C Extent and Patterns of Use
ANNEX 1
EXTENT OF MULTIPLE DRUG USE

Multiple drug use is the rule rather than the exception among those who use drugs, whether their use is medical or non-medical, licit or illicit. Data drawn from Commission surveys in the spring of 1970 provide an overview of drug use in the Canadian population aged 12 and over in terms of seven classes of drugs: hashish and marijuana are combined as cannabis; 'pep pills' and 'diet pills' form a second group; 'sedatives', tranquilizers and 'sleeping pills' form a third group; LSD and 'other hallucinogens' form a fourth group; and tobacco, alcohol, and solvents are each treated as a distinct class.142, 143, 144 The Commission surveys did not gather any data on the use of opiate narcotic drugs, and consequently, there is no 'opiates' category in this classification. Our data do not address themselves directly to the hypothesis of 'progression' from one drug to another. Rather, they provide information regarding the context of what may be termed normal multiple drug use in the Canadian population.

A major difficulty with most of the data published on multiple drug use is that frequency of use is not taken into account. The regular user of several drugs is often lumped together with, for example, someone who has used the same substances only once. The number of respondents in the Commission surveys does not allow a particularly fine analysis of frequency of use. We have, however, distinguished use of a substance ever from use once a month or more in the last six months. For most substances, this latter frequency cannot be called 'frequent use'. For example, someone who uses tobacco, alcohol and cannabis, but each only once a month, cannot be described as a heavy user of any of these substances.

Our data involve six sets of patterns of multiple drug use: the two sets of patterns given by two different definitions of multiple drug use (`ever used' and used 'once a month or more in the last six months') for each of three populations (adults, college and university students, and high school students). Table C.8 shows the number of classes of substances used in each of these six sets. Tables C.9 to C.11 show the most common patterns of multiple drug use, and Tables C.12 to C.14 show the correlations between pairs of classes of substances.

The only general statement that can be made about these data is that the choice of combinations of substances is not at all random. There are 128 possible patterns of multiple drug use for the seven classes of substances. If respondents chose among these seven classes on a random basis, we would expect each pattern to represent about ten respondents in the high school and university samples and about 22 respondents in the adult sample. In fact, only one out of ten patterns contains at least these numbers of respondents. In all cases, the four most common of the 128 patterns account for approximately 50 per cent or more of the respondents in each sample, involving no more than three of the seven classes of drugs. In no case does it require more than the top 12 of the 128 patterns (involving no more than five of the drug classes) to cover 80 per cent of the sample. Multiple drug use, according to either level-of-use definition, and in any of the three populations, can thus be largely accounted for in terms of a relatively restricted number of patterns of multiple drug use. But the particular patterns that occur, and the proportions of the population that they cover, differ from sample to sample, and, equally importantly, differ in terms of whether one is using the 'ever used' or used 'once a month or more in the last six months' level-of-use definitions.

We are dealing with seven classes of drugs. Flow many are used by our respondents? Not surprisingly there is a sharp difference in all three samples between the number of classes of drugs ever used, and the number used on the average at least once a month in the last six months. Table C.8 shows that in the high school sample, 93 per cent of the respondents have never used more than three types of drugs, and 89 per cent only one type or none at all on a more regular basis. In the college and university and national adult samples, the numbers of drug classes used by 90 per cent of the sample or more are four 'ever' and two on a more regular basis.

These data indicate the importance of distinguishing levels-of-use. Even our relatively weak measure of levels-of-use halves the number Of respondents who are defined as multiple drug users when we move from those who have ever used any of these drug classes to those who have used the classes once a month or more in the last six months.

The most notable features of the national adult sample (see Table C.9), in terms of use once a month or more in the last six months, are: the primary position of 'no use' of any of the seven classes of substances (30 per cent of the sample); the secondary position of alcohol and tobacco, alone or in combination (a total of 48 per cent of the sample); followed by sedatives, tranquilizers or sleeping pills alone or together with alcohol or tobacco (16 per cent of the sample); followed by pep pills or diet pills, alone or together with sedatives, tranquilizers or sleeping pills, alcohol, or tobacco (three per cent of the sample).

Patterns of multiple drug use in the college and university sample (see Table C.10), with use defined as use once a month or more on the average in the last six months, are similar to those of the national adult sample, with one notable exception. Patterns involving cannabis appear where patterns involving sedatives, tranquilizers or sleeping pills are found in the adult sample. Table C.11 indicates that, in terms of more regular use, high school students are remarkable for their abstemiousness when compared to adults and college and university students. Furthermore, high school students remain primarily committed to the traditional non-medical drugs of our society, tobacco and alcohol. Only eight per cent of the sample uses any other substance on a relatively frequent basis, medically or non-medically, licitly or illicitly.

The probability of future use of any given drug is greater among persons who have at some time used a drug than among those who have not. This observation is represented statistically by a positive correlation coefficient—a numerical summary measure of the degree to which two quantitative variables are interrelated such that an increase in one variable is associated with a corresponding increase in the other variable. Tables C.12 to C.14 present matrices of point correlations for all possible pairs of drug classes in the three Commission surveys. Most of the pairs of drugs are positively correlated, with the degree of correlation varying from very weak to moderate. The highest correlation is 0.55, that between cannabis use and LSD use in the high school survey (see Table C.14). Correlations of this order, although indicating a strong relationship betwen two variables, do not suggest that the variables are so closely related as to make it possible to predict an individual's use of one substance on the basis of his use of another substance.

As we move from the high school survey to the college and university and national surveys, we note that corresponding correlation coefficients tend to be lower. High school students are at an age where they are developing adult patterns of drug use, and older students tend to have more experience with a variety of drugs than younger ones. Thus, the variation in age in this high school sample, correlated as it is to the use of most drugs, would account for most of the stronger relationships found in the high school sample than in the other samples.

We note further that correlations tend to be lower when we define drug use as use 'once a month or more in the last six months', rather than 'ever used'. The phenomenon of multiple drug use changes its character when we more strictly define the level-of-use of the substances involved. The relationships among the drugs tend to be diminished since there are proportionately fewer frequent users of any combination of drugs than there are persons who have 'ever used' these combinations. Consequently, it becomes less possible—rather than more possible—to explain the use of one drug in terms of the use of another drug. Hence, correlations that do not consider level-of-use data are not only unsophisticated but, more importantly, can often prove misleading.

The Commission data indicate that multiple-drug use is in fact normal drug use in our society. This observation has also been made by other researchers who have observed that all drug use is related to all other drug use, and that an individual's use of any one psychotropic substances makes more likely his use of any other psychotropic substance.28, 82• 190 The values of these correlations, however, are never so strong as to serve as adequate predictors of drug use.