Paradox lost: Explaining the hispanic adult mortality advantage

Abstract

We tested three competing hypotheses regarding the adult "Hispanic mortality paradox": data artifact, migration, and cultural or social buffering effects. On the basis of a series of parametric hazard models estimated on nine years of mortality follow-up data, our results suggest that the "Hispanic" mortality advantage is a feature found only among foreign-born Mexicans and foreign-born Hispanics other than Cubans or Puerto Ricans. Our analysis suggests that the foreign-born Mexican advantage can be attributed to return migration, or the "salmon-bias" effect. However, we were unable to account for the mortality advantage observed among other foreign-born Hispanics.

The Impact of Salmon Bias on the Hispanic Mortality Advantage: New Evidence from Social Security Data

“…Their deduction was based on the finding that the mortality advantage of foreign-born Mexicans relative to NH whites was greater at older ages and the slope of the mortality curve for foreign-born Mexicans was consistent with selective return migration. In addition, the authors found the self-reported health status of elderly Mexicans who had returned from the United States to Mexico to be poorer than the health status of elderly Mexican-born individuals residing in the United States; a finding that is consistent with the salmon bias hypothesis (Palloni & Arias 2004). The potency of this finding, which is based on two cross sectional surveys, the Mexican Health and Aging Study (MHAS) in Mexico and the National Health Interview Survey (NHIS) in the United States, of course, depends on the comparability of the two data sources.…”

“…These findings are puzzling given the lower socioeconomic status of Hispanics, their lesser access to health care, and their lower self-reported health status (Hummer et al 2004). Several explanations have been proposed for this paradox: cultural factors that influence health behaviors, family dynamics, and social support (Palloni & Arias 2004;LeClere, Rogers & Peters 1997); data problems such as a lack of comparability in reporting of Hispanic origin in vital statistics and census records, age misreporting, and difficulties in linking persons of Hispanic origin among various data sources (Elo et al 2004;Rosenberg et al 1999;Elo & Preston 1997); and healthy in-migrant and unhealthy out-migrant selection effects (Jasso et al 2004;Palloni & Arias 2004;Franzini, Ribble & Keddie 2001;Abraido-Lanza et al 1999;Markides et al 1997). Nevertheless, uncertainty remains about the relative importance of the various explanations.…”

“…This type of selective emigration results in mortality for the US Hispanic population that is lower than if the mortality experience of the return migrants had not been omitted. Because the salmon bias hypothesis proposes that deteriorating health triggers return migration, one would expect its effect on mortality to be more pronounced at older ages when health problems become more prevalent (Palloni & Arias 2004;Palloni & Ewbank 2004).…”

“…That is, when survey data are prospectively linked to U.S. death registration records, deaths of individuals who emigrated during the follow-up period will be missed. The resulting downward bias in mortality due to this omission will be greatest for subgroups most likely to emigrate, notably the foreignborn (Palloni & Arias 2004;Abraido-Lanza et al 1999;Elo & Preston 1997). This type of error, in contrast to the salmon bias effect, will occur even when emigration is not selective of unhealthy individuals.…”

“…However, these explanations do not rule out the possibility that the salmon bias is an important factor in the mortality advantage of other Hispanic groups. In a recent study, Palloni and Arias (2004) concluded that the salmon bias hypothesis may well explain the mortality advantage of foreign-born Mexicans, although not that of other foreign-born Hispanics. Their deduction was based on the finding that the mortality advantage of foreign-born Mexicans relative to NH whites was greater at older ages and the slope of the mortality curve for foreign-born Mexicans was consistent with selective return migration.…”