The Norwegian Ministry of Health and Social Affairs recently introduced activity-based financing for hospitals partly based on diagnosis-related groups (DRG). We soon observed that there seemed to be a considerable discrepancy between the reimbursement amount and the real cost of allogeneic haemopoietic stem cell transplantation. It was therefore decided to undertake a prospective micro-cost analysis to define a more realistic reimbursement. To identify real costs, we undertook a registration of pre-transplant procedures, transplantation and 1 year follow-up costs, including harvesting, personnel costs, clinical and laboratory procedures, together with blood products and drugs related to patients and donors. These data were compared to hospital DRG reimbursement. This information was registered for 17 consecutive patients, with a mean age 40 years (range 17-58 years). Ten patients had chronic myeloid leukaemia, three had acute lymphatic leukaemia, two had acute myeloid leukaemia and two had myelodysplastic syndrome. The data analysis showed a mean cost of US$ 106825 (NOK 901982), (range US$ 42376-362430). The average actual hospital revenue (50% DRG reimbursement + income related to length of stay + special procedure funding) was US$ 36404 (range US$ 26228-55998). Activity-based financing as applied in Norway, under-compensates hospital costs for allogeneic bone marrow transplantation. The government should make realistic estimates of real costs before introducing financial reforms in the health care system.