Factors associated with support for smoke-free policies among government workers in Six Chinese cities: a cross-sectional study

Study participants were 6,646 employees from 160 government agencies targeted for intervention by city-level grantees (Table 1). Of the 6,646 who completed the survey, 5,405 had complete data on key variables of interest and are included in the analyses reported here. Those excluded from analyses did not differ significantly on any of the demographic variables reported in Table 2. A broad range of government agencies were included, such as Departments of Education, Health Bureaus, and the Office of Chinese Peasants & Workers Democratic Party. The most common types of agencies can be classified as economic/finance, policy, human resources/social services and districts. Health departments, transportation, environment/natural resources, public safety/security, education, research, and information technology agencies were also included in the sample. At the time of data collection in 2011, none of the cities had city-level smoke-free policies and all were covered by the national regulations for smoke-free health care facilities and national guidelines for schools, although neither was fully implemented yet.

Data were collected by local grantees, which were most commonly city-level Chinese Centers for Disease Control (CDCs) and Health Education Institutes. Recruitment methods were determined by local grantees. Although stratified cluster sampling with random selection of employees was encouraged, sites varied in their recruitment approach. Study protocols were reviewed and approved by local Institutional Review Boards (IRB) in China, including the Jilin Provincial Medical Association IRB for Changchun, the Kelamayi Municipal Medical Association IRB, the Guangxi Provincial Health Bureau IRB for Nanning, the First Affiliated Hospital IRB of Suzhou University, the North China Coal Medical School Affiliated Hospital IRB for Tangshan, and the Yinchuan Municipal Medical Association IRB. Data were collected using self-administered surveys developed by the Emory team that provided technical assistance and training to the cities on evaluation. Surveys were developed for each sector targeted by the grantees, with a core set of common measures across all survey tools. Measures reported here were from the China Global Adult Tobacco Survey (GATS) [2, 29]. The Emory team trained grantees on evaluation methods, and cities were responsible for recruitment, data collection and data entry using a standardized database. Grantees provided Emory with de-identified data sets which were then pooled for the analyses reported here.

The dependent variable, support for smoke-free policies, was assessed by asking: Do you think smoking should be permitted in the interior spaces of the following types of places: hospitals, work settings, restaurants, bars, elementary and middle schools, colleges, public transportation (e.g., airports, train stations, buses), religious sites and hotels [29]. Response options were: should be permitted and should not be permitted. The worksite policy question, from GATS, asked: “Which of the following best describes the indoor smoking policy where you work?” Response options included: smoking is allowed anywhere, smoking is allowed only in some indoor areas, smoking is not allowed in any indoor areas, there is no policy, don’t know. Beliefs about harms of SHS were also assessed using items from GATS. Demographics variables included gender, age, ethnicity, education, and length of time working for current employer.

For each of the dependent variables of interest, we conducted unadjusted analyses to assess univariate associations with each of the major independent variables and demographic characteristics. We used multivariate logistic regression to identify factors associated with support for policies in four settings with lower levels of support and/or high potential for tobacco control initiatives: worksites, bars, hotels and colleges. Interaction between smoking status, gender and each explanatory variable was assessed using backwards elimination from full models. In the final multivariable logistic regression models, we accounted for clustering within sites through use of Generalized Estimating Equations (GEEs). We used PROC GENMOD in SAS Version 9.3 (SAS Institute, Cary, NC) to fit the models.

Overall, 54.6% of the participants were men and 25.9% were current smokers (1.9% of women and 45.9% of men) (Table 2). The majority had worked for their agency over five years (58.7%). Of note, less than half (47.5%) reported that smoking was not allowed in any indoor areas at their worksite. Knowledge of the harms of SHS exposure ranged from 68.4% believing it causes heart disease in adults to 91.2% believing it causes serious illness among nonsmokers. Smokers were largely men (96.7%), older and more likely to have a high school education or less. Relative to nonsmokers, smokers were significantly less likely to report a smoke-free worksite and less likely to believe SHS was harmful to nonsmokers.

Table 3 shows support for smoke-free policies in a range of settings. Support was highest for smoke-free elementary and middle schools (97.8%), smoke-free hospitals (96.7%), smoke-free religious sites (96.9%), and smoke-free colleges (97.1%). Support was lower for smoke-free bars (79.0%), smoke-free hotels (82.3%) and smoke-free restaurants (90.0%). Smokers were significantly less supportive of smoke-free policies than nonsmokers across all settings. Smokers were particularly non-supportive of smoke-free policies in bars (54.4%) and hotels (56.6%).