The Latino Paradox - does it really exist?

Table of contents

1. Definition, history and extension of the Latino Paradox

2. Current body of evidence

3. Approaches to explain the phenomenon
3.1 Health behavior and acculturation
3.2 Social context
3.3 Genetic and biological approaches
3.4 Data errors: miscount of Hispanic deaths and Hispanic population
3.5 Return migration (“Salmon Bias”)
3.6 Selection processes (“Healthy migrant effect”)
3.7 Other explanations

4. Discussion and conclusion

5. References

1. Definition, history and extension of the Latino Paradox

The Latino Paradox is a phenomenon commonly referring to the finding that Americans of Hispanic origin do surprisingly well in respect of various health indicators, despite their unfavorable socioeconomic profile consisting of low income, low levels of education and less favored living environments. Such kinds of findings were first reported in the 1950s and 1960s and focused on mental health among Mexican Americans (Jaco, 1973 cited in Markides & Coreil, 1986; Karno & Edgerton, 1969 cited in Markides & Coreil, 1986). Also in the 1960s, first studies revealed surprisingly favorable infant mortality rates for Americans with Spanish surnames (Teller & Clyburn, 1974 cited in Palloni & Morenoff, 2001). Generally, Hispanics seemed to be clearly better off than non-Hispanic blacks, although the two groups share a lot of similar socioeconomic features.

In the following decades, the Latino Paradox was intensely studied and transferred from mental health and infant mortality to a whole range of different health indicators such as low birth weight (Buekens, Notzon, Kotelchuck, & Wilcox, 2000; Chung, Boscardin, Garite, Lagrew, & Porto, 2003; Collins & Shay, 1994; Fuentes-Afflick, Hessol & Perez-Stable, 1999; Gaffney, 2000; Rosenberg, Raggio, & Chiasson, 2005), pre-term births and other birth complications (Brown, Chireau, Jallah, & Howard, 2007; Buekens, Notzon, Kotelchuck et al., 2000; Fuller, Bridges, Bein, Jang, Jung, Rabe-Hesketh, Halfon, & Kuo, 2009), all-cause and cause-specific adult mortality (Abrafdo-Lanza, Dohrenwend, Ng-Mak, & Turner, 1999; Hunt, Resendez, Williams, Haffner, Stern, & Hazuda, 2003; Hunt, Williams, Resendez, Hazuda, Haffner, & Stern, 2002; Markides & Eschbach, 2005; Patel, Eschbach, Ray, & Markides, 2004; Smith & Bradshaw, 2006; Turra & Goldman, 2007), success of transplant surgery (Gordon & Caicedo, 2008), infant cognitive development (Fuller, Bridges, Bein, et al., 2009), and many more. But, of course, there was still ongoing research on infant mortality and mental health as well (Alegria, Canino, Shrout, Woo, Duan, Vila, Torres, Chen, & Meng, 2008; Farley, Galves, Dickinson, & Diaz Perez, 2005; Hummer, Powers, Pullum, Gossman & Frisbie, 2007).

Something that changed almost constantly from one study to the next, was the definition of “Latino” or “Hispanic” as well as the definition of the respective comparison group.

In few studies, “Hispanic” referred only to foreign-born immigrants from Mexico, and / or other Spanish-speaking Central or South American countries (e.g. Gaffney, 2000). In most studies, also Mexican Americans and / or other Hispanics born in the United states were included (e.g. Farley, Galves, Dickinson, et al., 2005; Fuentes-Afflick, Hessol & Perez-Stable,1999; Hummer, Powers, Pullum, et al, 2004; Turra & Goldman, 2007). Some studies only focused on Hispanics of Mexican origin (Buekens, Notzon, Kotelchuck, et al., 2000; Hunt, Resendez, Williams et al. 2003; Burgos, Schetzina, Dixon, & Mendoza, 2005; Xiong, Buekens, Vastardis, et. al 2006), whereas other studies analyzed differences between Mexicans and other Hispanic subgroups such as Puerto Ricans or Cubans (e.g. Abrafdo-Lanza, Dohrenwend, Ng-Mak, et al. 1999; Fuentes-Afflick, Hessol & Perez-Stable,1999; Turra & Goldman, 2007). A considerable number of studies just did not try at all to differentiate between Hispanics of different national (and supposedly cultural and racial) origin (e.g. Brown, Chireau, Jallah, et al., 2007; Chung, Boscardin, Garite et al., 2003; Smith & Bradshaw, 2006; Gordon & Caicedo, 2009). As to the choice of the reference group, one can also observe an obvious heterogeneity. In some studies, non-Hispanic whites alone formed the comparison group (e.g. Abrafdo-Lanza, Dohrenwend, Ng-Mak, et al., 1999; Fuentes-Afflick, Hessol & Perez-Stable,1999), other studies referred to both non- Hispanic whites and non-Hispanic blacks (Crimmins, Kim, Alley, Karlamangia, & Seeman, 2007; Burgos, Schetzina, Dixon, et. al, 2005), or additionally even non- Hispanic Asians (Fuller, Bridges, Bein et al., 2009; Gordon & Caicedo, 2009). Finally, some studies only made comparisons between different Hispanic subgroups without including a non-Hispanic reference group (Gaffney, 2000; Rosenberg, Raggio, & Chiasson, 2005).

All in all, one must conclude that research in the context of the Latino paradox has been quite diverse and quite heterogeneous, which renders it difficult to draw firm conclusions. Nevertheless, the next section aims at summing up the existing body of evidence concerning the Latino Paradox.

2. Current body of evidence

There are basically two big currents within the investigations around the Hispanic Paradox. On the one hand, there is the bulk of studies which comes to the conclusion that there really is a kind of paradox (e.g. Abrafdo-Lanza, Dohrenwend, Ng-Mak, et al., 1999; Brown, Chireau, Jallah, et al., 2007; Fuentes- Afflick, Hessol & Perez-Stable,1999; Hummer, Powers, Pullum, et al., 2007; Turra & Goldman, 2007). All these studies describe a health advantage on the part of Hispanic immigrants and / or American citizens with Hispanic roots that would not have been expected taking into account the unfavorable socioeconomic living conditions these people face. In general, Hispanics show health outcomes equivalent or superior to those of non-Hispanic whites and are clearly better off than the non-Hispanic black comparison group. Depending on the health measure under study, these differences are more or less stable across studies. There is a general trend that Mexicans seem to have the greatest health advantage of all Hispanic subgroups. Cubans or Puerto Ricans, e.g., seem to be less favored (Alegria, Canino, Shrout, et al., 2008; Collins & Shay, 1994; Fuentes-Afflick, Hessol & Perez-Stable,1999; Markides & Eschbach, 2005; Turra & Goldman, 2007). Moreover, foreign-born Hispanics seem to fare better than Hispanics born in the U.S., and Hispanics new to the United States do better than Hispanic having already lived there for several years (Alegria, Canino, Shrout, et al., 2008; Collins & Shay, 1994; Hummer, Powers, Pullum, et al., 2004; Turra & Goldman, 2007). With respect to adult mortality, especially elderly persons and especially men seem to profit from their Hispanic origin (Markides & Eschbach, 2005; Patel, Eschbach, Rey, et al., 2004; Turra & Goldman, 2007). The health advantage of Hispanics is commonly explained by factors such as healthy behaviors inherent to the “Hispanic culture”, social support / social networks, and biological differences (Morales, Lara, Kington et al., 2002; Scribner, 1996; Turra & Goldman, 2007). Despite this multitude of studies supporting the existence of a real health paradox, there are also critical voices doubting the convincing power of the research undertaken up to now (Palloni & Morenoff, 2001; Smith & Bradshaw, 2006). These authors - and also authors in favor of the existence of a Hispanic paradox - explain the seemingly paradoxical findings either completely or in part by means of possible data errors, flawed study designs, and hitherto undetected and untested selection processes (Smith & Bradshaw, 2006; Palloni & Morenoff, 2001; Patel, Eschbach, Rey, et al., 2004; Turra & Goldman, 2007). The number of studies denying the existence of a health paradox is clearly smaller, but generally, these studies dispose of more sophisticated and thought-out study designs.

To sum up, there is some empirical evidence that a Latino Paradox might exist. On the other hand, several authors have competing explanations for what might seem to be paradoxical. In the next section, these competing chains of explanation are presented in detail.

3. Approaches to explain the phenomenon

There are two big categories of theories regarding the Hispanic Paradox. The ones that assume that there really is an unexpected health advantage which Hispanics have especially in comparison to non-Hispanic blacks. And the ones that do not necessarily deny such a health advantage, but try to explain the findings at least in part by specifying possible distortions to the data. First, the theories trying to explain the existence of a real health paradox are presented (3.1 - 3.3). Then, alternative theories explaining the data under the assumption that there is not necessarily a health paradox are set out (3.4 - 3.5).

3.1 Health behavior and acculturation

This explanation assumes that the health advantage Hispanics seem to have is due to a number of culturally shaped behaviors which protect them from health problems. For instance, Hispanics tend to smoke less, drink less alcohol and on average have a healthier diet containing less fat and more fibers than non-Hispanic whites or blacks do (Abraído-Lanza, Dohrenwend, Ng-Mak, et al., 1999; Gordon & Caicedo, 2008; Morales, Lara, Kington, et al., 2002; Scribner, 1996; Turra & Goldman, 2007). Another study showed that Hispanics supposedly use advantageous strategies to cope with psychosocial stress (Farley, Galves, Dickinson, et al. 2005). In this study, Hispanics used positive reframing of the situation more often than non-Hispanic whites and tended to avoid unhealthy coping styles like substance abuse. All these differences in behavior could in principle explain a lot of the paradoxical findings, e.g. regarding birth outcomes, mental health, or adult mortality due to chronic diseases of the heart.

[...]