Objectives: to assess the impact of coding causes of death with the ICD-10 2016 version and the software Iris on Italian official statistics on mortality.
Design: coding of a sample of death certificates with two different coding systems (bridge coding).
Setting and participants: a sample of 63,525 deaths occurred throughout 2015 among people aged over one year, already coded using the ICD-10 2009 version and the Mortality Medical Data System (MMDS) software, was re-coded through the ICD-10 2016 version and the Iris software.
Main outcome measures: the transition matrix between the two coding systems was realized and the agreement percentages between the two coding systems, the comparability ratios, and the relative 95% confidence intervals were calculated. Comparability ratios have been calculated for both the underlying cause of death and the multiple causes.
Results: overall, 79% of deaths showed exactly the same underlying cause of death (ICD-10 code, 4 digits) in the two coding systems. On the three-digit level, the agreement was 89%; on ICD-10 chapter level, the agreement was 95%. At the chapter level, the most important changes were observed for: • certain infectious and parasitic diseases (-18% in ICD-10 2016/Iris); • diseases of the genitourinary system (-17%); • diseases of the respiratory system (+7%); • diseases of the nervous system and sense organs (+5%); • external causes of morbidity and mortality (+5%). Analyzing the multiple causes, the most important changes were observed for: • certain infectious and parasitic diseases (-19% in ICD-10 2016/Iris); • external causes of morbidity and mortality (+28%); • symptoms, signs, and abnormal clinical and laboratory findings, not elsewhere classified (+7%).
Conclusion: the results are very useful to explain any change in the Italian statistics on mortality comparing 2015 with the following years.