A web-based system for clinical decision support and knowledge maintenance for deterioration monitoring of hemato-oncological patients.

We introduce a web-based clinical decision support system (CDSS) and knowledge maintenance based on rules and a set covering method focusing on the problem of detecting serious comorbidities in hemato-oncological patients who are at high risk of developing serious infections and life threatening complications. We experienced that diagnostic problems which are characterized by fuzzy, uncertain knowledge and overlapping signs, still reveal some kind of patterns that can be transferred into a computer-based decision model. We applied a multi-stage evaluation process to assess the system's diagnostic performance. Depending on how system behavior was compared to presumably correct judgment of a case the correctness rate for closed cases with all data available varied between 58% and 71%, the overall rate after critical review was 84%. However, the real time behavior of our approach which data becoming available as time passes still has to be evaluated and observational studies need to be conducted.