The relationship between workers’ self-reported changes in health and their attitudes towards a workplace intervention – lessons from smoke-free legislation across the UK hospitality industry.

Background
The evaluation of smoke-free legislation (SFL) in the UK examined the impacts on exposure to second-hand smoke, workers’ attitudes and changes in respiratory health. Studies that investigate changes in the health of groups of people often use self-reported symptoms. Due to the subjective nature it is of interest to determine whether workers’ attitudes towards the change in their working conditions may be linked to the change in health they report.

Methods
Bar workers were recruited before the introduction of the SFL in Scotland and England with the aim of investigating their changes to health, attitudes and exposure as a result of the SFL. They were asked about their attitudes towards SFL and the presence of respiratory and sensory symptoms both before SFL and one year later. Here we examine the possibility of a relationship between initial attitudes and changes in reported symptoms, through the use of regression analyses.

Results
There was no difference in the initial attitudes towards SFL between those working in Scotland and England. Bar workers who were educated to a higher level tended to be more positive towards SFL. Attitude towards SFL was not found to be related to change in reported symptoms for bar workers in England (Respiratory, p = 0.755; Sensory, p = 0.910). In Scotland there was suggestion of a relationship with reporting of respiratory symptoms (p = 0.042), where those who were initially more negative to SFL experienced a greater improvement in self-reported health.

Conclusions
There was no evidence that workers who were more positive towards SFL reported greater improvements in respiratory and sensory symptoms. This may not be the case in all interventions and we recommend examining subjects’ attitudes towards the proposed intervention when evaluating possible health benefits using self-reported methods.