Quality of reporting and risk of bias: a review of randomised trials in occupational health