A newly built nursing home was analysed to establish why there had been a significantly lower attack rate in this building during a type A influenza epidemic. Measurements were taken of the number of respiratory illnesses and influenza cultures in consenting symptomatic residents, and of building characteristics. Of the four buildings in the home, Building A had significantly fewer culture-confirmed cases than the other buildings. Building A's ventilation system was unique, it had more public space per resident, and did not have office space serving the whole complex.