Protocol control of severely ill ICU patients seems feasible. A satisfactory computer infrastructure makes protocol control practical. A reported four-fold survival rate increase associated with protocol control of ventilatory management of adult respiratory distress syndrome patients suggests that it is not harmful. Protocol control represents a medical decision-support approach for standardizing therapy. Computerized protocols might provide a solution to the nonuniformity of care in clinical ICU practice and investigation. My colleagues and I can envision a multicenter consortium of hospitals equipped to carry out computerized protocol-controlled care. Such a consortium could rapidly complete large, randomized, clinical trials under computerized protocol control. This arrangement could provide much more definitive results than are currently possible. Interpretation of outcomes research results should thereby be made easier and conclusions should be more credible and more likely to contribute to medical policy formulation.