[Design and comparative analysis of an inventory of avoidable mortality indicators specific to health conditions in Colombia]

To develop a list of indicators of avoidable mortality (LIAM) in order to analyze failed efforts to control the mortality risks prevalent in Colombia, and to compare its results to those of two widely-used approaches.The official mortality records of Colombia for 1985-2001 were reviewed; the basic causes of death were classified according to the ICD-9. Indicators of avoidable mortality (AM) were selected using an algorithm that combined the lists of Holland and Taucher, the definition of Rutstein and colleagues, and the principle of Uemura. The proportions of avoidable deaths that resulted from applying the LIAM and the two AM lists were compared to a database containing Colombia's official death records from 1993-1996.Of the 680 617 deaths registered during the study period, 18.2% were classified as avoidable according to Holland's list, and 51.3%, by Taucher's list. According to the LIAM, avoidable mortality rose to 76.7%. This pattern remained relatively stable in 1993-1996. The differences observed between the proportions of avoidable deaths according to the LIAM and that of the two other lists were related to the local epidemiological profile and the conceptual approach of each list.The differences between the LIAM and the AM lists of Taucher and of Holland attest to the consequences of using one or another classification in the Colombian context. The LIAM can be a valuable resource for undergirding and evaluating health policies, but must be adjusted to the specific situation in which it is applied.