Moving beyond self-report in characterizing drug addiction: Using drug-biased behavior to prospectively inform treatment adherence in opioid use disorder.

Natalie McClain,Ahmet O. Ceceli,Greg Kronberg,Nelly Alia-Klein,Rita Z Goldstein
DOI: https://doi.org/10.1101/2025.01.01.25319860
2025-01-02
Abstract:Drug addiction is accompanied by enhanced salience attributed to drug over nondrug cues. This objectively measured bias is reliable yet underutilized in informing clinical endpoints, as clinical trials largely employ subjective (i.e., self-report or interview-based drug use and craving) or simple categorical (e.g., drug in urine) measures, with limited success. Having previously demonstrated their utility in cocaine addiction, we investigated whether behavioral picture choice (a lab-simulated drug seeking measure) and verbal fluency similarly reveal drug bias in 59 abstinent, inpatient individuals with opioid use disorder (iOUD) compared to 29 healthy controls (HC). Using a hierarchical regression, and compared to subjective measures, we then tested whether these objective markers can better inform prospective treatment completion—a clinically relevant and measurable outcome. As expected, results showed that the iOUD exhibited higher simulated drug seeking (ps<0.036) and drug fluency (p=0.008) compared to the HC. Importantly, after dimensionality reduction, while the self-reported years of regular opioid use and cue-induced craving showed null results (|β|<0.47, p>0.290), and controlling for demographics, drug choice was associated with treatment completion (β=-0.75, p=0.036), explaining greater variability in its likelihood compared to the subjective measures (model comparison: ΔR =0.102, p=0.027). Extending drug-biased choice and fluency from cocaine to opioid addiction, results further indicate that these objective measures of drug bias outperform the commonly employed subjective drug use and craving in informing a clinical outcome; unlike drug urine tests, they show important variability in abstinent iOUD. Results implicate these cognitive-behavioral tasks as powerful markers of drug bias and predictors of treatment outcome.
What problem does this paper attempt to address?