In 2006, we set up a surveillance programme with a large UK supermarket employing almost 4000 “scratch” bakers (who mix dough and bake bread from scratch using raw ingredients) in around 350 stores. We report here the analysis of the programme through to 2010. The company occupational health provider screened all bakery workers for respiratory symptoms every other year, using a initial questionnaire (Level 1), with positive responders completing a subsequent, more detailed telephone-administered questionnaire (Level 2). Those who reported work-related nasal or respiratory symptoms were asked to provide a serum sample for specific IgE to bakery antigens. Those with positive specific IgE to flour or α-amylase (>0.35 kU/l) were directly referred for a specialist opinion. Abstract S6 table 1 shows the results of three surveillance rounds in 2006, 2008 and 2010. The frequency of work-related symptoms, sensitisation and disease across the three rounds of surveillance were remarkably constant. Measured prevalence is low (0.3–2 in 1000 bakers) although this figure is likely to be an underestimate; a previous study in the same workforce has demonstrated a reluctance to report symptoms and incomplete response rates.1 This system of surveillance is efficient but has thus far not been effective in reducing the incidence of occupational allergy.