Abstract
BACKGROUND: In Mexico, low numbers of overdose deaths reported by government agencies, coupled with evidence of increased substance use in certain regions, suggest underreporting of overdose mortality. We explore the potential miscoding and underreporting of overdose deaths. METHODS: We assembled a mortality dataset using a combination of publicly available data and systematic freedom of information requests to Mexican state jurisdictions. Using data from post-mortem toxicology analyses, we identified proxy International Classification of Diseases (ICD)-10 codes for causes of death that were significantly more common (using p < 0.05 as a threshold) in decedents with positive toxicology for opioids, methamphetamine, cocaine, or benzodiazepines compared to decedents with negative toxicology for these substances. We then estimated annual crude overdose mortality rates, for individuals aged 20 to 39 with both standard overdose-specific codes (i.e., X40-44, X60-64, X85, and Y60-64), and derived proxy ICD-10 codes. RESULTS: We identified acute myocardial infarction (I21), acute pulmonary edema (J81), and acute respiratory failure (J96) as proxy ICD-10 codes for potential overdose deaths. Using these codes, the estimated crude overdose mortality rate per 100,000 aged 20 to 39 in Mexico increased from 4.14 in 2005 to 6.92 in 2019, compared to estimates of 0.31 to 0.37 using standard overdose codes for the same period. CONCLUSIONS: Our findings suggest a potential underestimate of overdose deaths due to possible misclassification and underscore the critical need for improved forensic capabilities to more accurately identify overdose deaths. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s12889-025-25249-9.