Background 2/3 of all fatal occupational injuries (OI) in Norway are recorded by the Norwegian Labour Inspection Authority (NLIA). The other 1/3 is recorded by authorities of oil production, shipping, fishing and aviation. The register at NLIA has been regarded as being complete. The validity of this statement was studied.

Material and method 183 fatal OI for 2000–2003 were found at NLIA, 171 related to persons resident in Norway. The records of these were found in the national death register (NDR) by the unique personal number. In addition 43 fatal OI within the sphere of authority of NLIA were identified in NDR. The real number of fatal OI was estimated by a capture-recapture model.

Results 214 fatal OI were identified, 98 in both registers. Missing injuries in NLIA were related to national defence (1 of 9), health and social services (3 of 7), road traffic injuries (36 of 52), fatalities occurring in North of Norway and Spitsbergen (17 of 28). The highest risks for fatal injury (per 100 million working hours) were found in agriculture and forestry (7.0), and in transport and communication (4.1). The number of fatal OI to should have been recorded by NLIA was found to be 246.

Conclusion The risk for a fatal OI in Norway seems to be 44% higher than shown in official statistics. Insufficient registration, especially of transport injuries, might result in misjudgements of size and trends of the occupational injury problem, and the need for prevention.