Injuries would probably be a better statistic to use, as you much more likely to die being hit by a car than a bicycle. I'd also guess injuries in bicycle on pedestrian collisions are nearly impossible to measure accurately, as most of them are minor and go unreported.