Multiple-choice questions (MCQs) are popular among learners, but are often criticized for emphasizing recognition over active recall, which makes them less effective for long-term memory retention. Well-designed competitive distractors that are similar to the correct answer in meaning (semantically related) or wordform (orthographically related) can prompt deeper cognitive processing and overcome this deficit. However, creating high-quality distractors traditionally requires a significant time investment and domain expertise. Recent advances in artificial intelligence enable the generation of distractors at scale, but are insufficient when it comes to creating distractors that specifically target learners' misconceptions about word forms. In this study, we present a scalable, data-driven method for automatically generating orthographically related distractors, based on common incorrect responses from open-answer retrieval practice, supplemented with rule-based generation of common misspellings where necessary. We apply this method to a large dataset of learners' errors in vocabulary learning, demonstrating that it is feasible to create distractors that align with learners' shared misconceptions in a real-world setting. This work contributes to making MCQs a more effective pedagogical tool.
Injuries as a result from bicycle accidents occur frequently in the Netherlands, annually 76.000 persons. In this article we focus on the consequences for daily practice in ten questions on this subject. Different injuries per category of bicycle (e-bike, fat-bike, regular bike) are addressed, with focus on eventual red flags for injuries that can potentially be missed in the acute phase by general practitioners or emergency physicians. Also, information on risk factors for participation in traffic, specifically for older persons are discussed and preventative measures to reduce the risk of injuries while participating in traffic by bicycle.
Introduction and Objective: KidneyIntelX is a composite risk score incorporating biomarkers and clinical variables at baseline for diabetic kidney disease (DKD) progression. We sought to determine the clinical relevance of KidneyIntelX and thresholds in a large cohort of patients with type 2 diabetes and a broad range of CKD. Methods: We measured tumor necrosis factor receptor (TNFR)-1), TNFR-2, and kidney injury molecule (KIM-1) on banked plasma samples from CANVAS and CREDENCE participants, and executed kidneyintelX.dkd at baseline and year 1. We assessed the association of baseline and changes in kidneyintelX.dkd with a composite kidney outcome of 40% decline in eGFR or kidney failure. Hazard ratios were estimated using multivariate Cox regression. Results: There were 4677 participants (mean eGFR 69.4 mL/min/1.73m2; median UACR 77.0 mg/g) with available plasma samples. At baseline, kidneyintelX.dkd scored 867 (18.5%) as high, 1520 (32.5%) as moderate, and 2290 (49.0%) as low risk. The adjusted HR per doubling in predicted probability was 2.26 (95% CI 1.80-2.84). At year 1, the median change in KidneyIntelX predicted probabilities was 0.0%, with >10% reduction in 25.6% of canagliflozin- vs. 17.0% of placebo-treated patients. The adjusted HR for the change in predicted probability from baseline to year 1 was 2.24 (95% CI 1.65-3.04). Canagliflozin led to more patients shifting to lower risk at year 1 (32.1% vs. 16.2% moderate to low; 38.9% vs 19.7% high to moderate or low). Low risk at year 1 was associated with very low kidney outcome risk of 0.3 events per 100 person-years, while staying at moderate had 1.4, moving from high to moderate had 2.2, and high risk at year 1 had 8.9. Conclusion: Baseline and year 1 KidneyIntelX risk assessments in patients with type 2 diabetes and CKD robustly stratified patients for DKD progression independently of classic predictors and should be considered for enriching clinical trials and assessing treatment response. E. Moedt: None. S. Coca: Consultant; Renalytix, Bayer Pharmaceuticals, Inc, Alexion Pharmaceuticals, Inc, Vera Therapeutics, Mylan. F. Fleming: Employee; Renalytix. H.L. Heerspink: Consultant; Alnylam Pharmaceuticals, Inc, Alexion Pharmaceuticals, Inc, AstraZeneca, Bayer Pharmaceuticals, Inc, Boehringer-Ingelheim, Eli Lilly and Company, Janssen Pharmaceuticals, Inc, Novartis AG, Novo Nordisk A/S, Roche Pharmaceuticals, Traveere Pharmaceuticals, Menarini.
