Evaluation of coding ill-defined and unknown causes of death in the Republic of Serbia.

One of the most frequently used indicators for the quality of mortality statistics is the use of R-codes as the cause of death (COD) in death certificates. The aim of this study was to analyse the coding of ill-defined and unknown causes of death in the Republic of Serbia (RS). The data for this descriptive study, covering a ten-year period (2006-2015), were obtained from the National Mortality Register. Since population and economic features differ widely between regions in RS, we compared the Belgrade region (BR) with other regions (Or). We estimated the frequency of certain types of death investigation methods regarding R00-99-coded deaths. The frequency of R-codes, by subcategories, and code-specific mortality rates were calculated by region for each year. The use of R-codes was significantly lower in BR than in Or (OR 0.85; 95% CI, 0.84-0.66; p < 0.001). The most frequent method of determining COD was external examination of the body, both in BR and Or (60.7% and 85.5%, respectively). The rate of forensic autopsies in BR was 38.1% while in Or it was only 3.6%. Clinical autopsies were performed in 1% of deaths in BR, compared to 0.5% in Or. Our results suggest that in BR the use of R-codes compared to other CODs has decreased over the past years while there has been an increasing trend in autopsy proportions; in Or the frequency of R-codes, as well as of autopsy proportions, has remained unchanged.