Measures of physical function have been developed primarily to assess health status, prognosis, and service needs. They are now, increasingly, being used as outcome measures in studies seeking to determine the causes of disability. However, the extent to which these standardized measures, as they currently are constituted, are meaningful for the evaluation of underlying pathophysiology is not defined. To assess evidence for an etiologic rationale for these measures, we evaluated self-report of difficulty in physical function in the Cardiovascular Health Study, a study of 5201 men and women 65 years and older in four U.S. communities. We determined (by factor analysis) that self-reported difficulty with each of 17 tasks of daily life aggregates in four groups; i.e. difficulty in one task is associated with having difficulty in the other tasks in the group. These groups include (1) activities primarily dependent on mobility and exercise tolerance; (2) complex activities heavily dependent on cognition and sensory input; (3) selected basic self-care activities; and (4) upper extremity activities. Groups 2 and 3 are similar, but not identical, to Instrumental Activities of Daily Living (IADL) and Activities of Daily Living (ADL), respectively. We then tested whether these groupings were associated with different underlying impairments. Multiple logistic regression analyses indicate that there are constellations of physiologic and disease characteristics significantly (p < 0.01) associated with difficulty in each of these four groups of activities, among 15 chronic diseases and conditions ascertained. Some diseases are uniquely associated with difficulty in one group of tasks; some overlap, and are associated with 2, 3 or 4 groups of tasks. The associations found with difficulty in performing tasks in groups 2 and 3 were frequently stronger than those with the larger groups of ADL or IADL tasks, suggesting increased specificity of associations found with these new groupings. These results suggest that re-grouping of tasks of daily life may provide a more refined physiologically-based outcome measure for use in evaluating causes of disability. The ability to define risk factors for disability may be enhanced by choosing outcome measures with a demonstrated physiologic rationale.