[Access to health data for research]

Secondary use of individual-level health data for research is cost-efficient but challenging. In the PCBase Xtend project, we collected PSA values, biopsy reports, MRI results, and chemotherapy data from all 21 regions and private healthcare providers in Sweden. The process required 46 applications over a period of almost 4 years and revealed major variability in IT systems, administrative routines, and costs. The time to data delivery ranged from weeks to years. Early quality checks were essential to correct errors in the data extraction. Despite challenges, >11 million PSA values, 550 000 prostate biopsies, 170 000 MRI results, and >6 million drug administrations could be added to PCBase for use in epidemiological studies of prostate cancer. Better administrative routines, data extraction integrated in IT systems, incentives for data migration, and national support structures are needed to improve access to health data for research.