From patient's perspective to predictive modeling: cerebrovascular events and chronic kidney disease

Abstract

Patients with chronic kidney disease (CKD) are at increased risk for both complications related to their condition and a variety of other health issues, particularly vascular events. Despite the uncertainty surrounding these risks, patients express a strong desire to be informed. Prediction models can provide personalized risk assessments tailored to individual needs. This thesis is divided into two parts: (1) we explore which outcomes are prioritized by patients with CKD, and (2) we evaluate the predictive performance of models predicting the risk of ischemic stroke. Each part includes a methodological chapter that details the specific methods used. Our research underscores the importance of understanding CKD patients' priorities for outcomes and care planning in order to deliver patient-centered care. Given the high risk of misclassification with conventional ischemic stroke models - primarily due to the exclusion of CKD patients from development studies - we recommend using the Modified-CHADS score instead of the CHA2DS2-VASc score for this population.