On-line monitoring of dialysis adequacy using DiaSens optical sensor: Accurate Kt/V estimation by smoothing algorithms

Objective: develop signal processing algorithm for the removal of disturbances not related to treatment course and compare with available clinical data. Methods: The study included 20 treatments from 19 uremic patients, 12 males and 4 females. Treatment instability was assessed with instability coefficient. Finally, Kt/V values obtained from optical signals were compared with laboratory results. Results: The value (mean ± SD) of instability coefficient for non-processed data was 324,75 ± 39,78, for AVRG algorithm 9,73 ± 4,01 and for SMART 12,06 ± 7,06. Deviation from available clinical data (RMSE) for raw data was 1,42 ± 0,20, AVRG 0,07 ± 0,03 and 0,07 ± 0,04 for SMART. Conclusion: Data processing algorithms show good results in removal of disturbances with significant precision increase.