This retrospective study used text mining to analyze the free text descriptions in 72,390 medication administration incident reports in the National Reporting and Learning System in England and Wales to identify terms most frequently associated with risk that might otherwise remain buried within other non-relevant text. The authors identified the most common medications described in free text (insulin, antibiotics, paracetamol and morphine) and presented the most common free text terms associated with these medications. Results indicate that checking patient allergies and medication doses, especially for intravenous and transdermal medications, should be a focus of efforts to increase medication administration safety.