Comparison of Artificial Intelligence and Multidisciplinary Team Assessment for Evaluating Biologic Decisions in...

The management of severe asthma with biologics has become more complex involving multidisciplinary team (MDT) meetings. Artificial intelligence (AI) has been proposed as an adjunct to aid health care professionals in other areas of the medical field.

To determine whether there is a role for AI, as an adjunct to the severe asthma MDT meeting, to offer improved consistency and efficiency in biologic prescribing.

This is a retrospective, unblinded single-center, multisite cohort study including patients with uncontrolled severe asthma (Global Initiative for Asthma-defined stage 5 asthma) for biologic consideration who were biologic-naive. They were discussed at joint MDT meetings between secondary care referral centers in the health board NHS Tayside, Scotland, between January 2024 and August 2025. The MDT biologic decisions were compared with 5 large language model recommendations (ChatGPT, Gemini, Copilot, PerplexityAI, and DeepSeek) using Cohen κ, with the MDT meeting as the reference.

Among 114 patients, the mean age was 54.5 ± 15.0 years, Asthma Control Questionnaire score was 2.98 ± 1.1, and annualized exacerbation rate was 5.21 ± 2.1. MDT decisions and AI recommendations showed poor agreement (κ = 0.143-0.213). Agreement between AI models was weak to moderate (κ = 0.348-0.586), with frequent decision-making errors identified with all AI models.

At the time of analysis, AI models show poor concordance, highlighting their unsuitability as either a standalone or an adjunctive decision-maker for biologic therapy. MDT meetings remain the reference standard.