Decrease in maximum paced walking speed may predict hospitalization in community-dwelling older people with disabilities: Assessment of change over time is more useful than a single time point assessment

Abstract Purpose Walking speed is a useful predictor of hospitalization for community-dwelling older people. However, whether it is an effective predictor for disabled community-dwelling older people has not been clarified. This study aimed to investigate the association of walking speed with unexpected hospitalizations in community-dwelling older people with disabilities. Methods The participants were ambulatory, community-dwelling, older adults aged 65 years or older with disability. Comfortable and maximum walking speeds were measured at two time points, baseline and 3 months later. Furthermore, the change over time at 3 months in walking speed was also calculated. If the change in walking speed decreased more than 0.1 m/s, it was defined as walking speed decreased. The primary outcome was unexpected hospitalization during 4-year follow-up. The associations among baseline walking speed, walking speed decline, and hospitalization were analyzed using Cox regression analysis adjusted for potential confounding factors. Results Ninety-three people were included, and unexpected hospitalization occurred in 47 people during 4-year follow-up. On Cox regression analysis adjusted for potential confounding factors, only the maximum walking speed decrease was significantly associated with hospitalization (Hazard ratio = 2.53, 95% confidence interval: 1.23–5.21], but not baseline walking speed and comfortable walking speed decrease. Conclusion As for the assessment of walking speed for prediction of unexpected hospitalization in disabled older people, measurement at a single time point is not useful, whereas change over time is. Monitoring of change over time in maximum walking speed appears to be one of the indicators for the health management of disabled older people.