Development of a risk stratification algorithm to improve patient-centered care and decision making for incident elderly patients with end-stage renal disease.

A significant number of elderly patients die during their first 3 months of dialysis. Because dialysis can impair the quality of both life and death, a personalized care plan based on both early prognosis and patient choices is required. We developed a prognostic screening tool to identify older patients in need of specific care based on a multidisciplinary approach. Our study included 24,348 patients aged 75 years and older from the French national renal epidemiology and information network (REIN) registry who began dialysis between 1 January 2005 and 30 September 2012. Our primary outcome was overall mortality during the first 3 months of renal replacement therapy. Multivariate logistic regression was used to construct a scoring system in a random half of the cohort (training set). This score, which included age, gender, specific comorbidities, albumin levels, and mobility, was then applied to the other half (validation set). In all, 2548 patients died during the first 3 months after dialysis initiation, 22% after dialysis withdrawal. Three risk groups were identified: low risk (score under 12 points, 3-month expected mortality under 20%), intermediate risk (score from 12 to 16, mortality between 20 and 40%, 9.5% of patients) and high risk (score 17 or more, mortality over 40%, 2.5% of patients). We developed a decision-making process that classifies patients according to their risk of early death in view of their potentially imminent need for supportive care or treatment.