Prediction of Cardiovascular Risk in Hemodialysis Patients by Data Mining

 

 Buy Article Permissions and Reprints

Summary

Objectives: The objective of this work was to contribute to the development, validation and application of data mining methods for prediction in decision support systems in medicine. The particular focus was on the prediction of cardiovascular risk factors in hemodialysis patients, specifically the interventricular septum (IVS) thickness of the heart of individual patients as an important quantitative indicator to diagnose left ventricular hypertrophy. The work was based on data from 63 long-term hemodialysis patients of the KfH Dialysis Centre in Jena, Germany.

Methods: The approach applied is based on data mining methods and involves four major steps: data based clustering, cluster based rule extraction, rulebase construction and cluster and rule based prediction. The methods employed include crisp and fuzzy algorithms. At each step, logical and medical validation of results was carried out. Different sets of randomly selected patient data were used to train, test and optimize the clusterbases and rulebases for prediction.

Results: Using the best clusterbase/rulebase combination designed, the IVS thickness cluster (‘small’ or ‘large’) was predicted correctly for 30 of the 35 patients with known IVS values in the training data set; no patient was predicted incorrectly and 5 were parity predicted. For the test data set, 4 of the 6 patients with known IVS values were predicted correctly, no patient incorrectly and 2 parity. These results did not substantially differ from those obtained using the second best clusterbase/rulebase combination which was finally recommended for use based on further performance criteria. The prediction of the IVS thickness clusters of the 22 patients with unknown IVS values also yielded good results that were (and could only be) validated by a medical individual risk assessment of these patients.

Conclusions: The approach applied proved successful for the cluster and rule based prediction of a quantitative variable, such as IVS thickness, for individual patients from other variables relevant to the problem. The results obtained demonstrate the high potential of the approach and the methods developed and validated to support decision-making in hemodialysis and other fields of medicine by individual risk prediction.

• References

• 1 Pfaff M, Guthke R, Pohlmeier R, Schroeder K, Stein G, Vienken J. Using Data Mining Methods to Assess Factors Associated with Dialysis Outcome. ASN American Society of Nephrology Renal Week 1999, 32nd Annual Meeting & Scientific Exposition, November 5-8, 1999, Miami, USA. Abstract: Journal of the American Society of Nephrology (JASN) Supplement September. 1999
  Google Scholar
• 2 Pfaff M, Schroeder K, Woetzel D, Weller K, Guthke R, Pohlmeier R, Vienken J, Stein G. Dialysis Data Mining to Examine Relations between Therapy and Outcome. ASN American Society of Nephrology Renal Week 2000, 33rd Annual Meeting & Scientific Exposition, October 13-16, 2000, Toronto, Canada. Abstract: Journal of the American Society of Nephrology (JASN) Supplement August. 2000
  Google Scholar
• 3 Pfaff M, Woetzel D, Weller K, Schroeder K, Vienken J, Pohlmeier R, Stein G, Guthke R. Data Mining to Optimize Hemodialysis Treatment. ESIT 2000 European Symposium on Intelligent Techniques, September 14-15, 2000, Aachen, Germany. Abstract, Paper and Talk: Session on Biomedical Applications. Final Programme & Proceedings (Abstracts + CD), Published by Verlag Mainz, Wissenschaftsverlag, Aachen, 2000, ISBN 3-89653-797-0. Programme: Booklet p. 7; Abstract: Booklet pp. 39/40; Paper: CD-ROM. 134-69.
  Google Scholar
• 4 Schroeder K, Woetzel D, Weller K, Pfaff M, Stein G. Comparison of Methods of Data Mining and Statistics as Applied to the Influence of ACE Inhibitor Therapy on the Dose of Recombinant Human Erythropoietin in Hemodialysis Patients. 108th Congress, Deutsche Gesellschaft fuer Innere Medizin, April. 6-10. 2002. Wiesbaden, Germany: Abstract in the Proceedings;
  Google Scholar
• 5 Schroeder K, Woetzel D, Pfaff M, Weller K, Pohlmeier R, Vienken J, Stein G. Blood Pressure during Hemodialysis Sessions in the First and Fourth Year of Treatment. HYPERTONIE 2000 24th Scientific Meeting, Deutsche Liga zur Bekaempfung des hohen Blutdruckes und Deutsche Hypertonie Gesellschaft, November 22-25, 2000, Heidelberg, Germany. Abstract and Poster: Programme p. 27, P2-16. and Journal Supplement DMW. 2000
  Google Scholar
• 6 Pfaff M, Weller K, Woetzel D, Schroeder K, Pohlmeier R, Stein G, Guthke R, Vienken J. From Data to Knowledge: Hemodialysis Data Mining. 1st CSVHC International Conference on Software Validation for Health Care, April 4-6, 2001, Bonn, Germany. Invited Talk. Extended Abstract and Presentation Slides in the Conference Proceedings, Published by SQS Software Quality Systems AG. Cologne, Germany: 2001: 1-34.
  Google Scholar
• 7 Woetzel D, Pfaff M, Weller K, Schroeder K, Pohlmeier R, Vienken J, Stein G, Guthke R. Data Mining and Prediction Tools to Support Patient Care in Hemodialysis. In: Intelligent Systems in Patient Care. Adlassnig K-P. (ed) Proceedings of the EUNITE Workshop, October 5, 2001, Vienna, Austria. Oesterreichische Computer Gesellschaft. 2001. ISBN 3-85403-154-8 books@ocg.at Vol. 154 163-5.
  Google Scholar
• 8 Pfaff M, Woetzel D, Weller K, Schroeder K, Pohlmeier R, Vienken J, Stein G, Guthke R. Cluster and Rule Based Prediction in Hemodialysis – Adaptivity vs. Robustness. EUNITE 2001 European Symposium on Intelligent Technologies, Hybrid Systems and their Implementation in Smart Adaptive Systems, December 13-14, 2001, Puerto de la Cruz, Tenerife, Spain. Abstract, Full Paper and Talk: Session on Computational Intelligence in Medicine. Final Programme & Proceedings (Abstracts + CD), Published by Verlag Mainz, Wissenschaftsverlag, Aachen. 2001. ISBN 3-89653-916-7. Programme: Booklet p. 13; Abstract: Booklet pp. 65/66; Full Paper: CD-ROM 286-335.
  Google Scholar
• 9 Pfaff M, Woetzel D, Weller K, Schroeder K, Pohlmeier R, Vienken J, Stein G, Guthke R. Cluster and Rule Based Prediction in Hemodialysis – Adaptivity vs. Robustness. In: Adaptive Systems and Hybrid Computational Intelligence in Medicine. Dounias GD, Linkens DA. (eds) Special Session Proceedings of the EUNITE 2001 Symposium. Organized by IBA D: Human, Medical & Health Care, December 13-14, 2001, Tenerife, Spain. Published by the University of the Aegean, Greece, in Collaboration with EUNITE, the European Network on Intelligent Technologies for Smart Adaptive Systems, ISBN 960-7475-19-4 2001: 14-31.
  Google Scholar
• 10 Pfaff M, Weller K, Woetzel D, Guthke R, Schroeder K, Stein G, Pohlmeier R, Vienken J. Prediction of Cardiovascular Risk in Hemodialysis Patients by Data Mining. BSI 2002 4th International Workshop on Biosignal Interpretation, June 24-26, 2002, Como, Italy. Abstract, Paper and Poster. In: Proceedings of the BSI 2002 Workshop. Cerutti S, Akay M, Mainardi LT, Sato S, Zywietz C. (eds) Published by the Polytechnic University of Milan, Italy, 2002. Programme Poster Session P2: Proceedings p. x; Abstract and Paper: Proceedings 445-50. (This contribution was awarded the 3rd Poster Prize of the Workshop and invited for publication in Methods Inf Med.)
  Google Scholar
• 11 Dunn JC. A Fuzzy Relative of the ISODATA Process and its Use in Detecting Compact, Well Separated Clusters. Journ Cybern 1974; 3: 95-104.
  PubMedGoogle Scholar
• 12 Bezdek JC. Pattern Recognition with Fuzzy Objective Function Algorithms. Plenum Press; New York, USA: 1981
  Google Scholar
• 13 Guthke R, Ludwig B. Generation of Rules for Expert Systems by Statistical Methods of Fermentation Data Analysis. Acta Biotechnol 1994; 14: 13-26.
  CrossrefPubMedGoogle Scholar
• 14 Guthke R, Pfaff M, Meyer F. Optimized Fuzzy Rule Generation for Fermentation Control. CAB7 7th International Conference on Computer Applications in Biotechnology. May 31 to June 4, 1998, Osaka, Japan. In: Conference Preprints. Yoshida T, Shioya S. (eds) 1998: 303-8.
  Google Scholar
• 15 Jessen H, Slawinski T. Test and Rating Strategies for Data Based Rule Generation. In: Computational Intelligence, Sonderforschungsbereich 531, Paper No CI-39/98, ISSN 1433-3325. Dortmund, Germany: 1998
  Google Scholar