Sustained human performance is critical to job and mission success in many federal agencies including national defense, aerospace exploration, and transportation. For the responsible agencies, applications of the basic biomedical and applied human factors science provide the best available solutions to help individuals perform more effectively and with increased safety. Key products of this research are biomathematical models that predict periods of impaired performance, with applications in planning tools, real time monitoring, and intervention decision aids. Since it is difficult to quantify the number of judgment errors or accidents averted, metrics of success for fatigue management systems must be largely based on the accuracy of performance predictions derived from laboratory-based research studies and the extent to which such results can be generalized to the field environment. Performance metrics must, at a minimum, be correlated with occupational task performance to demonstrate relevance to real-world applications. This paper outlines broad goals for human effectiveness research related to fatigue, alertness, and performance. To advance from the present state of knowledge to useful predictive models requires a well-coordinated commitment from federal agencies. Users should be made aware that current models and tests are not likely to encompass all of the aspects of human performance that are relevant to field environments and occupations.