Refine
Has Fulltext
- no (2)
Year of publication
- 2021 (2) (remove)
Document Type
- Article (2) (remove)
Language
- English (2)
Is part of the Bibliography
- yes (2) (remove)
Keywords
- Validation (2) (remove)
Despite advances in machine learning-based clinical prediction models, only few of such models are actually deployed in clinical contexts. Among other reasons, this is due to a lack of validation studies. In this paper, we present and discuss the validation results of a machine learning model for the prediction of acute kidney injury in cardiac surgery patients initially developed on the MIMIC-III dataset when applied to an external cohort of an American research hospital. To help account for the performance differences observed, we utilized interpretability methods based on feature importance, which allowed experts to scrutinize model behavior both at the global and local level, making it possible to gain further insights into why it did not behave as expected on the validation cohort. The knowledge gleaned upon derivation can be potentially useful to assist model update during validation for more generalizable and simpler models. We argue that interpretability methods should be considered by practitioners as a further tool to help explain performance differences and inform model update in validation studies.
A fast high performance liquid chromatography tandem mass spectrometry multi-method based on an ACN-precipitation extraction was developed for the analysis of 41 (modified) mycotoxins in beer. Validation according to the performance criteria defined by the European Commission (EC) in Commission Decision no. 657/2002 revealed good linearity (R2 > 0.99), repeatability (RSDr < 15%), reproducibility (RSDR < 15%), and recovery (79–100%). Limits of quantification ranging from 0.04 to 75 µg/L were obtained. Matrix effects varied from −67 to +319% and were compensated for using standard addition. In total, 87 beer samples, produced worldwide, were analyzed for the presence of mycotoxins with a focus on modified mycotoxins, whereof 76% of the samples were contaminated with at least one mycotoxin. The most prevalent mycotoxins were deoxynivalenol-3-glucoside (63%), HT-2 toxin (15%), and tenuazonic acid (13%). Exposure estimates of deoxynivalenol and its metabolites for German beer revealed no significant contribution to intake of deoxynivalenol.