OBJECTIVES:  To construct a patients grouping methodology based on risk adjustments, allowing categorize each patient in only one group, through a combination of demographic, epidemiological and risk factors variables. METHODS:  Through health resources consumption database, ambulatory, hospital, emergency and home care of one year in the Colombian health system, a classification of patients into mutually exclusive categories was constructed, using algorithms and rules on the identification of demographic variables, primary and secondary diagnosis (ICD-10). This classification prioritizes the identification of each patient with chronic diseases and risk factors. The consistency of the group was analyzed using a multivariate model run by Stata 13.0. RESULTS:  The group of patients allowed to classify each member within a single group, between 59 categories, summarized in three main groups. (I) patients without chronic disease, grouped by life cycle and gender; (II) patients with a chronic disease classified into 9 disease groups and each one of them with two levels of severity and finally (III) patients with chronic multi-disease distributed according to the number of diseases that the patient has (two, three, four or more chronic diseases). The multivariate model that explains the behavior and interaction between variables was highly significant. For a population of 3,256,380 users, 1.31% of users were classified as multi-chronic and their consumption was 12.47% of the total cost, compared to 90.78% patients classified without chronic disease, representing 57.68% the cost. CONCLUSIONS:  The grouping model allows characterization of total population with a minimum set of data available within the Colombian health system and its use allows the construction of proper follow-up strategies for people and improvement of resources allocation.