Aims & Objectives: Aim: In pediatric critical care, is necessary to use tools for prediction of mortality and compare to own outcomes. Paediatric index of mortality 3 (PIM 3) is one of the scores used for this. It is necessary to assess its performance in predicting mortality, according to the population in study. Objective: To validate PIM 3 score in pediatric critical care units at 2500 meters over the sea level in Colombia, and assess how other variables such as lactate and partial oxygen saturation (SpO2)/FIO2 ratio work in this model. Methods: Prospective, observational, national multicenter study. There were included six pediatric critical care units located at 2500 meters over the sea in Colombia with all the patients admitted consecutively during eight months. All the necessary information to calculate PIM 3, lactate and SpO2/FIO2 ratio was included. Results: A total of 2069 patients were analyzed. The mortality was 4.5% and the standardized ratio of mortality for PIM 3 was 1,19 (CI 95% 1,00 – 1,37). The score showed an inadequate calibration (p=0,01) and a good discrimination (AUC 0,84 CI: 0,83-0,86). A new proposed model using SpO2/FIO2 had a good discrimination and calibration. Conclusions: Mortality in Colombian children living at 2500 meters over the sea is similar to other countries. PIM 3 has a good discrimination but an inadequate calibration, this finding has many causes.The SpO2/FIO2 works for a better predictive value but is necessary to validate it with other population.