I found really interesting how some results we may find in analyses can be actually explained by migration, which is something I am not currently used to be looking for.
Hummer (2007) mentions that since it has been suggested that good outcomes in Mexican immigrant populations could be explained by the out-migration of this population, and therefore less reporting of their outcomes, a way of getting around this issue could be estimating infant mortality of Mexican immigrant populations in the US. Interestingly, they found that even though this group has lower health insurance coverage (among other characteristics) they had lower infant mortality compared to white non-hispanic population. The three mechanisms they consider that can be causing this include: immigration selectivity, cultural factors that encourage healthy behaviors, and data quality issues.
They ran simulations and they found that it would be required that 300,000 out-migrating women and infants between 1995-2000, with all of the infant deaths happening in Mexico and never being recorded in the US which is highly unlikely. They mention that they observe something similar in population from Cuba/Puerto Rico, which supports the explanation that cultural factors can encourage healthy behaviors among Hispanics in the US. I am curious if this same paradox would be observed in other Hispanic populations (El Salvador, Venezuela, Guatemala).
I also am not very familiar with how medical attention in the US works, but would it be possible that Mexican immigrant populations that receive medical attention during birth, could somehow be identified if they are illegally in the US, and being deported? Or is this does not happen?
Ro (2014)
As also mentioned by Hummer (2007) one potential explanation for some of these results can be in immigration selection. Ro (2014) explores how the changes in health profiles of origin countries also impact health trends in immigrating populations in the US (therefore making it more appropriate to compare them to populations from their origin countries likely to migrate and not the population in the US). Their results suggest that health selection increased as obesity increased in Mexico since obesity prevalence increased in Mexico in the period analyzed and the difference between this population and recent migrants to the US increased. I am still thinking about which would be the proper comparison group in Mexico given how the likelihood of migration was estimated, its association with socioeconomic status and obesity, and immigration selection.