Liu Q, et al. - Given that the detection of new biomarkers is important to advancing diagnosis and treatment of CKD, but depends profoundly on the statistical methods used, and inappropriate methods can result in both false positive and false negative links between biomarkers and outcomes, therefore, researchers focused on the accuracy of methods using computer simulations and compared biomarker association estimates in a prospective cohort investigation of patients with glomerular disease, named as the NEPhrotic syndrome sTUdy NEtwork. The three methods that were compared were time-invariant average, time-varying last observation carried forward (LOCF), and time-varying cumulative average. Markedly different results were obtained when different analytic methods were used. Inaccurate association estimates were generated using the time-invariant average, whereas additive (cumulative average) or instantaneous (LOCF) associations could be estimated by other methods, relying on the hypothesized underlying association mechanism and research question.