A comprehensive computerized database of diving activity for a Norwegian offshore diving contractor [Stolt-Nielsen Seaway (SNS)] covering the years 1983-1990 has been established. The database contains detailed dive information about 12,087 surface-oriented and 2,622 saturation dives. During this period a majority of the divers were permanently employed. Preliminary analysis had suggested that decompression sickness (DCS) might be the result of exposure to factors causing pathophysiologic effects which accumulate over the course of a single dive or a series of dives. This concept evolved into the HADES (Highest Accumulated Decompression Score) theory which assumes that DCS is predictable once the underlying exposure factors are understood. The incidence of DCS among the SNS divers from saturation diving in the North Sea was studied by use of a "nested" case-control design. Twenty-one case dives (i.e., dives where DCS occurred) were compared with 41 randomly selected control dives. For these dives, several saturation dive characteristics were established. The relative pressure change between maximum and minimum storage depths was significantly greater among the cases. For each 1% increase in the relative pressure change there was a 5% increase in the probability of a saturation dive resulting in DCS. Significantly more cases than controls performed a saturation dive with more than one storage depth, and the data suggested that there were more and greater ascending and descending changes in storage depth conditions among the affected divers.