Aims: To compare different socio-demographic predictors of quitting smoking in a cohort representative of adult smokers in the British population using appropriate models that take into account the clustering of smoking behaviours at the household and area levels.
Design: A longitudinal, population representative survey of British adults (the British Household Panel Survey, BHPS) from 1991 to 2000.
Setting and participants: At wave 1 of the BHPS, 10264 adults living in 5511 households were interviewed. Around 30% of the wave 1 respondents reported smoking cigarettes. Of these, 21% had quit smoking over a 10-year period.
Measurements: 'Quitters' (quitting smoking) were defined as smokers who had subsequently described themselves as non-smokers for at least 2 consecutive waves (years) of the BHPS. Degree of dependence was indexed using the number of cigarettes currently smoked per day.
Findings: Degree of dependence was the strongest predictor of quitting smoking, followed by occupational social class, social support, marital status and the proportion of smokers in the household. There was some evidence of clustering of quitting smoking behaviour within households-members of the same household had similar quitting smoking behaviours. This clustering at the household level appeared to be explained by mechanisms related to the household level. However, there was little evidence for clustering of smoking behaviour within areas.
Conclusions: In addition to reducing addiction to cigarettes, policies designed to encourage smokers to quit may need to take into account factors related to household support and employment relations, in order to encourage smokers from all socio-demographic groups to quit