Smoking behaviour predicts tobacco control attitudes in a high smoking prevalence hospital: A cross-sectional study in a Portuguese teaching hospital prior to the national smoking ban

Background: Several studies have investigated attitudes to and compliance with smoking bans, but few have been conducted in healthcare settings and none in such a setting in Portugal. Portugal is of particular interest because the current ban is not in line with World Health Organization recommendations for a "100% smoke-free" policy. In November 2007, a Portuguese teaching-hospital surveyed smoking behaviour and tobacco control (TC) attitudes before the national ban came into force in January 2008. Methods: Questionnaire-based cross-sectional study, including all eligible staff. Sample: 52.9% of the 1, 112 staff; mean age 38.3 +/- 9.9 years; 65.9% females. Smoking behaviour and TC attitudes and beliefs were the main outcomes. Bivariable analyses were conducted using chi-squared and MacNemar tests to compare categorical variables and Mann-Whitney tests to compare medians. Multilogistic regression (MLR) was performed to identify factors associated with smoking status and TC attitudes. Results: Smoking prevalence was 40.5% (95% CI: 33.6-47.4) in males, 23.5% (95% CI: 19.2-27.8) in females (p < 0.001); 43.2% in auxiliaries, 26.1% in nurses, 18.9% among physicians, and 34.7% among other non-health professionals (p = 0.024). The findings showed a very high level of agreement with smoking bans, even among smokers, despite the fact that 70.3% of the smokers smoked on the premises and 76% of staff reported being frequently exposed to second-hand smoke (SHS). In addition 42.8% reported that SHS was unpleasant and 28.3% admitted complaining. MLR showed that smoking behaviour was the most important predictor of TC attitudes. Conclusions: Smoking prevalence was high, especially among the lower socio-economic groups. The findings showed a very high level of support for smoking bans, despite the pro-smoking environment. Most staff reported passive behaviour, despite high SHS exposure. This and the high smoking prevalence may contribute to low compliance with the ban and low participation on smoking cessation activities. Smoking behaviour had greater influence in TC attitudes than health professionals' education. Our study is the first in Portugal to identify potential predictors of non-compliance with the partial smoking ban, further emphasising the need for a 100% smoke-free policy, effective enforcement and public health education to ensure compliance and promote social norm change.