The data were analysed by independent experts in the fields of conventional biostatistics using multivariate Cox regression and experts in the “novel” self learning rule-based modelling technique. We also compared their results with the established modified Charlson scoring system. For predicting survival the multivariate Cox regression explained 14.6%, the modified Charlson scoring system 12.9% and the rule based modelling 3.95% of the variation which indicated that Cox regression and the Charlson index had the highest predictive accuracy. A more detailed comparison of model performance in predicting 1- and 5-year survival is shown in the table.

These results are only available because of the high quality of the data which were returned annually by colleagues. As well as providing useful epidemiology information, this study proved the value of international collaboration, at which nephrologists excel. The citation for the published report is given below.

Citation: Colin C. Geddes, Paul C. W. van Dijk, Stephen McArthur, Wendy Metcalfe, Kitty J. Jager, Aeilko H. Zwinderman, Michael Mooney, Jonathan G. Fox, and Keith Simpson. The ERA-EDTA cohort study - comparison of methods to predict survival on renal replacement therapy. Nephrol Dial Transplant 2006; 21:945-956.