- Purpose: Mechanical efficiency can assess motor performance in individuals with physical disabilities. The purpose was to determine the utility of predicting it from heart rate (HR) during a self-paced stair-climbing test in adults with chronic hemiparesis after stroke and to determine the minimal detectable change of net mechanical efficiency (MEnet) measured by this exercise. Methods: First, 15 subjects with chronic hemiparesis participated in a validation study (A) and then 28 took part in a repeatability study (B). In study A the MEnet was calculated from external work and oxygen uptake above rest (dVO2), as directly measured and as predicted from body weight and increase in heart rate (dHR). In study B, predicted dVO2 was used to obtain MEnet for duplicate stair-climbing tests (T1, T2) with >30 min rest between. Results: Measured MEnet was closely related to predicted MEnet (r = 0.97, p < 0.001). In study B predicted MEnet for T2 and T1 were closely related (r = 0.91, ICC = 0.90). Conclusion: With a minimal detectable change of 0.6% (0.053 of average MEnet score of 10.4%), MEnet values from the stair-climbing test seem sufficiently meaningful to estimate ambulatory ability and its changes with interventions or walking aids in adults with hemiparesis. Implications for rehabilitation: Ambulatory ability can be estimated from mechanical efficiency, obtained from a 5-min stair-climbing test utilizing a 4-step stair, to measure external work, and the change in heart rate above rest to estimate the metabolic cost of the task. A change of > 0.6% in mechanical efficiency by this stair-climbing test indicates a significant change in ambulatory ability of persons with hemiparesis.