BACKGROUND: In 1988 a new patient insurance system was introduced in Norway. It was initially described as an 'objectified' system, similar to one based on the no-fault principle. Early doubts were raised about the system's status, as it contains rules stating that compensation will not be given if the medical intervention is adequate and the involved risk is acceptable. This study was undertaken to examine the practice of these rules. METHODS: An archival study was performed on the 41 shoulder dystocia cases that had been closed in the decade from 1988-1997. These cases were selected as shoulder dystocia was found to be the obstetrical event most often leading to a decision on acceptable risk. RESULTS: The most common injury in these cases was Erb's palsy, but fatalities and brain injuries were also observed. Compensation was given in nine cases, whereas it was denied due to an acceptable medical risk in the remaining cases. Indications of inconsistency among the reached decisions were found, and judged to be a result of differences of opinion between expert witnesses on the adequacy of the obstetrical practice. CONCLUSIONS: Doubts are raised as to whether similar decisions are reached in similar cases. Shoulder dystocia may be an acceptable risk in the sense that it is hard to predict and prevent. Whether the consequences of such a risk should be compensated, remains a political and economical issue. Present thinking leads to decisions that create a divide between the lucky unlucky and the plainly unlucky.