This study evaluates whether peer-assisted learning (PAL) can be used to improve students' clinical examination skills. Four year 4 students trained in PAL techniques and musculoskeletal (MSS) examination used the Gait, Arms, Legs and Spine (GALS) system in a five-week student selected module. These students then recruited and trained 28 second-year trainees. Trainees were evaluated using pre/post confidence questionnaires (100 mm visual analogue scale), a course experience questionnaire (five-point Likert scales) and end-of-year objective structured clinical examination (OSCE) scores. Baseline data from the experimental group were no different from a separate control group, but after training a statistically significant difference in confidence levels was observed in all parts of GALS, <38 to >73 (p < 0.0001). Course experience questionnaires demonstrated benefits in all parameters including communication skills and group work with all students recommending PAL training. In end-of-year OSCE 93% of PAL-trained students passed the MSS examination station compared with 67% for those participating in the standard curriculum alone (p < 0.0001). Examination results for other clinical skill stations showed no difference in performance between the two groups. This study shows that PAL is a useful adjunct to MSS training, and could be incorporated into medical curricula to enhance clinical skills.