Abstract Introduction/background: Teaching statistics to medical students is hampered in two ways. Firstly, as a “Cinderella” subject, statistics is perceived to be less glamorous than other apparently more clinical disciplines. Secondly, as it involves mathematics students often view it as difficult, unpleasant or just plain boring. Since 2005, first and second year medical students, undertaking the new integrated undergraduate medical curriculum at the University of New South Wales (UNSW), Sydney, have been learning medical statistics and evidence-based medicine (EBM) through a mainly online learning environment of nine consecutive online modules supported by lectures and practicals). Issues for exploration/ideas for discussion: This paper examines numeracy issues and threshold concepts as impediments to learning medical statistics and EBM and shows how online learning and support activities can be adapted to target these problematic areas. Purpose/objectives: Early feedback on the online statistics modules revealed that there was a problem with student engagement and also poor understanding of the major assessed topic areas. In an attempt to improve the modules, the impediments to student engagement and learning were identified. Prominent among these problem areas were major threshold concepts and also any content containing mathematics or formulae. Further research aimed to find a more engaging and successful way to approach these problem areas from both teaching and learning points of view. Sections in the modules with major threshold concepts and statistical formulae were redesigned to present information using more than one teaching approach (visual and narrative as well as mathematical) and by targeting online exercises and quizzes, lectures and face to face tutorials on hese more difficult areas. Several other changes were also implemented to support these initiatives, including a more comprehensive evaluation. Results: Subsequent evaluation using online evaluation forms, quiz and exam results revealed that the students’ knowledge in the targeted areas improved and that student engagement in the course had increased considerably. Discussion: The significance of these findings for other “Cinderella” subjects and for online statistical courses is discussed and suggestions are made for the application of this process to other areas of learning. However, the overall impact of changes is difficult to assess as numeracy issues involve emotional responses that are difficult to gauge and the measurement of student understanding of threshold concepts is complex and maximal evaluation remains elusive.