Objectives EU directives can lend themselves to evaluation through natural experiments because of the freedom regarding their implementation within individual countries. Natural experiments may be criticised because of the lack of direct casual evidence for a relationship between the intervention and the exposure or disease outcome. Such experimental designs could be strengthened by an a priori estimate of the impact of an intervention. Here we aim to predict the impact of EU Directive 2002/44/EC on VWF, CTS and sensorineural symptoms (HAVS).

Methods Previously we described the development of a Markov chain Monte Carlo accelerated failure time model to predict the incidence and prevalence of HAVS. The model was developed using longitudinal data from Italian workers and validated using published data for Swedish and UK workers, and compensation data from the Czech Republic. For the next step we have used the exposure data from the Eurofound European Working Conditions Survey from 2000 to 2015 and population level demographic data to predict the impact of the directive on HAVS in Europe.

Results The model predicted that a reduction in vibration exposure to 5 m/s2 (i.e. complete success of the directive) would result in a reduction in lifetime prevalence of VWF of around 25% in the UK. Predictions of the variation in impact according to differing reductions in exposure and across different age groups will be presented for the UK and other countries.

Conclusion Future work will compare these estimates of the impact of the directive using routinely collected data in European countries.