To determine the incidence of aspergillosis in patients with leukemia or bone marrow transplants during a construction-associated outbreak, and the effect of an environmental control program for Aspergillus. Clinical, microbiological, and pathological records were reviewed retrospectively once the outbreak was appreciated, and prospectively thereafter, to determine the presence or absence of aspergillosis and duration of neutropenia. A university tertiary-care center with a single designated hematology-oncology unit. From January 1988 to September 1993, there were 141 patients with leukemia or bone marrow transplants identified as being neutropenic during 231 admissions to this specialized unit. Installation of wall-mounted portable high-efficiency particulate air (HEPA)-filter air purifiers, application of copper-8-quinolinolate-formulated paint, replacement of perforated ceiling tiles with nonperforated type, sealing of all windows, replacement of horizontal, dust-accumulating blinds with vinyl, opaque, roller shades, and systematic and regular cleaning of surfaces. Thirty-six cases of nosocomial aspergillosis were diagnosed during this period. The incidence density (ID) in the preconstruction period was 3.18 per 1,000 days at risk. During construction activity-before the implementation of a control strategy-the ID increased dramatically to 9.88 per 1,000 days at risk. With infection control measures implemented and continued construction work, the ID decreased to 2.91 per 1,000 days at risk, comparable to the preconstruction baseline rate. An environmental control strategy incorporating widely available technology may have played an important role in controlling this outbreak of construction-associated invasive aspergillosis.