Do Peers Support or Subvert Recovery from Substance Use Disorders

We study peer effects in recovery from substance use disorders. We focus on peers who share an inpatient treatment episode and who reside in the same county, reflecting the salience of geographic proximity for peer influence in risky behaviors, and examine peer effects on posttreatment mortality. We access linked administrative data on death for the universe of individuals who are admitted to inpatient treatment for a substance use disorder in Norway in 2009-2010. The impact of peers is identified using variation in the timing of admissions into treatment, which institutional factors ensure is conditionally exogenous. Patients exposed to a greater share of peers from their home-county have a lower mortality risk. A standard deviation increase in the share of home -county peers reduces mortality by 36% relative to the mean, with one additional peer leading to a 5% reduction. The peer-induced reduction in mortality is concentrated amongst individuals admitted for treatment for a drug use disorder (as opposed to an alcohol use disorder). This is driven by peers who themselves receive treatment for a drug use disorder, and is consistent with peer effects working through two potential channels; reduced illicit drug use and safer illicit drug use. Examining hospital episodes for intoxication and (nonfatal) overdose indicates a limited role for safer drug use, suggesting that peers primarily reduce mortality by reducing drug use. We conclude that peers from inpatient treatment episodes can be instrumental in supporting recovery outside of clinical settings.