Outcomes research, quality improvement, and performance measurement are important activities in the cancer research community. They are, indeed, closely related activities because indicators of quality and performance can logically be regarded as the outcomes of interest in outcomes research. Considerable progress has been made in the past decade in broadening the definition and strengthening the measurement of the important outcomes of health care in general, and cancer care specifically. The real value of studying outcomes, according to the Donabedian paradigm, lies in understanding their relation to the structure and processes (eg, the health care) that have produced them, because it is these latter factors which we can control. Therefore, the methods that we have available to us by which we can infer this relation of causality become very important. Because the systems that we study in health care research are usually complex, we will need to invest more of our resources in the future in the development of methods of inference beyond what we have available now if we are to realize the full potential of health outcomes research. This presents a unique opportunity for leadership by the cancer outcomes research program.