Random left censoring: a second look at bone lead concentration measurements