Despite the human and financial cost of repetition strain injury (RSI), comprehensive incidence data have been lacking. A unique opportunity exists to obtain such data in South Australia, where since 1980-81 the Australian Bureau of Statistics has assigned all injuries, not explicitly diagnosed as diseases but stated as having been caused by repetitive movement, to a unique "type of accident" code, and has subclassified them according to bodily location. The statistical profile of diseases and accidents affecting the upper limb resulting from repetitive movement is not simply one of a keyboard operators' epidemic. Rather, it has revealed a problem which is endemic in sections of the blue-collar workforce, in whom both the numbers and the incidence rates are higher than in keyboard operators, and were higher even when the incidence in keyboard operators peaked in 1984-85. These conditions have been especially frequent in particular sections of the female blue-collar workforce, and interventions which have resulted in (or coincided with) benefits to keyboard operators have failed to improve the situation in the former group. It is suggested that the groups most at risk are female workers performing unfulfilling, unskilled tasks, and that interventions to benefit these workers will have to give attention to more fundamental issues than those hitherto addressed.