JMIR Publications

ABSTRACT

Background:

Drug-related mortality is monitored using operational definitions that vary across surveillance systems and settings, usually based on ICD-10 codes. These definitions vary in scope and inclusiveness, which may affect the magnitude of estimates and their comparability across countries.

Objective:

This study evaluates how five ICD-10-based surveillance definitions used by national and international monitoring agencies affect estimates and temporal trends in drug-related mortality in Spain.

Methods:

We conducted a repeated cross-sectional analysis of all deaths among Spanish residents recorded in the national mortality register of Spain (INE) during 2001–2022; drug-related deaths were identified separately under each operational definition. Five operational definitions that combine specific subsets of ICD-10 codes were applied: Spain’s National Plan on Drugs (PND), the European Union Drugs Agency (EUDA), the US Centers for Disease Control and Prevention (CDC), the UK Office for National Statistics (ONS), and an Australian/Penington Institute framework (AUS). Annual counts, total deaths, annual means, average annual change (AAC), and Spearman correlations were computed overall and stratified by gender and age (<35, 35–64, ≥65).

Results:

All five definitions showed broadly concordant temporal trends, with a marked increase after 2016. Differences between definitions affected primarily the absolute number of deaths, with PND yielding the lowest counts and AUS yielding the highest. In the general population, total deaths ranged from 14,298 under PND to 28,217 under AUS, and Spearman correlations with PND ranged from 0.84 to 0.95. Across strata, correlations were generally high, although lower values were observed for CDC in some groups. By gender, trends were coherent across definitions, with moderate increases in men and faster growth in women. By age, deaths declined among people aged <35 years and increased among those aged 35–64 and ≥65 years, with the largest absolute increases in the 35–64 group.

Conclusions:

Applying different operational definitions to the same Spanish mortality dataset substantially changed the estimated burden of drug-related mortality, nearly doubling counts under the broadest definition. However, temporal patterns were broadly consistent across definitions. Transparent and harmonized ICD-10 code sets are needed to improve international comparability.