Context: Burnout is prevalent amongmedical students and is a predictor of subsequent serious consideration of dropping out of medical school and suicide ideation. Understanding of the factors that protect against burnout is needed to guide student wellness programmes. Methods: A total of 1321 medical students attending five institutions were studied longitudinally (2006-2007). The surveys included standardised instruments to evaluate burnout, quality of life, fatigue and stress. Additional items explored social support, learning climate, life events, employment status and demographics. Students who did not have burnout at either time-point (resilient students) were compared with those who indicated burnout at one or both time-points (vulnerable students) using a Wilcoxon-Mann-Whitney test or Fisher's exact test. Similarly, the differences between those who recovered and those who were chronically burned out were also compared in students with burnout at the first time-point. Logistic regression modelling was employed to evaluate associations between the independent variables and resiliency to and recovery from burnout. Results: Overall, 792 (60.0%) students completed the burnout inventory at both timepoints. No differences in demographic characteristics were observed between resilient (290 /792 [36.6%]) and vulnerable (502 / 792 [63.4%]) students. Resilient students were less likely to experience depression, had a higher quality of life, were less likely to be employed, had experienced fewer stressful life events, reported higher levels of social support, perceived their learning climate more positively and experienced less stress and fatigue (all p < 0.05) than vulnerable students. On multivariable analysis, perceiving student education as a priority for faculty staff, experiencing less stress, not being employed and being a minority were factors independently associated with recovery from burnout.