Title : Beyond abstinence: The shape of compulsivity, recovery and relapse
Abstract:
Background: Addiction is commonly measured through endpoints: substance use or abstinence, symptom severity, treatment attendance, and relapse. Yet compulsive behavior unfolds as a moving process. Vulnerability, distress, salience, executive control, resilience, treatment effects, environment, and social forces change over time. The clinically important question may therefore be less what happened? than what changed, when and why?
Methods:This presentation draws on more than 25 years of naturalistic clinical observation and a longitudinal behavioral health dataset involving more than 5,400 patients, 25,000 encounters, and 45,000 contemporaneous narratives. Each encounter began with a simple directional question: “Since I saw you last, are you better, worse, or the same?” Subsequent enquiry explored what changed, when it changed, the patient’s explanation, treatment exposure, positive function, behavioral instability, compulsivity, relapse, and recovery.
Results: Repeated enquiry revealed clinically meaningful patterns that were less obvious in isolated encounters. Patients described predictable periods when control over compulsive eating began to erode, when agitation or distress became harder to withstand, and when executive function began to decline before medication was perceived as fully “worn off.” These observations suggested a broader model in which compulsive behavior emerges when accumulating pressure to act exceeds the individual’s changing capacity to resist. At a larger timescale, relapse may represent the final point of a trajectory involving deterioration in sleep, mood, motivation, resilience, executive function, and behavioral control rather than an isolated event.
Conclusions: The value of longitudinal behavioral data lies not in time alone, but in a line of enquiry that considers time: ask repeatedly, seek narrative, identify patterns, test them, and learn. This approach contributed to the development of Continuous Behavioral Intelligence (CBI), designed to turn clinical pictures into behavioral movies by preserving sequence, context, and trajectory. This is also fundamental to artificial intelligence. AI cannot infer information that was never elicited or distinguish antecedent from consequence without temporal sequence. Understanding the motion between abstinence and relapse may therefore be more informative than measuring either endpoint alone.
