Background: The international comparability of data from population-based cancer registries depends strongly on the completeness of case ascertainment. Furthermore, Austrian observed incidence rates suggest that the completeness of case ascertainment differs between Austrian federal states. Completeness of case ascertainment is to be investigated on national and regional level. Methods: We used the flow method to evaluate the completeness of the Austrian National Cancer Registry. This method is based on the logical flow of data in the registration system, and on the time distribution of various probabilities inherent in this flow. Results: Overall completeness of the Austrian cancer incidence data 2005 was 94.2% after a registration period of 5 years. The flow method found striking differences in completeness between the federal states, which are contrary to the time series analyses. Conclusion: Overall completeness of the Austrian National Cancer Registry is in concordance with estimates from international registries. The biggest part of the decrease of incidence rates in the past 2 published years seems to be a result of incompleteness. The importance of the registration date of a cancer case and the survival time on completeness estimation using the flow method has become apparent. Further investigation into the comparability of registration date between the federal states and into the quality of survival time estimates is recommended.