Artificial intelligence (AI) is finding ways to help people with schizophrenia, a serious mental illness. While new tech has pros and cons, AI can predict how well patients might respond to treatment.
Unlike chatbots (LLMs) that use language, scientific AI uses data—like demographics or test scores—to make numerical predictions. These models focus on positive goals, such as preventing relapses.
A recent Korean study tested this. Researchers used data from 299 patients (average age 36.7, in treatment for 8.8 years) who switched to extended-release antipsychotics (easier to take than multiple daily doses). They wanted to find out who would stick with the new treatment.
What predicted success? BMI (body mass index) was the top factor, followed by positive attitudes toward meds. Early on, social relationships and symptom severity mattered too. By 24 weeks, education level, symptom severity, and illness duration also played a role.
The goal? Use simple, cheap factors (like BMI or drug attitudes) instead of expensive tests (like brain scans or genetics) to guide treatment. AI could match patients to the best treatments faster, making care more personalized.
Right now, this AI is mostly for researchers, but the hope is to bring it to daily clinics. Used well, AI can help patients recover without worsening symptoms—using tech to make treatment better.
READ MORE
