The selection of primary clinical outcome by which to evaluate the success of a therapy is discussed. Outcome should be relevant to the treatment, important to the patient and to the public health, and easy to measure with little associated variance. The odds of death is chosen as the primary outcome by which to evaluate chronic dialysis treatment because the purpose of dialysis is to preserve life. The odds of death is then correlated with the processes of care, which themselves may be measurable and therefore viewed as outcome of other processes. A general model for outcome research using existing data and for evaluating the relationships among processes and outcome is discussed using a large patient cohort analysis as an example. The analysis suggests that two dimensions of the care process account for most of the explainable death odds difference among patients: a dialysis intensity-related dimension and a nutrition-related dimension. Experience has suggested that knowledge of these principles combined with simple measurements has, over time, led dialysis professionals to increase the intensity of the dialysis treatment they deliver. That was not the case for the nutrition-related dimension. We speculate that these professionals understand, from prior knowledge, how to control dialysis intensity but do not possess similar knowledge about nutritional therapeutics for dialysis patients. Therefore, a better understanding of nutritional therapeutics will be required to achieve maximum reduction of mortality rates.