Performance tests and disability: developing an objective index of mobility-related limitation in older populations

Abstract

Background: Disability reflects physiological limitations, social and environmental barriers, and "sickness" behavior. Being able to measure these influences separately would greatly assist interpretation of disability comparisons over time or between populations. This study aimed to identify an index of mobility-related limitations composed of physiological measures that are most closely associated with reported mobility disability in elders.

Methods: Data from the Third National Health and Nutrition Examination Survey (NHANES III) were used. Participants aged 60 and older were included in this analysis. Participants included 6596 respondents who were interviewed in their homes, and 5724 (87%) of these attended a further examination. Domains of measurements included body measurements, bone densitometry, physical examination, spirometry, fundus photography, and physical performance measures. Multivariate models were developed on a random half subsample of the data and were validated on the other half. Receiver operating characteristic (ROC) areas and logit rank slopes were used to evaluate sets of measures.

Results: In weighted logistic regression models, six and five measures were significantly associated with difficulty and inability in walking a quarter of a mile, respectively. These mainly included measures of lower extremity and lung function. A relatively minimal loss of sensitivity and specificity occurred from using more economical models, employing a subset of the identified measures.

Conclusions: Subsets of measures associated with reported mobility disability could provide objective indices of mobility-related limitation for comparing populations or long-term population health monitoring.