Measuring Ethnic Mortality in Colombia: The Under-registration Gap

Background Under-registration of vital statistics is common in low- and middle-income countries, espe- cially among under-privileged groups. We examine the level of completeness of death counts in the Colombian vital registration records relative to those in the Census 2018. Our aim is to estimate the bias caused by under-registration of deaths, and how this bias varies across age and ethnicity. Methods We use data from the 2018 National Population Census and 2017 Vital Registration System from Colombia. Under-registered deaths by ethnic groups were estimated using two methods. First, we compared the number of deaths in the census with those in the mortality records. Second, we fitted a logistic regression model to estimate the relationship between mortal- ity under-registration and ethnicity, using a census variable indicating whether a death in the household was registered or not. Results A comparison of census and vital registration data indicates that the majority of deaths to both Indigenous and Afro-descendent households were not registered. Indigenous under- registration was considerably higher, with 77.8% of deaths reported in the census not captured in the vital registration system. The probability of under-registration remains high among the Indigenous and Afro-descendent population, even after adjusting for household characteristics. Conclusions Vital registration data in Colombia severely underestimate the deaths to Indigenous and Afro- descendants, reinforced by a clear geospatial patterning which correlates with the distribution of ethnic populations. Contribution This study is the first to quantify the magnitude of under-registration of deaths among ethnic groups in Colombia using census micro-data and vital registration data.