Measuring the impact of interventions

Many hazardous agents in the workplace are subject to controls aimed at reducing total exposure and thereby the burden of occupational disease. Armstrong and Darnton1 set out to clarify how such reductions in health burden can be estimated and illustrate the methods using real data on lung cancer cases attributable to polycyclic aromatic hydrocarbons. “Moderate” extensions of standard expressions for attributable fractions appear to be the way forwards, but information is required on the distribution of exposure in the population and the shape of the exposure-response relationship. Their analysis shows that for linear threshold relations and left skewed …