Table 1 Description of SOTM studies

Document review

Adoption

Design: Review of online, official Park Board meeting records, City Council minutes and commentaries, and the results of the Park Board’s pre-law public opinion survey from 2007–2012.

Analysis strategy: Documents were analyzed thematically to understand the reasons for adopting the bylaw.

Key informant interviews

Adoption

Design: Semi-structured interviews.

Sample: Eight key informant interviews with civic officials, public health advocates and health care providers conducted from May to December 2011.

Analysis strategy: Recorded interviews were transcribed verbatim, and analyzed thematically to generate a chronological account of the introduction of the bylaw and to understand the reasons the informants had for supporting or opposing the bylaw.

Social and built environment study [59]

Support

Design: Semi structured interviews and focus groups between March 2010 and February 2011 (prior to the implementation of the smoke-free bylaw)

Sample: 40 telephone interviews (with 21 women and 19 men in Greater Vancouver) and focus groups with seven additional participants who were exposed to secondhand smoke daily or almost daily.

Analysis: Recorded interviews were transcribed verbatim and analyzed thematically to obtain information on support for the bylaw.

Media analysis [42]

Support

Design: Content analysis of print news media from January 1, 2010 to December 31, 2011.

Sample: 90 articles from the Canadian Newsstand Database and independent newspapers.

Analysis: Articles were coded in two stages, first using a custom Perl script and then with a set of content variables. The articles were further coded using 45 content variables into the categories of relevance, geographic focus, slant, primary approach, theme, and tobacco control topics.

Park User Telephone Survey [43]

Support

Design: A cross-sectional survey using a random digitalized calling sampling process between September 15^th and 25^th, 2011.

Sample: 496 Vancouver residents (446 nonsmokers and 50 smokers) who had visited a beach or park in the previous year (from Sept 2010 to Sept 2011) —the first year of the smoking bylaw.

Analysis: Data obtained from respondents included demographic information, smoking status, support for the smoke-free bylaw, and opinions regarding the smoke-free bylaw. Unadjusted and adjusted logistic regression analyses were used to examine the correlates of supporting the bylaw.

Park and Beach Observation Study [60]

Compliance

Design: Observations of parks and beaches at nine time-points (pre-bylaw, one-week, one-month, 8-months, 9-months, 10-months, 12-months, 22-months, and 24-months after bylaw implementation) from August 2010 to September 2012.

Sample: Purposively selected parks (n = 3) and beaches (n = 3) in Vancouver, Canada.

Analysis: Observed smoking in each venue was recorded during a 30-min time period. Observation sessions were limited to afternoons and evenings on the weekends (Friday-Sunday). Information on the maximum number of persons, total number of smokers, duration of time spent, and average daily temperature were recorded per venue. Friedman’s tests were used to assess the changes in the total smoking rates in venues over time. Wilcoxon signed rank tests were used to assess the differences between prelaw and each subsequent observation time point smoking rates. Mann–Whitney tests were used to examine the differences in smoking rates between parks and beaches.

Beach Litter Study

Compliance

Design: Secondary analysis of observational data from the Great Canadian Shoreline Cleanup (see http://www.shorelinecleanup.ca/) which comprised park and beach litter data from one year before and two years after the implementation of the bylaw from 2010–2012.

Sample: Litter (from cigarettes/cigarette filters, tobacco packaging, cigarette lighters, and cigar tips) among 40 sites from which litter was consistently obtained in all 3 data collection periods.

Analysis: For each venue, information on number of volunteers, distance cleaned, and litter (cigarettes/cigarette filters, cigarette lighters, cigar tips, tobacco packaging) was obtained. Repeated measures analysis for negative binomial regression was based on the generalized estimating equation (GEE) approach and was used to evaluate differences in the amount of litter obtained between parks and beaches over the 3-year study period. Each model included the factors of venue type (park vs. beach), year (2010, 2011 and 2012) and the interaction between venue type and year as well as the time-dependent covariates for number of volunteers and kilometers covered.

Park Ranger Focus Group and Citation Information

Enforcement

Design: Two semi-structured focus groups in October 2011 (13 months following implementation) and then again a year later in August 2012. Citation data was obtained from the metro police department.

Sample: Twelve individuals participated in the focus groups (6 individuals participated in both groups). Rangers who participated in the focus groups included novices and senior officers (8+ years), and both seasonal and permanent employees. (The permanent Park Ranger contingent is tiny, consisting of a full-time Lead Ranger, a full-time Homeless Liaison, and four part-time Rangers; in the summer months, when park and beach usage peaks, 36 seasonal auxiliary Rangers join the permanent staff).

Analysis: Focus group data were analyzed with a 23-item coding frame created by the Principal Investigator (PI) and a team member. Inter-coder reliability with a third team member was .849 (Krippendorff's alpha). A saturation of themes was demonstrated when no additional codes were created during the coding process. A narrative summary was compiled for each code for each focus group, and the number and density of responses analyzed; the two focus groups were also compared for changes from year one to year two. The identification of themes was formulated by the PI in conjunction with the research team. Citation data from Municipal Ticket Information system reported as frequencies.