One advent of the widespread adoption of Electronic Health Records (EHR) in the US is the secondary use of the information to improve patient safety and quality of care. Learning Health Systems can leverage the power of big data, accelerating the process of translating newly generated knowledge to clinical practice. China is in the process of adopting EHR systems. We plan to utilize a limited data set derived from regional EHR data from Yinzhou to test the feasibility of building a Learning Health System to improve chronic kidney disease (CKD) care, specifically focusing on anemia. A new Anemia-in-CKD Research Database will help us develop several predictive models for outcomes of common anemia treatments, such as changes in hemoglobin from iron therapy or from using erythropoietin-stimulating agents (ESAs). In addition, we plan to represent, in fully computable form, the actionable recommendations published in the current KDIGO guideline recommendations in both English and Chinese. We will use the open-source Knowledge Grid technology from U-M to package, manage and execute the predictive models we develop and the computable KDIGO guideline recommendations we build. Finally, as proof of concept, we will use real-world patient data from Yinzhou to test the potential impact of the predictive models and computable guidelines as the potential logic of a future clinical decision support system for use in Yinzhou.