Background
Evidence of the ill effects of smoking has been growing since the 1950’s. In 1964, the U.S. Surgeon General issued the landmark report Smoking and Health: Report of the Advisory Committee to the Surgeon General, which stated unequivocally that a link between smoking and certain cancers exists (U.S. Department of Health, Education, and Welfare, 1964). Cigarette smoking and breathing secondhand smoke (SHS) have now been causally linked to an increased risk for multiple cancers and chronic diseases and is the leading preventable cause of death in the United States (U.S. Department of Health and Human Services [USDHHS], 2010, 2014). In Wyoming, this leads to approximately 800 deaths each year and nearly $258 million in annual healthcare costs that can be directly attributed to smoking (Centers for Disease Control and Prevention [CDC], 2014a).
The Wyoming Tobacco Prevention and Control Program (TPCP) works to reduce tobacco use in Wyoming by using a comprehensive, multi-strategy approach. The TPCP aligns its strategies around four goals that it has shared with the CDC for many years:
- Preventing initiation of tobacco use (CDC, 2014b)
- Eliminating nonsmokers’ exposure to secondhand smoke (CDC, 2017)
- Promoting quitting among adults and young people (CDC, 2015)
- Identifying and eliminating tobacco-related disparities (CDC, 2014b, 2015, 2017)
As part of monitoring progress on these goals, the TPCP tracks the prevalence, consumption, and use of tobacco products including cigarettes, electronic nicotine delivery systems (ENDS; also known as e-cigarettes or vaping), and other forms of tobacco.
A key issue in evaluating tobacco prevention and control is monitoring the same indicators of behaviors and attitudes related to tobacco over time. Because tobacco prevention outcomes result from state government programs including the one implemented in Wyoming, the efforts of multiple federal agencies—most prominently, the U.S. Food and Drug Administration [FDA], Substance Abuse and Mental Health Services Administration [SAMHSA], CDC, and other groups (e.g., the Robert Wood Johnson Foundation, Campaign for Tobacco-Free Kids, American Nonsmokers’ Rights Foundation, American Cancer Society, and American Lung Association)—changes over time reflect the cumulative impact of many sustained and new efforts to affect the key indicators. These indicators have not been revised drastically since at least 2005 (see CDC, 2007, 2014a, 2014b, 2015, 2017; Starr et al., 2005).
The Wyoming Adult Tobacco Survey (ATS) is a standardized telephone survey administered by the Wyoming Survey & Analysis Center (WYSAC) at the University of Wyoming under contract to the Wyoming Department of Health, Public Health Division (PHD). Its purpose is to collect state- and county-level data pertaining to the prevalence of tobacco use, the four TPCP/CDC goals, and the broader goal of reducing tobacco-related disease and death. In addition to conducting analyses on the 2017 data, WYSAC merged data from the 2017 ATS with data from previous iterations of the survey to analyze trends. WYSAC has formatted this report similarly to previous iterations of the report to facilitate readers’ efforts to compare the current results to those in previous reports. The previous report is available here: https://wysac.uwyo.edu/wysac/reports/View/5553.
2017 ATS Methods
In this section, we provide a general summary of the methods used to collect and analyze the data for the 2017 ATS. Additional technical details are in Appendix B. The CDC protocols for the 2017 ATS, the 2010 National Adult Tobacco Survey, and the previous iterations of the ATS (2002, 2004, 2006–2009, 2012, 2015) were generally similar, which allowed WYSAC to perform analyses of trends for comparable questions on the surveys and reflecting the stability in key indicators related to tobacco prevention work.
Questionnaire Development
WYSAC developed the 2017 ATS items based on CDC’s core and supplemental ATS items. The Wyoming TPCP and WYSAC selected some optional questions and developed some Wyoming-specific questions based on the indicators most directly related to TPCP efforts in Wyoming. Because the national and Wyoming tobacco prevention programs have been stable since the 2015 iteration of the ATS, few changes to the survey questionnaire were required. Key changes for the 2017 ATS included adding questions about whether dual cigarette and ENDS users first used cigarettes or ENDS, adding questions about use of flavored ENDS and different ENDS brands, simplifying the questions about smoking cessation to focus on the Wyoming Quit Tobacco Program (WQTP), adding an item to assess barriers to quitting smoking, adding items to assess efforts to quit using ENDS, adding items to assess perceived harmfulness of ENDS use (overall and relative to smoking), adding an item to assess the perceived relative benefits of switching from smoking to ENDS use, adding an item to assess mental health as a key disparity in the burden of tobacco use, and shortening the survey (in an effort to improve the response rate) by eliminating questions about children in the adults’ homes.
Survey Administration
The random digit dialing (RDD) landline and RDD cell phone samples for the 2017 ATS were disproportionately stratified to produce county-level data. The goal was to complete roughly 200 surveys in each county for a statewide total of 4,600 surveys. Both samples were pre-screened to avoid calling numbers that were known to be disconnected, businesses, or on WYSAC’s list of people who have asked not to be surveyed. WYSAC made 356,204 call attempts for the 75,755 numbers that were left after the pre-screening process. Some numbers were called up to 27 times before they were assigned a final disposition code, which resulted in an average of 4.7 call attempts per record. Calling began on April 30th, 2017, and ended on December 20th, 2017. The final dataset contained a total of 4,647 completions (meeting the goal of 4,600 completes statewide), including 2,152 landline completions and 2,495 cell phone completions. The response rate for the landline sample was 27%, the response rate for the cell phone sample was 38%, and the overall response rate was 33%. After the completion of the data collection, the CDC contractor weighted the 2017 ATS data to make the results more representative of the Wyoming adult population. Weighting variables included selection probability; nonresponse adjustment; gender by age, race/ethnicity; educational attainment; county; and phone usage (cell phone only, landline only, and dual phone use). Weighting data does not change responses. It makes the dataset generally more reflective of the entire Wyoming adult population than the unweighted data. WYSAC received the 2017 ATS data file from the CDC contractor on February 26, 2018.
Analysis
WYSAC analyzed the data using Stata, version 12.1 with the complex sample survey methods available in that statistical package. In the tables and figures of this report, WYSAC used weighted data to calculate estimates and associated confidence intervals. WYSAC used logistic regressions to test for trends for time periods longer than two years. Generally, trends reported in this document are based on the earliest year a comparable question was asked through 2017. WYSAC also used logistic regression to identify statistically significant associations between outcomes and their correlates. The body of this report summarizes results for evaluation indicators for the Wyoming TPCP, including the estimates from 2017 and trend analyses (when possible). Results for each survey item are reported in Appendix A as a supplement to this summary. Relationships noted as significant in the body of the report are significant, p< .05. Confidence intervals for estimates to responses for the 2018 survey are in Appendix A. Key statistical results are detailed in Appendix C.
Data Limitations
Most ATS survey items have been tested and validated by the CDC and reused over time. However, because ATS provides self‐reported data, respondents’ recollection of events and interpretation of the survey items are estimates and might include respondent bias (e.g., underreporting undesirable behaviors). Also, not all estimates have the same level of precision due to survey skip pattern, analysis of subgroups, and the combination of both. For example, questions asked of smokers or other tobacco users tend to have sample sizes smaller than other questions asked of nonsmokers or non-tobacco users. Estimates for small subgroups, e.g., African Americans in Wyoming, would be also less precise than estimates for larger subgroups, e.g., White Americans.
