Bridging the gap between visual assessments and self-reported outcomes for optimal mobility modeling in Geriatrics.#

Authors#

Alan Castro Mejia, Stefano Sapienza, Patricia Martins Conde, Jean-Paul Steinmetz, Lukas Pavelka, Rejko Krüger, Jochen Klucken

Abstract#

Mobility, one of the four primary domains assessed in geriatrics, is frequently affected in persons above 60 through several geriatric syndromes, including frailty. While changes in gait and posture are well known across geriatrics, including neurodegenerative disorders, understanding how patients perceive changes in their mobility is beneficial to assess their overall health status and quality of life.

Self-reported changes in mobility provide an added dimension to clinical evaluation, as they can be linked to adverse medical outcomes regardless of the underlying cause. Accounting for these changes is especially important in geriatric medicine, as elderly patients are frequently treated with motor-modifying medications, like anti-arthritics, sedatives, and levodopa, that affect activities of daily living and motor capacity.

While different mobility assessments are conducted in geriatric units, the current assessments fall short in reflecting the patient’s perception. Existing methods rely on rater-dependent visual scales (e.g., Performance-Oriented Mobility Assessment, Berg Balance Scale) that often miss subtle mobility changes reported by the patient. Moreover, these assessments fail to capture the complexities of daily life, where cognitive and motor functions are combined, making comprehensive mobility and functional assessments challenging—highlighting the need for methods that capture the subtleties that the patient reports.

Previous studies have laid the groundwork for using and predicting self-reported mobility in community-dwelling elderly patients. A recent literature review by Nicolson et al. identified clinical, patient-reported, and demographic predictors that inform self-reported mobility. As a follow-up, Sanchez-Santos et al. effectively predicted self-rated mobility decline in community-dwelling geriatric patients with a limited set of easily collectible variables.

Sensor-derived gait analysis offers a solution by bridging the gap between subjective visual assessments and self-reported mobility (SRM) by quantifying impairments that are reported but not visually apparent during the clinical visit. With the advent of wearable, continuous, daily-life mobility tracking, geriatric neurological research has yet to fully explore the capacity of gait features to predict SRM, including understanding how scripted gait assessments affect the capacity of digital gait outcomes to be adequate predictors.

This manuscript defines the best gait paradigm to model self-reported mobility when various sources of clinical information, including clinical, patient-reported, demographic, and digital gait outcomes, are available. The objective is to determine the gait characteristics captured by different assessment paradigms and to identify which parameters are the best predictors of self-reported mobility.