A comparison was presented of results from a collaborative project undertaken by the Australian National Institute of Occupational Health and Safety in Australia and by NIOSH in the United States. The purpose for the comparison was two fold. First, it explored methodologic differences between the two evaluations to identify ways of improving the quality of fatal occupational injury surveillance on a national level. Second, it examined difficulties of comparing surveillance data from different sources, resulting in a discussion of preconditions for meaningful comparison. Despite differences in data acquisition methods and in the structure of the labor and work environment, the results of the two evaluations were remarkably similar. The mean annual rates of data injury while working for persons in the employed civilian work force during 1982 through 1984 were 5.9 per 100,000 in the United States and 6.7 per 100,000 in Australia. The fatality rates for age, sex, and employment groups were also similar. The authors stress that valid fatality surveillance requires that all relevant cases are detected and that all cases included in the collection comply with the stipulated definition of work related fatality; that definitions, coding systems and data acquisition methods should be documented in detail for all published fatal injury data; and that the language be simplified as much as possible in the definition of the terms used by the individuals involved in the identification and classification of cases.