Influence of American acculturation on cigarette smoking behaviors among Asian American subpopulations in California

Introduction

As the leading cause of preventable death in the United States, smoking causes more than 438,000 deaths annually (Armour, Woollery, Malarcher, Pechacek, & Husten, 2005). Smoking among Asian Americans remains understudied because, in part, of its overall relatively low prevalence compared with the general U.S. population (13.3% vs. 22.5%) (Centers for Disease Control and Prevention, 2004). While the rate of smoking appears low when examined for Asian Americans as a whole, this approach may obscure the subgroup differences based on the heterogeneity of this population in immigrant history, U.S. acculturation, socioeco-nomic status, and other cultural characteristics, many of which documentably affect health behavior practices ( U.S. Census Bureau, 2003). The limited scientific literature on tobacco use by Asian Americans has indicated varying smoking prevalence for Asian American subgroups (Lew & Tanjasiri, 2003; Ma, Shive, Tan, & Toubbeh, 2002; Shelley et al., 2004). Recent research also indicates that smoking disparities accounted for much of the disparities in cancer death rates among the Asian and Pacific Islander subpopulations in the United States (Leistikow, Chen, & Tsodikov, 2006). Closer examination of smoking prevalence and its predictors in Asian American subgroups may reveal potential modifiable risk factors for tobacco use and help to design tobacco interventions that are culturally appropriate for the subgroups at higher risk.

The impact of acculturation on immigrants’ health has gained increasing attention in public health research (Arends-Toth & van de Vijver, 2006). The effect of acculturation on cigarette smoking has been studied among Hispanic and selected Asian American immigrants in the United States in recent years (Ji et al., 2005; Ma et al., 2004; Perez-Stable et al., 2001; Tang, Shimizu, & Chen, 2005). However, the strength and direction of acculturation effects vary from study to study because of differences in measurement of acculturation, study populations, and study methods. A review of recent studies among Hispanic Americans found that 7 out of 11 selected studies used acculturation scales and the rest used language spoken and/or country of birth to measure acculturation status (Bethel & Schenker, 2005). A positive association between acculturation and smoking among women was found in nine of these studies and a negative association was found in one study. Among seven acculturation and smoking studies focusing on specific Asian American subgroups, acculturation was measured using diverse instruments. One study used an acculturation scale (Moeschberger et al., 1997). The remaining studies used either one or a combination of the following measures: English language skills/preference, preference for traditional food, length of stay in the United States, and country of birth (Juon, Kim, Han, Ryu, & Han, 2003; Ma et al., 2004; Maxwell, Bernaards, & McCarthy, 2005; Rahman et al., 2005; Shelley et al., 2004; Tang et al., 2005). Despite the diversity of acculturation measures, most of these studies found a negative association between acculturation and smoking among men. A trend towards a positive association among women was found by several studies but the results did not reach statistical significance (Ma et al., 2004; Maxwell et al., 2005). The majority of these studies examined only Southeast Asian, Chinese, Filipino, or Korean American women and men. Little information is available for other large ethnic subgroups.

To better inform our study of the effect of acculturation on smoking, we elaborate below the concept of acculturation and discuss limitations of commonly used acculturation measures. The acculturation concept originated in anthropology and has evolved in other disciplines such as sociology and psychology. In epidemiology, acculturation is usually described as the process by which each individual’s norms, values, attitudes, beliefs and behaviors change during his/her continuously direct contact with different cultures (Ma et al., 2004). Measures of acculturation tend to be a combination of measures from other disciplines (Salant & Lauderdale, 2003). Acculturation conditions refer to social contextual and individual factors that can affect the acculturation process. These factors include characteristics of the society of origin (home culture) and the receiving society (host culture), type of immigration, length of stay, generation status, and socioeconomic status (Arends-Toth & van de Vijver, 2006). These acculturation conditions not only can influence the pace and orientations of the acculturation process but also affect acculturation-related outcomes. Acculturation orientations can be categorized into one of three classifications: unidimensional model, bidimensional model, and fusion model. The unidimensional model assumes that the acculturation process is a shift from maintenance of the home culture to full adaptation to the host culture (Gordon, 1964). Most smoking related research has adopted this unidimensional model. The bidimensional model treats cultural maintenance and adaptation as two independent dimensions and suggests four acculturation orientations: separation (high maintenance and low adaptation), integration (high maintenance and high adaptation), marginalization (low maintenance and low adaptation), and assimilation (low maintenance and high adaptation) (Berry, 1997). Each orientation may have different effects on individuals’ health. The fusion model assumes an individual mixes both home and host cultures in a new “ integrated culture.” The validity of this model is not certain. Different Asian American ethnic subgroups have different acculturation conditions such as differences in exposure to English and to Western influences prior to immigrating to the United States. Therefore, it is likely that different Asian American subgroups have different acculturation orientations and outcomes. Examining the ethnic subgroup-specific effect of acculturation on smoking is therefore desirable among Asian Americans.

With reference to acculturation measures, some researchers have tried to capture every aspect of acculturation in one acculturation scale and used a single comprehensive score to summarize individuals’ overall acculturation status. A major criticism of using this method is that while extreme scores may be interpretable, middle scores may not be. Different acculturation conditions and orientations may have different effects on specific study outcomes (e.g., psychological distress, smoking status) and yet generate similar middle scores. A middle score may therefore not be helpful in predicting a specific acculturation outcome (Arends-Toth & van de Vijver, 2006). Other researchers go to another extreme and use a single narrowly defined proxy indicator to measure acculturation status, such as English proficiency. A single measure may not be able to capture the larger picture of the acculturation process and its potential impact on health behaviors and may not be equally valid for different ethnic subgroups. Most Asian Indian immigrants, for example, are already well schooled in English prior to immigrating whereas most Korean immigrants are not but they may be equally unacculturated. Given the limited information about acculturation included in the California Health Interview Surveys, we propose to use a combination of three commonly used single acculturation measures, in an effort to capture the diverse acculturation conditions experienced by different Asian American subgroups.

The current study had three goals: (a) to estimate the independent effects of multiple acculturation indicators on both current smoking and quitting behaviors for the six largest Asian American subgroups separately; (b) to compare gender- and ethnic subgroup-specific current smoking prevalence and quit rates; (c) to examine the independent effects of other potential predictors of smoking behavior for both men and women in these subgroups.

Methods

Results

Discussion

Consistent with previous studies (Koch-Weser, Grigg-Saito, & Liang, 2004; Lew et al., 2001; Ma et al., 2002; Shelley et al., 2004; Yu, Chen, Kim, & Abdulrahim, 2002), our findings indicate gender and ethnic differences in cigarette smoking patterns among the six largest Asian American subgroups in California. Men’s current smoking prevalence was uniformly higher than women’s, especially in the first generation. Korean, Vietnamese, and Filipino American men had higher current smoking prevalence than other ethnic subgroups of men. Japanese, Koran, and Filipino American women had higher smoking prevalence than other ethnic subgroups of women. Additionally, the quit rates of men of Korean, Vietnamese, Filipino, and South Asian descent and of women of South Asian and Korean descent were lower than quit rates in the U.S. general population (Centers for Disease Control and Prevention, 2004). These results are noteworthy in light of the many excess cancer deaths ascribable to tobacco use in the subgroups shown to have higher smoking prevalence (Leistikow et al., 2006). Reducing the smoking prevalence among these Asian American subpopulations should therefore be a public health priority in California.

Our findings also show that the effect of acculturation indicators on smoking behaviors varied across gender and ethnic groups. Among men, English language preference was associated with a lower smoking prevalence and a higher quit rate in most ethnic subgroups except Filipino and South Asian Americans. Higher generational status was associated with a lower current smoking prevalence among men in all ethnic subgroups. Among women, by contrast, English language preference was associated with a higher smoking prevalence among most ethnic subgroups except Japanese Americans but had little effect on quitting. Preferred language spoken at home and generational status were found to be stronger predictors of smoking status than years lived in the United States. Despite a few exceptions, these findings corroborate and extend existing literature on the effects of acculturation on Asian American tobacco use. For example, researchers recently found that English language proficiency was associated with lower smoking prevalence among Chinese men and Asian American men as a whole (Fu, Ma, Tu, Siu, & Metlay, 2003; Tang et al., 2005) and with higher smoking prevalence among Asian American women (Tang et al., 2005). Juon and colleagues found that living in the United States for more than 20 years was associated with lower smoking prevalence among Korean American men (Juon et al., 2003), consistent with our findings. Other researchers, however, found a negative effect of acculturation on quit rates among Korean American men and women, contrary to our findings (Ji et al., 2005). Different measurements of acculturation status used between studies may explain this discrepancy.

We reviewed possible explanations for our findings. First, there are significant differences in the smoking-related social norms between Asian countries and the United States. The sex role-specific smoking-related norms of their countries of origin may continue to influence Asian immigrants’ smoking behaviors in the United States. For example, in most Asian countries smoking by men, but not women, is socially acceptable. Male smoking prevalence in these countries is therefore high; female smoking prevalence is very low except among Japanese women (Khang & Cho, 2006; Sato, Araki, & Yokoyama, 2000; Yang et al., 1999). By contrast, even though smoking is less tolerated overall in California, women’s smoking is more socially tolerated in the state than in most Asian countries. In California, the anti-smoking environment may not be strong enough to completely offset the greater tolerance of women’s smoking that Asian American women find in the United States. This results in a convergence of men’s and women’s smoking rates over time with increasing acculturation. In addition, smoking prevalence was relatively high among women in Japan than in other Asian countries. The smoking prevalence among Japanese women ranged from 22% (20–29 years) to 14.1% (40–49 years) in 1995, suggesting greater social acceptance of women’s smoking in Japan than in other Asian countries (Sato et al., 2000). As a result, more acculturated Japanese American women in California smoke less than their less acculturated peers. This pattern is contrary to that observed for women from other Asian subgroups but consistent with the suppressive effect that California’s anti-smoking norms appear to have on immigrants whose country-of-origin peers smoke at high rates.

Second, different acculturation conditions including type of immigration and unique cultural characteristics of their home countries may contribute to the inconsistent impact of different acculturation indicators on smoking behaviors in specific ethnic subgroups. The finding of no decreased smoking with increasing English language preference in Filipino and South Asian American men is consistent with the view that each Asian subgroup’s immigrant history has to be examined separately for insight into that group’s risk of smoking in the United States. Unlike other ethnic subgroups, Filipino and South Asian American immigrants come from countries with long colonial histories influenced by western countries. Most Filipinos and South Asians had been already deeply exposed to western culture, religion, and language even before they came to the United States. Thus American acculturation, as reflected exclusively by English language skills, may be less relevant as a proxy for acculturation in these subgroups, even though it is highly useful as a marker for acculturation in other Asian subgroups.

Third, the acculturation indicator, “years living in the U.S.,” has a weaker and less consistent effect on smoking status than English language preference and generational status. One common hypothesis is that the longer an immigrant stays in a host culture, the more likely the person will adopt the host culture’s social norms and language. Our data showed an insignificant correlation between years living in the United States and English language preference. This weaker effect of duration in the United States on smoking may reflect the experience that many first generation immigrants live in the United States for many years but live in linguistically isolated enclaves such as Koreatown in Los Angeles. In such enclaves, they can do well economically without learning to speak English and learning U.S. social conventions. The tendency for first generation immigrants from some Asian subgroups to settle in enclaves of Asian communities may thereby blunt the effect of duration in the United States on changing their social norms and smoking behaviors.

The results reported here are subject to important limitations. First, some Asian groups, notably South Asian American women, use smokeless tobacco primarily rather than cigarettes. However, the CHIS questionnaire excluded questions about smokeless tobacco use. Thus, the relatively low rates of smoking among South Asian women could be an underestimate of their total tobacco use. Second, our data were cross-sectional. A true test of the causal effects of acculturation on smoking would require prospective data. Our study was designed to reveal potential causal associations between acculturation and smoking but to leave the more rigorous hypothesis-testing of causal direction to future longitudinal research efforts. Third, the use of existing data constrained the investigators’ measures of acculturation. The available measures did not do justice to the multidimensional nature of acculturation.

Our findings have important implications for implementing tobacco control programs as well as helping policy makers determine which Asian American subgroups might be at greater risk for smoking related illnesses. First, future tobacco-control programs in California need to focus on higher-risk Asian American subgroups such as Japanese women and Filipino men (regardless of acculturation), less acculturated Korean and Vietnamese men and more acculturated Chinese, Korean, and Filipino women. Second, anti-smoking social norms should be strengthened in all ethnic subgroups. This strategy may carry even greater significance for those ethnic subgroup immigrants who live in relatively segregated ethnic communities that lack anti-smoking social norms. Third, culturally competent strategies are called for, especially for less acculturated Asian American men and women. Multilingual, culturally tailored anti-smoking health education materials and quitting services are needed. Additionally, further exploration of specific determinants of quitting in ethnic subgroups could help to better target interventions so that sustained abstinence could be achieved. Fourth, future studies should explore the use of more comprehensive measures of acculturation in order to capture different acculturation dimensions and conditions and their effects on health-related behaviors.