In 1981-1982, a Massachusetts Hospital Association and Massachusetts Blue Cross-Blue Shield task force reviewing ancillary services found that the use of respiratory therapy at New England Deaconess Hospital far exceeded the statewide average. These findings led to a hospital-wide effort to reduce ancillary services. As part of this effort, we studied the effect of the reductions in respiratory therapy on patient outcome. At the beginning of the study, senior respiratory therapists advised physicians about the optimal use of various respiratory-therapy services, including their discontinuation when no longer necessary. One year after the intervention began, we analyzed the use of respiratory therapy by employing data-collection techniques that were identical to those used by the task force. Marked reductions in all categories of respiratory therapy had occurred, but morbidity and mortality from pulmonary disorders had not increased. In the largest group studied--patients undergoing coronary-artery bypass surgery--the charges for respiratory therapy, the length of hospital stay, and pulmonary complications had all decreased. We conclude that consistent application of prescribed guidelines for respiratory therapy results in marked decreases in its use and that such decreases can be achieved without a reduction in the quality of care.