Global Health
Ill-defined mortality in Brazil during the COVID-19 pandemic: population study by region and race/ethnicity Doris Duran* Doris Duran Renzo Calderon-Anyosa Jay S. Kaufman
Ill-defined or garbage codes (GC) are uninformative ICD-10 codes in mortality statistics. Smaller GC proportions indicate better data quality. The Global Burden of Disease (GBD) group classified GC into four classes by their health planning impact and redistribution potential. Class 1 is the most critical, as reassignment to a specific cause is difficult (e.g., sepsis), while class 4 is more specific and can be reallocated to disease groups. During COVID-19, with overwhelmed health systems and high excess mortality in Brazil, GC deaths likely varied by region and race/ethnicity.
Aim: To describe GC mortality rates in Brazil by race/ethnicity before and during the COVID-19 pandemic.
Methods: Mortality data were from vital statistics (SIM Brazil). Population data came from census-based projections. Mortality rates for deaths with a GC code were estimated by region, race/ethnicity, and GBD class from 2018–2022 for ages 25+. Annual percentage changes were calculated by GC class.
Results: There were 7,158,724 deaths during the study. Deaths peaked in 2020–2021. GCs accounted for 35.5% (n=2,514,756) of all deaths. The South had the lowest GC proportion (27%), and the Northeast the highest (39.2%). Blacks had the highest GC proportion (38.8%), Whites the lowest (33.6%). GC mortality in the North was highest among Mixed-Brown individuals; in the South, it was among Whites and Blacks. GC mortality rose during the pandemic for all groups, except in the North region in 2020 (Fig1A). Class 4 GCs made up 39.4% of GC deaths; Class 1, 29.9%. Classes 1–2 increased for all groups until 2021 (Fig1B). Classes 1–3 saw slight improvement in 2022.
Conclusion: GC mortality is high in Brazil, and it increased during COVID-19 particularly for classes 1-2. Northern and Southern regions showed opposing trends for the Mixed-Brown group, Brazil’s second largest race/ethnicity group. GC deaths challenge cause-specific analysis and should be redistributed for accurate assessments by region and race.
