Model-Based Decision Support System for Individual Prescription of the Dialysate Bicarbonate Concentration in Hemodialysis

A decision support system has been developed that determines the optimal dialysate bicarbonate concentration in hemodialysis therapy for each patient individually. The knowledge about the behavior of the acid-base state during treatment has been provided by a mathematical model for the description of dynamic exchange processes during hemodialysis. This model simulates the sodium and water distribution, the acid-base state as well as the ventilation. The decision support system uses the model for the prediction of the end-dialysis acid-base state and calculates by means of linear optimization the dialysate bicarbonate concentration which is necessary to reach a specified end-dialysis state. If the aspired acid-base state can not be reached, the system varies the dialysate sodium concentration and the treatment time. The whole program can be used on a PC and is easy to use. One decision making process lasts between 10 seconds and 5 minutes depending on the computer.

Dieter Falkenhagen (1942–2015): A Multifaceted Scientist

Dieter Falkenhagen was born in 1942 in Dresden, Germany and died in 2015. He specialized in internal medicine and nephrology. Focusing on artificial organ research, he investigated various aspects of the efficacy and safety of hemodialysis and adsorption technologies, including biocompatibility issues related to blood versus surface interactions and the adverse effects of endotoxin contamination. He studied various mathematical models to analyze efficacy and safety, and animal models to help clarify uncertainty issues. Through his studies, adsorbents were developed, resulting in Prometheus, an artificial liver support device. Anticoagulation models, including citrate perfusion, were improved and made safer by his work. He also stepped into bioreactor research to increase efficacy of liver support devices.

Improvement of decisions regarding hemodialysis patients by urea kinetic modeling

The potentials for improving decisions about adequacy of dialysis (AD) and daily protein intake (DPI) by urea kinetic modeling (UKM) were examined. Four nephrologists evaluated AD, DPI, and metabolic stability in 62 patients. UKM was done three times; but the results were not revealed. Clinicians' decisions were then compared with UKM measures of effective dialysis (Kt/V) and protein catabolic rate (pcr). Detection of inadequately treated patients by the clinicians was poor (28%, Kt/V less than 1.0; and 40%, Kt/V less than 0.9). Specificities of the clinicians' decisions were 0.96 and 0.92, respectively. Combining pcr and consensus decisions, 59 patients could be assigned a probable DPI. Using this as "gold standard," the average clinician detected 77% of 13 low DPIs. Single or triple pcr determinations alone detected 77% or 62%, respectively. Specificities were 0.91, 0.85, and 1.0. Simulated decision making suggested that combining pcr with clinical evaluation in a logical way would lead to detection of most patients with low DPI. Qualitative data from individual patient cases causing controversies are presented and discussed. It is concluded the UKM should be used routinely to assess the adequacy of dialysis and daily protein intake.