Nicotine Dependence among Rural-Urban Migrants in China

We employed a stratified multi-stage systematic sampling procedure to recruit participants. In Stage 1, we selected three metropolises based on broad geographical representation. They were Chengdu in southwest China, Shanghai in southeast China, and Beijing in north China. Beijing, the Chinese capital, has a population of approximately 17.4 million [15]. This population includes an estimated 5.4 million migrants. Shanghai is the most populous city in China with 18 million, including approximately 7.1 million migrants [16]. Chengdu is the capital of Sichuan province, and has a population of 11.2 million, including some two million migrants [17]. In Stage 2, we selected one residential district with a high density of rural-urban migrants in each of the three study cities. In Stage 3, we selected as the sampling unit within each district medium-sized or large worksites (operationalized as employing 200 or more workers) that employed migrants. We then randomly sampled 10% of these sites by type [18]: (1) factory (manufacture, machinery, electronics, textile, printing); (2) construction (public and private, transportation and public works infrastructure, and maintenance operations); and (3) service (tourist bureaus, hotels, restaurants, barbershops, beauty salons, bath houses, night clubs, karaoke venues, dance halls, and bars) and commercial sites (markets, street vendors, and small retail shops). In Stage 4, the final stage, we selected as eligible participants from each district cluster of worksite types all rural-urban migrant workers who were ages 18 years or older, held rural "hukou," (that is, were registered permanent residents in a rural area), and had resided in a destination city for at least six months. The number of participants disaggregated by worksite type approximately corresponded to their estimated overall distribution across each district cluster of worksites [9, 10, 18].

Data were collected by trained interviewers through face-to-face interviews in the households of participants. Various strategies were used to reach migrants at the sample sites. First, we contacted employers at sampling units for permission to conduct the survey on their premises. Upon receipt of this permission, we then contacted recognized group "leaders" who mobilized and encouraged their peers to participate in the study. The study purpose was explained to all eligible participants and verbal consent was obtained from all who elected to participate. The Ethics Committee of the Medical Center at Zhejiang University approved the study protocol. All surveys were administered by local medical professionals, who were trained in research survey methods, and investigators were on-hand to answer questions. Questionnaires were completed by eligible and willing participants and were checked by investigators for completeness. Assistance was provided to participants who experienced difficulty in completion; typically because of an educational deficit. Following completion, all participants were given a token of appreciation (tooth brush and tooth paste; value US$0.50) and asked to provide their names and phone numbers for verification purposes.

Current smoking status: current frequency and quantity of smoking. Participants reporting that they used tobacco products, including cigarettes, cigars, rolled cigarettes, and Chinese pipes, were considered smokers. They were divided into daily and occasional smokers. Quantity was calculated by units for pipe smokers and number of cigarettes for the remainder.

Nicotine Dependence: assessed using the six-item Mandarin Chinese Version of the Fagerström Test for Nicotine Dependence (CFTND) [13]. Items covered (F1) number of units smoked per day, (F2) time to first cigarette after awakening in the morning, (F3) early excess smoking (smoking status during the first hours) (F4) reluctance to give up the first cigarette in the morning, (F5) difficulty in refraining from smoking in forbidden areas, and (F6) smoking even when ill in bed.

Data were entered using Microsoft Excel and transferred to SAS. All analyses were conducted using SAS 6.12 Version and in several stages. First, Cronbach's alpha for the CFTND was calculated to test scale reliability. Secondly, we calculated the CFTND total score by summing responses to the six individual items, ranging from 0 (not nicotine dependent) to 10 (most nicotine dependent), and reported scores across demographics. Thirdly, we calculated the arithmetic mean (with standard deviation) to reflect the nicotine dependence level for the sample, and then performed MANOVA to examine the association between nicotine dependence on the one hand and smoking status, demographics, and migration variables on the other. Nicotine dependence was our dependent variable. Our independent variables were age, education, ethnicity, marital status, region of origin, age at first migration, length of migration (years), time spent in city in each year (months), number of cities ever lived in, work type, personal income, type of city residence, and type of smokers. They were all entered into the MANOVA equation. We considered values of p = < 0.05 to be statistically significant.

A total of 4,825 individuals were identified as eligible participants for our study. Of those, 4,198 were approached and agreed to be interviewed, resulting in a participation rate of 87%. Complete data were reported on the participants due to immediate remedial followup as needed.

Of the 4,198 participants, 53% were male. The great preponderance was under 40 years of age (81%) and earned a personal monthly income of 2,000 RMB or less (81%). Forty-five percent were employed in construction, 32% in the service sector, and 21% in factories. Sixty-two percent of participants had junior school or less education, 47% had worked in the city of destination for five or fewer years, and 25% had worked in two or more cities.

The current smoking rate was 28.3%, the daily smoking rate 21.2%, and the occasional smoking rate 7.1%. Among daily smokers (894) the Cronbach alpha for CTFND was 0.67, which indicated acceptable scale reliability. The average number of cigarettes smoked daily was 15.82(SD: 8.09) and 86% of participants reported smoking 10 or more. The total CFND score for all daily smokers was 3.39(SD: 2.32).

Table 1 shows the mean and standard deviation of CFTND scores for daily smokers pertaining to respondent sociodemographic and migrant characteristics. Scores increased with age, and tended to vary inversely with education. Scores were lowest among the never married than the married and the previously married, and among construction workers relative to other workers. CFTND scores tended to rise with age at first migration, length of migration, and number of cities of migration.

Table 2 shows the results of the MANOVA analyses of CFTND scoresamong daily smokers. Score were higher among construction workersthan factory and service workers. Age at first migration, length ofmigration, and number of cities of migration were all associated withnicotine dependence.