# Moving Beyond Self-Report in Characterizing Drug Addiction: Using Drug-Biased Behavior to Predict Treatment Completion and Dropout in Heroin-Primary, Medication-Maintained Opioid Use Disorder

## TL;DR

This study shows that objective behavioral tests, not self-reports, better predict treatment success in opioid use disorder.

## Contribution

First demonstration that drug-biased behavioral measures predict treatment outcomes better than self-reports in opioid use disorder.

## Key findings

- Individuals with opioid use disorder showed higher drug-biased choice and fluency compared to healthy controls.
- Drug-biased behavioral measures predicted treatment completion and dropout more reliably than self-report measures.
- Behavioral tasks demonstrated strong test-retest reliability and significant predictive power for clinical outcomes.

## Abstract

Drug addiction is accompanied by enhanced salience attributed to drug over nondrug cues. This bias can be objectively measured and is reliable but underutilized in informing clinical end points, where self-report measures are most commonly used, with limited success.

We investigated whether behavioral picture choice (laboratory-simulated measure of drug seeking) and verbal fluency (drug and nondrug words generated) revealed drug-biased processing in 59 individuals with opioid use disorder (iOUDs) compared with 29 healthy control (HC) individuals; assessed twice, we also inspected the test-retest reliability of these tools. All iOUDs were heroin primary, abstinent (160.58 ± 188.18 days), and stabilized on medication for OUD at an inpatient treatment facility at baseline. Then, we tested whether, compared with self-report measures, these drug-biased behavioral measures could better predict prospective outcome measures in the iOUDs, i.e., study treatment completion as further validated using dropout from inpatient treatment.

Results revealed that the iOUDs exhibited higher drug choice (ps < .036) and drug fluency (p = .008) compared with the HC individuals; task performance demonstrated the strong test-retest reliability of these measures. Controlling for cognitive demographics, the self-report drug-use severity and craving measures did not show significant associations with study treatment completion (|β| < 0.47, ps > .290), but drug-biased choice did (β = −0.75, p = .036; model comparison: ΔR2 = 0.10, p = .027). Importantly, these results were validated using inpatient treatment dropout as the outcome (drug-biased choice: β = 0.81, p = .049; model comparison: ΔR2 = 0.11, p = .035).

This study is the first to demonstrate reliable drug-biased choice and fluency in iOUDs. Compared with traditional self-reported drug-use and craving measures, the objective drug-biased cognitive behavioral measure was a significant predictor of treatment-related outcomes.

In this study, we examined whether behavioral markers of drug bias provide better prediction of treatment outcomes compared with traditional self-report measures in heroin-primary individuals with opioid use disorder (iOUDs). Results revealed that, compared with healthy control individuals, iOUDs exhibited drug-biased choice and fluency behavior; task performance also revealed high test-retest reliability of these measures. Importantly, drug-biased behavior predicted whether iOUDs completed study treatment or dropped out of inpatient care, while self-reported drug-use severity and craving measures did not. These findings suggest that objective behavioral tasks may offer more reliable markers of severity and indicators of future treatment engagement in OUD.

In this study, we examined whether behavioral markers of drug bias provide better prediction of treatment outcomes compared with traditional self-report measures in heroin-primary individuals with opioid use disorder (iOUDs). Results revealed that, compared with healthy control individuals, iOUDs exhibited drug-biased choice and fluency behavior; task performance also revealed high test-retest reliability of these measures. Importantly, drug-biased behavior predicted whether iOUDs completed study treatment or dropped out of inpatient care, while self-reported drug-use severity and craving measures did not. These findings suggest that objective behavioral tasks may offer more reliable markers of severity and indicators of future treatment engagement in OUD.