Relationship between timed ‘up and go’ and gait time in an elderly orthopaedic rehabilitation population

Abstract

Objective: To analyse the relationship between the timed ‘up and go’ (TUG) and gait time in an elderly orthopaedic population, in order to determine whether additional useful information is obtained by measuring both.

Design: Observational study.

Setting: Jewish Rehabilitation Hospital, Laval, Quebec.

Subjects: Seventy-nine consecutive inpatients in the orthopaedic programme with a primary admitting diagnosis of either total hip replacement (THR), total knee replacement (TKR), or hip fracture repair.

Outcome measures: Timed ‘up and go’ and time to walk 10 metres.

Results: The admission correlation between gait time and TUG was r = 0.745. The correlation at discharge (r = 0.816) was higher than that seen on admission. The relationship between gait time and TUG was linear both at admission and discharge. The correlation between gait time and TUG was strong for patients with TKR at admission (r = 0.868) and discharge (r = 0.878), and for patients with THR, both at admission (r = 0.809) and discharge (r = 0.879). However, the correlation on admission was weaker for patients with hip fracture (r = 0.497). For slow walkers (people with a gait speed below 0.5 m/s) on admission, the correlation was moderate (r = 0.649). However, for those with gait speeds faster than or equal to 0.5 m/s, the correlation was weaker (r = 0.484). This discrepancy was no longer evident on discharge. Likewise, for patients with a fast TUG score (< 30 seconds) on admission, there was virtually no relationship between TUG and gait time (r = 0.084), although a good correlation was present for those with ‘up and go’ times longer than 30 seconds (r = 0.634). As with gait speed, this difference disappeared by discharge.

Conclusions: The relationship between gait time and TUG in an elderly orthopaedic rehabilitation population is good, and its strength varies by specific diagnosis, mobility, and time point in the course of therapy. The two measures are not redundant in this population.