A0289
Title: Risk factors and transitional probability of clinical events in Korean CKD patients using the multi-state model
Authors: Jinheum Kim - University of Suwon (Korea, South) [presenting]
Abstract: Since the KNOW-CKD study is a longitudinal study, CKD patients are not only exposed to fatal events such as death during the study period but also to non-fatal events such as cardiovascular disease (CVD) and end-stage kidney disease (ESKD). Such non-fatal events are called intermediate events. By ignoring intermediate events, estimating the survival probability of CKD patients only with information on whether they survive or are censored may lead to misleading results. In situations where intermediate events can be experienced in a patient's disease progression, do not consider a two-state model consisting of only the initial state and death state as an analytical model but use so-called multi-state models created by adding intermediate states to the two-state model. It is better to consider the most used multi-state model is the so-called illness-death model, which consists of a healthy state corresponding to an initial state, an illness state corresponding to an intermediate event, and a death state corresponding to an absorbing state. The advantage of this illness-death model over the two-state model is that it can compare the effects of illness on patient mortality. A four-state model that extends the illness-death model is proposed to compare the effects of CVD or ESKD experience on patient death. Through this multi-state model, the disease progression of CKD patients can be explained more adequately.
