Kevin Sabet’s recent Washington Post op-ed, co-authored with Charles Fain Lehman, on new data undermining the central premise of so-called “safe” injection cites generated a lot of interest–and some pushback. (For a fuller picture of the data, check out the FDPS whitepaper Kevin and Charles co-wrote on the issue.)
Part of the pushback came from Professors M-J Molloy and Dan Werb of the University of British Columbia and UCSD, respectively:
Charles Fain Lehman and Kevin A. Sabet’s April 21 op-ed, “‘Safe drug sites’ don’t work. The data proves it.,” asserted that a “large body of research” has found no effect of these sites on overdose risk. As public health researchers whose studies evaluating supervised consumption facilities were reviewed in the op-ed, we disagree. For example, unmentioned is one study that tracked nearly 1,000 of Vancouver’s highest-risk drug users for more than 10 years and found that frequent users of the city’s main supervised consumption facility had an approximately 50 percent lower risk of death after accounting for differences in age, gender and drug use patterns. Also unmentioned is the study estimating the combined impact of these facilities, free access to medications to reverse overdoses, and opioid dependence treatment, which found that more than 3,000 overdose deaths were averted in the first year of our current overdose crisis. The authors also did not acknowledge data from Toronto showing that, over a two-year period, neighborhoods with supervised consumption facilities experienced a lower rate of overdose deaths over time.
These facilities have seen millions of high-risk drug use events over the past two decades but not a single death by overdose. For that reason alone: How can they not have a place in our response to this crisis?
The authors bring up three studies.
The first study compares people who use an injection site more versus less. Naturally, the people who use SIFs more could be better connected to community resources, less afraid to use emergency services, and/or more careful with their drug use. This data tells us more about what kinds of people frequent injection sites than whether injection sites reduce overdoses.
The second says nothing direct about the value of injection sites. It’s a modelling exercise where the authors estimate how many overdoses that happened in an injection site would have been fatal otherwise. Essentially, their argument is that if you’re going to overdose, it’s better to be at an injection site than not. That’s may well be true but it is not the relevant question. The question is whether those overdose reversals are or are not outweighed by increased overdoses–that study calculates the benefits without considering the potential costs.
The third study is the closest to telling us anything useful about whether injection sites help or hurt overdose rates, but it’s missing an essential piece. It assumes that the control and treatment areas were on parallel trajectories of overdoses had it not been for injection sites being present, but provides no evidence to support that assumption.
So, yes: injection sites have reversed overdoses.That observation is true but misleading–and it doesn’t mean they save lives. Indeed, if injection sites increase overdose rates by encouraging or enabling drug use, and that increase outweighs any directly reversed overdoses, then they could increase mortality.