[Colombian death certificate quality, 2002-2006]

Mortality study is useful for the estimation of cancer incident cases; but a quality assessment of the mortality information is required previous to the process of cancer incidence estimation. The aim of this study is to evaluate the quality of death-certification in Colombia.

Overall quality of death certification and quality of cancer mortality certification was analyzed for the period 2002-2006. Classic validity indexes were described through simple percentages for each district and four cities, where operating population-based cancer registries: Bucaramanga, Cali, Manizales, and Pasto. A principal component analysis was performed in order to identify relationships that might not be evident in the descriptive analysis.

952 104 registries were analyzed, 150 163 corresponding to cancer deaths. A high quality index was observed for overall death certification, with 92.8 % of the overall mortality properly certified. Most errors were due to ill defined causes of death. A high quality index was also observed for cancer death registration with 91.5 % of the cancer deaths properly certified. Ill-defined, secondary and unspecified cancer sites certification prevailed as the most frequent causes of error.

Colombia has high quality standards for overall death and cancer death certification. Mortality data can be used confidently in the cancer incidence estimation process, without requirements for any specific corrections.