When a pedestrian suffers a reduction in mobility, do they become less safe when crossing roads, and if so do they adapt their road-crossing behaviour to account for this reduction? An experiment was conducted to investigate the effects of mobility impairment, induced via a leg brace, and self reports of risk-taking on road-crossing safety in a virtual reality environment. A line of 11 vans created 10 randomised gaps of differing temporal size. Three initial distances between the vans were used so gap size was determined by the velocity of each van. There were 60 trials divided into 10 blocks. For the first and last block the participants were unimpaired while for the intervening 8 blocks they were impaired by the brace. Safety, indexed by ratio of the time available to cross to safety relative to the time taken to cross to safety (the safety ratio), decreased when the brace was attached and increased when the brace was removed. There was no significant change in the percentage of unsafe crossings (either a collision or being within 0.5 s of being hit) when the brace was either attached or removed, indicating that the reduction in the safety ratio did not necessarily mean the participants were more likely to have an unsafe crossing. There was a decrease in safety when the vans were further apart, indexed by both the safety ratio and unsafe crossings, which replicates previous research. Only limited conclusions could be drawn from the analysis of risk-taking, as although the general trends were in the expected directions the relationships between measures of risk-taking and the performance variables were generally small. Overall, the participants adjusted to their reduced walking speed, surpassing their original unimpaired safety ratios after approximately 3 blocks of impaired trials. The reduction in safety from mobility impairment was only temporary.