Influence of the general data protection regulation on data quality in registry data collection : Selection bias due to data protection requirements and the impact on key indicators for measuring the quality of care in a level 1 trauma center

Due to the new general data protection regulations (GDPR), the requirements for correct patient information on the documentation of pseudonymized data in a registry have increased enormously. In particular, written consent applies to the TraumaRegister DGU® as it is not always possible to get written permission from severely injured patients in acute situations. Therefore, the study aimed to investigate the influence of undocumented cases due to a lack of clarification on the standardized mortality rate (SMR). In 2019, 274 patients meeting the criteria of the baseline dataset were retrospectively recorded. In the remaining 197 patients, the RISC II score could be calculated in all cases. In addition, due to state-specific law, all deceased patients were documented in our trauma center. In this study with 197 primary care patients (72% male), 147 (74,6%) were informed and gave permission or died and were subsequently documented. The predicted mortality, actual mortality and SMR were 18.5%, 19.0% and 1.03, respectively. For patients who were not informed (n = 50), the predicted mortality, actual mortality, and SMR were 7.0%, 0% and 0. When these cases are included, the SMR is significantly more favorable at 0.93. Due to the lack of written consent from surviving patients, only about 75% of all patients at Leipzig University Hospital could be documented for the TraumaRegister DGU®. On the other hand, since the local legal situation permits registry documentation of deceased patients, this has a detrimental effect on the standardized mortality rate (SMR), which is about 10% higher in our collective than it actually is.