22.8% of adults who enter formal substance use disorder treatment drop out before completing it — the action-side regret proxy drawn from 649,479 discharge episodes in the TEDS-D dataset (Baird, Cheng, and Xia 2022). Dropout is the most tractable available signal for treatment dissatisfaction: people who found the programme insufficient or incompatible enough to leave. The figure sits alongside a 33.8% completion rate, meaning roughly one in three entrants finishes the programme as designed, and the remaining third transfer to other programmes, are terminated, or leave for logistical reasons. No published study has directly asked treatment-seekers whether they regret entering formal care; dropout is used as a proxy under full disclosure. What the evidence does establish clearly, via the 2020 Cochrane review of 27 trials and 10,565 participants, is that AA and 12-step facilitation produce the highest continuous abstinence rates of any tested formal intervention — 24% at one year in Project MATCH versus 15% for cognitive-behavioural therapy and 14% for motivational enhancement therapy — suggesting that among those who remain engaged, treatment pays off substantially.
For adults who avoided or delayed formal treatment, the Tucker et al. (2020) re-analysis of NESARC-III (n=7,785 US adults with AUD) found that 34% had persistent AUD with no recovery trajectory at the time of the survey. Since only 23% of the NESARC-III sample had ever received alcohol treatment, the 34% persistent-AUD rate predominantly reflects untreated pathways. The Liu et al. JAMA Network Open analysis of NSDUH data (2013–2023, n=657,583) adds a cross-sectional layer: among all untreated adults with any SUD, only 5.7% acknowledged a perceived need for treatment at the moment of survey, and only 2.7% had taken steps to seek it. This low perceived-need figure is consistent with the anosognosia literature on substance use disorders and does not imply that the remaining 94% were satisfied with their untreated status — it more likely reflects the insight deficits that characterise active addiction. The 34% persistent-AUD proxy captures the long-run outcome for those who remained on the untreated path.
Under Gilovich and Medvec’s temporal regret framework, actions generate more regret shortly
after the decision, while inaction regret accumulates over years as the forgone alternative
becomes salient. The pattern for addiction aligns with inaction_dominates: the consequences
of remaining untreated — continued dependence, relationship loss, occupational decline,
health deterioration — compound over time in ways that are not apparent at the moment of
avoidance. Treatment dropout (action-side regret) tends to be experienced acutely but may
attenuate as individuals transition to other recovery pathways (natural recovery, mutual help,
medication-only). The persistent-AUD outcome for untreated individuals (34%) exceeds the
treatment-dropout rate (23%), and that gap, while based on non-equivalent denominators and
disclosed proxy assumptions, is directionally consistent with the full body of longitudinal
treatment-outcome evidence.