AI-based analysis of social media language predicts addiction treatment dropout at 90 days

Springer Nature

In-person treatment for substance use disorders is an incredibly important tool, but there’s a high failure rate — more than 50 percent of people who enter drop out within the first month. There hasn’t been a highly accurate method of identifying who might leave and who might succeed, and knowing this could help centers allocate resources to give the right type of assistance to the right people at the right time. One tool available is called the Addiction Severity Index, which is used to help identify the severity of the addiction and thus customize treatment, but it wasn’t developed to gauge whether a patient might drop out entirely. So a team of researchers decided to mine something known as a digital phenotype. Dr. Brenda Curtis is a clinical researcher at the National Institute on Drug Abuse Intramural Research Program, and she’s one of the paper’s authors. Read the full study here: https://www.nature.com/articles/s41386-023-01585-5