- This table gives an overview of the performance of indicators based on the main results of this study. The colors indicate poor (red), moderate (orange), or good (green) performance on feasibility and discriminability (adherence rates or between-centre variation). The adherence rates and event rates are shown as the median indicator scores across centres. For the determination of the feasibility, we calculated the amount of available data at patient level
- Discriminability is determined by adherence rates and between-centre variation: high adherence rates for structures and processes are considered as low discriminability. Discriminability is also reflected in the IQR (unadjusted) and the MOR (adjusted for case-mix and random variation). For outcome indicators, the statistical uncertainty (median number of events) was determined
- Feasibility: we determined that > 70% available data reflects good performance
- Discriminability: the potential for quality improvement was determined by the percentage adherence of centres to structure and process indicators (i.e. with high adherence rates, quality of care cannot be improved that much). We set the threshold for moderate potential for quality improvement at 80–90% and for poor potential at 90–100%. In addition, we considered a low (unadjusted) interquartile range on scores (IQR < 10) or non-significant (adjusted) between-centre differences as poor performance
- BTF Brain Trauma Foundation, DVT deep venous thrombosis, EDH epidural hematoma, GCS Glasgow Coma Scale, ICU intensive care unit, GOSE Glasgow Outcome Scale Extended, IQR interquartile range, MOR median odds ratio, OR odds Ratio, SDH subdural hematoma, TBI traumatic brain injury
- *Statistical uncertainty for outcome indicators (the less complications, the better the quality of care) was determined by the median number of events/median number of included patients per centre. We set the threshold for poor potential at less than 10 events
- aPharmaceutical or mechanical
- bBased on the IQR
- cBased on non-significant between-centre differences
