Injury indicators can be used to give policy makers an estimate of the scale of injuries and their long-term effects. They can help compare injury levels in different areas and countries and can be used to help measure the effectiveness of interventions. Work on severity related indicators is promising. However there are no perfect indicators to date as many are hampered with difficulties in case definition and under reporting. For example, mortality rates are affected by improvements in care even if the incidence of an injury remains the same, the abbreviated injury scale (AIS) takes 10-20 minutes to code and so is not used in health service databases, surveys have problems with recall bias, definition of injury and response rates. If we accept that we need to make the best out of imperfect indicators and imperfect data then we should use multiple sources of data and accept that no one indicator can be used universally but needs to be selected for the purpose. For example, one possible new indicator of the incidence of non-fatal injury might be fracture data in the emergency department. Fractures are painful and so nearly always end up with a hospital attendance. This might give a means to compare incidence of non-fatal injury in different areas and countries. In conclusion, we need injury indicators to progress in injury prevention. Imperfect indicators can be used for targeting and evaluating interventions as long as we know and adjust for their limitations.