On 4 October 1992, a large cargo plane crashed into an apartment building in the Bijlmermeer quarter of Amsterdam. In the years following the accident, an increasing number of people started reporting health complaints, which they attributed to exposure to dangerous substances after the crash. Since the aircraft had been carrying depleted uranium as counterbalance weights and about 150 kg uranium had been found missing after clearance of the crash site, exposure to uranium oxide particles was pointed out as the possible cause of their health complaints. Six years after the accident, a risk analysis was therefore carried out to investigate whether the health complaints could be attributed to exposure to uranium oxide set free during the accident, The scientific challenge was to come up with reliable results, knowing that -considering the late date - virtually no data were available to validate any calculated result. The source term of uranium was estimated using both genetic and specific data. Various dispersion models were applied in combination with the local setting and the meteorological conditions at the Lime of the accident to estimate the exposure of bystanders during the fire caused by the crash. Emphasis was given to analysing the input parameters, inter-comparing the various models and comparing model results with the scarce information available, Uranium oxide formed in the fire has a low solubility, making the chemical toxicity to humans less important than the radiotoxicity, Best-estimate results indicated that bystanders may have been exposed to a radiation dose of less than 1 mu Sv, whereas a worst-case approach indicated an upper limit of less than 1 mSv, This value is considerably less than the radiation dose for which acute effects are to be expected. It is therefore considered to be improbable that the missing uranium had indeed led to the health complaints reported.