Background Quality of data is essential to best understand the real magnitude and consequences of road traffic injuries (RTI). The objective of this study was to estimate the potential underestimation of RTI mortality at the subnational level in Mexico for a period of 15 years and to identify social and economic variables at the state level potentially associated with the quality of statistical classification of deaths.
Methods We conducted secondary analysis of validated mortality databases for the period 1999–2013. Five categories of relevant “garbage codes” pertinent for RTI were identified and the percentage they represent of the total was estimated. Using multiple imputation models, registries statistically likely to be due to RTI were estimated and the potential underestimation of mortality was quantified. We explore correlation between health resource availability, social and economic variables with the percentage of underestimation of mortality caused by RTI at the state level using the Kendall’s rank correlation test.
Results 1.99% of all deaths were assigned to “R” codes; 2.40% were injuries of undetermined intent; in 22.96% of unintentional injuries the external cause is not codified; and 0.11% of transport injuries did not specify the means of transport. In over 40% of RTI, the specific road user deceased was unknown. The potential underestimation of deaths from RTI during the period was 18.85% at the national level, with significant variations amongst the 32 Mexican states, varying from 5.32% in Queretaro to 51.49% in Baja California. From the data analysed, there was no statistical evidence of any association of the percentage of RT deaths underestimation with variables analysed.
Conclusions Performance in terms of mortality classification is different at the state level, but more analysis is needed to better understand underlying reasons of garbage coding. This information is useful for targeting interventions to improve recording of deaths in Mexico.
- road traffic injuries
- unintentional injuries
- mortality underestimation