A figure commonly used to compare the risk of death from pregnancy compared with the risk of death from abortion might be based on outdated data, a new study suggests.