Method
Results
Discussion
Bibliography
Annex
Author: Reginald G. SMART, Paul C. WHITEHEAD
Pages: 39 to 47
Creation Date: 1972/01/01
Common concern attaches to all aspects of illicit drug use. This concern has been shown by the increase in drug use surveys (see Berg, 1970) and the development of treatment facilities for drug abuse cases. What is still uncertain are the preventive measures for drug abuse. Most preventive programmes are based on public or specialized education programmes (Ungerleider and Bower, 1969; Smith, Kline and French, 1967). However, even where these programmes have been extensively used as in certain high school populations, both experimental and heavy drug use is still increasing (e.g. Smart, Fejer, and White, 1970). So far, efforts to prevent drug use and abuse have not been based upon knowledge of the drug using population. Epidemiological aspects of use although well known have not been employed in constructing preventive programmes. This paper argues that the distribution of drug use suggests that prevention of drug abuse can only be achieved by per capita reductions in drug use by the population at large. Efforts to reduce drug abuse directly without reducing per capita drug use will fail, according to these principles.
An earlier paper (Smart and Whitehead, and Laforest, 1971) showed that five large scale surveys of drug use generated log 'normal distributions when the total drug use of each user is considered. The log normal distribution is a smooth one with no discontinuities. In it, infrequent users are most numerous, moderate users less numerous, and heavy users are least numerous. It might be considered that this distribution, with different means in four cities at different points in time, gives a picture of total drug use under many circumstances. This indicates that to reduce the proportion of heavy users it will be necessary to reduce the per capita consumption of drugs in the whole population of users. The data available, so far, on this proposition are for total drug use among users. The present study is concerned with the distribution of consumption of each of the drugs separately.
* A revised version of this paper was presented to the International Conference on School Drug Use Surveys, Newark, New Jersey, 1971.
It is of considerable interest to know whether different drugs have the same distribution of consumption. If drugs with such varying effects as marihuana, alcohol, LSD, tranquillizers, etc., have the same pattern of consumption, this suggests a general drug using phenomenon. It leads to a consideration of the similarities in users of various drugs and to an interest in whether the same factors predispose to the use of different drugs. Also, it has a bearing on problems of prevention. If all types of drug use show differentdistributions of consumption then all will probably require different techniques of prevention. However, if the various mood-modifying drugs do not vary in their distributions, a single type of preventive programme can be considered. Essentially it could be argued that "speed" 1 abuse and hashish abuse constitute similar problems and that similar techniques of preventing them would be effective.
The aim of the present paper is to examine the distribution of drug use for each of the following drugs in three different high school populations: marihuana, LSD, solvents barbiturates, tranquillizers, speed and stimulants other than "speed". Students were asked to state on exactly how many occasions they had used each of the above drugs.
Data on the frequency of drug use were obtained from surveys conducted in Toronto, Halifax, and the Niagara counties in 1970. The details of these investigations have been reported elsewhere (Smart, Fejer and White, 1970; Smart, Fejer and Alexander, 1970; and Whitehead, 1971) and only a few relevant details will be given here. All surveys were based on large, carefully chosen samples.
1. Generally understood to mean methamphetamine.
The sample sizes were 8,568 in Toronto, 5,900 in the Niagara counties, and 1,526 in Halifax. In total, some 15,994 students were involved in the studies. The Toronto sample was chosen to include 120 students from grades 7, 9, 11 and 13 in 20% of the high school districts in Toronto. The sample in Halifax was a 25% random sample of classes from grades 11 and 12 and a 12.5% random sample of classes from grades 7-10. In the Niagara counties a 25 % sample of all classes from grade 9 to 13 was drawn.
The data for this study were obtained in questionnaire investigations of drug use. Separate questions were asked about the use of marihuana, LSD, solvents, barbiturates, tranquillizers, "speed" and stimulants other than "speed ". Drugs such as opiates, alcohol and tobacco were omitted from these questions; the first because they are used by so few students and the others because information about their frequency of use has been so often collected.
The data for each type of drug use were arranged separately in a frequency distribution. This paper presents only the data on the number of times each drug was used by those who were users. The non-users of each drug have been left out of the analysis. This means, of course, that the total number of persons in the survey do not appear, only the numbers of users of each drug. These distributions of frequencies of use are shown in the Annex. It can be seen that these distributions are continuous and that they fit the general form of the log normal expectancy.
Measures of skewness were computed for all distributions in order to determine the fit to the log normal distribution. Skewness refers to the amount and character of the dispersion in the data, i.e., the extent to which the data are clustered together or spread out. Measures of skewness below .20 are taken (Croxton and Cowden, 1955) to indicate a fit to the log normal expectancy.
Toronto |
Halifax |
Niagara counties |
|
---|---|---|---|
Marihuana
|
.29
|
.14
|
.10
|
LSD
|
.16
|
.04
|
.17
|
"Speed"
|
.06 |
.22
|
.06
|
Solvents
|
.11
|
.08
|
.14
|
Stimulants
|
.16
|
.15
|
.0l
|
Tranquillizers
|
.01
|
.03
|
.04
|
Barbiturates
|
.07
|
.19
|
.09
|
The measures of skewness are shown in the table. It can be seen that for only two of the distributions is the measure of skewness over .20. For both of these distributions, the major discontinuity is in the piling up of responses at the frequencies of 50, 100 and 150, indicating some difficulty in remembering large numbers of drug experiences. All of the distributions for LSD, solvents, stimulants, tranquillizers and barbiturates fit the log normal expectancy. Only marihuana use in Toronto and "speed" use in Halifax do not fit, although both are very close to fitting.
The results of the present investigation show that, in general, the log normal distribution describes psychoactive and hallucinogenic drug use of many types. These findings apply to high school students in several Canadian cities. The drugs involved are LSD, solvents, barbiturates, tranquillizers, stimulants, marihuana and "speed ". It would be ideal to demonstrate that similar distributions occurred in different countries at different times, thus establishing the generality of the findings; however, such data are not yet available though we are attempting to obtain them.
The findings indicate that with two minor exceptions the 21 sets of drug use data describe log normal distributions. The extent of use varies from place to place for different drugs but the general character of consumption does not. This suggests a strong link between the various types of drug use, both legal and illegal. The psychoactive drugs such as tranquillizers, barbiturates and stimulants are most often acquired on prescriptions whereas solvents represent a "legal" high though never prescribed. Possession and sale of marihuana and LSD are both illegal at present in Canada, as is the case with the stronger forms of opiates such as heroin. It is also remarkable that drugs with such different effects as stimulants and barbiturates, marihuana and "speed" should have essentially the same consumption patterns.
It has been argued elsewhere (Smart and Whitehead, and Laforest, 1971) that to reduce the abuse of drugs for which consumption is log normal it will be necessary to reduce per capita drug consumption. O'Neill and Wells (1971) have shown that log normal curves do not necessarily require that the proportion of heavy users depend upon the average or per capita consumption value. However, it has been shown (Schmidt and DeLint, 1971) that the frequency of problems due to alcohol varies directly with the level of per capita consumption. This suggests that per capita alcohol use may have to be decreased in order to decrease problems from alcohol use (such as liver cirrhosis). Similarly, there may be no way to cut off the heavy use part of the drug use distribution without lowering the whole level of the curve and reducing average consumption. Demonstration that per capita consumption must be reduced before heavy consumption is reduced depends upon empirical studies which have not been attempted as yet.
The data also indicate that the problem of prevention is similar for the seven psychoactive and hallucinogenic drugs studied here. If a preventive programme based on modifying the distribution works for one of the drugs, e.g. marihuana, it is likely to work for all of them. The present study suggests that any legal or social measures which reduce drug consumption by large numbers of persons will lower average drug use and eventually reduce heavy use. Such measures might involve removing some drugs from the pharmacopoeia or limiting pres criptions for psychoactive drugs. For illegal drugs it might involve reducing experimentation and recreational drug use by campaigns of persuasion and information of the abuser population, supported by more strict enforcement.
The data also show a total lack of bi-modality in the distributions. There is no clear differentiation of normal users and abusers in terms of frequency of use. Characterizing any particular part of the distribution as entailing "abuse" or "excessive" use will be arbitrary. A series of studies could be made, as have been done with alcohol (Schmidt and DeLint, 1970) to establish what types of pathology are associated with what levels of consumption.
Berg, Dorothy. " Extent of illicit drug use: a compilation of studies, surveys and polls." Unpublished report. Bureau of Narcotics and Dangerous Drugs, 1970.
Croxton, F. R. and Cowden, D. J. Applied General Statistics. Englewood Cliffs, Prentice-Hall, 1955.
O'Neill, B. and Wells, W. T. "Blood alcohol levels in drivers not involved in accidents and the log normal distribution." Quarterly Journal of Studies on Alcohol, 32 ,1971, 798-803.
Schmidt, W. and DeLint, J. E. "Estimating the prevalence of alcoholism from alcohol consumption and mortality data." Quarterly Journal of Studies on Alcohol, 31 (December): 957-964.
Smart, Reginald G., Fejer, Dianne, and White, J. "The Extent of Drug Use in Metropolitan Toronto Schools: A Study of Changes from 1968 to 1970." Addiction Research Foundation, Toronto, 1970.
Smart, Reginald G., Whitehead, Paul C., and Laforest, Lucien. "The prevention of drug abuse by young people: an argument based on the distribution of drug use." Bulletin on Narcotics, XXIII (June): 11-15.
Smart, Reginald G., Fejer, Dianne, and Alexander, Eileen. "Drug Use Among High School Students and Their Parents in Lincoln and Welland Counties." Addiction Research Foundation, Substudy 1-7 & Jo & AI-70, 1970.
Smith, Kline and French, "Drug Abuse: Escape to Nowhere: A Guide for Education." National Educational Association, 1967.
Ungerleider, J. T. and Bower, H. L." Drug abuse and the schools." American Journal of Psychiatry, 1969, 125, 1691-1697.
Whitehead, P. C. "The epidemiology of drug use in a Canadian city at two points in time: Halifax, 1969-1970", pages 520-533, in Boydell, Craig L, Grindstaff, Carl F., and Whitehead, Paul C. (eds.) Critical Issues in Canadian Society. Toronto: Holt, Rinehart and Winston, Ltd., 1971. ( British Journal of Addiction, in press.)