The prevalence of anemia in children from Santander (Colombia), is nearly 28 %. Economic conditions make it difficult to perform the hemoglobin test in all children with suspected anemia. The palmar pallor sign has been recommended by the World Health Organization (WHO) and the Pan-American Health Organization for the diagnosis of anemia in children aged less than 5 years old; however, the yield of the test has not been established in Colombian communities. The aims of the present study were to establish the efficiency of the test and to determine its reproducibility. The palmar pallor sign was evaluated in a sample of 167 children attending a low complexity hospital located in a deprived area of Bucaramanga city. Two pediatricians determined the presence of the sign, according to criteria from the "Integrated Management of Childhood Illness" strategy of the WHO and the Pan-American Health Organization. A hospital laboratory determined the concentration of hemoglobin in each of the children. Based on these data, the sensitivity, specificity, positive predictive value and negative predictive value were determined and the agreement between observers was established. The prevalence of anemia in the sample evaluated was 28.74 %. Sensitivity was 33.33 %, specificity was 79.83 %, the positive predictive value was 40 % and the negative predictive value was 74.8 %. The reproducibility of the test was low. The yield of the test substantially improved when the hemoglobin cut-off point was reduced from 11 g/dl, which is the value proposed by the WHO, to 10.5 g/dl. Anemia is frequent in children aged less than 5 years old in Bucaramanga city. The Palmar pallor sign is a valuable tool for the diagnosis of severe anemia, but is much less useful in mild or moderate anemia. If the hemoglobin level proposed by the WHO were lower, the palmar pallor sign would be more useful.