Failure to complete performance-based measures is associated with poor health status and an increased risk of death

Sir—Mobility impairment is common in elderly people, often leads to adverse outcomes [1–4] and is intertwined with frailty [5, 6]. Three types of standardised mobility assessments [self-reported, gait laboratory and performance-based measures (PBMs)] are used, and each has its own advantages and disadvantages, including variable feasibility. PBMs attempt to optimise the practicality of clinical and self reported assessments, and the precision of the gait laboratory [7]. The Timed Up and Go (TUG) [8] and the Functional Reach (FR) [9] are used widely, [10, 11] but often cannot be used for a large proportion on whom the tests are attempted [12]. Such difficulty in undertaking the tests yields missing data, most frequently in those who are ill [13], or frail [14]. We therefore studied infeasibility in PBMs and whether missing data were informative. We compared the characteristics of people from the clinical examination in the Canadian Study of Health and Aging-2 (CSHA-2) who were able to perform both the FR and TUG, with those who could not, and then tested the predictive validity of missing data for both these tests.