[Substance use, affective problems and personality traits: test of two association models]

The International Consortium of Psychiatric Epidemiology has confirmed the high comorbidity in community-drawn samples between substance use disorders and anxiety or depression. In the same way, associations between substance use and specific personality traits (such as novelty seeking, harm avoidance or antisocial personality) have also been extensively documented. Self-medication and social deviance are among the most commonly evoked explanatory models for these forms of comorbidity, and are based on findings that affective disorders and specific personality traits often precede the onset of substance use disorders. The self-medication model postulates that an individual chooses a specific substance according to its psychopharmacologic action on the given psychological state of the person. By contrast, the social deviance model posits that this form of comorbidity is due to the fact that persons consuming certain substances may have affective or personality characteristics that are more severe or more deviant than non-consumers (or than consumers of socially well-accepted substances). In this way, the individual does not use a particular substance to assuage pre-existing disorders but, due to a more deviant personality, is less influenced by social norms and may more easily turn to using illicit substances or to polyconsumption. However, a major limitation of the current scientific literature concerning tests of these models is that previous investigations have been based in overwhelming majority on clinical populations. The examination only of clinical samples renders difficult the identification of causal (or primary) risk factors for the emergence of substance use disorders from the potential consequences of substance use itself. The goal of the current study was therefore to simultaneously compare both models of association using a non clinical population of substance users. In addition to selecting subjects based on use (rather than abuse or dependence), multiple comparisons were corrected with a Bonferroni adjustment.

A two-phase sampling plan was used with post-stratification on substances use. In the first stage, an initial sample of 685 students was pre-selected based on responses to a battery of self-questionnaires, including information concerning recent consumption of substances (alcohol, cannabis, cocaine, heroin, acid, solvents, and so on), anxiety levels measured by the State and Trait Anxiety Inventory (STAI, Spielberger, 1983) and depression levels evaluated by the Center of Epidemiologic Studies Depression Scale (CES-D, Radloff, 1977). Based on responses to these questionnaires, 98 subjects were selected in the second phase to compose four groups of substance users: non consumers (those who did not use any substance during the last month); consumers of alcohol only, consumers of cannabis (with or without alcohol) and consumers of other illicit substances (with or without cannabis or alcohol). These subjects were then invited to participate in a brief laboratory-based meeting where they completed the Temperament and Character Inventory (TCI, Cloninger, 1992), which assessed different personality characteristics such as novelty seeking (NS), harm avoidance (HA) or antisocial personality disorder (APD).

The hypotheses concerning self-medication were tested by multiple logistic regression by comparing each group of substance consumption to the non-consumer group relative to levels of anxiety, depression and scores of novelty seeking and harm avoidance. The social deviance model was tested by ANOVAs using contrasts which allowed for a test of a linear tendency across the four groups of consumption relative to each of the personality traits (novelty seeking, harm avoidance and antisocial personality).

Results of multiple logistic regressions showed no difference between non-consumers and any group of consumers with regard to anxiety, depression and harm avoidance. However, consumers of other illicit substances significantly differed from non-consumers for novelty seeking trait (qOR=8.4; p<0.05). Results of the ANOVA also showed no differences between the four groups with regard to scores of harm avoidance and level of antisocial personality but again a comparison of novelty seeking scores was significant, F(94)=6.46, p<0.05. Moreover, the contrast method demonstrated that novelty seeking scores increased linearly and significantly (p<0.001) from the group of non-consumers to the group of the consumers of the most deviant substances.

The results obtained in this non-clinical sample are in favor of social deviance model which posits that the personality trait of novelty seeking is associated to the consumption of the most illicit and deviant substances (such as heroin or cocaine). On the other hand, no support was found for the hypothesis of self-medication which assumes that specific substances should be particularly associated with specific psychological characteristics or vulnerabilities.