Method
A chemical incident was defined pragmatically as any unforeseen event involving chemicals that causes ill health or that was considered by public health agencies to have the potential to cause ill health and that therefore necessitated an immediate response. Incidents that involved
only occupational exposures within industrial premises were excluded from surveillance. All 37 local authority environmental health departments, both port health authorities and one of the three fire brigades in Wales participated in a full three year study (1993–1995). In addition, daily press cuttings from the national Welsh newspapers were reviewed to identify incidents. The Welsh National Poisons Unit participated in 1994 and 1995.Each agency received a monthly prompt to complete a specially designed reporting form for all incidents or to provide “nil returns”. The variables collected were date; nature and location of incident; the source of initial notification;other agencies involved, the numbers exposed and with symptoms, chemicals released,
and details of environmental monitoring and health surveillance undertaken. In most incidents actual numbers of people who were or who might be exposed could only be estimated roughly as biomarker surveys of exposure and dispersion modelling are seldom carried out. Therefore we have recorded the estimates made by local responders of the numbers of people at risk of exposure.