A recent review paper covers animal models of adolescent drug taking, which is in and of itself an interesting read. Human adolescence tends to be a time when people first encounter psychoactive substances taken for recreational purposes. Unsurprisingly, problematic drug taking which emerges later in life often has antecedent roots in adolescent drug taking. The epidemiology goes further in identifying age-graded risk such that the earlier one starts using some drugs, the greater the chances of problematic drug use later in life.
Given the inescapable limitations human epidemiology (lack of random assignment means you cannot eliminate risk-associated variables from the genetic to the environmental) animal models are required to determine if there are neurobiological sensitivities in the adolescent brain which confer increased risk of developing dependence on a given recreational drug. This paper reviews much of the animal studies to date.
In providing the background context, Schramm-Sapyta et alia did something dear to my heart, as my readers will quickly appreciate. They created an interesting graphical depiction of the conditional probability of dependence on several drugs from the US National Survey on Drug Use and Health (NSDUH, 2007).
Fig. 1 Percentages of the US population over the age of 12 years who have ever tried the indicated drug (top number, light gray circle); who used the indicated drug in the past month (middle number, darker gray circle); who meet criteria for dependence on the indicated drug (bottom number, black circle). Numbers in the center of each diagram represent the percentage of people who have ever tried the indicated substance who are currently dependent. Data obtained from the NSDUH 2007, lifetime use, past month use, DSM-IV dependence criteria (for all drugs except tobacco), and daily cigarette use (for tobacco)
Yes, these are cross-sectional data much like the Anthony et al 1994 paper to which I frequently refer. But these are also large representative sample data so trite dismissals on this basis are hard to support.
What draws my eye? Two things. First, of course, is the number in the center of each set of concentric circles. This represents the percentage of the sample of individuals who have sampled a given drug in their lifetime who are currently dependent. I emphasize "currently" because sometimes you will see data on "ever met dependence criteria"; this latter would result in a higher estimate of dependence risk, obviously. The Anthony et al 1994 data use lifetime rates, not current rates for example. Nevertheless this central number is another estimate of the conditional probability of becoming dependent and more importantly gives us relative relationships for several drugs of interest. [ As always in science, the more replication the better. Slightly different sampling / survey methods will give us different numbers of course. So let us not make the common mistake of viewing every scientific paper in isolation. These numbers need to be integrated with other estimates of conditional probability of dependence that we can find...in a thoughtful way. ]
The second thing that draws my attention can be found in the three numbers listed to the left of the circles. The bottom one represents the rate of dependence in the general population. To my view this is a critical bit of information for those who want to have an informed discussion of public health related policy. The number of citizens affected by a given condition is as assuredly relevant as is the severity of that condition.
These data make it emphatically clear that the number of people in the US who are currently dependent on cannabis slightly exceeds the number who have ever so much as tried heroin. The number dependent on cannabis is three times as large as the number dependent on cocaine and 16 times as large as the number of people dependent on heroin.
h/t: dr_leigh for passing this along