JUMPING TO CRACK CONCLUSIONS

One of the caveats against a harm reduction approach to drug abuse is
that you have to be pretty sure the steps you are taking will actually
reduce harm. Researchers in Vancouver have taken an audacious step
this week by calling for crack cocaine use to be considered an
independent source of HIV risk, based on a new local study in the
Canadian Medical Association Journal, and recommending the possible
adoption of "safer crack kits" and supervised inhalation rooms at
existing drug-injection sites as strategies for reducing the spread of
the virus.

This will not, it should go without saying, make the debate over
Vancouver's existing harm-reduction infrastructure any less
complicated than it already is. The federal government is fighting in
court for the ability to shut down the city's Insite program; now, at
the same moment, many of Insite's traditional defenders are
effectively suggesting that the program ought to be expanded, in order
to deal with an increasingly prevalent crack problem that is more
dangerous to public health than previously thought.

The CMAJ study was a retrospective re-analysis of the nonrandom sample
of participants in the longitudinal Vancouver Injection Drug Users
Study (VIDUS) which began in May 1996. All the participants are people
who had used some illicit injectable drug less than one month before
entering the study; the researchers broke out the 1,048 who were
HIV-negative when enrolled, and compared the daily crack users going
forward with the other drug users. What they found was that, after
controlling for other possible confounding variables, the daily crack
smokers appeared to be 2.7 times as likely to acquire HIV over the
next nine-and-a-half years as the other junkies.

The harm reductionists will, as is their habit, declare that this
means the evidence is in and that human decency demands the policy
interventions they propose. We are prepared to agree with the core
logic of their position: If the evidence truly supports a
harmreduction approach to crack, we should not let squeamishness or
superstition, or some idea that crack has subtle moral differences
from heroin and other injectable drugs, keep us from trying it. We
would like it if they would recognize that any harm-reduction step
always carries attendant dangers of encouraging or facilitating abuse,
however insubstantial, and that the tricky calculus of giving crack
users free crack pipes and a place to use them may have subtle
differences from that underlying Insite's established work.

The real problem is that their evidence, while raising a scary
possibility that needs to be investigated more closely -- namely, that
crack use is an independent risk factor for HIV on the same order as
drug injection -- is not yet a slam dunk. They leap from their
statistical findings to the suggestion that crack may facilitate HIV
seroconversion because crack pipes "produce wounds in and around the
mouth" which would create an infection route during subsequent
pipe-sharing or oral sex. But the caveats are abundant.

As the researchers acknowledge, they didn't test for this precise
hypothesis directly; VIDUS participants weren't checked for oral
health. There exists the obvious possibility that crack users may be
more prone to HIV primarily because they have more sexual partners and
more unprotected sex of other kinds. The only sexual-behaviour
variables the VIDUS data allowed the authors of the study to control
for were participation in sex work and unprotected sex -- as in, "Have
you had unprotected sex at least once in your entire life?"

An editorial commentary attached to the paper in CMAJ rightly
describes this as a "crude simplification" that "may have eliminated a
possible significant association." It notes the further logical
problem that "oral sex has not been shown to be significantly
associated with HIV transmission" in any way. And the statistical
power of their subset of VIDUS participants was not great. (The
95%-confidence limits on either side of that figure of 2.7 reach all
the way from 1.06 to 7.1, so there is still a fairly significant
chance that crack use alone has only a slight effect on HIV
vulnerability.)

The "sex-for-crack" lifestyle is already known to be one that
increases HIV/AIDS risk; the Vancouver researchers have really told us
nothing that is not already consistent with that fact of life. The
question is how much of that additional risk is the result of their
oral-health hypothesis. If the answer is "not very much," then
harm-reduction measures specifically intended to protect the oral
health of crack users won't accomplish much. We would need to consider
an intervention that stops crack addicts from having sex with many
strangers to obtain crack or money for crack. To which one is tempted
to add a cynical "good luck."