Injury risk assessment based on post-mortem human subject (PMHS) data is essential for informing safety standards. The common ‘matched-pair’ method, which matches energy-based inputs to translate human response to dummy, consistently results in less conservative human injury risk curves due to intrinsic differences between human and dummy. Generally, dummies are stiffer than PMHSs, so force and displacement cannot be matched simultaneously. Differences in fracture tolerance further influence the dummy risk curve to be less conservative. For example, translating a human lumbar injury risk curve to a dummy of equivalent stiffness using matched-pair resulted in a dummy injury risk over 80% greater than the PMHS at 50% fracture risk. This inevitable increase occurs because the dummy continues loading without fracture to attenuate energy beyond the ‘matched’ PMHS input selected. Human injury response should be translated using an iso-energy approach, as strain energy is well associated with failure in biological tissues. Until PMHS failure, dummy force is related to PMHS force at iso-energy. Beyond PMHS failure, dummy force is related to PMHS force through failure energy. This method does not require perfect PMHS/dummy biofidelity and ensures that energy beyond PMHS failure does not influence the injury risk function.