Judgment analysis in clinical nephrology.

To ameliorate the clinical performance of nephrologists, improving their clinical judgment is crucial. No methodology for judgment analysis in nephrology is currently available. Therefore, we designed a trial to assess the intraphysician consistency of the judgment of typical non-end-stage renal disease (ESRD) patients by 24 board-certified nephrologists. The participants were asked to analyze cases to determine the interobserver variability with respect to diagnosis, therapy, prognosis, and strategy of follow-up. They were unaware that every patient was presented on 2 occasions separated by a period of 6 months. Of the 1,288 questionnaires that were completed, 28 cases belonged to 1 of the following 3 groups: (A) patients once with, once without renal histology, (B) patients twice without histology, and (C) patients twice with histology. Only cases of group (A) differed at the 2 occasions of assessment with respect to knowledge of histology. The results from the first and second assessment were compared and analyzed. The median (95% confidence interval) percentages of changed diagnoses were 64% (59% to 68%), 50% (44% to 62%), and 33% (26% to 47%) in groups A, B, and C, respectively, indicating large intraobserver variability. The frequency of changes in diagnoses declined with the degree of confidence in the first diagnosis in all 3 groups. The subjective desire to know the histology was without impact on the frequency of changes in diagnoses. However, a knowledge of the histology enhanced the degree of confidence in the diagnoses. Interestingly, the enormous variability in changing diagnoses from one analysis to the other was not reflected by corresponding changes in the judgment of prognosis, therapy to be prescribed, or strategy of follow-up. The individual judgment with respect to diagnosis of clinical cases is inconsistent and highly dependent on the subjective degree of confidence in the diagnosis. The practical relevant consequences traditionally derived from a diagnosis (therapy, prognosis, and strategy of follow-up) are only marginally, if at all, affected by changing the diagnosis. Thus, the utility of "diagnosis" for judgment analysis in clinical nephrology should be reconsidered.