The introduction of SFL around Europe and beyond, has enabled a novel investigation of the effects of public health interventions. The collection of contemporaneous health and attitudinal data in the evaluation of legislative changes in Scotland and England has allowed an investigation of the relationship between workers’ attitudes towards an occupational intervention and their self-reported health changes.
Initial attitude did not have an effect on the change in symptoms reported by those in England. There was, however, a relationship between the change in reported respiratory symptoms and initial attitude in Scotland. The biggest improvement in respiratory symptoms, from P1 to P3, was reported by those who were initially negative towards the SFL.
We found that the initial attitude is more likely to be associated with the symptoms reported initially, with those who were initially more positive towards the legislation being more likely to report no symptoms than those who had a negative attitude.
Our study found no evidence that those who were more in favour of the proposed intervention were more likely to report greater improvements in their health a year later, indeed the opposite appears to be true. Initially this was thought to be affected by the difference in attitudes between smokers and non-smokers with smokers being generally more negative towards the proposed change and also having poorer reported health at the study baseline thus giving them more opportunity to experience health improvement than the more positive non-smoking participants. We found though that there was no association between smoking status and change in reported health.
The same interviewers collected information on attitudes and health at all phases of the study, within each country. One of these interviewers collected data in both the Scotland and England study. The same, straightforward, questionnaire was used for both studies and all of the interviewers were trained in its administration and advised to be neutral so as not to influence response. The protocol for carrying out the interviews was designed to minimise any interviewer effect on the responses obtained from the participants.
The fact that both studies used the same methods meant that the data could be amalgamated in order to examine the effects of the SFL in the UK. It also means that the results of any analysis are directly comparable thus allowing comparison of the attitudes towards the SFL and the effect this possibly had on changes in self-reported symptoms between the two countries.
At the time of implementation there was significant press coverage and public debate surrounding the SFL .This may have, potentially, affected a number of areas of this study. The high level of agreement that SFL was needed to protect bar workers’ health may be due to the comprehensive information campaign (for example; NHS Health Scotland, The Scottish Executive and Cancer Research UK) in the months leading up to the legislation. The generally positive experience of the legislation in Ireland may also have influenced Scottish bar workers’ expectations and attitudes towards smoking restrictions. The response rate of bars that were asked to join the study could possibly have been affected by the attitude of the bar managers towards the ban, with a lower proportion of bars in England agreeing to participate (18% England; 45% Scotland), but as data were not available on the reasons for non-participation this cannot be confirmed.
The follow-up study was hampered, to some extent, by attrition of a significant proportion of participants (65% in England and 49% in Scotland). As discussed by Ayres et al.
] bar staff often consist of students who only work during term-time/holidays and the populations in cities in general are typically transient. This could explain the higher rate of loss to follow-up in Edinburgh, Glasgow, London and Newcastle.
The SFL implemented in England and Scotland has been evaluated via a number of routes. From previous work on the bar workers in Scotland it was evident that smokers were more negative towards SFL, initially, than non-smokers [29
] and that the attitudes of the smokers underwent a greater improvement following the implementation of the legislation. It is likely that this may be due to smokers realising that the legislation did not have such a negative impact on them and their workplaces as they feared.
The bar workers were initially quite negative regarding the financial impact of the legislation. It did, however, appear that a high proportion of bar workers felt that SFL would make the bars more comfortable to work in and be better for workers’ health. So while they felt there was the possibility that the legislation would have a negative impact on business they did see that it was needed to protect their health.
There was no real difference in initial attitudes between England and Scotland. This is perhaps surprising, due to the publicity surrounding the effects that the Scottish legislation had on health and exposure to SHS in the year following implementation (the year leading up to its introduction in England). It might have been hypothesised that the positive experience of Scotland as a result of the legislation would improve the initial attitudes of those in England when the legislation was introduced a year later, but this does not seem to have been the case. The bar workers in England were, however, more negative, initially, towards the financial aspects of the legislation than those in Scotland. This could have been impacted by reports of bars closing in Scotland where the blame was given to SFL.
The results reported here examine health symptoms reported in the run up to the introduction of SFL and again a year later. Results were collected a year later in order to take account of seasonality, which would likely have an effect on symptoms being reported, due to the weather. Those with a cold at either phase were excluded from the analysis of self-reported health symptoms to attempt to exclude those who were suffering from any or all of these symptoms due to an illness.
In both countries symptoms reduced from P1 to P3, but there was a bigger decrease in England. It is possible that there is higher chance of having cold-like symptoms in the Spring (when Scotland data was collected) than in the Summer (England data collection), although the removal of those who reported having a cold should have reduced the impact this would have had. Previous analyses also demonstrated that smokers had more symptoms than non-smokers at P1 [21
] and so it is possible that these smokers thus had more capacity to experience health improvement than the non-smoking group.
The majority of studies investigating the effect of SFL around the world make use of self-reported health symptoms. Many of these studies also collected information on attitudes towards the legislation, but most do not report any investigation of the effect that attitudes may have on changes in self-reported health.
One study [23
] found that, among hospitality workers in Sweden, there was a larger decline in the prevalence of symptoms for those who felt more positive towards the legislation. They suggested that selection bias could have contributed to this as their study sample consisted of volunteers. This study used a very different measure for the attitudes and the participants were generally more positive towards the legislation being implemented. Another study investigating the impact of SFL in Norway reported that initial attitudes towards the legislation influenced subjective reports of economic effects of the SFL [26
]. This study of over 500
bar and restaurant workers found a negative pre-ban attitude towards SFL significantly increased the odds for reporting a negative economic impact post-ban.
It is clear that there are many factors which could affect the bar workers’ feelings towards SFL including their age and their own smoking habits. Their personal feelings towards SFL may have also been affected by positive and/or negative publicity concerning the health effects of SHS and SFL, as well as the possible financial effects to their workplace.
This study highlights the complex interplay of knowledge, attitudes and perceptions of changes to the working environment. Making workers more aware of the hazards associated with their working environment and of the potential benefits to their health and working life that a proposed intervention could have will all play a role in bringing about compliance with control measures.
In general, workplace interventions which are aimed at improving health and safety tend to be evaluated by looking at changes in health and behaviours using self-reported information, ideally in conjunction with some objective measure of change. Although we have reported here that the initial attitudes of workers tended not to introduce bias in the reported changes in health, it is clear that this should be considered when designing evaluation studies of complex occupational interventions.