Key points are not available for this paper at this time.
As noted in the previous paper of this series, 6 the purposes of prognostic stratification are to improve the efficiency of research design and data analysis, to enh: mce confidence in prediction, and to demarcate phenomena for scientific explication. Without such stratifications, erroneous or distorted results will continue to be produced when conventional statistical “models” are applied to the data of medical cohorts. The statistical “models” are based on the analysis of response to the maneuver of a single experiment, but in most medical forms of cohort research, the investigator observes the results of a double experiment 2 : a maneuver arranged by man is imposed on the response to a previous maneuver arranged in essence by nature.
Alvan R. Feinstein (1972) studied this question.