Physical activity interventions are delivered consistently across hospitalized older adults but multimorbidity is associated with poorer rehabilitation outcomes: A population-based cohort study.

Background: Older adults live with multimorbidity including frailty and cognitive impairment often requiring hospitalization. While physical activity interventions (PAIs) are a normal rehabilitative treatment, their clinical effect in hospitalized older adults is uncertain.

Objective: To observe PAI dosing characteristics and determine their impact on clinical performance parameters.

Design: A single-site prospective observational cohort study in an older persons' unit.

Subjects: Seventy-five older persons' unit patients ≥65 years.

Intervention: PAI; therapeutic contact between physiotherapy clinician and patient.

Measurements: Parameters included changes in activities-of-daily-living (Barthel Index), handgrip strength, balance confidence, and gait velocity, measured between admission and discharge (episode). Dosing characteristics were PAI temporal initiation, frequency, and duration. Frailty/cognition status was dichotomized independently per participant yielding 4 subgroups: frail/nonfrail and cognitively-impaired/cognitively-unimpaired.

Results: Median (interquartile range) PAI initiation occurred after 2 days (1-4), frequency was 0.4 PAIs per day (0.3-0.5), and PAI duration per episode was 3.75 hours (1.8-7.2). All clinical parameters improved significantly across episodes: grip strength median (interquartile range) change, 2.0 kg (0.0-2.3) (P < .01); Barthel Index, 5 (3-8) (P < .01); gait velocity, 0.06 m.∙s (0.06-0.16) (P < .01); and balance confidence, -3 (-6 to -1) (P < .01). Physical activity intervention dosing remained consistent within subgroups. While several moderate to large associations between amount of PAIs and change in clinical parameters were observed, most were within unimpaired subgroups.

Conclusions: PAI dosing is consistent. However, while clinical changes during hospital episodes are positive, more favourable responses to PAIs occur if patients are nonfrail/cognitively-unimpaired. Therefore, to deliver a personalized rehabilitation approach, adaptation of PAI dose based on patient presentation is desirable.