Cigarette smoking in an adolescent psychiatric population

To examine the relationship between cigarette smoking and psychiatric symptomatology in an outpatient psychiatric population of adolescents.

A retrospective analysis was done of 934 patient charts at an outpatient psychiatric centre.

48.4% of the psychiatric sample reported regular smoking behaviour, which is substantially more than the 18.1% prevalence found in a local epidemiological study. In comparing smokers and non-smokers within the psychiatric sample, it was noted that smokers were significantly younger and scored somewhat higher on depression rating scales than non-smokers. A logistical regression, using quasi-Newton estimation, was chosen as the most suitable statistical method for building a classificatory model of smoking. Two continuous variables, age and the Hamilton depression score, along with 39 discrete variables, were chosen for modelling purposes. Model building was conducted in a hierarchical fashion, starting with demographic variables, the variable selection being controlled by using chi-square tests of model differences. A predictive model of smoking with nine variables was finally selected.

As a whole the results support the strong association between smoking and psychiatric problems, but in this adolescent sample smoking is more likely to be part of a general risk-taking behaviour pattern than an attempt to medicate depression. Anti-tobacco campaigns that highlight the risks of smoking are therefore open invitations for adolescents to take up the habit.