BACKGROUND: This study was designed to assess the current incidence of tuberculosis (TB) in National Health Service hospital staff in the West Midlands region of England and to evaluate the effectiveness of occupational health screening and surveillance procedures. METHODS: A four year prospective study (1992-5) was carried out with case ascertainment through occupational health departments and the Midlands Thoracic Society registry of rare respiratory diseases. RESULTS: No case was known to an occupational health department that was not notified by the registry and, in addition, the registry yielded a further six cases. There were 26 cases of active TB, of which 24 presented with symptoms and 12 had no pre-employment screening. All except one doctor were foreign born--mainly from the Indian subcontinent (ISC)--and of recent UK entry, whilst all except one nurse were white females. The following are annual TB incidences per 10(4) (95% confidence interval): ISC and non-white doctors 17.0 (7.8 to 26.2), not significantly greater than in the local ISC and non-white social class 1 working age population (11.8 (8.5 to 15.1)); white nurses 0.6 (0.2 to 1.0), significantly greater (p < 0.05) than that of white women in the local social class 2 working age population (0.2 (0.1 to 0.3)); others (mortuary attendant, physiotherapist, radiographer and theatre technician) 1.0 (0.02 to 2.0), not significantly greater than in the local social class 2 working age population (0.6 (0.5 to 0.7)). CONCLUSIONS: Case ascertainment via the registry was complete. Most cases of TB were in nurses and doctors, more commonly in doctors. However, all but one of the doctors were foreign born and probably had reactivation of infection acquired abroad rather than infection occupationally acquired in the UK. All but one of the nurses, however, were UK born and may have had occupationally acquired disease, their incidence being higher than in the community. As most cases present with symptoms, there needs to be continued emphasis on prompt reporting of suspicious symptoms both before and during employment, and pre-employment BCG vaccination where appropriate.