An Empirical Model for the Prediction of Hemodialysis System Performance

The examination of steady state behavior of hemodialysis treatment using the analytical model created by Olson et al. revealed that for each set of hemodialysis conditions, a fixed time constant exists that dictates how quickly the patient’s waste level cycle reaches a steady state condition. They also revealed that initial waste level does not affect the final steady state waste level. In this study, an empirical model for the time constant and the final steady state maximum waste level were found that lumps hemodialysis inputs such as flow rates, dialyzer properties, and patient waste generation by conducting a parametric study on a previous hemodialysis model from Olson et al. [1] The empirical model is validated by comparing the curve that predicts how the peak waste level of each cycle changes over time with the analytical model’s results. For all the tested input values which cover most of practical hemodialysis treatments, the curve closely matched the numerical model’s results with R2 value higher than .9973. The empirical model created in this study provides a much simpler prediction method without the use of complex numerical simulations. In addition, the Olson model cannot be used to run an inverse simulation to determine optimal inputs for desired outputs. This limitation is overcome by our empirical model, which further allows much easier and more extended exploration of different therapies (dose length and schedule) for both doctors and renal replacement system designers.Copyright © 2011 by ASME