How chronic self-regulatory stress, poor anger regulation, and momentary affect undermine treatment for alcohol use disorder: Integrating social action theory with the dynamic model of relapse

We tested the hypothesis that indices of chronic self-regulatory stress specified by Social Action Theory (SAT; Ewart, 2011) and indices of affect specified by the Dynamic Model of Relapse (DMR; Witkiewitz & Marlatt, 2004) predict daily alcohol use by persons in outpatient treatment for alcohol use disorder (AUD). Participants (69 men, 50 women; aged 19-76) in AUD outpatient programs completed 2 baseline laboratory assessments, then provided daily ecological momentary assessment (EMA) reports of state affect and alcohol use for 28 days. SAT indices of chronic self-regulatory stress were obtained in the laboratory with the Social Competence Interview (SCI) and the Anger Transcendence Challenge (ATC). Chronic regulatory stress was indexed by blood pressure responses to both tasks. Emotion regulation capabilities were indexed by anger regulation behavior observed during ATC, and by self-ratings of recent positive social interactions and goal thoughts. DMR indices of trait and state affect were obtained with the General Psychiatric Distress scale of the Brief Symptom Inventory, the trait anger scale of the Buss-Perry Aggression Questionnaire, and intensive EMA ratings of state affect (anger, sadness, fear, happiness, stress) during each day. The outcomes were: (a) not drinking on a given day, and (b) number of drinks consumed when drinking. Statistical models accounting for problems of zero inflation in drinking data and missing data evaluated the hypothesis that "tonic" (cumulative) and phasic (situational) processes specified by the SAT and DMR models together shape drinking in a given moment. No drinking was reported on 91% of days; reports were missing on 7.2% of days. Both drinking outcomes were predicted by SAT indices of SCI self-directive stress, ATC anger regulation, positive goal cognitions and social interactions, and by DMR indices of general distress and state affect (sad, angry, stressed). Unexpected and potentially informative associations emerged also. The findings were robust under models that accounted for different missing data assumptions. Findings suggest that SAT and DMR models and methods may advance scientific understanding of drinking risk during outpatient AUD treatment.