Response to Editor and Reviewers’ comments
BMC Public Health
Ref: 2289653541377849
Title: Impact of a long-term tobacco-free policy at a comprehensive cancer center: a series of cross-sectional surveys
October 23rd, 2014
L’Hospitalet de Llobregat, Barcelona.
Dr. Raymond Niaura
BMC Public Health
Dear Dr Raymond Niaura,
We would like to thank you very much for the opportunity to resubmit our manuscript entitled “Impact of a long-term tobacco-free policy at a comprehensive cancer center: a series of cross-sectional surveys” which has been modified in line with the useful reviewers’ comments.
We greatly appreciate the reviewers’ comments and suggestions. We enclose a point-bypoint response. As requested, modifications in the text of the manuscript have been marked using the "tracked changes" option.
Sincerely,
Cristina Martínez, RN, BA, PhD
Tobacco Control Unit, Cancer Prevention and Control Programe,
Institut Català d’Oncologia
1

Editorial Requests: a) Line numbering: Please revise your manuscript to include line and page numbers.
Authors are asked to ensure that line numbering is included in the main text file of their manuscript at the time of submission to facilitate peer-review. Once a manuscript has been accepted, line numbering should be removed from the manuscript before publication. For authors submitting their manuscript in Microsoft Word please do not insert page breaks in your manuscript to ensure page numbering is consistent between your text file and the PDF generated from your submission and used in the review process.
Response: We have included line numbering in both the marked copy and the nonmarked copy. b) Please provide an abstract within the main manuscript file.
Response: Done. c) Please change 'contribution statement' to "Author's contribution".
Response: Done
2

Response to Reviewer 1
This is an interesting study evaluating changes in smoking prevalence and tobaccorelated behaviors and attitudes in the context of a long-term hospital-based tobacco control interventions and to smoking legislations, one banning smoking in public places and the second banning smoking in all public places including hospitality venues and outdoor spaces in hospitals.
We thank Reviewer 1 for general and useful comments. Our response below is focused on addressing specific critiques. Because review comments may sometimes be similar, we number each reviewer comment and present them in the same order that they appeared in the original critique.
1. A table showing participants characteristics (age, sex, job status, and other relevant information) by survey year is needed to compare the changes over time. They are mentioned in the text but I think it would be better to have a table.
Response: As suggested, we have included information regarding demographic participants’ characteristics in a new table by three study periods (baseline, after 1 st law, after 2 nd law) and mentioned it (Table 1) in the first paragraph of the Results section.
2. It is unclear why the distribution of type of job status changed over time since the survey was conducted as a representative sample of the hospital employees in each year. Was there a major change in the organization and type of workers in the hospital?
Response: Our organization is a Comprehensive Cancer Center devoted to clinical care and research. In the last 10 years the organization has opened new departments and units committed to research. This could be a reasonable explanation to the increase of the “others employers” in our hospital (mainly statisticians, non-medical researchers such as biologists, pharmacists, etc.).
Being aware of this fact, we have included in the limitations the following information:
“hospitals have a high staff turnover rate, mainly among younger workers and the professional group "others".
3. Something that requires a little bit more thought is the high prevalence of “never smokers” in the 2012 survey. The prevalence of never smokers changed from 41.6% to
49.7%. In a stable population, the expectation is that the smoke-free legislation would decrease current smoking prevalence by increasing former smoking prevalence. Do you have any possible explanation for those changes? Could it be related to the change in the population distribution (see comment above)? To changes in the questionnaire?
Something interesting is that the percentage of smokers <10 cig/day is actually a bit lower than before, this would suggest that the light smokers have quit and heavier smokers are the ones who remained, except that the prevalence of former smoking has
3

decreased instead of increased. Could it be that light smokers are interpreting the questions differently now and are classifying themselves as never smokers?
Response: We agree that in a stable population the expectation is that a smoke-free legislation would decrease the prevalence of current smokers by increasing the prevalence of former smokers. However, hospital staff usually has a high level of turnover. This might be the cause that never smokers have increased from 41.6% to
49.7%. We are confident that this is not due to changes in the questionnaire, because the smoking status question was exactly the same for all the surveys.
Regarding light smokers (< 10 cig/day) we contemplate that part of them have quitted and others have shifted from light smokers to occasional smokers. We have commented on this in the discussion section:
“Another remarkable point is the substantial increase in the number of occasional smokers in our hospital worker population. However, this finding is in line with other studies that have reported a similar increase in the number of the occasional smokers in countries where tobacco consumption in the overall population is decreasing".
4. It is unclear if the survey of all oncology workers was done in the context of the main employee survey or through a specific survey. Are those workers included in the main survey? How many are they in each year?
Response: As pointed out above, our center is a Comprehensive Cancer Center, providing oncologic care, research, and training. Thus, all our staff is considered “oncology workers”. As mentioned in the results, about 200 workers were surveyed every year.
Table 1 now details the number and proportion of workers per profession per study periods (baseline, post 1st law, post 2nd law). To do not mislead we have deleted the term oncology in the following paragraph on page 6 (reviewed document).:
“Finally, we surveyed all oncology hospital workers to assess their support for the
‘Hospital Tobacco Control Project’, their agreement (…)”
5. The lack of impact in smoking prevalence among the group <35 years is a concern and should be highlighted in the discussion together with some potential strategies that could be used.
Response: We have highlighted the lack of effect among the youngest group (< 35 y) in the discussion by adding a comment that includes two possible explanations. (Added text in italics).
“Decreases in tobacco consumption were observed mainly in hospital workers ≥
35 years old, doctors, and women. We hypothesize that the smoke-free legislation has had lower impact on the youngest group (< 35 y) because young smokers tend to trivialize the harmful effects of smoking, and, in our context, there are insufficient initiatives addressed to motivate cessation among young
smokers, even for health professionals”.
4

6. Page 8, paragraph 2: sentence “with adjusted PRs of 0.20 (95%CI: 0.05-0.87) and 0.78
(95%CI: 0.42-1.46), respectively (Table 1).” should indicate “… and 0.78 (95%CI: 0.42-
1.46), respectively, compared to baseline.
We appreciate the suggestion. We have included it into the manuscript.
7. When mentioning the outdoor spaces that are covered by the 2nd law it is unclear if it refers to all public places or only hospitals.
We have rewritten the paragraph in the introduction which explains the 2nd smoke-free law passed in Spain in 2010. This law ruled: a) Smoke-free indoors for the hospitality sector (bars and restaurants) without exception. b) Smoke-free outdoors in some public areas including: hospital grounds, educational campuses, and playgrounds.
The final paragraph at the end of page 3 (in the new reviewed version) poses that:
“In 2010, the law was amended providing: (1) smoke-free indoors for the hospitality sector (bars and restaurants) without exceptions, and (2) Smoke-free outdoors in some public areas including hospital grounds, educational campuses, and playgrounds [22]”.
8. The decrease in the perception of tobacco smoking as a health concern over time is actually a concern. One would have expected that hospitality workers are now more informed and aware about the risks. It seems that some additional interventions are needed to increase concern regarding smoking and secondhand smoke exposure. Are there surveys of general populations in Spain to compare to? Is it similar? Better, worse?
As the reviewer pointed out, smokers’ concern on the harmful effects that their smoke could cause on others decreased during the study period significantly. We believe that this decrease is caused by two possible reasons:
1) Since the passage of two smoke-free laws (one banning smoking in indoor locations and the other in outdoors locations of hospitals), smoker hospital workers never smoke in the hospital premises, so they are not in contact with non-smokers. So, theoretically non-smokers are never exposed to second hand smoke in the hospital (both indoors and outdoors).
2) In our hospital, those hospital workers who still smoke after the passage of the two laws have been described mainly as occasional or heavy smokers. These two types of smokers may understate the harmful effects of smoking and second hand smoke.
Although these are two possible hypotheses, we are aware that we cannot confirm neither of them because we did not explore them in this study. However, this it is an
5

interesting element to be explored more in detail not only among our smoker population
(hospital workers) also in the rest of the smokers in Spain. To our knowledge, the information on the degree of knowledge and concern of the population about smoking and secondhand smoke in Spain is scant, except for two cross-sectional studies, one from our group conducted in 2002 in a city in the Metropolitan area of Barcelona and another of national scope conducted in 2010. However, only information on risk perception for smoking was gathered. Thus, we are inclined not to comment on this point given the lack of data.
9. In the discussion, minimize the repletion of the results that are already in the tables and text.
The purpose of repeating some of the data was to remind the main findings to the reader. We have deleted some data repetition in the discussion that we consider could be redundant as suggested by the reviewer.
See our first paragraph on the discussion that has been shorten avoiding repetition.
Moreover, on page 12 (in the new reviewed version) we have deleted information on brackets.
“Our study reveals lower smoking rates among oncology nurses (…..).”
6

Response to Reviewer 2
General comment: This is a rather relevant study assessing the impact of a comprehensive tobacco control (TC) programme in a workplace and healthcare setting- a comprehensive cancer setting in Catalonia, Spain-by implementing several FCTC measures; while simultaneously Spain has notoriously advanced its tobacco control agenda. The study methodology is sound: authors have used standardised questionnaires and present the findings of surveys consistently repeated over time; also include a multivariate model analyzing time trends. I make several suggestions that authors should take into account in order to improve the manuscript’s readability and also to highlight the quality and interest of the paper.
Response: As with Reviewer 1, we thank Reviewer 2 for general and positive comments showing her understanding of this work. We also appreciate the several suggestions that authors have given us in order to improve the quality and soundness of the manuscript.
Our response below is focused on addressing specific critiques:
1.
Authors should indicate that the first law was a non-comprehensive smoke-free policy (SFP) or a partial ban allowing smoking in hospitality venues; and that the second one was comprehensive ban banning smoking in all indoor places, “and extending the ban to outdoor areas of certain premises such as hospitals”. Authors should also indicate when these 2 bans came into force. Lastly, while authors have written “This study assesses the impact of a long-term tobacco control policy project on hospital workers at a comprehensive cancer center before and after two national smoke-free laws were passed.”, the study aims are not clear. Please clarify the study
main outcomes.
Response: As the reviewer points out two smoke-free laws have been passed in Spain in the last years. Due to the space limitations in the abstract, we rewritten part of the background section including the years when the laws were passed and how affected to the healthcare sector (our study setting for this study). We have reserved the details of the two laws for other sectors for the main manuscript. Please see below how we provided information in the abstract and in the manuscript (also rewritten by suggestion of reviewer 1)
Abstract:
“Background: Spain has passed two smoke-free laws in the last years. In 2005, the law banned smoking in indoor places, and in 2010 the ban was extended to outdoor areas of certain premises such as hospitals.”
7

Manuscript body last paragraph of the introduction (page 3 of the reviewed document):
“In 2005, Spain passed its first comprehensive smoke-free law, which banned smoking in indoor public places, workplaces, and health care services, except in hospitality venues [21]. In 2010, the law was amended providing: (1) smoke-free indoors for the hospitality sector (bars and restaurants) without exceptions, and amended by extending the prohibition of smoking to all hospitality venues without exception and (2) Smoke-free outdoors in some public areas including hospital grounds, educational campuses, and playgrounds to certain outdoor public areas, including hospital premises, educational campuses, and playgrounds [22]”
Regarding the comment about improving the way the aim is posed:
We agree with the reviewer that some clarification on the aim was needed. So, we have included that we explore smoking consumption among hospitals workers after the passage of two smoke-free laws. Please see below the final paragraph:
“This study assesses the impact on smoking consumption among hospital workers at a comprehensive cancer center after the passage of two national smoke-free laws”
2.I suggest substituting workers for hospital workers or hospital staff.
Response: We have included the suggestion in the abstract.
3.Regarding the sentence “Logistic regression was used to compare differences in the odds of smoking after the laws took effect (baseline vs. 1st law; 2nd law vs. 1st law)”, I suggest substituting:
• 1st law by partial ban and 2nd law by comprehensive ban; this is more clear about the policy change in hospital and in Spain.
• Logistic regression was used to compare differences between the odds of smoking after the bans came into force.
Response: As commented above the study was only run in the healthcare sector, in particular, we studied smoking consumption among workers in a Comprehensive Cancer
Center.
Although the first law (passed in 2005) can be considered partial because smoking was still permitted in bars and restaurants in Spain, for the sector we are exploring in this study –the healthcare sector- smoking was forbidden without exceptions in acute hospitals since 2005. Thus, we consider that including all these details are out of the scope of the study and could mislead the reader. However, we provide more information in the body of the manuscript regarding the rulings of each law.
4. Authors have written “Baseline smoking prevalence was 33.1%.” Baseline smoking prevalence should be breakdown by gender and also indicating 95%CI. The findings related to gender are important and should be highlighted in the abstract. Therefore I
8

strongly recommend that authors would indicate smoking prevalence and PR by gender
(baseline and post-comprehensive ban).
Response: In the abstract we have mentioned that women were one of the groups that decreased more smoking prevalence. However, due to space constrains (250 words according to the journal requirements) we are not able to include the PR of all participants characteristics. This has been much more highlighted on the body of the manuscript and in the discussion.
5. Regarding the sentence “Changes included an increase in occasional smokers, …” I
suggest substituting changes by observed trends over time included…
Response: We have included this specification in the abstract.
6. Authors should briefly emphasise that the decrease in smoking prevalence was significantly more pronounced after the comprehensive ban.
Response: As suggested, we have emphasized that the outdoor smoke-free ban was more determinant in helping to decrease smoking consumption among hospital workers.
“Conclusions: A long-term tobacco control project combined with two smoke-free national laws reduced smoking rates among health workers and increased their support for tobacco control policies. The decrease was more significant after the passage of the outdoor smoke-free ban.”
7.I recommend that authors systematically use the term partial ban when reporting about the 1st law/ban and comprehensive ban instead of 2nd law/ban, as it is more clear about the policy change in hospital and in Spain and helps the reader to easily understand which policy/period authors are reporting about. I recommend that authors briefly frame the implementation of tobacco control FCTC policies during the study period (2001-2012), as well the evolution of the main TC indicators in Spain on this time frame.
Response: We have explained in more detail the two smoke-free bans passed in Spain after the approval of the FCTC. See response#7 to reviewer 1. On the other hand, as explained above - in response #3- we cannot consider the first law as partial for the healthcare sector because banned smoking in all indoor spaces of the acute premises.
Thus, we prefer to maintain the terminology used.
8. I suggest to include the setting on the sentence beginning “The benefits of smokefree policies in…..include…
Response: We have included healthcare services such as the setting of our concern in this study.
9

9.On the third paragraph beginning with “In 2005, Spain passed its first comprehensive smoke-free law…” authors should indicate that this was a noncomprehensive or partial ban and then specify the exemptions of such a ban. Following the sentence beginning with “In 20110 that law was amended…”authors should indicate that this was a comprehensive smoke-free law and then specify the amendments done on the former law.
Response: A more extended explanation has been given in this paragraph. Moreover, a comprehensive answer is provided when we replied reviewer #1 in question #7.
10. On page 3, while authors describe the hospital TC project, I suggest specifying the following issues:
• Focus of “awareness campaigns” because is not clear -SHS or tobacco health hazards?; benefits of cessation?......
• Promotion activities should be health promotions activities ?
• Evaluation efforts means what? -monitoring policies implementation and impact assessment?
• All group professionals were targeted by TC policies and engaged with TC equally? If not, briefly describe the differences.
• Incentives and type of cessation support for hospital staff
• Training programmes-briefly describe who was targeted and the dissemination plan
• Describe any strengthening of the programme during the implementation of the comprehensive ban; and how did you manage with the outdoors ban to overcome the compulsory abstinence during work hours-NRT was available for all smokers to deal with temporary abstinence?
This is important to guide other healthcare providers (HCPs), and also hospital managers and policy makers; and to highlight the need to implement a sustained and comprehensive TC policy in order to engage hospital staff and reduce smoking among them.
Response: Following the reviewers’ suggestions we have included a bit more of information about the several policies recommended in the ENSH model.
See 1 st paragraph on page 4 line 6 (reviewed version):
“ENSH model have been implemented, including: awareness campaigns on the hazards of SHS, smoke-free policies to protect people from SHS exposure, tobacco cessation services (including psychological support and pharmacotherapy if needed), and training courses for professionals, promotion activities, and evaluation efforts”
11.Same suggestions regarding the terminology of smoke-free policies.
Response: We have answered this point before.
10

12. On page 5, first paragraph, while describing the questionnaire changes over time, authors should specify whether the survey tools assessing the study core variables remained the same.
Response: This information was already present in the Methods section (page 6 line 3, of the reviewed document):
“(…) maintained the following core variables: social and demographic data, profession, smoking status, and attitudes towards active and passive smoking.”
13.On page 6, while describing the staff categories, it is not clear if authors have considered other HCPs besides nurses and doctors. I am not aware of the type of HCPs involved in a cancer center, but I suppose that there are other HCPs besides D+N, such as psychologists, dietitians, physiotherapists, etc. Where they included as a specific professional group category such as others HCPs?
Response: Professional categories included doctors, nurses, administrative and “other hospital workers”. The category “other hospital workers” included mainly technicians, statisticians, researchers, and a very small number of workers who are phycologists (one in our organization), nutritionists (one in our organization). To simplify and help to draw conclusions of our stud, doctors and nurses were categorized as oncology health care professionals, and the rest were consider non-oncology health professionals. We have expanded the paragraph in the Methods section to provide more details on “other hospital workers”: See on page 6 in line 21:
“Professional categories included doctors, nurses, administrative employees, and other hospital workers (included mainly technicians, statisticians, researchers, and a very small number of workers who are phycologists (one in our organization), nutritionists (one in our organization). included technicians and support staff).”
14.Education and socio-economic status (SES) are important determinants of smoking behaviour. Did authors include education and SES on the regression models? if not, specify why.
Response: We agree with the reviewer that SES is an important determinant of smoking.
However, in this study this variable is highly correlated with the professional status
(doctors and nurses are all university graduates; administrative staff have at least secondary education
 and some of them have a university diploma, and “other hospital workers” include a variety but mainly with university degree). Thus, to avoid overadjustment (and biased estimates) we preferred not to include education in the models.
11

15.Socio-demographics data (SDD) Authors indicate important SDD changes over time.
This should be depicted in a table.
Response: According to this comment and that of Reviewer 1 (comment #1), we have included a new table showing this information (new Table 1).
16.Did authors assessed the type of tobacco product consumed by the staff and did any change occur over time?
Response: The questionnaire includes the type of tobacco (cigarettes, cigars and pipes) but 99.6% of smokers were cigarette users. We have included this detail in the Methods section when describing the variables. Page 7 (reviewed document):
“The prevalence (%) of daily and occasional smokers, former smokers, and neversmokers of cigarettes, cigars or pipes, and 95% confidence intervals (CI) were computed. 99.6% of smokers were cigarette users.”
17.The study findings suggest that smoking prevalence was reduced because of cessation among smokers. I suggest that authors would compute cessation rates (exsmokers/ex-smokers +current smokers) over time, and according to the independent variables; and, If possible, quitting age trends, i.e to assess changes over time on the age of quitting (see Ravara et al, 2014).
Response: We appreciate this comment. However, the manuscript already includes a lot of data to address its primary goal and adding this information would complicate both the presentation of results and the Discussion. Thus, we prefer not to lengthen the manuscript (at least one table and one figure would be necessary) to address this secondary objective. If the Editor finds this point essential, we may consider how to extend the manuscript accordingly.
18.On page 11, while authors compare the decrease in smoking prevalence between the hospital staff and general population, authors should interpret and emphasise this based on the differences of implementation of FCTC measures between the general population and hospital workers and specially article 14 and 8 and 12.
Response: This point has been already mentioned in the discussion a bit later. We have pointed out that the tobacco control model of the ENSH integrates ten policies in agreement with Article 8, 12, 14 and 21 of the WHO FCTC. See below the paragraph on page 13 (reviewed document):
“Nevertheless, despite the clear benefits of smoke-free policies [4], the WHO
FCTC encourages organizations and governments to do more than just implement restrictions, advocating for the development of a broad tobacco control approach [1]. In this regard, our ‘Hospital Tobacco Control Project’ has developed a comprehensive tobacco-free model based on the ENSH-Global
Network for Tobacco-free Health Care Services. The ENSH model integrates ten policies in agreement with Article 8, 12, 14, and 21 of the WHO FCTC
(Article 8: “smoking bans in public places,” Article 12: “consumer
12

information,” and Article 14: “access to treatment for quitting smoking,”
Article 21: “research, surveillance and exchange of information”). The ENSH concept follows an organizational and cultural change model for implementing innovations [43] that has shown that a gradual implementation improves tobacco control policies [10, 44].”
19.When authors state that “catalan hospitals provide tobacco cessation services with the support of the regional government…and our center has offered tobacco cessation aid to workers since 2005…” this means that this programme is free of charge and that medication is free or reimbursed?-please specify this.
Response: Our cessation program has been supported by the Catalan government since its beginnings. For four years (2005 to 2008) tobacco cessation treatments were free of charge to all smoker hospital workers who were interested to quit. However, smokers who need treatment aids have to pay them out of their pockets. We have detailed more this information on the following paragraph on page 13 (reviewed document):
“Our comprehensive cancer center has offered tobacco cessation aid to workers since 2005, including behavioral support and free pharmacological treatment from 2005 to 2008. Afterwards, smoker workers should pay their own pharmacological treatment, and professional tobacco cessation consultation remains out of charge.”
20.In spain, besides Catalonia, other regions have implemented comprehensiveTC programs and SFPs in healthcare and have documented a downward trend over time in smoking, for example Galicia and Andaluzia (see Smoking tendencies in physicians and surgeons from Andalusia, M.J. Rius, ERJ abstracts conference 2013). Authors should compare their findings with them; and also with other international studies that have assessed trends over time (Fitzpatrick et al, 2009; Wheeler JG et al, 2007, Kannegaard N et al, 2005, etc), even if these observational studies have shorter time-frames.
Response: In the discussion we have included information comparing the results from the study of Fitzpatrick et al. with ours. Please see on the last paragraph on page 12
(reviewed document):
“A similar trend study conducted in Ireland showed also a striking decrease in smoking staff rates but with a stronger occupational gradient than in ours [41].
Our study reveals lower smoking rates among oncology nurses than among administrative and general population.
21. Although the baseline smoking prevalence was higher in females than in males, the downward trend observed over time was more notorious and more significant in females. The gender gap regarding smoking prevalence also narrowed over time. These findings are important and should be highlighted.
13

How do authors interpret this gender effect? Did authors implement gender-tailored TC programmes?
Response: Women represent about 75% of our hospital staff, and nurses about the 40% of our staff work power. As mentioned in the study
“Although, the hospital has not launched special campaigns addressed to them, we believe that the several cessation training programs addressed to nurses - who are 40% of our work force and from them 90% are women- may have had a stronger impact on quitting among this group. In addition, nurses smoking rates have had an appreciable decrease mainly after the passage of the 2nd law.
22.On page 13, when authors discuss the increase in occasional smoking and state that this pattern has occurred in general population as a consequence of strong TC policies; other studies specifically targeting HCPs have observed the same (MJ Rius, 2012;
Ravara et al, 2014).
Response: We appreciate the suggestion and we have included Ravara’s study in the following paragraph on page 14, line 12:
“However, this finding is in line with other studies that have reported a similar increase in the number of the occasional smokers in countries where tobacco consumption in the overall population is decreasing [52] especially among some
role model professions such as healthcare providers [53].”
23.While discussing the study findings and trends over time, the authors should highlight the additional impact of TC policies implemented at a national level during
2001-2012; also emphasise the importance of adequately promote and support cessation -assisted quitting is successful quitting; finally a policy-time-effect.
Response: We have devoted one paragraph to talk about smoking cessation programs on page 15. We argue about the importance of providing them in a regular basis and the most important barriers identified so far in healthcare services. We consider that this point has been already addressed and increase it will length the manuscript in excess. We have included the paragraph below: Page 14, line 15 (reviewed version):
“Smoking cessation care in hospitals continues to present a challenge in many organizations [10, 47, 48]. Worldwide, the most commonly-identified barriers to smoking cessation efforts include: lack of resources, knowledge, time, and support [55-57]. The deficit in adequate tobacco cessation knowledge starts at the university level. According to a recent study, few health sciences degrees include tobacco cessation training in their curricula [58]. Also, constraints on financial and staffing resources may threaten the suitability of innovative projects
[50,51]. As a result neither health professionals nor hospital administrators see providing tobacco cessation services as part of their responsibilities [59]. In our study, hospital workers as a group increased their support for hospital tobacco control policies; however, agreement about the exemplary role of health professionals is still lower than desirable.”
14