The Connection Model: Evidence, Limits, and What It Gets Right

“The connection model is not a refutation of pharmacology. It is a correction of the pharmacological-only model. Both things are true: addiction has chemical dimensions and social dimensions, and the policy implication is that you have to address both.”

The central claim of Chasing the Scream, 'the opposite of addiction is connection', is a formulation that clarifies and overstates simultaneously. It clarifies because it redirects attention from the drug to the person and the person's environment. It overstates because pharmacological dependence is real, withdrawal is a physiological process, and social connection, however necessary, is not sufficient to reverse all forms of addiction.

Bruce Alexander and colleagues at Simon Fraser University ran the Rat Park studies in the late 1970s and published them between 1978 and 1981 (Alexander, Coambs and Hadaway, Psychopharmacology, 1978; Hadaway and colleagues, Psychopharmacology, 1979; Alexander and colleagues, Pharmacology Biochemistry and Behavior, 1981). Rats kept alone in standard cages drank more morphine solution than rats living together in a large open colony. Later attempts gave mixed results. Bozarth, Murray and Wise (1989) found that rats housed alone and rats housed in groups both learned to give themselves heroin and reached similar levels after a few weeks. Petrie (1996) could not repeat the housing difference: the isolated rats in his studies drank less than Alexander's had, and he suggested the supplier's rat strain had changed. The original result has not been reliably replicated, although the broader idea that surroundings can change drug taking still has support from other animal and human research.

The Vietnam natural experiment provides the strongest human data. Lee Robins's studies of Vietnam veterans (Archives of General Psychiatry, 1975; Addiction, 1993) found that most men who had been addicted to heroin in Vietnam did not become addicted again after returning home: about 5 percent relapsed to addiction within the first year and about 12 percent within three years, most of them without formal treatment. The most important variable was not pharmacological, it was the radical change in environment. The drug was the same; the context was different; the behavior changed.

This finding, combined with the Rat Park data, supports the general claim that social disconnection is a major driver of addiction. But the phrase 'the opposite of addiction is connection' obscures an important nuance: the Vietnam veterans were not simply getting more connection at home. They were getting a different set of environmental affordances, work, family, housing, purpose, that made drug use both less necessary and less available. 'Connection' is a partial description of a richer social ecology.

The pharmacological critique of Hari's strong formulation is that opioid withdrawal is a genuine physiological process that does not reverse simply because a person feels more connected. George Koob and Nora Volkow's framework, which emphasizes the neuroadaptive changes that chronic drug use produces in reward, stress, and executive function circuits, shows that the brain changes measurably with sustained use, and that these changes persist after environmental improvement. The social context model and the neurobiological model are not contradictory; they are complementary accounts of different aspects of the same phenomenon.

The policy implication of this more nuanced picture is: both. Treatment for opioid use disorder needs both medication-assisted treatment (buprenorphine, methadone) and social support (housing, employment, community). Portugal's model, which Hari cites as the strongest evidence for the connection approach, actually implements both: decriminalization removes the legal barrier to treatment-seeking, and the social investment provides the environmental conditions that support recovery.

Where Portugal's results have been weaker, particularly through the 2015-2023 period, when austerity cut social service budgets, the model has underperformed. Not because decriminalization failed but because the social investment was insufficient. The lesson is not that connection is the opposite of addiction; it is that connection is necessary but not sufficient, and that the pharmacological and the social interventions work best in combination.

Hari's formulation, for all its incompleteness, accomplished something the more nuanced version would not have: it made the argument accessible and memorable, it redirected public attention from moral failure to social conditions, and it built the political case for harm reduction and decriminalization that the evidence supports even if the slogan overstates. The compression of a complex finding into a memorable formulation is not always wrong. Sometimes it is the thing that moves policy when the nuanced version would not.

Sources

• Hari, Chasing the Scream (2015)
• Alexander, The Globalization of Addiction (2008)
• Gage & Sumnall, Addiction (2019)
• Koob & Volkow, Lancet Psychiatry (2016)
• Robins et al., Vietnam Veterans Three Years After Vietnam, Am J Addict (2010 reprint); Robins, Addiction (1993)
• Alexander, Coambs & Hadaway, Psychopharmacology (1978)
• Hadaway et al., Psychopharmacology (1979)
• Alexander et al., Pharmacology Biochemistry and Behavior (1981)
• Bozarth, Murray & Wise, Pharmacology Biochemistry and Behavior (1989)
• Petrie, Psychological Reports (1996)
• Robins, Helzer & Davis, Archives of General Psychiatry (1975)