Diagnosing substance use disorder can be challenging due to patient denial and the stigma surrounding addiction. However, a recent study conducted by the University of Cincinnati has introduced a novel artificial intelligence system that predicts behaviors associated with substance use with remarkable accuracy.
Key Findings from the Study:
- The AI can predict substance use defining behaviors with up to 83% accuracy.
- It can also assess the severity of addiction with an accuracy of 84%.
- This technology could enable clinicians to deliver treatment more swiftly to those in need.
The study is pioneering in its use of a computational cognition framework, allowing researchers to analyze how human judgment can inform predictions about substance use disorder. The clinical standard for diagnosing this disorder includes:
- Impaired control
- Physical dependence
- Social impairments
- Risky use of substances
Professor Hans Breiter from the UC College of Engineering and Applied Science emphasized that this AI represents a cost-effective initial step for triage and assessment of mental health conditions, including addiction.
Study Methodology:
The research involved 3,476 participants aged 18 to 70, who provided informed consent and completed questionnaires. Participants also rated their preferences for 48 mildly emotional images, which were used to quantify their judgments. This data, combined with demographic information, was analyzed using AI algorithms to predict substance use disorder behaviors.
Implications of the Research:
The AI system demonstrated the ability to identify the type of substance used (such as stimulants, opioids, or cannabis) with up to 82% accuracy. Additionally, it revealed behavioral patterns associated with higher severity of substance use disorder, such as:
- Increased risk-seeking behavior
- Lower resilience to losses
- Less variance in preferences
This innovative approach could broaden the assessment of various addictions, potentially including behavioral addictions like excessive social media use or gaming.
The study was published in the journal Mental Health Research and was supported by grants from the Office of Naval Research and UC alumnus Jim Goetz.