Large cities concentrate a substantial part of the educated, highly qualified, and economically active populations. Such social “selection” with the peculiarities of lifestyle determines the distinctive characteristics of the level and structure of mortality. Even though data on deaths by causes of death for the large cities are available in Ukraine, very few studies have analyzed cause-specific mortality in these cities. The objective of the study is to make a comparative analysis of mortality from the most influential causes of death in large cities. The novelty lies in the comparative analysis done for Dnipro, Kyiv, Lviv, Odesa, and Kharkiv for the first time. The study uses the direct method of standardization to calculate standardized death rates by sex in 2005-2019. The results indicate lower all-cause mortality rates for the large city residents compared to the corresponding average country-level indicators. Kyiv, Lviv, and Odesa have lower death rates compared to Dnipro and Kharkiv. In Kyiv and Lviv, this is attributed to lower mortality from almost all major causes of death, while in Odesa this mainly resulted from the extremely low ischemic heart disease mortality. Relatively high mortality from circulatory diseases is observed in Kharkiv and Dnipro. However, in Dnipro, this is associated with a high death rate from coronary heart disease and a very low contribution of cerebrovascular disease, whereas in Kharkiv coronary and cerebrovascular disease death rates are quite high. Mortality rates from diseases of the digestive system in the large cities are found to be the closest to the average in Ukraine (except for Lviv). The neoplasms are the only large group of diseases with a mortality rate that exceeds the average level in Ukraine, in particular for women. Overall, the death rates from most of the causes of death in the large cities demonstrated a positive trend in 2005-2019, with some exceptions. External causes and infectious diseases showed the most decrease while mortality from AIDS and ill-defined causes increased. Also, there were uncertain dynamics of deaths due to suicide and injuries with undetermined intent. Given some specific mortality differences between the cities, some concerns have been raised over the accuracy of the coding of diagnoses. In particular, unusually low mortality from ischemic heart disease was found in Odesa and from cerebrovascular disease in Dnipro, very rare deaths from alcoholic liver disease in Odesa, accidental alcohol poisoning in Kyiv, and a group of other liver diseases in Dnipro. We also assume misclassification of suicides as injuries with undetermined intent in Kharkiv. Our findings highlight the importance of the implementation of automated coding and selection of causes of death that can minimize the number of subjective decisions made by coders and lead to significant improvements in the quality of data.
How much do the physician review and InterVA model agree in determining causes of death? a comparative analysis of deaths in rural Ethiopia
