Will this trial be stopped early?

Look up any Phase 1–3 drug trial on ClinicalTrials.gov that started in 2008 or later. You'll see its termination-risk score, how it ranks against the trials running now, and what moved the score.

Or try one of these

How good is it?

0.710ROC AUC on 10,242 trials it never saw (95% CI 0.696–0.722)
2.1×termination rate in the top-scored 10% (32% vs 15%)
0.788AUC for termination due to slow enrollment

Checked against leakage: scored with registry records as they stood at each trial's start (from 25 archived snapshots), the AUC changes by +0.001 (95% CI -0.001 to +0.003) on 18,568 trials. Fields found edited after start are left out of the model.

When, not just whether: a competing-risks survival model ranks which trials terminate within 2 years at time-AUC 0.649 (95% CI 0.629–0.672), ahead of the yes/no model (0.613), with still-running trials handled as censored. Termination risk is rising: 5.1% of trials started in 2013–14 were terminated within 2 years, against 7.4% of those started in 2019–20.

Tested on trials that started in 2015–2016, after training on 2008–2014. The score ranks trials; read it as a percentile, not a promise. Reasons describe the model, not causes of termination.