Adverse events reported in oncology clinical trials don’t always match what patients experience. What is behind this disconnect and is it possible to reconcile?