Naloxone and opioid deaths in the U.S.

Two weeks ago, I wrote a post about vaccinations against drug addiction. Some readers might have been sceptical that introducing such a vaccine "might reduce drug addiction, but come at the cost of more drug overdoses". The mechanism I suggested there was that drug addicts would be more likely to overdose by accident if they had received the vaccine, because they would require a higher does in order to get high.

Well-meaning interventions can often lead to such unintended consequences. In a related case, I offer this 2018 working paper by Jennifer Doleac (Texas A&M) and Anita Mukherjee (University of Wisconsin-Madison). Doleac and Mukherjee looked at the impact of increasing access to Naloxone on opioid-related deaths in the U.S. Naloxone is "a drug that can reverse an opioid overdose if administered quickly". So, the availability of Naloxone reduces the risks of consuming opioids, meaning that drug users can use more, or use opioids more riskily, with the knowledge that their risk of overdose has been reduced. This is an example of a Peltzman effect (which I have previously discussed here and here and here) - when you reduce the risks associated with risky behaviour, people engage in more of that behaviour.

Doleac and Mukherjee indeed find evidence of offsetting behaviour:

After naloxone access laws take effect, Google searches for "drug rehab" (a proxy for interest in drug treatment4) fell by 1.4%, arrests for possession and sales of opioids increased by 17% and 27%, respectively, opioid-related visits to the emergency room increased by 15%, and opioid-related theft increased by 30% (though the effect on overall theft rates is much smaller). Meanwhile, expanding access to naloxone had no effect on opioid-related mortality, on average.

In other words, after Naloxone became readily available, people responded by reducing their demand for drug treatment (why seek treatment when opioids are now less dangerous), increasing their demand for opioids (once they realised that opioids were less risky, people consumed more), and increased opioid-related thefts (in order to fund their increased drug habit). The net effect on opioid-related deaths was close to zero. The authors propose two channels to explain the offsetting behaviour:

First, the reduced risk of death makes opioid abuse more appealing, leading some to increase their opioid use - or use more potent forms of the drug - when they have naloxone as a safety net. Some of those abusers may become criminally active to fund their increased drug use. Second, some opioid abusers are saved by naloxone and may continue their previous drug use and criminal behavior (this intended effect of naloxone will mechanically increase rates of both behaviors).

Doleac and Mukherjee also do a great job of testing the robustness of their analysis, which holds up quite well. There are also regional differences, and differences in the effects between areas that have greater access to drug treatment and those that have less access. Specifically:

...naloxone access laws increased mortality more in places with fewer drug treatment facilities per capita, or more limited eligibility for Medicaid (which covers substance abuse treatment). In other words, easier access to treatment is associated with more beneficial policy effects.

That provides a useful policy prescription for complementing Naloxone access and reducing the negative impacts: provide more access to drug treatment. However, whether that would work requires some further investigation, since Doleac and Mukherjee show (as noted above) that increased Naloxone access reduced the demand for drug treatment. So, increasing supply of drug treatment by itself might be ineffective.

In any case, as this research shows, combating drug-related deaths is clearly not an easy fix (pun intended!).

[HT: Eric Crampton at Offsetting Behaviour, last year]

Read more:

• This couldn't backfire, could it?... Vaccine against drug addiction edition