GlobalData claims 85% accuracy in AI-driven approval forecasting

UK analytics group GlobalData (LSE: DATA) says its artificial intelligence-powered Likelihood of Approval model has achieved 85.4% accuracy in predicting whether experimental drugs will secure US market authorization.

The model uses 108 machine learning algorithms trained on 24 years of data covering successful and failed development programs. It calculates the probability of a drug progressing through each stage of clinical testing and ultimately reaching approval. GlobalData said the upgrade, which shifts fully to machine learning methods, enhances precision for portfolio management and risk assessment.

The company pitches the tool as useful not only for drug developers but also for investors, consultants, and business development teams who want a clearer picture of regulatory risk.