Publications by authors named "Florian D van Leeuwen"
Background: External validations are essential to assess the performance of a clinical prediction model (CPM) before deployment. Apart from model misspecification, also differences in patient population, the standard of care, predictor definitions, and other factors influence a model's discriminative ability, as commonly quantified by the AUC (or c-statistic). We aimed to quantify the variation in AUCs across sets of external validation studies and propose ways to adjust expectations of a model's performance in a new setting.
View Article and Find Full Text PDFPractices for controlling intracranial pressure (ICP) in traumatic brain injury (TBI) patients admitted to the intensive care unit (ICU) vary considerably between centres. To help understand the rational basis for such variance in care, this study aims to identify the patient-level predictors of changes in ICP management. We extracted all heterogeneous data (2008 pre-ICU and ICU variables) collected from a prospective cohort (n = 844, 51 ICUs) of ICP-monitored TBI patients in the Collaborative European NeuroTrauma Effectiveness Research in TBI study.
View Article and Find Full Text PDFTo compare the incremental prognostic value of pupillary reactivity captured as part of the Glasgow Coma Scale-Pupils (GCS-P) score or added as separate variable to the GCS+P, in traumatic brain injury (TBI). We analyzed patients enrolled between 2014 and 2018 in the Collaborative European NeuroTrauma Effectiveness Research in Traumatic Brain Injury (CENTER-TBI, = 3521) and the Transforming Research and Clinical Knowledge in Traumatic Brain Injury (TRACK-TBI, = 1439) cohorts. Logistic regression was utilized to quantify the prognostic performances of GCS-P (GCS minus number of unreactive pupils) and GCS+P versus GCS alone according to Nagelkerke's .
View Article and Find Full Text PDFArticle Synopsis
- Interrupted time series (ITS) designs are valuable for studying the impact of events in natural experiments, but their effectiveness can vary with different data structures and time points.
- A Monte Carlo simulation study revealed that the ability to detect changes in data (power) largely depends on sample size for step changes, while slope changes rely more on the number of time points.
- The research shows that detecting a significant step change requires a larger sample size and more time points compared to detecting a slope change, highlighting the need for researchers to thoroughly evaluate their data and model choice before using ITS designs.