Prevalence and Predictors of Tobacco Use Among Asian Americans in the Delaware Valley Region

Abstract

Objectives. This study examined tobacco use rates and potential predictors of use among Asian Americans residing in the Delaware Valley region.

Methods. A cross-sectional survey design was used. The sample consisted of 1174 Chinese, Koreans, Vietnamese, and Cambodians.

Results. Findings indicated that the mean age at initiation of tobacco use was 18.3 years. Among the respondents, 40.2% had a history of tobacco use, and 29.6% were current users. Men were more likely than women to smoke. There were significant differences between never smokers, current smokers, and ex-smokers in sex, ethnicity, educational attainment, and marital and employment status.

Conclusions. The findings suggest that tobacco use is still a serious public health problem among Asian Americans, especially men.

Tobacco use is the single most important preventable cause of illness and death in the United States, accounting for more than 430 000 deaths annually.1 In the states of Pennsylvania and New Jersey in 1999, for example, cigarette smoking accounted for 36 000 deaths, and death rates attributed to smoking were 346/100 000 and 372/100,00, respectively.2 In the same year, New Jersey ranked 14th among the states in number of deaths from smoking, whereas Pennsylvania ranked 21st. Approximately 480 368 years of potential life were lost in the 2 states in 1999, or an average of 13.3 years for each death due to smoking, and the estimated cost of tobacco use in regard to related medical expenses was $4.7 billion.2

In 2000, the US Behavioral Risk Factor Surveillance Survey (BRFSS) revealed that median smoking prevalence rates were 24.4% among men and 21.2% among women.3 State-specific versions of the 2000 survey showed that adult smoking prevalence rates were 24.3% in Pennsylvania (25.4% of men and 23.3% of women) and 21% in New Jersey (23.5% of men and 18.6% of women).3

Asian Americans and Pacific Islanders, the fastest growing US ethnic/racial group, represent approximately 3% of the residents of the greater Philadelphia region of Pennsylvania and New Jersey (the site of the present study).4 By and large, city and state tobacco prevention programs are failing to reach members of the Asian American community, whose smoking behaviors remain largely unchanged. In turn, the community lacks an understanding of the health risks associated with tobacco use.5

Analysis of data on Pennsylvania's 2.3 million smokers reveals that nicotine addiction is associated with early smoking initiation.6 Data on male high school seniors derived from the Monitoring the Future Survey indicates that Asian youth smoking rates increased from a moderately low 16.8% in 1989 to a high of 20.6% in 1994.7,8 A study conducted in an ethnically and racially diverse section of Philadelphia showed that 18.2% of White adolescents, 9.0% of African American adolescents, and 14.1% of Asian American adolescents had smoked daily since they began smoking.9 In general, however, Asian Americans begin smoking later in life than members of other racial/ethnic groups, and this is particularly true for Asian American men.10–12

It has long been asserted that sociodemographic characteristics have a significant impact on tobacco use prevalence patterns. This relationship has been the subject of scrutiny in the general population13–15 but not in the Asian American population.11,16–18 Studies on the latter, substantiated by data from the Centers for Disease Control and Prevention,19–21 show that smoking prevalence rates among Asian Americans vary by sex.

Despite conclusive evidence of a relationship between age and smoking behaviors among Asian Americans, current studies indicate variability in smoking prevalence rates among men. For example, a study of Vietnamese men in Massachusetts conducted by Wiecha et al.11 indicated that the highest smoking rates were among men aged 25 to 44 years.

It has been shown that smoking rates among Asian American men also vary according to ethnic subgroup: 35% to 56% among Vietnamese, 33% among Koreans, 71% among Laotians, 71% among Cambodians, and 55% among Chinese.22,23 A more recent national study found that 34% of Vietnamese American men and 31% of Korean American men were current smokers.24 Collectively, these findings clearly indicate significantly higher smoking rates for Southeast Asian men than among men in other racial or ethnic groups studied to date.7,10,11,16–19,25,26

Research on the relationship between smoking and educational level has led to the conclusion that smoking rates decrease as a function of educational attainment. That is, the highest prevalence rates have been observed among those with a high school education or less, and the lowest rates have been observed among those with a graduate school education.11,17,18

Recent studies have found a high prevalence of certain types of cancer among Asian Americans and Pacific Islanders. According to Chen and Hawks,27 lung cancer, the leading cause of death among Asian Americans and Pacific Islanders, is 18% more prevalent among Southeast Asian men than among White men, and rates are expected to increase as a result of this group's smoking patterns.

Doong28 pointed out that cancer is the leading cause of death among Asian American women, and Chu29 concluded that cancer is highly associated with smoking in both male and female Chinese Americans. Chu noted that Chinese American men and women have high rates of nasopharyngeal cancer relative to other racial/ethnic groups (e.g., 14 and 11 times higher, respectively, than their counterparts in the White population).

In response to the need for tobacco control and prevention and in an attempt to reduce the incidence of and risk for cancer among Asian Americans in the Delaware Valley region of Pennsylvania and New Jersey, the Asian Tobacco Education, Cancer Awareness, and Research (ATECAR) project was established at Temple University in Philadelphia. The goal of ATECAR is to form a sustainable public health infrastructure by building strategic partnerships between health service agencies and Asian community organizations. A comprehensive needs assessment questionnaire was developed as a preliminary step toward initiation of tobacco control, prevention, and intervention programs.

The primary purpose of the study described in this article, conducted in 2000, was to determine tobacco use rates among the Asian American community in the Delaware Valley region. A secondary purpose was to determine the demographic variables that are potential predictors of current tobacco use among members of this ethnically and racially diverse community.

METHODS

We used a cross-sectional survey design because of the advantages provided by this method30–33; also, we adapted a stratified-cluster proportional sampling technique.34 Initially, we identified Asian American community organizations (n = 52) located in the 7 counties of the Delaware Valley region of Pennsylvania and southern New Jersey using information provided by the Asian Community Cancer Coalition, a health service agency, and ATECAR project staff. These community organizations were located throughout the 7 counties thereby maximizing the opportunity to obtain a representative sample of Asian Americans across ethnic subgroups, ages, and socioeconomic status.

From the list of 52 organizations, 26 were randomly selected and divided into clusters. The selected organization clusters were stratified according to 4 race/ethnicity or language groups (Chinese, Korean, Vietnamese, Cambodian). A proportional allocation procedure was used in which sample sizes were assigned proportionally to subgroups.33

Instrument

A 77-item questionnaire was developed for the survey. Some of the questionnaire items were modified from other instruments, including the 2000 National Health Interview Survey, the 1998 National Household Survey on Drug Abuse, the 1999 Youth Risk Behavior Survey, the Florida Youth Tobacco Survey, and the American Indian Cancer Control Project instrument. Average time required to complete the questionnaire was 20 to 25 minutes.

A pilot test was conducted to establish the reliability and validity of the instrument and to verify data collection methods. Appropriateness of questionnaire format, validity of content, level of difficulty, and length of time required to complete the survey were also determined. The face validity and content validity of the questionnaire were tested with 50 Asian American adults who did not participate in the study. The Guttman split-half reliability coefficient was 0.67, indicating that, overall, participants responded consistently to questionnaire items.

Data Collection Procedures

Surveys were administered in both one-to-one and group formats. ATECAR staff, in conjunction with community organization leaders, administered the self-report survey to Asian American participants on site in the community organizations' facilities. Participants had the option of responding in English or in Chinese, Korean, Cambodian, or Vietnamese. Translators were provided when needed for individual or group translations.

Data Analysis

Data were analyzed with SPSS 10.0 Chicago, IL: SPSS, Inc.; 1999. We used descriptive statistics to summarize data collected for demographic variables (described subsequently). The χ² test for independence was used to determine Asian subgroup differences in demographic characteristics and smoking rates. This test was also used to determine significant differences between smoking status and demographic variables. A one-way analysis of variance was used to determine mean differences between the subgroups in age and age at initiation of tobacco use.

Logistic regression analyses were used to identify the potential predictor variables of current tobacco use. Independent variables included sex, age, country of birth, length of time residing in the United States, educational attainment, occupation, marital status, children younger than 18 years living at home, employment status, and immigration status. The dependent variable was current tobacco use, defined as “having used tobacco within the last 30 days.”

RESULTS

Sample Characteristics

The overall sample consisted of 1174 respondents, distributed as follows: Chinese, 34.9%; Korean, 37.1%; Vietnamese, 16.7%; Cambodian, 8.4%; and “other,” 2.7% (Table 1 ▶). This last category consisted of Laotian, Hmong, Filipino, and Japanese Americans. There were more men (55.2%) than women (44.8%; χ²₄ = 15.6, P < .01) in the 4 Asian American subgroups. The mean age was 41.4 years. Approximately 70% of the sample was between 23 and 60 years of age. About 16% of the respondents were younger than 22 years; 14% were older than 60 years. Significant age differences were found between the groups (F_4,1133 = 73.4, P < .001).

TABLE 1.

—Demographic Characteristics of the Sample

	Chinese (n = 410)	Korean (n = 436)	Vietnamese (n = 196)	Cambodian (n = 100)	Other (n = 32)	χ2	df	Total (n = 1174)
Sex, %						15.6*	4
Male	53.8	51.4	67.3	56.6	46.9			55.2
Female	46.2	48.6	32.7	43.4	53.1			44.8
Country of birth, %						2374.4*	16
China	77.2	0.2	0	1.3	6.9			26.8
Korea	0	98.1	2.4	0	0			38.8
Vietnam	8.6	0.7	95.9	16.5	0			20.0
Cambodia	.08	0	0	40.5	0			3.3
Other	13.4	1.0	1.8	41.8	93.1			11.0
Length of time in US, %						148.9**	12
≤ 2 y	24.8	13.9	10.9	8.1	40.6			17.5
3–5 y	13.5	7.4	8.3	4.0	9.4			9.5
6–10 y	16.2	12.7	30.2	14.1	12.5			16.9
>10 y	36.6	64.9	43.8	58.6	28.1			50.0
Entire life	8.8	1.2	6.8	15.2	9.4			6.2
Level of education, %						209.8**	20
High school or less	10.5	5.2	12.8	23.5	12.5			10.1
High school	32.3	35.5	44.4	54.1	31.3			37.3
Trade school	3.5	4.0	3.7	4.1	3.1			3.8
Associate degree	5.3	4.5	15.0	5.1	0			6.4
College	17.0	38.8	19.3	13.3	34.4			25.7
Graduate/professional	31.5	12.0	4.8	0	18.8			16.8
Occupation, %						113.8**	16
Student	19.9	19.0	37.1	62.7	47.6			27.2
Executive, professional	24.5	10.2	0.8	0	9.5			13.1
Businessperson, manager	29.5	37.6	25.0	13.7	28.6			29.9
Manual laborer	11.1	12.2	21.2	11.8	9.5			13.4
Semiskilled laborer	14.9	21.0	15.9	11.8	4.8			16.4
Marital status, %						143.4**	16
Married	70.5	79.9	59.8	33.3	32.3			68.6
Separated	2.0	1.7	1.6	1.3	3.2			1.8
Divorced	1.5	2.0	1.6	1.3	3.2			1.7
Widowed	3.0	5.7	0.5	1.3	3.2			3.5
Never married	22.9	10.8	36.4	62.7	58.1			24.4
Children <18 y at home, %	55.6	45.3	46.6	45.8	21.9	18.7**	4	48.3
Employment status, %						50.1**	8
Full time	51.8	60.6	52.8	36.8	31.3			53.4
Part time	21.1	12.4	18.1	10.5	37.5			17.0
Unemployed	27.1	27.0	29.0	52.6	31.3			29.7
Immigration status, %						80.5**	8
US citizen	49.5	49.9	68.4	42.7	35.5			51.9
Permanent resident	25.3	36.8	28.9	43.8	25.8			31.8
Noncitizen	25.3	13.3	2.6	13.5	38.7			16.3
Age, y, %
14–17	7.5	2.1	5.7	37.1	13.8			7.9
18–22	6.0	3.1	11.5	29.9	24.1			8.3
23–40	37.7	24.6	47.4	24.7	44.8			33.6
41–60	35.7	48.0	27.6	4.1	13.8			35.6
≥ 60	13.2	22.2	7.8	4.1	3.4			14.6
Age, y, mean (SD)a	41.4 (15.8)	48.5 (15.1)	36.9 (13.9)	23.9 (12.7)	28.7 (11.7)			41.4 (16.5)

^aGroup differences were significant (F = 66.6, P < .001).

*P < .01; **P < .001.

Overall, 8.8% of Chinese, 1.2% of Koreans, 6.8% of Vietnamese, and 9.4% of Cambodians were born and had spent their entire lives in the United States. Substantial percentages of Chinese (36.6%), Korean (64.9%), Vietnamese (43.8%), and Cambodian (58.6%) respondents had lived in the United States for more than 10 years. Approximately 52% of respondents reported US citizenship, 31.8% reported permanent residency, and 16.3% reported that they were not US citizens.

In terms of education, 10.1% of respondents had less than a high school education, 37.3% had graduated from high school, and 42.5% had a college education. There were significant differences in educational attainment between the groups (χ²₂₀ = 209.8, P < .001). Chinese (31.5%) exhibited the highest rate of professional or graduate education, followed by Cambodians (18.8%), Koreans (12.0%), and Vietnamese (4.8%).

Approximately 53% of the respondents were employed full-time, 17.0% were employed part-time, and 29.7% were unemployed. We observed significant subgroup differences in employment status as well (χ²₈ = 50.1, P < .001). Koreans (60.6%) were more likely to be employed full-time than were Chinese (51.8%), Vietnamese (52.8%), or Cambodians (36.8%). In regard to type of occupation, 29.9% of respondents reported that they were employed in a business; 29.8% reported that they were manual or semiskilled laborers, 27.2% reported that they were students, and 13.1% reported that they were executives or professionals. Significant differences between the groups were observed in regard to occupation (χ²₁₆ = 113.8, P < .001).

In terms of marital status, 68.6% of respondents were married, 3.5% were divorced or separated, 3.5% were widowed, and 24.4% had never been married. We observed significant subgroup differences in this area as well (χ²₁₆ = 143.4, P < .001). Koreans (79.9%) were more likely to be married than were Chinese (70.5%), Vietnamese (59.8%), or Cambodians (33.3%). In addition, there were significant differences among Chinese (55.6%), Koreans (45.3%), Vietnamese (46.6%), and Cambodians (45.8%) in regard to children younger than 18 years living at home (χ²₄ = 18.7, P < .001).

Smoking Prevalence

Approximately 40% of respondents reported ever having used tobacco, and 30% reported current tobacco use. Rates of lifetime use varied among the subgroups: Chinese, 33%; Koreans, 43%; Vietnamese, 47%; and Cambodians, 50% (χ²₁ = 21.1, P < .001). Current tobacco use rates reflected similar variations: Chinese, 24.1%; Koreans, 26.8%; Vietnamese, 40.3%; and Cambodians, 42.4% (χ²₄ = 27.9, P < .001).

We observed significant differences among ever smokers by sex and ethnicity. A larger proportion of men than of women reported tobacco use (odds ratio [OR] = 4.30; 95% confidence interval [CI] = 3.21, 5.74). Among those who had ever smoked, 81.4% were men, and rates varied according to ethnic group, ranging from 75% to 87%. Among ever smokers, 18.6% were women, ranging from 13% to 25% according to ethnic subgroup.

Similar trends were observed for current smokers. Men were more likely to use tobacco than were women (χ²₁ = 105.1, P < .001). Approximately 78% of current smokers were men, and subgroup rates among men ranged from 71% to 85%. Overall, 21.8% of current smokers were female, ranging from 15% to 29% according to ethnic subgroup.

Also, we observed significant differences between current smokers, never smokers, and ex-smokers in ethnicity, educational attainment, marital status, and employment status (Table 2 ▶). The differences in smoking status between ethnic subgroups were significant (χ²₈ = 47.1, P < .001). Southeast Asians (Cambodians and Vietnamese) were more likely to be current smokers or ex-smokers than were Chinese and Koreans. Never smokers and ex-smokers were more likely to have a higher level of education than current smokers. Never smokers (66.5%) and ex-smokers (87.5%) were more likely to be married than current smokers (65.6%). Furthermore, never smokers were more likely to be female, to never have been married, to be foreign born, and to be unemployed.

TABLE 2.

—Demographic Characteristics, by Smoking Status

	Never smokers (n = 700), %	Current smokers (n = 345), %	Ex-smokers (n = 127), %	χ2	df
Sex				223.5*	2
Male	37.6	78.3	89.8
Female	62.4	21.7	10.2
Country of birth				3.91	2
United States	5.5	7.7	9.7
Other	94.5	92.3	90.3
Length of time in US				12.1	8
≤ 2 y	18.3	17.7	12.8
3–5 y	9.1	10.8	8.0
6–10 y	17.4	18.3	10.4
>10 y	49.8	46.2	60.8
Entire life	5.5	7.0	8.0
Ethnicity				47.1*	8
Chinese	39.2	28.7	28.3
Korean	35.5	33.9	55.1
Vietnamese	14.8	22.9	10.2
Cambodian	7.1	12.5	5.5
Other	3.4	2.0	0.8
Level of education				33.9*	10
High school or less	10.7	11.0	4.8
High school	38.8	39.0	24.8
Trade school	4.0	3.9	2.4
Associate degree	5.6	8.6	4.8
College	23.1	25.0	40.8
Graduate/professional	17.9	12.5	22.4
Occupation				12.8	8
Student	27.3	28.6	23.5
Executive, professional	15.2	7.8	15.3
Businessperson, manager	28.9	29.2	36.5
Manual laborer	11.6	18.2	10.6
Semiskilled laborer	17.0	16.1	14.1
Marital status				28.0*	8
Married	66.5	65.6	87.5
Separated	1.5	2.8	0.8
Divorced	1.4	2.8	0.8
Widowed	4.3	3.1	0.8
Never married	26.3	25.6	10.0
Children <18 y at home	49.6	46.0	48.4	1.1	2
Employment status				22.9*	4
Full time	48.5	58.8	65.6
Part time	18.0	18.0	8.2
Unemployed	33.5	23.2	26.2
Immigration status	2.0	4		2.0	4
US citizen	52.0	50.0	56.8
Permanent resident	32.1	32.8	27.2
Noncitizen	15.9	17.2	16.0

*P < .001.

Current Tobacco Use

Among current cigarette smokers, 31.6% reported daily use. We observed a significant difference in frequency of tobacco use among the ethnic subgroups (χ²₁₂ = 33.8, P < .001; Table 3 ▶). Koreans (85.3%) were more likely to report daily use than were Chinese (73.8%), Vietnamese (70.0%), or Cambodians (62.1%). Cigarettes were the most popular type of tobacco used (90.2%), followed by cigars (37.9%), pipes (30.8%), and chewing tobacco (20.5%). Koreans (96.3%) were more likely to be cigarette smokers than were Vietnamese (90.7%), Chinese (86.9%), or Cambodians (81.0%; χ²₄ = 9.9, P < .05).

TABLE 3.

—Demographic Characteristics of Current Tobacco Users by Ethnic/Racial Subgroup

	Chinese (n = 99)	Korean (n = 117)	Vietnamese (n = 79)	Cambodian (n = 43)	Other (n = 7)	χ2	df	Total (n = 345)
Age of initiation, y, mean (SD)a	18.1 (5.3)	19.8 (6.4)	16.9 (5.1)	13.9 (4.3)	18.7 (3.8)			18.3 (5.9)
Sex, %						9.72*	4
Male	70.7	79.5	84.8	81.0	71.4			78.2
Female	29.3	20.5	15.2	19.0	28.6			21.8
Frequency of smoking, %						33.8***	12
Within the last year	4.6	6.3	13.3	3.4	20.0			7.5
Monthly	3.1	2.1	5.0	13.8	40.0			12.2
Weekly	18.5	6.3	11.7	20.7	0			75.2
Daily	73.8	85.3	70.0	62.1	40.0			31.6
Type of tobacco, %
Cigarettes	85.5	96.1	90.1	80.0	85.7	9.97*	4	89.5
Pipe	26.9	25.7	53.8	25.0	0	8.14	4	31.9
Cigar	40.0	42.1	40.0	28.0	50.0	1.64	4	38.5
Chew	28.6	15.6	26.3	20.0	0	2.31	4	21.7
Country of birth, %						23.7***	4
United States	11.7	1.8	8.2	13.5	14.3			7.7
Other	88.3	98.2	91.8	86.5	85.7			92.3
Level of education						89.6***	20
High school or less	10.4	6.0	13.7	23.3	0			11.0
High school	34.4	32.8	47.9	53.5	14.3			38.8
Trade school	1.0	6.0	1.4	9.3	0			3.9
Associate degree	7.3	5.2	19.2	4.7	0			8.7
College	19.8	41.4	12.3	9.3	57.1			25.1
Graduate/professional	27.1	8.6	5.5	0	28.6			12.5
Length of time in US, %						37.5**	16
≤ 2 y	26.3	19.0	7.6	14.0	14.3			17.7
3–5 y	12.1	12.1	8.9	7.0	14.3			10.8
6–10 y	13.1	13.8	34.2	11.6	28.6			18.3
>10 y	37.4	52.6	44.3	55.8	28.6			46.2
Entire life	11.1	2.6	5.1	11.6	14.3			7.0
Initial source of tobacco, %						23.7	16
Friends	52.8	70.4	63.5	67.6	80.0			64.0
Family	9.7	3.1	1.6	0	0			4.0
Vending machine	1.4	0	0	0	0			0.4
Store	33.3	18.4	23.8	17.6	20.0			23.5
Other	2.8	8.2	11.1	14.7	0			8.1
Occupation, %						27.2*	16
Student	28.1	13.1	36.0	61.1	33.3			28.6
Executive, professional	14.0	8.2	2.0	0	16.7			7.8
Businessperson, manager	29.8	34.4	26.0	16.7	33.3			29.2
Manual laborer	17.5	23.0	20.0	5.6	0			18.2
Semiskilled laborer	10.5	21.3	16.0	16.7	16.7			16.1
Marital status, %						45.2***	16
Married	60.8	75.9	71.8	44.1	14.3			65.6
Separated	5.2	1.9	1.4	2.9	0			2.8
Divorced	2.1	4.6	1.4	0	14.3			2.8
Widowed	1.0	5.6	0	2.9	14.3			2.8
Never married	30.9	12.0	25.4	50.0	57.1			25.9
Have children <18 y at home, %	46.1	45.6	50.7	45.0	0	6.6	4	45.8
Employment status, %					21.8**	8
Full time	47.3	72.1	56.4	53.8	57.1		58.9
Part time	27.5	12.6	20.5	10.3	0		18.1
Unemployed	25.3	15.3	23.1	35.9	42.9		23.0
Immigration status					34.5***	8
US citizen	48.9	43.4	67.1	37.5	42.9		49.8
Permanent resident	21.1	38.9	31.6	45.0	28.6		32.8
Noncitizen	30.0	17.7	1.3	17.5	28.6		17.3

^aGroup differences were significant (F = 9.7, P < .001).

*P < .05; **P < .01; ***P < .001.

Approximately 29% of current smokers were businesspeople; 28.6% were students. We observed significant group differences in occupation among current smokers (χ²₁₆ = 23.7, P < .05). Cambodian students (61.1%) were more likely to be smokers than were Chinese (28.1%), Vietnamese (36.0%), or Korean (13.1%) students. Among those who reported business-related or managerial occupations, Koreans (34.4%) were more likely than Chinese (29.8%), Vietnamese (26.0%), or Cambodians (16.7%) to be smokers.

We also found significant differences in employment status between the subgroups (χ²₈ = 21.8, P < .01), with Cambodians exhibiting the greatest likelihood of being unemployed. Finally, there were significant group differences among smokers in marital status (χ²₁₆ = 45.2, P < .001). Korean smokers were more likely to be married (75.9%) than were Vietnamese (71.8%), Chinese (60.8%), or Cambodian (44.1%) smokers.

Tobacco Initiation

A Tukey test revealed significant differences between the subgroups in age at initiation of smoking (F_4,412 = 9.7, P < .001). Cambodians (mean = 14 years, SD = 4.3) began using cigarettes at a significantly earlier age than did Chinese (mean = 18.1 years, SD = 5.3). Koreans (mean = 19.8 years, SD = 6.4) initiated use significantly later than did Vietnamese (mean = 16.9 years, SD = 5.1) and Cambodians. As shown in Table 4 ▶, friends (64.0%) were the most frequently reported initial source of tobacco, followed by stores (23.5%), family members (4.0%), and vending machines (0.4%). Differences in initial source of tobacco across subgroups were not significant.

TABLE 4.

—Final Logistic Regression: Predictor Variables of Current Smoking (n = 550)

	Coefficient (SE)	Odds Ratio	95% Confidence Interval
Male (vs female)	1.68 (0.26)***	5.36	3.21, 8.93
Executive/professional (vs student)	–1.28 (0.44)**	0.28	0.12, 0.66
Divorced (vs married)	1.74 (0.75)*	5.68	1.30, 24.85
Widowed (vs married)	2.26 (1.07)*	9.54	1.17, 77.79
Never married (vs married)	–0.89 (0.37)*	0.41	0.20, 0.84
Age	–0.03 (0.12)**	0.97	0.95, 0.99

*P < .05; **P < .01; ***P < .001.

Acculturation and Prevalence

We also observed significant differences between subgroups in regard to immigration status, length of time residing in the United States, and country of birth (Table 3 ▶). Current smokers were more likely to be foreign born. Vietnamese smokers (67.1%) were significantly more likely to be US citizens than were smokers in the other 3 subgroups. Across all subgroups, the majority of smokers had resided in the United States for 6 or more years.

Correlates of Current Tobacco Use

A logistical regression analysis was undertaken to determine the independent variables that best predicted current tobacco use after consideration of the effects of the other variables (Table 4 ▶). The following variables were not associated with current tobacco use and were eliminated from the regression model: country of birth, length of time in the United States, educational attainment, children younger than 18 years of age living at home, employment status, and immigration status. Nagelkerke's “max-rescaled” R² value, an estimate of variations in outcome variables explained by a logistic regression model, was calculated. The max-rescaled R² was 0.21, indicating that 21% of the variance in current tobacco use was explained by the logistic regression model.

The following variables were positively associated with current tobacco use: sex, occupation, marital status, and age. Women, persons who had never been married, executives or professionals (vs students), and older respondents were less likely to be smokers.

DISCUSSION

Our findings revealed differences not only in tobacco use but also in demographic characteristics across the Asian American subgroups examined, including differences in sex, age, length of time residing in the United States, educational attainment, occupation, and marital, employment, and immigration status. Previous studies have shown that subgroups of Asian Americans and Pacific Islanders differ in language, religion, culture, immigration and generation histories, socioeconomic status, and the extent to which they are acculturated or assimilated into the White or Anglo-American culture.35,36

The overall smoking rate among our sample was 29.4%; rates among the ethnic subgroups ranged from 24.1% to 43.0%. This finding is consistent with results of previous studies revealing smoking prevalence rates of between 25.3% and 71%.1,19,23–25,37 In the current study, the rate of lifetime smoking was 40.2%, ranging from 32.9% to 50.0% across the subgroups studied. This rate is substantially higher than overall smoking rates among adults in Pennsylvania, New Jersey,3 and Philadelphia38 and rates among White and Black adults, but it is comparable to the rate for Hispanic adults in Pennsylvania.39

Our results indicate that smoking may be as serious a problem among Asian Americans as it is in the general population, if not more so, especially when the ethnic subgroup differences presented here are taken into account. Considered in the light of the 2010 national goal of reducing the rate of tobacco use among adults to 12% or less,40 these high rates may require the special attention of state public health agencies and an infusion of federal funds into local prevention, intervention, and cessation programs.

Mean age at initiation of tobacco use among the sample studied was 18.3 years, a finding that is also consistent with previous studies.10,11,41 Subgroup differences, however, provide a more accurate assessment of needs. Cambodians and Vietnamese were oversampled to ensure their adequate representation in the Asian American population studied. Both of these subgroups exhibited low mean ages at smoking initiation—14 and 17 years, respectively—corroborating findings of earlier studies.42

Differences in tobacco use among the Asian American population and other ethnic groups and differences within Asian American subgroups lend support to the idea that cultural differences affect tobacco use patterns. For example, our study showed that current smokers were more likely to have been born outside the United States. Although it is generally assumed that immigrants' tobacco use will eventually conform to mainstream patterns of smoking, the current study clearly indicates that any prevention or intervention programs for immigrants must promote protective factors that the subculture provides and reduce factors that lead to heavy tobacco use.

Our findings suggest that intervention and prevention efforts need to focus on men and on the role of the family in influencing smoking behavior. The focus on men is particularly critical, because younger male Asians tend to emulate adult Asians' behaviors. For example, studies have shown that those who are heavy smokers are more likely to give cigarettes as gifts.43 Because smoking among Asian men is more tolerated among Asian Americans than smoking among women, Asian male adolescents may perceive this exchange of cigarettes as gifts as both an acceptance of smoking as a cultural norm and a reflection of one's affection or respect for others inside and outside the family circle. These and similar cultural mores have a tangible influence on smoking behavior, especially among young people, and require special attention by prevention, intervention, and cessation program planners.

Cause and effect cannot be determined from the findings of this study. However, family structure, marital status, occupation, and age seem to be related to current tobacco use, a circumstance that affords opportunities for further exploration of the relative impact of these variables on smoking behavior as well as on program planning.

The comprehensive needs assessment questionnaire developed for the study further corroborated the findings of other studies regarding tobacco use in Asian American populations—in particular, lower perceived risks regarding smoking-related cancers and chronic diseases and a pervasive lack of readiness for change in smoking behavior among Asian American smokers.44,45 The current findings prompted ATECAR to respond to these special needs by developing a series of community-based, culturally tailored comprehensive tobacco prevention, intervention, and cessation programs for the target Asian communities.

Among these programs are 2 research-based community tobacco prevention and intervention programs for adults and young people, a culturally adapted Asian youth smoking cessation program, a theory-based Asian adult smoking cessation intervention trial, a physician-based smoking intervention program to be provided in conjunction with intensive motivational behavior counseling, and a pilot study addressing cervical cancer among Vietnamese women, the group with the highest cervical cancer morbidity rate in the United States. This last component, although unrelated to the objectives of the study, was a by-product of the comprehensive assessment to which ATECAR felt obligated to respond. (ATECAR's experience in developing and implementing health surveys in Asian American communities is the subject of an upcoming article.)

Our study involved 3 major limitations. First, as a result of the cross-sectional design,30–33 cause and effect relationships could not be determined. For example, it is difficult to determine cultural influences on tobacco use with such a design. Second, the self-report procedure did not allow determination of the relative veracity of cohort responses. Finally, certain cultural factors and characteristics of the sample made simple random sampling difficult. Most community organizations, for example, would not provide membership lists because of confidentiality issues, and Asian Americans are generally reluctant to provide researchers with personal information. Such obstacles mandated modifications of the simple random sampling design to facilitate greater access to these communities and subgroups. Despite these limitations, we consider our approach viable and the data appropriate for the design of more culturally appropriate prevention, intervention, and cessation programs for our study community.