer missing dosing information and improve the completeness of datasets used for research.
Improving the quality and usability of EHR data is particularly important for research into rare diseases, where patient populations are often small and comprehensive datasets can be difficult to obtain. By enabling missing treatment information to be reconstructed from existing clinical data, this work can help create more robust datasets for research while supporting the development of privacy-preserving machine learning models.
