Problems with trials and intervention studies on barrier creams and emollients at the workplace.

The potential effect of barrier creams or emollients in the prevention of work-related hand eczema has mostly been documented in a laboratory setting on experimentally damaged skin. Publications on real intervention studies of barrier creams or emollients in a workplace setting are scarce: only four such studies could be identified. The relatively short follow-up of most studies, and the small numbers of subjects enrolled, may reflect the organisational complexity of such studies. The observed effect is a combination of the intervention effect to be measured, and a number of disturbing variables. These disturbing variables are not just measurement error, but also phenomena such as the course of the disease, confounding and effect modification. In hand-eczema studies, the outcome parameters that should form the basis of the intervention effect are ill defined. Moreover, the study itself (i.e. the investigators) influences the results because there is rarely a single intervention: unknowingly, there are improvements in glove wearing, hand-washing habits or occupational hygiene in general. The performance of a blind study is often difficult or impossible. Large numbers of subjects have to be enrolled if a meaningful reduction in incidence of hand eczema is to be detected. One can demonstrate this by calculating the required enrollment in a study among employees in a nursing home.