The study is based on a secondary analysis of the longitudinal data of the National Population Health Survey (NPHS) of Statistics Canada which includes five cycles from 1994鈥?995聽to 2002鈥?003. A panel of 6585聽people from 15聽to 55聽years old and employed at cycle 1聽and nested in 1413聽neighbourhoods was selected. Multilevel models of regression were estimated on three levels: repeated measures (level 1 = 24,785聽observations) were nested in the individuals (level 2 = 6585聽individuals) and the individuals nested in the local communities (level 3 = 1413聽neighborhoods).
The prevalence of multiple episodes (two episodes and more between cycle 1聽and cycle 5) of psychotropic drugs use was 6.7%(95%CI=6.0鈥?.4%). Only occupation and the number of working hours showed a significant contribution. Family and individual variables like marital status and personality traits (locus of control and sense of coherence) had a significant contribution, in addition to time, gender, age, physical health, number of cigarettes and stressful childhood events.
Work contributes weakly to the risk of chronic psychotropic drugs use, whereas individual characteristics make a much more important contribution to the phenomenon.