Doctors need statistical skills. These skills matter for interpreting research data critically and for understanding and explaining statistical information to patients. However, historically statistical teaching and learning has been poorly integrated into medical training. The discipline is not popular with many students, and it is sometimes poorly understood and retained. One commonly cited reason is that it seems irrelevant to medicine.
The aim of this miniproject was to investigate the usefulness of a PBL approach to the learning of statistics for medical students, i.e. non-statistics undergraduates. The project team focused on how a PBL approach might usefully be employed to teach a large group of students the statistical skills necessary for them to apply an evidence-based approach to their medical development. Students and academics in Medical Statistics, Medical Education and the Learning Development Unit were consulted.
Anticipated project outcomes included:
1. a summary of the usefulness of PBL when teaching statistics to non-statistics undergraduates;
2. new information about the fit between PBL and statistical learning;
3. guidelines for designing PBL modules for statistics teaching of non-statistics students; and,
4. a PBL module in statistics for non-statistics students.
The project began by surveying both current students and practicing doctors in order to design a core curriculum that addressed their needs. This was then used to develop PBL objects to deal with these, such that statistical skills could be learnt through addressing medical problems requiring statistical solutions. Clearly articulated statistical learning outcomes underpinned the development of each problem. These PBL objects were tested for both usability and acceptability, with a particular focus on how well they worked as a means of learning key statistical concepts.
The understanding produced provides valuable knowledge for both medical education and other disciplines where learning statistics is an essential component.
In general, with one or two exceptions, both the doctors and students felt that a course based on teaching core concepts rather than particular techniques was most appropriate. They felt that an understanding of confidence intervals and p-values was important, as was risk. The main theme for improving the teaching was ensuring that the teaching was relevant to future practice with lots of relevant examples.
The understanding produced thus far has provided valuable knowledge for both medical education and other disciplines where learning statistics is an essential component and it is hoped that the further work that is being undertaken will extend this to the benefit of future students and statistical educators across a range of disciplines.