Text S1: Mortality and causes of death.

Differential underreporting of childhood and adult deaths from across regions or countries may lead to deceptive geographic patterns of mortality rates calculated from reported deaths. In addition, both the quality of medical information available at the time of a death and cause-of-death recording practices vary among and within countries. In order to address variability in cause-of-death coding practices and gaps in reporting of deaths in Mexico, vital registration data were corrected.

All-cause mortality rates

The total number of deaths must be corrected for factors such as underreporting of deaths, misreporting of age at time of death, and migration. In addition, mortality rates calculated using census data would be inaccurate if the census were incomplete. To address these potential sources of bias, we adjusted total mortality by age, sex, and state, using standard methods as described in the main manuscript.

Disease-specific deaths

In Mexico, deaths are medically certified and the cause of death is recorded on the death certificate. Based on the death certificate, deaths are assigned International Classification of Disease (10th revision, ICD-10) codes by two government agencies, the Instituto Nacional de Estadística, Geografia e Informática (INEGI), who compile official death statistics, and the Secretaría de Salud, who manage a database called Sistema Estadístico de Defunciones (SEED). SEED data include multiple causes of death, while INEGI data are more complete for injuries because they include police data. We used 2004 SEED data for maternal and communicable (Group I) and noncommunicable (Group II) diseases in order to take into account additional information provided by multiple cause of death codes. For injuries, we used 2004 INEGI data in order to avoid undercounting total injuries.

We made two types of corrections to improve the comparability of cause-of-death data. First, regional variations in selecting an underlying cause of death when multiple specific causes contributed to the death can confound geographical mortality patterns. For example, whether and where on the death certificate diabetes is listed as a cause of death for cardiovascular deaths among diabetics varies based on community and individual characteristics in Mexico; a similar problem exists for ischemic heart disease and other cardiovascular causes of death. We used regression analyses to quantify these biases and calculate standardized rates of death based on community and individual characteristics for which we expect the most valid cause-of-death assignment (Table S2). Second, for some deaths, insufficient medical information about the underlying cause of death is recorded on the death certificate, and the death is assigned an ICD code for one of several ill-defined conditions. Following Global Burden of Disease (GBD) practice, we reassigned these deaths to specific causes of death based on the age of the decedent and the ill-defined ICD code used (Table S2). Once the cause-of-death corrections were performed, deaths were assigned to a condensed list of 136 GBD causes [1].

Disease-specific death rates

We estimated disease-specific death rates by applying the adjusted distribution of the causes of death to the adjusted total mortality figures (Table S3).

References:

1. Mathers CD, Lopez AD, Murray CJL (2006) The burden of disease and mortality by condition: data, methods, and results for 2001. In: Lopez AD, Mathers CD, Ezzati M, Jamison DT, Murray CJL, editors. Global burden of disease and risk factors. Washington, D.C.: Oxford University Press and The World Bank. pp. 45-240.