Toward an integrative model for alcohol use and dependence

Background: If the alcohol use and dependence disorders are differentiated by the pharmaco-dependence, they share complex relationships with other clinical disorders and personality disorders. The purpose of this paper is to produce a model that reflects its relations both among users than among addicts. Method: Data from questionnaires measuring key variables selected for this study have been collected from people with alcohol misuse (n = 83) and alcohol-dependent (n = 81) in rehab. A model of drug dependence has been produced from these data that gives complete satisfaction to the criteria of SEM. Discussion: This model reflects the shift from abuse to a dependent consumption by the presence of feedbacks involving pharmaco-dependence, disturbance of the alcohol consumption by psychological distress and depressive traits. To further guarantee its validity, however it should be tested by collecting data from other surveys.

Alcoholic behaviours have been studied through the motives for drinking.They are implied in the disturbance of the alcoholic behaviours and can lead, as Goldsmith, Tran, Smith and How (2009) [10] have explained recently to major problems in alcohol drinking among younger people.Of course, other motives for drinking have been identified according to Pelc (1978) [11]: Search for social contacts, taste for alcohol, psychological problems, sensation seeking, and pharmaco-dependency.This last factor is one of the main differential diagnostic between alcohol use (F10.1) and alcohol dependence (F10.2.x).
The aim of this study is to produce a model of this main symptom that includes both the alcoholic behaviours and their components, clinical and personality features, and that can explain these among people with alcohol abusing or depending disorders.

METHOD
To measure the disturbance of alcoholic behaviour and the motives for drinking, we have used the Questionnaire of the Habits of Alcoholic Drink, devised by Pelc [11].This self-questionnaire explores 5 dimensions of the motives for drinking, as described above, and provides a score for the disturbance of the alcoholic behaviour.According to Pelc, a score higher than 24 reflects an abnormal drinking behaviour.To measure the intensity of the clinical features we have used the French version (Pariente and Guelfi, 1998 [12]) of Derogatis's Symptoms Check List 90r [13] that estimates the intensity of such clinical features as somatisation, obsession, sensitivity, hostility, anxiety, depression phobia, psychosis or paranoia.Finally, for to estimate the personality disorders we have used the French version (Cottraux et al., 1985 [14]) of the Dysfunction Attitude Scale (DAS) by Weisman and Beck (1978) [15].
Then, we have presented these questionnaires to a large sample (n = 81) of persons in detoxification for their alcohol-dependencies and to a large sample (n = 83) of people who drink alcohol in public houses, bars or night clubs.After randomization, we kept 40 protocols of the first group and 52 of the second one.
The data were processed using statistical software, SPSS ® and the step by step regressions method led us to a model of the pharmaco-dependency that we have computed in a structural equation modelling software (Amos ® ), following Arbuckle's criteria [16] for validation of structural model: non significant chi-square; GFI and CFI closed to 1, RMSEA closed to 0.

RESULTS
The global outcomes are summed up in the following table (Table 1

):
Since the distribution of the outcomes on the DAS of Weisman and Beck is not normal, the intensity of the personality disorders could not be computed to our model.Regarding the comparison of those groups, one can note that, apart the item "drinking for searching social contact", all the variables are significantly different between the persons with an alcohol dependence and the alcohol users whose the mean of the disturbance of the alcoholic behaviours is higher than 24.This suggests a problem with alcohol drinking (alcohol abuse) among this sample.If we focus on the results for pharmacodependence (48.33 versus 0.44), it is clearly higher among the alcohol-dependent persons than among the alcohol abusers, that confirms what we expected.
If we follow the step by step regression method from the score of pharmaco-dependency, one can see that the level of the disturbance of the alcoholic behaviour is at first implied in the two groups, then if we continue, one can see that "drinking for sensation seeking" is implied in the score of pharmaco-dependency.Eliminating this variable leads us to a non-significant result for the both groups.Then, if we study the "disturbance of the alcoholic behaviour", three variables are implied in the groups: Sensation seeking and anxious and depressive features.The "sensation seeking" is explained by anxious features which are themselves influenced by the depresssive features.The study of the depressive features leads to an effect of "drinking because of psychological problems" that is explained first by the "disturbance of the alcoholic behaviour" and secondly by the "pharmacodependence".
Thus, we arrive at the following model (Schema 1) which provides for the two groups the conditions of the validity of a structural equation model: non-significant χ 2 , GFI and CFI closed to 1, RMSEA closed to 0. Also we can argue that to differentiate abusers of alcohol and dependent persons the intensity of the variables is more important than their combinations.But if the differences between alcohol using and dependency are a matter of intensity how one can explain that the abusing turns to dependency?
According to our outcomes, the sequence "pharmacodependence", "drinking alcohol because of psychological problems", "depressive features" and "disturbance of the alcoholic behaviour" is a retroaction that may explain the shift from a mean score of 0.44 among heavy drinkers in a score of 48.33 recorded in alcohol-dependent individuals (cf.Table 1).Another retroaction can be drawn that implies "anxious features" and "sensation seeking" that also can increase the pharmaco-dependence.

DISCUSSION
This model needs to be validated against other data with larger groups or other questionnaires.Therefore, if we test it with the data of the two groups that we left because of the initial randomization, the criteria of validity are still significant.
According to the model, there is a transition from alcohol abuse to alcohol dependency, or in other words, dependency is a form of alcohol abusing that does not need more variables, or factors, to explain.
If we focus on the relationships between alcoholic behaviours and depressive features the feed-back in this model lets us answer to the question of the primary or secondary depressive syndromes among alcoholic people.Depression is primary to disturbance of alcoholic behaviours and has direct and indirect (mediated by anxiety and sensation seeking) to the disturbance of the alcoholic behaviour; on the other hand, through the mediation of the drinking because of psychological problems, the depressives features are induced by the alcoholic behaviours and appear as secondary to alcohol related problems.Therefore, the depressive features are not the central basis of this model.If we control their intensity by using the partial correlation method, the correlations implied in this model keep their significance.
The model that can be constructed from our data leads us to the conclusion that the intensity of pharmaco-dependence depends on the intensity of two main variables: level of the disturbance of alcohol relationship and consumption of alcohol for sensation seeking.The feedback from the "disturbance of the alcohol relationship" to depressive features (mediated by the consumption of alcohol because of psychological problems) that involves the disturbance of the alcoholic behaviour directly and indirectly contributes to increasing pharmaco-dependence.
But how does alcohol abuse turn to dependence?This model let us understand the transition from an abusing drinking (m = 35.65) to a dependent consumption of alcohol (n = 70.32)thanks to the indirect incidence of the "alcohol drinking because of psychological problems" on the pharmaco-dependence (0.44 among the groups of alcohol users against 48.3 in the alcohol dependent group) and the retroaction of pharmaco-dependence on "alcohol drinking because of psychological problems".
Finally, the data confirm the incidence of the disturbance of the alcoholic behaviour on the alcohol-dependency.If we focus on the persons whose score of alcohol problems is higher than 24, our model becomes nonsignificant.Therefore the size of this sample (n = 21) does not allow us to get further.Also it would be interesting to look for the critical point where disturbance of the alcoholic behaviour turns a "normal drinking" into dependant alcoholism.
Other works especially using specific questionnaires for each factor should confirm this study.Such a data collect would let us re-test the validity of this integrative model which deals with both majors forms of alcoholic disorder: abusing and dependency.

Table 1 .
Global outcomes of this study.