Good preclinical models of ageing are needed to discover the molecular mechanisms behind declining human physical performance (Nature 511, 405–407; 2014). The latest animal models of frailty are a step in the right direction.

For example, a genetically modified frail-mouse model mimics the inflammation and weakness that often afflicts older people (see A. Akki et al. Age 36, 21–30; 2014). Frailty can also be modelled in naturally ageing mice as a frailty-phenotype score, graded by such performance measures as grip strength and walking speed (H. Liu et al. J. Gerontol. A http://doi.org/t6p; 2013).

We and others have quantified a clinical frailty index in mice by tracking the accumulation of age-related deficits. This index increases with age in the same way as it does in humans (J. C. Whitehead et al. J. Gerontol. A 69, 621–632; 2014).

More-sophisticated animal models of frailty will need to include a broader range of performance measures if they are to properly represent the condition in people.