Date of Graduation

8-2024

Document Type

Dissertation

Degree Name

Doctor of Philosophy in Health, Sport and Exercise Science (PhD)

Degree Level

Graduate

Department

Health, Human Performance and Recreation

Advisor/Mentor

Gray, Michelle

Committee Member

Schmitt, Abigail

Second Committee Member

Greene, Nicholas

Third Committee Member

Douglas, Michael

Keywords

Aging; Biomotor abilities; Functional assessment; Physical disability; Physical function; Sit-to-Stand

Abstract

Though poor physical function (PF) and disability are two of the most impactful problems facing older adults today, the field of kinesiology has not yet devised a method of measuring PF, thus limiting investigations with the potential to lead to improved PF interventions and assessments. This dissertation aimed to develop a single, continuous measure of PF using a latent variable modeling approach, and then use the generated scores to evaluate the role of lower body muscular power in PF assessment and to construct an improved PF assessment battery for predicting overall PF. Each aim of this dissertation was addressed through an independent research study, and each study recruited a sample of adults over the age of 50. The first study utilized Bayesian exploratory factor analysis (EFA) to construct a measurement model for PF which was determined to have a single-factor structure with all examined activity of daily living (ADL) movement task indicators loading on the PF latent factor (n = 41, p(θ|X) = 99.96%). A confirmatory factor analysis (CFA) model using the extracted factor structure implied by the EFA had excellent data-model fit (n = 42, PPP = .124, BRMSEA < .001, BCFI = .952, BTLI = 1.190) and performed well in all posterior model diagnostics. Taken together, the EFA and CFA results indicate that PF can be indirectly approximated via factor analysis using measurements of the ADL movement tasks affected by PF. Factor scores were extracted for all participants and this score was termed the Physical Function Quotient (PFQ). Study two assessed the ability of lower-body muscular power to predict the PFQ scores extracted in study one and the Physical Performance Test (PPT-7) score. Generally, muscular power was a very poor predictor of both scores (n = 42, R2 ≤ .042, p ≥ .327). The final study, used exploratory regression to assess which commonly used PF assessments, taken together, creates the best predictive model for PFQ and PPT-7 in an effort to produce an optimized PF testing battery. Nine biomotor assessments and four ADL movement assessments were chosen for inclusion based on their validated and common use in PF research. Models including biomotor assessment predictors were generally poor predictors of PFQ, explaining less that 30% of the variance (R2 ≤ .294, p < .001), but 8-foot up-and-go (8UG) on its own explained almost 100% of the variance in PFQ (R2 > .999, p < .001). On the other hand, neither biomotor nor ADL movement assessment performance generated models that predicted PPT-7 well (R2 ≤ .364, p < .001). Cumulatively, these results demonstrate the utility of measuring PF indirectly with a latent variable modeling approach, biomotor assessments are not good real-time assessments of overall PF in healthy adults over 50 years of age, and 8UG is indicated for use as a simple assessment to approximate total PF.

Download

Included in

Biomechanics Commons

Share

COinS